Systems and methods for monitoring, diagnosing, and supporting decision-making in diabetes in patients with kidney disease

Abstract

Certain aspects provide a monitoring system comprising a continuous analyte sensor configured to generate an analyte measurement associated with an analyte level of a patient, and a sensor electronics module coupled to the continuous analyte sensor and configured to receive and process the analyte measurement.

Description

Technical Field

[0001] Cross - Reference to Related Applications This application claims priority and benefit to U.S. Provisional Patent Application No. 63 / 365,702, filed on June 1, 2022; U.S. Provisional Patent Application No. 63 / 376,673, filed on September 22, 2022; U.S. Provisional Patent Application No. 63 / 387,078, filed on December 12, 2022; U.S. Provisional Patent Application No. 63 / 377,332, filed on September 27, 2022; U.S. Provisional Patent Application No. 63 / 403,568, filed on September 2, 2022; and U.S. Provisional Patent Application No. 63 / 403,582, filed on September 2, 2022, which are assigned to the assignee of this application and are hereby expressly incorporated by reference herein in their entirety as if fully set forth below and for all applicable purposes.

Background Art

[0002] The kidneys perform many important functions in the human body, including filtering waste products and excess fluid excreted in urine, and removing acids produced by the body's cells to maintain a healthy balance of water, salts, and minerals (such as sodium, calcium, phosphorus, and potassium) in the blood. In other words, the kidneys play a major role in homeostasis through the renal mechanisms that transport and regulate the secretion, reabsorption, and excretion of water, salts, and minerals. Additionally, the kidneys secrete renin (such as angiotensinogenase), which forms part of the renin-angiotensin-aldosterone system (RAAS) and mediates extracellular fluid and arterial vasoconstriction (such as blood pressure). More specifically, hypertension (such as hypertension) can be regulated through RAAS inhibitors such as angiotensin-converting enzyme (ACE) inhibitors and angiotensin receptor blockers (ARBs). When the kidneys are diseased or damaged, impairment or loss of these functions can cause significant damage to the human body.

[0003] Kidney diseases occur when the kidneys are diseased or damaged. Kidney diseases are generally classified as either acute or chronic based on the duration of the disease. Acute kidney injury (AKI) (also known as "acute renal failure") is usually caused by events that result in kidney dysfunction, such as dehydration, blood loss due to major surgery or injury, and / or the use of medications. On the other hand, chronic kidney disease (CKD) is usually caused by long-term diseases such as hypertension or diabetes that gradually damage the kidneys and reduce their function over time.

[0004] Conventional kidney disease diagnosis methods and systems include electrocardiogram (ECG) monitoring, albumin-to-creatinine ratio (ACR) tests, glomerular filtration rate (GFR) tests, and blood tests to monitor a patient's potassium level. CKD is classified into five stages based on the severity of kidney dysfunction measured by various methods and systems. Kidney diseases in stages 1 to 3a are mild to moderate kidney dysfunctions. Kidney diseases in stages 3b to 5 are moderate to severe kidney dysfunctions. End stage renal disease (ESRD) is total kidney dysfunction or renal failure.

[0005] The GFR method for diagnosing and staging kidney disease represents the flow rate of the filtered fluid through the kidneys. The creatinine clearance rate is the volume of plasma from which creatinine is removed per unit time and is used to estimate the GFR. The GFR can be measured (e.g., measured GFR (mGFR)) using the gold standard method (e.g., evaluated from urinary or plasma clearance measurements of exogenous filtration markers), or can be estimated (e.g., eGFR) using equations (e.g., from measured serum levels of exogenous filtration markers using the equation). The eGFR provides a more convenient and rapid analysis for evaluating kidney function.

[0006] In some cases, when CKD remains untreated, elevated potassium levels in patients with CKD can lead to hyperkalemia, while lower potassium levels in patients with CKD can lead to hypokalemia. Hyperkalemia is a medical term that describes blood potassium levels that are higher than normal (e.g., higher than the normal blood potassium level of 3.6 to 5.2 millimoles per liter (mmol / L) per liter). Hyperkalemia increases the risk of cardiac arrhythmia episodes and sudden death. On the other hand, hypokalemia is a medical term that describes blood potassium levels that are lower than normal. In particular, CKD patients can develop hypokalemia due to gastrointestinal potassium losses from diarrhea or vomiting, or renal potassium losses from non-potassium-sparing diuretics (e.g., diuretics used to increase the amount of fluid passing from the body into the urine, regardless of the amount of potassium lost from the body into the urine). Severe hypokalemia and hyperkalemia can result in severe symptoms of respiratory failure, sudden cardiac death, or other mortality-driving events.

[0007] The kidneys also play an important role in the regulation of blood glucose. The kidneys produce glucose via gluconeogenesis and release it into the blood, thereby increasing blood glucose levels. The kidneys also lower blood glucose levels by reabsorbing glucose in the proximal renal tubules. Additionally, the kidneys use glucose as an energy source.

[0008] Glucose is a simple sugar (e.g., a monosaccharide). Glucose can not only be ingested, but also be produced in the body from proteins, fats, and carbohydrates. Serum glucose is maintained at healthy levels (e.g., glucose homeostasis) through several mechanisms. High blood glucose (i.e., hyperglycemia) is reduced by insulin and removed by the kidneys. Additionally, the kidneys consume glucose, filter glucose from the body, produce glucose (e.g., through gluconeogenesis), and reabsorb the filtered glucose. Thus, as kidney function declines, glucose consumption, filtration, production, and reabsorption decline, which can lead to, for example, an increase in glucose levels in the body at rest and after consuming a meal. Low blood glucose (i.e., hypoglycemia) is caused by gluconeogenesis in the kidneys and liver.

[0009] An increase in blood glucose levels stimulates insulin release. Insulin causes cells to take up glucose, thereby reducing serum (e.g., extracellular) glucose levels to maintain glucose homeostasis. Insulin also stimulates potassium uptake by cells, thereby reducing serum potassium levels. In some cases, when a patient's glucose level increases and the rate of change of the glucose level in the patient's body is high, excessive insulin can be produced, thereby causing excessive movement of potassium into cells. On the other hand, when a patient's glucose level decreases and the rate of change of the glucose level in the patient's body is low, insulin secretion can be less. Low insulin can result in limited access of cells to glucose and potassium. Thus, extracellular glucose and potassium levels can increase.

[0010] Gluconeogenesis is the formation of glucose from precursor molecules (e.g., lactate, glycerol, and / or amino acids). Glucose is formed in the kidneys and liver and then released into the circulation. Gluconeogenesis is a mechanism for maintaining glucose homeostasis by preventing low blood glucose (i.e., hypoglycemia). When kidney function declines, gluconeogenesis in the kidneys decreases, and thus the ability of the kidneys to respond to a decrease in blood glucose is limited.

[0011] True diabetes is a disorder in which the pancreas cannot make insulin (type I or insulin-dependent) and / or insulin is not effective (type 2 or non-insulin-dependent). In a diabetic state, the patient is troubled by high blood sugar, which causes many physiological abnormalities associated with microvascular deterioration (e.g., kidney failure, skin ulcers, or bleeding into the vitreous humor of the eye). Hypoglycemic reactions (low blood sugar) can be induced by inadvertent over-administration of insulin, or after normal administration of insulin or glucose-lowering agents, or by insufficient food intake. The treatment of diabetes requires maintaining glucose homeostasis. Glucose levels can be controlled by various drug applications including exogenous insulin.

[0012] In some cases, a patient may suffer from insulin resistance. Insulin resistance occurs when the patient's muscle, fat, and liver cells do not respond adequately to insulin. Thus, glucose metabolism, as well as intracellular potassium movement, can be impaired. As a result, the patient's pancreas makes more insulin to help glucose and insulin enter the patient's cells. Furthermore, the effect of insulin resistance on glucose metabolism can vary among different patients.

Brief Description of the Drawings

[0013] To enable a detailed understanding of the features of the present disclosure enumerated above, a more specific description, briefly summarized above, may be made by reference to the embodiments, some of which are illustrated in the drawings. However, it should be noted that the attached drawings merely illustrate certain exemplary embodiments of the present disclosure and, therefore, should not be regarded as limiting the scope thereof since the description may be applicable to other equally effective embodiments.

Figure 1

Figure 2

Figure 3

Figure 4

Figure 5

Figure 6

Figure 7

Figure 8A

Figure 8B

Figure 8C

Figure 8D

Figure 8E

Figure 9A

Figure 9B

Figure 9C

Figure 9D

Figure 9E

Figure 10A

Figure 10B

Figure 10C

Figure 11

Figure 12A

Figure 12B

Figure 12C

Figure 12D

Figure 13A

Figure 13B

Figure 14

Figure 15A

Figure 15B

Figure 15C

Figure 16

[0014] For ease of understanding, the same reference numbers are used, where possible, to designate the same elements common to each figure. It is contemplated that elements disclosed in one aspect may be beneficially utilized in other aspects without specific recitation.

DETAILED DESCRIPTION

[0015] One of the greatest risks associated with chronic kidney disease (CKD) is hypoglycemia. 1.7% of annual hospitalizations are due to hypoglycemia in early CKD (CKD < stage 3). 3.6% of hospitalizations associated with end-stage renal disease (ESRD) are due to hypoglycemia, and the mortality rate is 30%. Specifically, a decrease in glucose homeostasis increases the risk of hypoglycemic events. Hypoglycemic events in kidney disease are due to a decrease in insulin clearance, impaired glucose clearance, and impaired renal gluconeogenesis. For ESRD-related hospitalizations, dialysis is also a worsening factor in the decrease of glucose homeostasis. Improved diabetes diagnosis and glucose management are needed for patients with kidney disease.

[0016] Renal dysfunction reduces the body's ability to maintain glucose homeostasis through decreased insulin clearance, impaired gluconeogenesis, and impaired glucose clearance. Insulin clearance decreases with decreasing renal function. The kidneys are unable to properly filter and remove (i.e., eliminate) insulin from the circulation. Therefore, both endogenous and exogenous insulin can have long-term and / or significant effects. This increases the risk of hypoglycemia in patients with renal dysfunction. As renal dysfunction progresses, insulin can have even longer-term and / or significant effects, which further increases the risk of hypoglycemia in patients with worsening CKD. For example, in patients with worsening CKD, an insulin dose that at one point reduces blood glucose to a healthy range can cause hypoglycemia at a later point.

[0017] Additionally, gluconeogenesis in the kidneys also decreases as a result of CKD. Therefore, the kidneys' ability to respond to decreased blood glucose or hypoglycemia is reduced. Furthermore, due to the decrease in insulin clearance, insulin remains longer and circulates, which further reduces glucose levels. In such cases, the renal response to decreased glucose, gluconeogenesis, is ineffective, glucose levels continue to decrease, insulin remains for a longer time, and hypoglycemia results. In addition, in CKD patients, glucose clearance is impaired because the kidneys are unable to effectively filter, reabsorb, and consume blood glucose. Therefore, high glucose levels are not reduced to appropriate levels and can cause hyperglycemia.

[0018] Current methods for providing decision support for diabetes, including diagnosis, monitoring, and treatment, are not reliable for patients with diabetes and CKD because such methods cannot account for the effect of renal dysfunction on blood glucose. Specifically, A1C, one methodology for diabetes diagnosis, is particularly unreliable in patients with advanced CKD (CKD > stage 3). A1C is a measure of the glycosylation of hemoglobin found in red blood cells. This is the percentage of glycated modified hemoglobin based on the assumed red blood cell half-life. Thus, A1C summarizes the duration of high blood glucose over the lifespan of red blood cells. However, in patients with renal dysfunction, the red blood cell half-life decreases. This shorter red blood cell half-life means that the clinically measured A1C is based on an inaccurate assumed red blood cell half-life. More specifically, the clinically measured A1C for patients with CKD is lower than the patient's actual A1C. For example, a CKD patient may have a glycated A1C level in the healthy range (e.g., 5.5%), but the actual blood glucose level may be elevated (e.g., 200 mg / dL). This can be particularly prominent as a difference with respect to postprandial times when the patient's glucose levels can rise significantly but are not as easily known through some diagnostic methods such as fasting glucose tests. Additionally, for patients with severe CKD (estimated GFR < 30 mL / min / 1.73 m2), the shortened red blood cell half-life is well-documented, particularly in patients receiving dialysis and erythropoietin-stimulating agents. Clinical practice guidelines for physicians treating patients with CKD recommend maintaining only slightly elevated A1C as a way to reduce the chance of severe hypoglycemic episodes in this patient population, despite literature demonstrating that the progression of kidney disease accelerates faster in patients with poor glycemic control compared to those with better glycemic control. Thus, A1C is not a reliable estimate of blood glucose in patients with CKD.

[0019] Thus, patients with CKD may remain undiagnosed with diabetes until diabetes progresses to a severity that can be detected by conventional methods. These patients suffer from a decline in glucose homeostasis, including an increased risk of adverse glycemic events such as hypoglycemia and hyperglycemia.

[0020] For patients with renal impairment, the diagnostic level of A1C can be adjusted downward to account for the effect of renal disease on A1C. For example, patients with known renal impairment have an artificially low A1C, such that a clinician may diagnose diabetes at a lower threshold (e.g., 6.0% instead of 6.5%) when calculating the clinical A1C to account for the effect of renal impairment on red blood cell half-life. However, because the actual red blood cell half-life in patients with renal impairment is unknown without additional analysis, this method remains error-prone. Additionally, impaired clearance of waste products such as urea can accelerate the glycation process and thus result in higher A1C values compared to healthy individuals. Thus, adjusted A1C is still calculated based on inaccurate assumptions without taking into account the variability between patients for individualized assessment.

[0021] Additionally, in patients with CKD, blood glucose may vary more widely due to renal impairment. Greater blood glucose variability results in increased glycemic variability. High glycemic variability can result from higher and / or longer elevated glucose levels, as well as lower and / or longer decreased glucose levels. Patients with CKD may experience greater glycemic variability because the kidneys are less effective at lowering high glucose levels and raising low glucose levels. Thus, renal impairment impairs glucose homeostasis and increases glycemic variability.

[0022] Additionally, in patients with advanced CKD during dialysis, blood glucose can have greater variability in fluctuations. The effects of dialysis, the concentration of carbohydrates or salts in the dialysis solution, the membrane charge or filtration size, the diafiltration rate, and other dialysis parameters can contribute to the variability of fluctuations as glucose is either filtered or added to the bloodstream during dialysis. In addition, although insulin is a molecule large enough not to be easily removed through the dialysis membrane, insulin molecules can adhere to the membrane such that they are partially removed through dialysis. Loss of insulin during dialysis can reduce the amount of insulin available during and immediately after the dialysis procedure, which can result in a post-dialysis blood glucose spike.

[0023] Disorders of glucose homeostasis and increased blood glucose variability increase the risk of hyperglycemia and / or hypoglycemia. Accordingly, patients with CKD are at increased risk of hyperglycemia and / or hypoglycemia. Also, patients with renal dysfunction have an atypical glucose profile compared to patients in a similar condition without renal dysfunction. For example, a patient without renal dysfunction may consume a meal and as a result experience an increase in glucose levels, while a patient with renal dysfunction consuming the same meal may experience a lower resulting increase in glucose levels. Renal dysfunction can also affect insulin clearance, so an atypical glucose profile may also appear after insulin administration. For example, a patient without renal dysfunction may consume a meal, receive a certain dose of insulin, and experience a resulting glucose profile, while a patient with renal dysfunction consuming the same meal and insulin dose may experience a different glucose profile.

[0024] Accordingly, even if a CKD patient is diagnosed with diabetes, conventional treatment methods may not be able to account for the impact of CKD on their diabetes and ability to maintain glucose homeostasis. Conventional methods and systems used to evaluate, monitor, and treat diabetes in patients with renal disease cannot account for the interaction between renal disease and diabetes. Current diabetes monitoring and treatment utilize continuous analyte monitoring systems that include at least continuous glucose sensors.

[0025] Some examples of continuous glucose monitors include glucose monitoring sensors. In some embodiments, the glucose monitoring sensor is an implantable sensor as described with reference to U.S. Patent No. 6,001,067 and U.S. Patent Publication No. 2011-0027127 (A1). In some embodiments, the glucose monitoring sensor is a transcutaneous sensor as described with reference to U.S. Patent Publication No. 2006-0020187 (A1). In some embodiments, the glucose monitoring sensor is a dual electrode analyte sensor as described with reference to U.S. Patent Publication No. 2009-0137887 (A1). In yet other embodiments, the glucose monitoring sensor is configured to be implanted in a host blood vessel or extracorporeally, such as the sensor described in U.S. Patent Publication No. 2007-0027385 (A1). These patents and publications are hereby incorporated by reference in their entirety.

[0026] As used herein, the term "continuous" can mean fully continuous, semi-continuous, periodic, etc. Such continuous monitoring of an analyte is advantageous in the diagnosis and staging of diseases, considering that continuous measurements provide information not only about the trend and rate of analyte changes over a continuous period, but also continuously provide the latest measurements. Such information can be used to make more informed decisions in the assessment of glucose homeostasis and the treatment of diabetes.

[0027] Overall, existing diagnostic methods have a first technical problem of being inaccurate and unreliable methods for diagnosing diabetes in patients with renal dysfunction due to the effect of renal dysfunction on glucose homeostasis, including the effect on red blood cell half-life and increased glucose variability.

[0028] Additionally, existing monitoring and treatment methods also suffer from a second technical problem of being unable to account for the effect of renal disease on diabetes and / or glucose homeostasis when providing health and intervention strategies.

[0029] As a result of these technical problems, diagnosing diabetes, hyperglycemia, and / or hypoglycemia, or the risks thereof, and maintaining glucose homeostasis in patients with renal disease using conventional techniques is not only inaccurate but also impossible, and in some cases, can prove life-threatening to patients with such diseases. Specifically, predicting the presence of diabetes in patients with renal dysfunction with reasonable accuracy in a personalized manner may be necessary due to the conflation, interaction, and coexistence of diabetes and renal disease. Further, considering that both hypoglycemia and hyperkalemia can result in severe and life-threatening conditions, it may be necessary to predict hyperglycemia and / or hypoglycemia and the risks of corresponding atypical glucose trends. Accordingly, an improved method for detecting and predicting diabetes in patients with renal disease is desired based on the interaction between such conditions in a patient.

[0030] Accordingly, certain embodiments described herein provide a technical solution to the above technical problems by providing an improved decision support and diagnostic system configured to take into account the impact of renal disease on glucose homeostasis in order to provide more accurate and effective decision support for patients having both diabetes and renal disease. As discussed in more detail herein, decision support may be provided in the form of risk assessment, diagnosis, staging, warnings, alarms, and / or recommendations for the treatment of diabetes in patients with renal dysfunction, as described in more detail herein. As used herein, risk assessment may refer to an assessment of the risks associated with the presence and progression of diabetes, an assessment or prediction of diabetes reaching a more advanced stage, the risk of death, the risk of being diagnosed with one or more other diseases, the risk of experiencing one or more symptoms, and the like.

[0031] In certain embodiments, a continuous analyte monitoring system may provide decision support to a patient based on various collected data including analyte data, patient information, secondary sensor data (e.g., non-analyte data), and the like. For example, the analyte data may include continuously monitored glucose data in addition to other continuously monitored analyte data such as potassium, lactate, pyruvate, insulin, and ketones.

[0032] Certain embodiments of the present disclosure provide techniques and systems for using measurements associated with other analyte sensor data, secondary sensor data, and / or other patient information in conjunction with glucose levels, as further described below. As described above, the collected data may also include patient information that may include information associated with age, gender, family history of diabetes, kidney disease, family history of kidney disease, other health conditions, and the like. The secondary sensor data may include accelerometer data, heart rate data (ECG, HRV, HR, etc.), body temperature, blood pressure, or any other sensor data other than analyte data.

[0033] According to certain embodiments of the present disclosure, the decision support system presented herein is configured to provide disease decision support for diagnosing patients with or at risk of diabetes and for assisting patients in managing their kidney disease or risk thereof. Providing diabetes decision support includes (1) automatically detecting abnormal patterns associated with various glucose metrics, (2) assessing the presence and severity of diabetes, (3) risk stratifying patients to identify those at high risk of diabetes, (4) identifying risks associated with current diabetes diagnoses (e.g., risk of death, hyperglycemia, hypoglycemia, dementia, end-stage renal disease, CKD stages 3-4, heart failure, myocardial infarction, chronic obstructive pulmonary disease COPD, cirrhosis, proliferative retinopathy, etc.), (5) making patient-specific treatment decisions or recommendations for glucose homeostasis management and kidney disease, and (6) using large amounts of collected data, such as the analyte data, patient information, and secondary sensor data described above, to provide information regarding the effects of interventions (e.g., the effects of patient lifestyle changes, surgical procedures, the effects of patients taking new medications, the effects of patients receiving insulin injections, etc.). In other words, the decision support system presented herein can provide information useful for directing and improving care for patients with or at risk of having renal dysfunction or a decrease in glucose homeostasis.

[0034] According to certain embodiments of the present disclosure, the decision-making support system presented herein is configured to predict at least one risk or likelihood of an atypical glucose trend occurring in real time for a patient or within a specified period (e.g., a future period such as the next few seconds, minutes, hours, days, etc.). In particular, the decision-making support system presented herein can be configured to predict, for example, the risk or likelihood that a patient with a kidney disease experiences hyperglycemia and / or hypoglycemia (in real time or later), as well as the effectiveness of various treatment options. As discussed, patients with kidney disease and diabetes can experience hyperglycemia, hypoglycemia, and / or other glucose trends (e.g., glucose variability, time in range (TIR), etc.) differently than patients with diabetes alone. Further, the effectiveness of treatment can differ in patients with kidney disease and diabetes as compared to patients with diabetes alone. Thus, recognition of atypical glucose trends and effective treatment options for patients with diabetes and kidney disease can be important for the success of glucose management in patients with kidney disease.

[0035] In certain embodiments, the decision-making support system described herein can use various algorithms or artificial intelligence (AI) models, such as machine learning models, trained based on patient-specific data and / or population data to provide real-time decision-making support to a patient based on the collected information about the patient. For example, certain aspects can target algorithms and / or machine learning models designed to assess the presence and severity of diabetes in a patient. The algorithms and / or machine learning models can be used in combination with one or more continuous analyte sensors, including at least a continuous glucose sensor, to provide a real-time assessment of diabetes. In particular, the algorithms and / or machine learning models can take into account parameters such as the patient's glucose metrics over time, as well as the patient's physiological parameters such as kidney disease and stage, when diagnosing, monitoring, and providing decision-making support for diabetes.

[0036] Based on these parameters, the algorithm and / or machine learning model can provide risk assessments for diabetes, type, and / or severity, CKD and its progression, type, and severity risk assessment, as well as various types of decision-making support. The algorithm and / or machine learning model can take into account population data, personalized patient-specific data, or a combination of both when diagnosing diabetes in a patient with kidney disease.

[0037] According to certain embodiments, prior to deployment, the machine learning model is trained with training data including, for example, user-specific data and / or population data. As described in more detail herein, population data can be provided in the form of a dataset containing data records of historical patients with kidney disease at various stages, having various types of diabetes, and / or experiencing a decrease in glucose homeostasis, different symptoms associated with diabetes in some cases. Each data record can be used as an input to the machine learning model to optimize such a model to generate accurate predictions associated with diabetes (e.g., predictions of the presence of diabetes, type of diabetes, severity of diabetes, etc.). A combination of a continuous analyte monitoring system with a machine learning model and / or algorithm for (1) predicting the presence and / or severity of diabetes in a patient with kidney disease and (2) providing decision-making support for glucose management for a patient with kidney disease (e.g., predicting optimal treatment, providing dietary and exercise recommendations, etc.). For example, the decision-making support system can be used to provide early warnings of diabetes, a decrease in glucose homeostasis, and / or deliver information regarding glucose levels. Maintaining glucose levels reduces the risk of severe health outcomes, particularly for patients with kidney disease. For example, detection and support for glucose management for patients with kidney disease reduces the risk of hospitalization, complications, and death in some cases. In addition, the glucose metrics and atypical glucose trends provided by the continuous analyte monitoring system can be used as inputs to the machine learning model and / or algorithm to triage patients for more urgent care.

[0038] Through a combination of a continuous analyte monitoring system and machine learning models and / or algorithms, the decision support system described herein is configured to provide the accuracy and reliability required by the patient. For example, when assessing the presence and severity of diabetes in a patient with kidney disease, bias, human error, and emotional influences can be minimized. Further, machine learning models and algorithms combined with an analyte monitoring system can provide insights into patterns and / or trends that may previously have been missed and that degrade the health of the patient, at least with respect to the kidneys and / or glucose. Accordingly, the decision support system described herein improves upon existing decision support systems and, more generally, the fields of disease monitoring, diagnosis, and treatment.

[0039] Exemplary Decision Support System Including Exemplary Analyte Sensors FIG. 1 illustrates an exemplary decision support system 100 for (1) predicting the presence and / or severity of diabetes in a patient with kidney disease and (2) providing decision support to a user 102 (referred to individually herein as a user and collectively herein as users). The decision support system 100 is configured to provide decision support to a user 102 using a continuous analyte monitoring system 104 that includes at least a continuous glucose sensor. The user may be, in certain embodiments, a patient or, in some cases, a caregiver of a patient. In certain embodiments, the decision support system 100 includes a continuous analyte monitoring system 104, a display device 107 that executes an application 106, a decision support engine 114, a user database 110, a history database 112, and a training server system 140, each of which is described in more detail below.

[0040] As used herein, the term "analyte" is a broad term used in its ordinary sense and includes, but is not limited to, substances or chemical components within a biological fluid (e.g., blood, interstitial fluid, cerebrospinal fluid, lymph, or urine) that can be analyzed. Analytes can include naturally occurring substances, artificial substances, metabolites, and / or reaction products. Analytes that can be measured by the present device and method include, but are not limited to, potassium, glucose, acarboxyprothrombin; acylcarnitine; adenine phosphoribosyltransferase; adenosine deaminase; albumin; α-fetoprotein; amino acid profile (arginine (Krebs cycle), histidine / urocanic acid, homocysteine, phenylalanine / tyrosine, tryptophan); androstenedione; antipyrine; arabinitol enantiomers; arginase; benzoylecgonine (cocaine); biotinidase; biopterin; c-reactive protein; carnitine; carnosinase; CD4; ceruloplasmin; chenodeoxycholic acid; chloroquine; cholesterol; cholinesterase; conjugated 1-β-hydroxy-cholic acid; cortisol; creatine kinase; creatine kinase MM isoenzyme; cyclosporin A; cystatin C; d-penicillamine; deethylchloroquine; dehydroepiandrosterone sulfate; DNA (acetylation polymorphism, alcohol dehydrogenase, alpha1-antitrypsin, glucose-6-phosphate dehydrogenase, hemoglobin A, hemoglobin S, hemoglobin C, hemoglobin D, hemoglobin E, hemoglobin F, D-Punjab, hepatitis B virus, HCMV, HIV-1, HTLV-1, MCAD, RNA, PKU, Plasmodium vivax, 21-deoxycortisol); desbutylhalofantrine; dihydropteridine reductase; diphtheria / tetanus antitoxin; erythrocyte arginase; erythrocyte protoporphyrin; esterase D; fatty acid / acylglycine; free β-human chorionic gonadotropin; free erythrocyte protoporphyrin; free thyroxine (FT4); free tri-iodothyronine (FT3); fumarylacetoacetase;Galactose / gal-1-phosphate; galactose-1-phosphate uridyltransferase; gentamicin; glucose-6-phosphate dehydrogenase; glutathione; glutathione peroxidase; glycolic acid; glycosylated hemoglobin; halofantrine; hemoglobin variant; hexosaminidase A; human erythrocyte carbonic anhydrase I; 17-α-hydroxyprogesterone; hypoxanthine phosphoribosyltransferase; immunoreactive trypsin; lactate; lead; lipoproteins ((a), B / A-1, β); lysozyme; mefloquine; netilmicin; phenobarbital; phenytoin; phytanic acid / pristanic acid; progesterone; prolactin; prolidase; purine nucleoside phosphorylase; quinene; reverse tri-iodothyronine (rT3); selenium; serum pancreatic lipase; sisomicin; somatomedin C; specific antibodies that recognize any one or more of the following that may be included (adenovirus, antinuclear antibody, anti-zeta antibody, arbovirus, pseudorabies virus, dengue virus, guinea worm, tapeworm, amoeba dysenteriae, enterovirus, giardiasis, Helicobacter pylori, hepatitis B virus, herpes virus, HIV-1, IgE (atopic disease), influenza virus, Donovan leishmania, Leptospira, mumps / epidemic parotitis / rubella, mycoplasma pneumoniae, myoglobin, guinea worm, parainfluenza virus, malaria parasite, poliovirus, Pseudomonas aeruginosa, respiratory syncytial virus, Rickettsia (scrub typhus), Schistosoma mansoni, Toxoplasma, Treponema pallidum, Trypanosoma cruzi / Langeri, vesicular stomatitis virus, Wuchereria bancrofti, yellow fever virus); specific antigens (hepatitis B virus, HIV-1); succinylacetone; sulfadoxine; theophylline; thyrotropin (TSH); thyroxine (T4); thyroxine-binding globulin; trace elements; transferrin; UDP-galactose-4-epimerase; urea; uroporphyrinogen I synthase; vitamin A; white blood cells;and zinc protoporphyrin. Salts, sugars, proteins, fats, vitamins, and hormones that are naturally present in blood or interstitial fluid can also constitute an analyte in certain implementations. An ion is a charged atom or compound that can include, for example, sodium, potassium, calcium, chloride, nitrogen, or bicarbonate. An analyte can be naturally present in a biological fluid, such as metabolites, hormones, antigens, antibodies, ions, etc. Alternatively, an analyte can be introduced into the body or be exogenous, such as a contrast agent for imaging, a radioisotope, a chemical agent, a fluorocarbon-based synthetic blood, a challenge agent analyte (e.g., introduced for the purpose of measuring an increase and / or decrease in the rate of change of the concentration of a challenge agent analyte or another analyte in response to the introduced challenge agent analyte), or exogenous insulin; glucagon, ethanol; cannabis (marijuana, tetrahydrocannabinol, hashish); inhalants (nitrous oxide, amyl nitrite, butyl nitrite, chlorinated hydrocarbons, hydrocarbons); cocaine (crack cocaine); stimulants (amphetamine, methamphetamine, Ritalin, Cylert, Preludin, Didrex, Prestate, Voranil, Sandrex, Plegine); antidepressants (barbiturates, methaqualone, Valium, Librium, Miltown, Serax, Equanil, Tranxene, and other tranquilizers); hallucinogens (fenciclovir, lysergic acid, mescaline, peyote, psilocybin); narcotics (heroin, codeine, morphine, opium, meperidine, Percocet, Percodan, Tussionex, fentanyl, Darvon, Talwin, Lomotil); synthetic narcotics (fentanyl, meperidine, amphetamine, methamphetamine, and analogs of fenciclovir, such as ecstasy); anabolic steroids;It may be a drug or pharmaceutical composition that contains nicotine and is not limited thereto. Metabolites of the drug and pharmaceutical composition are also intended analytes. For example, analytes such as ascorbic acid, uric acid, dopamine, norepinephrine, 3-methoxytyramine (3MT), 3,4-dihydroxyphenylacetic acid (DOPAC), homovanillic acid (HVA), 5-hydroxytryptamine (5HT), and 5-hydroxyindoleacetic acid (5HIAA), as well as intermediates of the citric acid cycle and other chemicals produced in the body, such as neurochemicals, can also be analyzed.;

[0041] Analytes measured and analyzed by the devices and methods described herein include glucose and potassium, and in some cases, lactate, pyruvate, and ketones, and other analytes, including but not limited to those listed above, are also contemplated and can be measured, for example, by the analyte monitoring system 104.

[0042] In certain embodiments, the continuous analyte monitoring system 104 is configured to continuously measure one or more analytes and transmit analyte measurements to an electronic medical records (EMR) system (not shown in FIG. 1). The EMR system is a software platform that enables the electronic input, storage, and maintenance of digital medical data. The EMR system is generally used throughout hospitals and / or other care facilities to document clinical information about patients over a long period of time. The EMR system organizes and presents data in a way that aids clinicians, for example, by interpreting the state of health and providing ongoing care, scheduling, billing, and follow-up. Reports for patient clinical care and / or disease management can also be created using the data contained in the EMR system. In certain embodiments, the EMR may communicate (e.g., via a network) with the decision support engine 114 to perform the techniques described herein. The communication can occur via various network connection data configurations including, but not limited to, web API protocols, HL7, FHIR, EDI, XML, CDA, etc. These data communication configurations may be sent directly to the EMR or may be sent through one or more intermediate systems including, but not limited to, an interface engine, enter the EMR system, and then be displayed. Patient data communicated to the EMR via any of these other means may be matched to patient records via probabilistic matching, manual human matching, or via an EMPI or MPI (Master Patient Index, Electronic Master Patient Index) to ensure that data entry into one system matches patient information in another system. Data from the analyte device may also be matched with data from an alternative device or system before being entered into the EMR, or the data may be sent in the reverse direction to the inventors' historical record database 112, user database 110, and / or decision support engine. These data transfers enable the system to perform optimized decision support via the means described herein.In particular, as described herein, the decision support engine 114 can obtain data associated with a user, use the obtained data as input to one or more trained models, and output a prediction. In some cases, the EMR can provide the decision support engine 114 with data that is used as input to one or more models. Further, in some cases, the decision support engine 114 can provide the output prediction to the EMR after making a prediction.

[0043] In certain embodiments, the continuous analyte monitoring system 104 is configured to continuously measure one or more analytes and transmit the analyte measurements to a display device 107 for use by an application 106. In some embodiments, the continuous analyte monitoring system 104 transmits the analyte measurements to the display device 107 through a wireless connection (e.g., a Bluetooth connection). In certain embodiments, the display device 107 is a smartphone. However, in certain other embodiments, the display device 107 can be any other type of computing device, such as a laptop computer, a smartwatch, a tablet, or any other computing device capable of running the application 106. In some embodiments, the continuous analyte monitoring system 104 and / or the analyte sensor application 106 transmit the analyte measurements to one or more other individuals interested in the patient's health (e.g., family members or physicians for the patient's real-time treatment and care). The continuous analyte monitoring system 104 can be described in more detail with respect to FIG. 2.

[0044] The application 106 is a mobile health application configured to receive and analyze analyte measurements from the analyte monitoring system 104. In particular, the application 106 stores information about the user, including the user's analyte measurements, in the user's user profile 118 for use by the decision support engine 114 for processing and analysis and for providing decision support recommendations or guidance to the user.

[0045] The decision-making support engine 114 refers to a set of software instructions having one or more software modules, including a data analysis module (DAM) 116. In certain embodiments, the decision-making support engine 114 executes entirely on one or more computing devices in a private or public cloud. In such embodiments, the application 106 communicates with the decision-making support engine 114 via a network (e.g., the Internet). In some other embodiments, the decision-making support engine 114 executes partially on one or more local devices such as the display device 107 and partially on one or more computing devices in a private or public cloud. In some other embodiments, the decision-making support engine 114 executes entirely on one or more local devices such as the display device 107. As will be discussed in more detail herein, the decision-making support engine 114 may provide decision-making support recommendations to the user via the application 106. The decision-making support engine 114 provides decision-making support recommendations based on the information included in the user profile 118.

[0046] User profile 118 may include information collected regarding the user from application 106. For example, application 106 provides a set of inputs 128 including analyte measurements received from continuous analyte monitoring system 104 and stored within user profile 118. In certain embodiments, inputs 128 provided by application 106 include additional data in addition to analyte measurements received from continuous analyte monitoring system 104. For example, application 106 may obtain additional inputs 128 via manual user input, one or more other non-analyte sensors or devices, other applications executed on display device 107, etc. Non-analyte sensors or devices include, but are not limited to, an insulin pump, an electrocardiogram (ECG) sensor or heart rate monitor, a blood pressure sensor, a sweat sensor, a respiratory sensor, a thermometer, a peritoneal dialysis device, a hemodialysis device, sensors or devices provided by display device 107 (e.g., an accelerometer, a camera, a global positioning system (GPS), a heart rate monitor, etc.), or other user accessories (e.g., a smartwatch), or any one or more of any other sensors or devices that provide relevant information regarding the user. Inputs 128 to user profile 118 provided by application 106 are described in further detail below with respect to FIG. 3.

[0047] The DAM 116 of decision support engine 114 is configured to process the set of inputs 128 to determine one or more metrics 130. Metrics 130, which are discussed in more detail below with respect to FIG. 3, may in at least some cases indicate the user's health or condition, such as one or more of generally the user's physiological state, trends associated with the user's health or condition, etc. In certain embodiments, metrics 130 may then be used by decision support engine 114 as inputs for providing guidance to the user. As illustrated, metrics 130 are also stored in user profile 118.

[0048] The user profile 118 also includes demographic information 120, disease progression information 122, and / or pharmaceutical information 124. In certain embodiments, such information may be provided through user input or obtained from a particular data storage device (e.g., an electronic medical record (EMR)). In certain embodiments, the demographic information 120 may include one or more of the user's age, body mass index (BMI), ethnicity, gender, etc. In certain embodiments, the disease progression information 122 may include information regarding the user's diseases, such as whether the user has previously been diagnosed with acute kidney injury (AKI), chronic kidney disease (CKD), and / or diabetes, or whether the user has a history of hyperkalemia, hypokalemia, hyperglycemia, hypoglycemia, etc. In certain embodiments, the information regarding the user's diseases may also include the length of time since diagnosis, the stage of the disease, the level of disease control, the level of compliance with disease management therapies, predicted renal function, other types of diagnoses (e.g., heart disease, obesity), or health metrics (e.g., heart rate, exercise, stress, sleep, etc.). In certain embodiments, the disease progression information 122 may be provided, for example, as the output of one or more prediction algorithms and / or trained models based on analyte sensor data generated via the continuous analyte monitoring system 104.

[0049] In certain embodiments, the medication information 124 may include information regarding the amount, frequency, and type of medication taken by the user. In certain embodiments, the amount, frequency, and type of medication taken by the user are time-stamped and correlated with the user's analyte levels, thereby indicating the impact of the amount, frequency, and type of that medication on the user's analyte levels. In certain embodiments, the medication information 124 may include information regarding the consumption of one or more drugs known to control and / or improve glucose homeostasis. Examples of one or more drugs known to control and / or improve glucose homeostasis may include medications for lowering blood glucose levels, such as insulin including rapid-acting and long-acting insulin, and other medications for lowering glucose, such as metformin. As will be described in more detail below, the decision support system 100 may be configured to use the medication information 124 to determine the optimal insulin administration to be prescribed to different users. In particular, the decision support system 100 may be configured to identify one or more optimal insulin administrations based on the patient's health, the patient's current condition, and / or the effectiveness of insulin administration.

[0050] In certain embodiments, the medication information 124 may include information regarding the consumption of one or more drugs known to damage the kidneys. Examples of one or more drugs known to damage the kidneys may include nonsteroidal anti-inflammatory drugs (NSAIDs), such as ibuprofen (e.g., Advil, Motrin) and naproxen (e.g., Aleve), vancomycin, iodinated radiocontrast agents (e.g., referring to any contrast agent used in diagnostic tests), angiotensin-converting enzyme (ACE), such as lisinopril, enalapril, and ramipril, aminoglycoside antibiotics, such as neomycin, gentamicin, tobramycin, and amikacin, anti-viral human immunodeficiency virus (HIV) drugs, zoledronic acid (e.g., Zometa, Reclast), foscarnet, and the like.

[0051] In certain embodiments, the medical information 124 may include information regarding the consumption of one or more drugs known to control complications of kidney disease. The one or more drugs known to control complications of kidney disease may include medications for reducing blood pressure and maintaining kidney function, such as ACE inhibitors or angiotensin II receptor blockers, medications for treating anemia, such as supplements of the hormone erythropoietin, medications used to lower cholesterol levels, such as statins, medications used to prevent weak bones, such as calcium and vitamin D supplements, phosphate binders, and the like.

[0052] In certain embodiments, at least a portion of the information stored in the user profile 118 may be modified over time and / or new information may be added to the user profile 118 by the decision support engine 114 and / or the application 106, such that the user profile 118 is dynamic. Thus, the information in the user profile 118 stored within the user database 110 provides an up-to-date repository of information related to the user.

[0053] The user database 110, in some embodiments, refers to a storage server operating in a public or private cloud. The user database 110 may be implemented as any type of data store, including a relational database, a non-relational database, a key-value data store, a file system including a hierarchical file system. In some exemplary implementations, the user database 110 is distributed. For example, the user database 110 may comprise a plurality of distributed persistent storage devices. Further, the user database 110 may be replicated such that the storage devices are geographically distributed.

[0054] The user database 110 includes user profiles 118 associated with a plurality of users that interact in the same manner as the application 106 running on the display devices 107 of other users. The user profiles stored in the user database 110 are accessible not only to the application 106 but also to the decision support engine 114. The user profiles in the user database 110 may be accessible to the application 106 and the decision support engine 114 via one or more networks (not shown). As described above, the decision support engine 114, more specifically the DAM 116 of the decision support engine 114, can fetch the input 128 from the user database 110 and then calculate a plurality of metrics 130 that can be stored as application data 126 in the user profile 118.

[0055] In certain embodiments, the user profiles 118 stored in the user database 110 may also be stored in the history database 112. The user profiles stored in the history database 112 may provide a repository of the latest and historical information for each user of the application 106. Thus, the history database 112 essentially provides all the data related to each user of the application 106, and the data is stored according to the associated timestamp. The timestamp associated with the information stored in the history database 112 can identify, for example, when the information related to the user was acquired and / or updated.

[0056] Furthermore, the history database 112 can maintain time-series data collected about a user over a period of time, including that for users who use the continuous analyte monitoring system 104 and the application 106. For example, the analyte data for a user who has used the continuous analyte monitoring system 104 and the application 106 over a five-year period for managing the user's health may have time-series analyte data associated with the user maintained over the five-year period.

[0057] Further, in certain embodiments, the history database 112 may include data for one or more patients who are not users of the continuous analyte monitoring system 104 and / or the application 106. For example, the history database 112 may include information (e.g., user profiles) regarding one or more patients who have not previously been diagnosed with diabetes and / or kidney disease, as analyzed by, for example, a health care provider (or other known method), and information (e.g., user profiles) regarding one or more patients who have previously been diagnosed with diabetes and / or kidney disease (of various types and stages), as analyzed by, for example, a health care provider (or other known method). The data stored in the history database 112 may be referred to herein as population data.

[0058] The data associated with each patient stored in the history database 112 may provide time series data collected over the patient's disease lifetime, where the disease may be diabetes and / or kidney disease. For example, the data may include information about the patient prior to being diagnosed with kidney disease, and information associated with each stage of kidney disease that progressed and / or regressed in the patient, and information associated with other diseases such as hyperkalemia, hypokalemia, diabetes, hypertension, heart conditions and diseases, or similar diseases that co-exist in relation to kidney disease. Such information may indicate, over the lifetime of the kidney disease, the patient's symptoms, the patient's physiological state, the patient's potassium level, the patient's glucose level, the patient's lactate level, the patient's pyruvate level, the patient's insulin level, the status / condition of one or more of the patient's organs, the patient's habits (e.g., activity level, food consumption, etc.), prescribed medications, and the like.

[0059] In another example, the data may include information related to a patient prior to being diagnosed with diabetes, hyperglycemia, hypoglycemia, and information associated with diabetes that has progressed and / or regressed in the patient, as well as information associated with other diseases such as hyperglycemia, hypoglycemia, kidney disease, hypertension, heart conditions and diseases, or similar diseases that coexist in relation to diabetes. Such information may indicate, throughout the course of the illness, the patient's symptoms, the patient's physiological state, the patient's glucose level, the patient's potassium level, the patient's lactate level, the patient's insulin level, the condition / status of one or more of the patient's organs, the patient's habits (e.g., activity level, food consumption, etc.), prescribed medications, medication compliance, etc.

[0060] Although depicted as separate databases for clarity of concept, in some embodiments, the user database 110 and the history database 112 may operate as a single database. That is, the history and current data related to users of the continuous analyte monitoring system 104 and the application 106, as well as the history data related to patients who were not previous users of the continuous analyte monitoring system 104 and the application 106, may be stored in a single database. The single database may be a storage server operating in a public or private cloud.

[0061] As described above, the decision support system 100 is configured to diagnose, stage, treat, evaluate the risk of diabetes, and predict the likelihood of experiencing an atypical glucose trend and / or an atypical glucose trend associated with kidney disease for a user who uses the continuous analyte monitoring system 104 including at least a continuous glucose sensor. For example, the decision support engine 114: (1) automatically detects abnormal patterns associated with various glucose metrics; (2) evaluates the presence and severity of diabetes; (3) risk-stratifies patients to identify patients at high risk of diabetes; (4) identifies risks associated with the current diabetes diagnosis (e.g., risk of death, risk of hyperglycemia, risk of hypoglycemia, etc.); (5) makes patient-specific treatment decisions or recommendations for glucose homeostasis management and kidney disease; and (6) collects information associated with the user in the user profile 118 stored in the user database 110 to provide information regarding the effect of an intervention (e.g., the effect of a patient's lifestyle change, the effect of a surgical procedure, the effect of a patient taking a new medication, the effect of a patient receiving an insulin injection, etc.) and performs an analysis. Further, in certain embodiments, the glucose metrics of each user recorded over time may be analyzed to provide an indication of improvement or deterioration of the patient's diabetes. In certain embodiments, the user's glucose metrics may include glucose level, rate of change of glucose level, glucose trend, average glucose, glucose management indicator (GMI), glucose variability, time in range (TIR), glucose clearance rate, and the like.

[0062] In certain embodiments, the decision support engine 114 may be used to collect information associated with the user within the user profile 118 and perform an analysis on that information for determining the probability of the presence and / or severity of diabetes for the user and, based on the determination, providing one or more recommendations for treatment. For example, the decision support engine 114 may perform an analysis on the collected information associated with the user within the user profile 118 to determine one or more glucose metrics and generate a diabetes diagnosis prediction based on the determined one or more glucose metrics of the user. In certain embodiments, the decision support engine 114 may collect information associated with the user within the user profile 118 to determine the likelihood that the user experiences one or more atypical glucose trends associated with diabetes and kidney disease and may be used to perform an analysis for providing one or more recommendations for treatment based on the determination.

[0063] The user profile 118 may be accessible to the decision support engine 114 via one or more networks (not shown) for performing such analysis. In certain embodiments, the decision support engine 114 is configured to provide real-time and / or non-real-time decision support regarding diabetes to users and / or others including, but not limited to, healthcare providers (HCPs), the user's family, the user's caregivers, researchers, and / or other individuals, systems, and / or groups that assist in learning from care or data.

[0064] In certain embodiments, the decision support engine 114 may utilize one or more trained machine learning models to (1) predict the presence and / or severity of diabetes in users with a renal disease (e.g., CKD), (2) predict the presence / severity of diabetes in users with an unknown but suspected renal disease, (3) predict the presence / severity of a renal disease (e.g., CKD) in users with prediabetes or “mild” type II diabetes, (4) predict the presence / severity of diabetes in users with known mild CKD but an unknown diabetes status, (5) predict the presence / severity of diabetes and / or CKD when both are in an unknown state, and (2) provide decision support to users with diabetes and renal disease based on the information provided in the user's profile 118. In the illustrated embodiment of FIG. 1, the decision support engine 114 may utilize a trained machine learning model provided by the training server system 140. Although depicted as a separate server for clarity of concept, in some embodiments, the training server system 140 and the decision support engine 114 may operate as a single server. That is, the model may be trained and used by a single server, or it may be trained by one or more servers and deployed for use on one or more other servers. In certain embodiments, the model may be trained on one or more virtual machines (VMs) that are at least partially executed on one or more physical servers in a relational and / or non-relational database format.

[0065] The training server system 140 is configured to train a machine learning model using training data, which may include data associated with (e.g., from a user profile) one or more patients previously diagnosed as having (1) no diabetes and no kidney disease, (2) no diabetes and having kidney disease at various stages, (3) having diabetes at various stages and no kidney disease, or (4) having diabetes at various stages and having kidney disease at various stages (e.g., users or non-users of the continuous analyte monitoring system 104 and / or the application 106). The training data may be stored in the historical record database 112 and may be accessible to the training server system 140 via one or more networks (not shown) for training the machine learning model. The training data may also, in some cases, include user-specific data for a user over time.

[0066] Training data refers to, for example, a characterized and labeled data set. For example, the data set may include a plurality of data records, each containing information corresponding to a different user profile stored in the user database 110, and each data record is characterized and labeled. In machine learning and pattern recognition, a feature is an individual measurable property or characteristic. Generally, the features that best characterize the patterns in the data are selected for creating a predictive machine learning model. Data labeling is the process of adding one or more meaningful and useful labels to provide context to the data for learning by the machine learning model.

[0067] As an illustrative example, each relevant feature of a user reflected in corresponding data records can be a feature used when training a machine learning model. Such features include the user's age; gender; various glucose metrics including glucose level; change in glucose level from a first timestamp to a second timestamp (e.g., delta); glucose level over time (e.g., glucose levels from two or more subsequent timestamps); average glucose level, change in average glucose from a first series of timestamps (e.g., average glucose from a first timestamp to one or more subsequent timestamps) to a second series of timestamps (e.g., average glucose level from a second timestamp to one or more subsequent timestamps) (e.g., delta); average glucose level over time (e.g., daily, weekly, monthly, etc.) (e.g., average glucose from two or more subsequent timestamps); average glucose level over the first day compared to the average glucose level over the second day (e.g., average glucose level from morning, afternoon, or evening); average glucose level during an event-specific time range on the first day (e.g., morning, before bedtime, during sleep, after exercise, after dialysis) compared to the average glucose level during an event-specific time range on the second day; glucose management indicator (GMI); change in GMI from a first timestamp to a second timestamp (e.g., delta); GMI over time (e.g., GMI from two or more subsequent timestamps); glucose variability (e.g., standard deviation of average glucose); change in glucose variability from a first series of timestamps (e.g., delta); (e.g., glucose variability from a first timestamp to one or more subsequent timestamps) to a second series of timestamps (e.g., glucose variability from a second timestamp to one or more subsequent timestamps); glucose variability over time (e.g., glucose variability from two or more subsequent timestamps); time in range (TIR) (e.g., glucose level at a threshold, above a threshold, below a threshold, or between thresholds); change in TIR from a first timestamp to a second timestamp (e.g., delta); TIR over time (e.g., TIR from two or more subsequent timestamps); glucose clearance rate;Changes in glucose clearance rate from the first timestamp to one or more subsequent timestamps (e.g., delta) Glucose clearance rate over time (e.g., blood and / or kidney glucose clearance rate from two or more subsequent timestamps; glucose metabolism, insulin activity / insulin resistance, change in glucose level from the first series of timestamps (e.g., from the first timestamp to one or more subsequent timestamps) to the second series of timestamps (e.g., from the first timestamp to one or more subsequent timestamps); information associated with changes (e.g., delta) in glucose metrics (e.g., average glucose, glucose management indicator (GMI), glucose variability, time in range (TIR), glucose clearance rate, etc.) from the first series of timestamps to the second series of timestamps; glucose metrics over time (e.g., glucose metrics from two or more subsequent timestamps); change (e.g., delta) in kidney disease stage or severity from the first timestamp to the second timestamp; change (e.g., delta) in diabetes stage or severity from the first timestamp to the second timestamp; derivative of the determined linear system of glucose metrics at a specific timestamp or a specific series of timestamps, and / or difference in derivatives to determine the rate of change of the slope of increase or decrease of glucose metrics; derivative of the measured linear system of glucose level and / or difference in derivatives at one or more specific timestamps to determine the rate of change of the slope of increase or decrease of glucose level; derivative of the determined linear system of average glucose at a specific timestamp or a specific series of timestamps, and / or difference in derivatives to determine the rate of change of the slope of increase or decrease of average glucose; derivative of the determined linear system of GMI at a specific timestamp or a specific series of timestamps, and / or difference in derivatives to determine the rate of change of the slope of increase or decrease of GMI; derivative of the determined linear system of glucose variability at a specific timestamp or a specific series of timestamps, and / or difference in derivatives to determine the rate of change of the slope of increase or decrease of glucose variability;The derivative of the determined linear system of the TIR at a specific timestamp or a specific series of timestamps, and / or the difference of derivatives for determining the rate of change of the slope of the increase or decrease of the TIR; it may include the derivative of the determined linear system of the glucose clearance rate at a specific timestamp or a specific series of timestamps, and / or the difference of derivatives for determining the rate of change of the slope of the increase or decrease such as the glucose clearance rate.

[0068] Other characteristics include the user's potassium level; the change in potassium level from a first timestamp to a second timestamp (e.g., delta); the potassium level over time (e.g., potassium levels from two or more subsequent timestamps); the derivative and / or difference of the derivative of the measured linear system of potassium measurements at one or more specific timestamps to determine the rate of change of the slope of the increase or decrease in potassium level; the potassium level over time (e.g., potassium levels from two or more subsequent timestamps); the average potassium level, the change in average potassium (e.g., delta) from a first series of timestamps (e.g., average potassium from the first timestamp to one or more subsequent timestamps) to a second series of timestamps (e.g., average potassium level from the second timestamp to one or more subsequent timestamps); the average potassium level over time (e.g., average potassium from two or more subsequent timestamps); the change in the variation of potassium from a first series of timestamps (e.g., the variation of potassium from the first timestamp to one or more subsequent timestamps) to a second series of timestamps (e.g., the variation of potassium from the second timestamp to one or more subsequent timestamps); the variation of potassium over time (e.g., the variation of potassium from two or more subsequent timestamps); the time in range (TIR) (e.g., potassium levels at, above, below, or between the threshold); the change in TIR from a first timestamp to a second timestamp (e.g., delta); the TIR over time (e.g., TIR from two or more subsequent timestamps); the potassium clearance rate; the change in potassium clearance rate (e.g., delta) from the first timestamp to one or more subsequent timestamps; the potassium clearance rate over time (e.g., blood and / or renal potassium clearance rate from two or more subsequent timestamps;Change (e.g., delta) in a potassium metric (e.g., average potassium, potassium variability, time in range (TIR), potassium clearance rate, etc.) from a first series of timestamps (e.g., change in potassium level from a first timestamp to one or more subsequent timestamps) to a second series of timestamps (e.g., change in potassium level from a first timestamp to one or more subsequent timestamps); may include potassium metrics over time (e.g., potassium metrics from two or more subsequent timestamps).;

[0069] Other characteristics include the user's lactate level; the change in lactate level from a first timestamp to a second timestamp (e.g., delta); the lactate level over time (e.g., the lactate level from two or more subsequent timestamps); the measured derivative and / or difference in derivatives of a linear system of lactate levels at one or more specific timestamps for determining the rate of change of the slope of an increase or decrease in lactate level; the lactate level over time (e.g., the lactate level from two or more subsequent timestamps); the average lactate level, the change in average lactate (e.g., delta) from a first series of timestamps (e.g., average lactate from the first timestamp to one or more subsequent timestamps) to a second series of timestamps (e.g., average lactate level from the second timestamp to one or more subsequent timestamps); the average lactate level over time (e.g., the average lactate from two or more subsequent timestamps); the variability of lactate (e.g., the standard deviation of average lactate); the change in variability of lactate from a first series of timestamps (e.g., the variability of lactate from the first timestamp to one or more subsequent timestamps); from a first series of timestamps (e.g., the variability of lactate from the first timestamp to one or more subsequent timestamps) to a second series of timestamps (e.g., the variability of lactate from the second timestamp to one or more subsequent timestamps); the variability of lactate over time (e.g., the variability of lactate from two or more subsequent timestamps); time in range (TIR) (e.g., lactate level at a threshold, above a threshold, below a threshold, or between thresholds); the change in TIR from a first timestamp to a second timestamp (e.g., delta); the TIR over time (e.g., the TIR from two or more subsequent timestamps); the lactate clearance rate; the change in lactate clearance rate from the first timestamp to one or more subsequent timestamps (e.g., delta); the lactate clearance rate over time (e.g., the blood and / or renal lactate clearance rate from two or more subsequent timestamps);Changes (e.g., deltas) in lactate metrics (e.g., average lactate, lactate management index (GMI), lactate variability, time in range (TIR), lactate clearance rate, etc.) from a first series of timestamps (e.g., changes in lactate level from a first timestamp to one or more subsequent timestamps) to a second series of timestamps (e.g., changes in lactate level from a first timestamp to one or more subsequent timestamps); and may include lactate metrics over time (e.g., lactate metrics from two or more subsequent timestamps).;

[0070] Other features may include the user's ketone level; changes in ketone level (e.g., deltas) from a first timestamp to a second timestamp; ketone levels over time (e.g., ketone levels from two or more subsequent timestamps); the derivative and / or difference in derivatives of a measured linear system of ketone levels at one or more specific timestamps to determine the rate of change of the slope of an increase or decrease in ketone level; pyruvate level; changes in pyruvate level (e.g., deltas) from a first timestamp to a second timestamp; pyruvate levels over time (e.g., pyruvate levels from two or more subsequent timestamps); the derivative and / or difference in derivatives of a measured linear system of pyruvate levels at one or more specific timestamps to determine the rate of change of the slope of an increase or decrease in pyruvate level; insulin level; changes in insulin level (e.g., deltas) from a first timestamp to a second timestamp; insulin levels over time (e.g., insulin levels from two or more subsequent timestamps); and the derivative and / or difference in derivatives of a measured linear system of insulin levels at one or more specific timestamps to determine the rate of change of the slope of an increase or decrease in insulin level.

[0071] In addition, other features may include non-analyte data; changes in non-analyte data from a first timestamp to a second timestamp (e.g., delta); non-analyte data over time (e.g., non-analyte data from two or more subsequent timestamps); the measured derivative of a linear system of non-analyte data and / or the difference in derivatives at one or more specific timestamps to determine the rate of change of the slope of an increase or decrease in non-analyte data. In addition, the data record is labeled with indicators related to diabetes diagnosis, assigned disease severity and / or type, identified risk of diabetes, kidney disease diagnosis, assigned disease severity and / or type, identified risk of other diseases, type, dosage and associated timestamps for started, stopped and / or administered medications and / or therapeutics, type, dosage and associated timestamps for started, stopped and / or administered insulin, consumed food, sleep, stress, heart rate, blood pressure, start and / or stop of exercise, etc., associated with the patient of the user profile.

[0072] The model is then characterized and trained by the training server system 140 using the characterized and labeled training data. In particular, the features of each data record can be used as input to the machine learning model, and the generated output can be compared to the label associated with the corresponding data record. The model can calculate a loss based on the difference between the generated output and the provided label. This loss is then used to modify the internal parameters or weights of the model. By repeatedly processing each data record corresponding to each historical patient, the model can be iteratively refined in certain embodiments to generate accurate predictions associated with diabetes risk, presence, progression, improvement, and severity in a patient. Further, in certain other embodiments, by repeatedly processing each data record corresponding to each historical patient, in certain embodiments, the model can be iteratively refined to generate accurate predictions of the risk and / or presence of one or more symptoms associated with a decrease in glucose homeostasis.

[0073] As illustrated in FIG. 1, the training server system 140 deploys these trained models to the decision support engine 114 for use during runtime. For example, the decision support engine 114 may obtain a user profile 118 associated with a user, use the information within the user profile 118 as input to a trained model, and output a prediction. The prediction may indicate the presence and / or severity of diabetes for the user, or may indicate the presence or risk of a user experiencing at least one non-typical glucose trend in real-time or within a certain time period (e.g., as shown as output 144 in FIG. 1). The output 144 generated by the decision support engine 114 may also provide one or more decision support recommendations regarding treatment based on the prediction. The output 144 may be provided to the user (e.g., through application 106), to the user's caregiver (e.g., parent, relative, guardian, teacher, nurse, etc.), to the user's physician, or in some cases, to any other individual interested in the user's health for the purpose of improving the user's health, such as by achieving the recommended treatment.

[0074] In certain embodiments, the user's own data is used to personalize one or more models that were initially trained based on population data. For example, a model (e.g., trained using population data) may be deployed for use by the decision support engine 114 to predict the presence or risk of a particular user experiencing an irregular glucose trend in real-time. After making a prediction using the model, the decision support engine 114 may be configured to obtain the user's actual glucose trend and calculate a loss between the prediction and the actual glucose trend, which loss may be used to retrain the model. Thus, the calculated loss between the prediction and the actual glucose trend can be used as an input to the model to continuously retrain and personalize the model for the user. In another example, a model (e.g., trained using population data) may be deployed for use by the decision support engine 114 to predict in real-time the effect of a treatment (e.g., medication administration, food consumption, etc.) on a particular user experiencing an irregular glucose trend. After making a prediction using the model, the decision support engine 114 may be configured to obtain the user's actual glucose trend and calculate a loss between the prediction and the actual glucose trend, which loss may be used to retrain the model. Thus, the calculated loss between the prediction and the actual glucose trend can be used as an input to the model to continuously retrain and personalize the model for the user.

[0075] In certain embodiments, the output 144 generated by the decision support engine 114 may be stored in the user profile 118. In certain embodiments, the output 144 may be a prediction regarding the presence, type, and / or severity of diabetes in the user. In certain embodiments, the output 144 may be a patient-specific treatment decision or recommendation for managing diabetes. In certain embodiments, the output 144 may be a prediction regarding the risk of users having an atypical glucose trend. In certain embodiments, the output 144 may be a harmful blood glucose event prediction. In certain embodiments, the output 144 may be a patient-specific treatment decision or recommendation for glycemic control. The output 144 stored in the user profile 118 may be continuously updated by the decision support engine 114. Thus, the previous diagnoses and / or the user's glucose metrics initially stored as the output 144 within the user profile 118 within the user database 110 and then passed to the history database 112 provide an indication of the progression of diabetes in the user over time, as well as an indication regarding the effectiveness of different treatments (e.g., medications) recommended to the user to maintain glycemic control.

[0076] In certain embodiments, the user's own historical data may be used to provide decision support and insights regarding the user's glycemic control and / or disease. For example, the user's historical data may be used as a baseline by an algorithm to indicate an improvement or deterioration in the user's glycemic control. As an illustrative example, the user's data from two weeks ago can be used as a baseline that can be compared to the user's current data to identify whether the user's glycemic control has improved or deteriorated. In certain embodiments, the user's own historical data can be used by the training server system 140 to train a personalized model that may further be capable of predicting or planning the user's glucose level or the user's glycemic control based on the user's recent data patterns (e.g., glucose metrics, exercise data, food consumption data, medication data, etc.).

[0077] In certain embodiments, the model is trained to provide lifestyle recommendations, exercise recommendations, diet recommendations, medication recommendations, medical treatment recommendations, and other types of decision support recommendations to help a user manage glucose homeostasis based on the user's historical data, including how different types of medications, foods, and treatments have affected the user's glucose homeostasis in the past. In certain embodiments, the model can be trained to detect the root cause of a particular improvement or deterioration in a patient's glucose homeostasis. For example, the application 106 can display a user interface having a graph showing the patient's glucose metric or score along with a trend line, which can show, for example retrospectively, the cause of a glucose homeostasis impairment at a particular point in time (e.g., excessive glucose intake, insulin dosage, decreased kidney function, etc.).

[0078] FIG. 2 is a diagram 200 conceptually illustrating an exemplary continuous analyte monitoring system 104 including an exemplary continuous analyte sensor having sensor electronics, according to certain aspects of the present disclosure. For example, the system 104 can be configured to continuously monitor one or more analytes of a user according to certain aspects of the present disclosure.

[0079] In the illustrated embodiment, the continuous analyte monitoring system 104 includes a sensor electronics module 204 and one or more continuous analyte sensors 202 associated with the sensor electronics module 204 (referred to herein individually as continuous analyte sensor 202 and collectively as continuous analyte sensors 202). The sensor electronics module 204 may wirelessly communicate (e.g., directly or indirectly) with one or more display devices 210, 220, 230, and 240. In certain embodiments, the sensor electronics module 204 may also wirelessly communicate (e.g., directly or indirectly) with one or more medical devices, such as medical device 208 (referred to herein individually as medical device 208 and collectively as medical devices 208), and / or one or more other non-analyte sensors 206 (referred to herein individually as non-analyte sensor 206 and collectively as non-analyte sensors 206).

[0080] In certain embodiments, the continuous analyte sensor 202 may include a sensor for detecting and / or measuring an analyte. The continuous analyte sensor 202 may be a multi-analyte sensor configured to continuously measure two or more analytes, or a single-analyte sensor configured to continuously measure a single analyte as a non-invasive device, a subcutaneous device, a transdermal device, a transdermal device, and / or an intravascular device. In certain embodiments, the continuous analyte sensor 202 may be configured to continuously measure a user's analyte level using one or more measurement techniques, such as enzymatic, chemical, physical, electrochemical, spectrophotometric, polarimetric, calorimetric, electrophoretic, radiometric, immunochemical. In certain aspects, the continuous analyte sensor 202 provides a data stream indicative of the concentration of one or more analytes within the user. The data stream may then include a raw data signal that is converted to a calibrated and / or filtered data stream used to provide an estimated analyte value to the user.

[0081] In certain embodiments, the continuous analyte sensor 202 may be a multi-analyte sensor configured to continuously measure multiple analytes within a user's body. For example, in certain embodiments, the continuous multi-analyte sensor 202 may be a single multi-analyte sensor configured to measure potassium, glucose, lactate, ketones, pyruvate, and insulin within a user's body.

[0082] In certain embodiments, one or more multi-analyte sensors may be used in combination with one or more single analyte sensors. By way of illustrative example, the multi-analyte sensor may be configured to continuously measure potassium and glucose and, in some cases, may be used in combination with an analyte sensor configured to measure only lactate levels. Information from each of the multi-analyte sensor and the single analyte sensor can be combined to provide diabetes decision support using the methods described herein.

[0083] In certain embodiments, the sensor electronics module 204 includes electronic circuitry associated with the measurement and processing of continuous analyte sensor data, including predictive algorithms associated with the processing and calibration of sensor data. The sensor electronics module 204 can be physically connected to the continuous analyte sensor 202 and can be integrated with (non-removably attached to) or removably attached to the continuous analyte sensor 202. The sensor electronics module 204 can include hardware, firmware, and / or software that enables the measurement at the display device level via the continuous analyte sensor 202. For example, the sensor electronics module 204 can include a potentiostat, a power source for powering the sensor, other components useful for signal processing and data storage, and a telemetry module for transmitting data from the sensor electronics module to one or more display devices. The electronics can be fixed, for example, to a printed circuit board (PCB) and can take various forms. For example, the electronics can take the form of an integrated circuit (IC) such as an Application-Specific Integrated Circuit (ASIC), a microcontroller, and / or a processor.

[0084] The display devices 210, 220, 230, and / or 240 are configured to display displayable sensor data, including analyte data, that may be transmitted by the sensor electronics module 204. Each of the display devices 210, 220, 230, or 240 can include a display such as a touch screen display 212, 222, 232, / or 242 for displaying sensor data to the user and / or receiving input from the user. For example, a graphical user interface (GUI) can be presented to the user for such purposes. In some embodiments, the display device can include another type of user interface, such as an audio user interface, instead of or in addition to the touch screen display, for communicating sensor data to the user of the display device and / or receiving user input. The display devices 210, 220, 230, and 240 can be examples of the display device 107 illustrated in FIG. 1 that are used to display sensor data to the user of FIG. 1 and / or receive input from the user.

[0085] In some embodiments, one, some, or all of the display devices are configured to display or otherwise communicate the sensor data when the sensor data is communicated from the sensor electronics module (e.g., in the data packages transmitted to each respective display device) without any additional expected processing required for calibration and real-time display of the sensor data.

[0086] The plurality of display devices can include custom display devices specially designed to display certain types of displayable sensor data associated with the analyte data received from the sensor electronics module. In certain embodiments, the plurality of display devices can be configured to provide warnings / alarms based on the displayable sensor data. Display device 210 is an example of such a custom device. In some embodiments, one of the plurality of display devices is a smartphone, such as display device 220 representing a mobile phone, configured to use a commercially available operating system (OS) to display graphical representations of continuous sensor data (e.g., including current and past data). Other display devices can include a display device 230 representing a tablet, a display device 240 representing a smartwatch, a medical device 208 (e.g., an insulin delivery device or a blood glucose meter), and / or other handheld devices such as a desktop or laptop computer (not shown).

[0087] Since the user interface varies depending on the display device, the content of the data package (e.g., the amount, format, and / or type of data to be displayed, alarms, etc.) can be customized for each display device (e.g., programmed to vary depending on the manufacturer and / or end user). Thus, in certain embodiments, multiple different display devices can communicate directly wirelessly with a sensor electronics module (e.g., an on-skin sensor electronics module 204 physically connected to a continuous analyte sensor 202, etc.) during a sensor session to enable multiple different types and / or levels of display and / or functionality associated with displayable sensor data. In certain embodiments, the type of alarm customized for each particular display device, the number of alarms customized for each particular display device, the timing of the alarms customized for each particular display device, and / or the threshold levels configured (e.g., for activation) for each of the alarms are based on an output 144 stored in the user profile 118 of each user (e.g., as mentioned, the output 144 can indicate the user's current health, the user's glucose status, and / or the current treatment recommended for the user). In certain embodiments, one or more of the display devices may have a user interface that can include various interfaces such as a liquid crystal display (LCD) for presenting UI features, a vibrator, an audio transducer (e.g., a speaker), a backlight (not shown), etc. Components with such a user interface can provide control for interconnecting with a user (e.g., a host). One or more UI features can enable, for example, toggles, menu selections, option selections, status selections, yes / no responses to on-screen questions, an "off" function (e.g., for an alarm), an "affirmative response" function (e.g., for an alarm), a reset, etc. The UI features can also provide the user with, for example, a visual data output.An audio transducer (e.g., a speaker) may provide an audible signal in response to the activation of certain alerts, such as the current and / or predicted state. In some exemplary implementations, the audible signal may be distinguished by a tone, volume, duty cycle, pattern, duration, etc. In some exemplary implementations, the audible signal may be configured to be muted (e.g., positively responded to or turned off) by pressing one or more buttons.

[0088] As mentioned, the sensor electronics module 204 may communicate with the medical device 208. The medical device 208 may be a passive device in some exemplary embodiments of the present disclosure. For example, the medical device 208 may be an insulin pump for administering insulin to a user. For various reasons, it may be desirable for such an insulin pump to receive and track glucose, potassium, lactate, insulin, ketone, and / or pyruvate values transmitted from the continuous analyte monitoring system 104, and the continuous analyte sensor 202 may be configured to measure glucose, potassium, lactate, insulin, ketone, and / or pyruvate.

[0089] Furthermore, as mentioned, the sensor electronics module 204 may also communicate with other non-analyte sensors 206. The non-analyte sensors 206 may include, but are not limited to, an altimeter sensor, an accelerometer sensor, a temperature sensor, a respiratory rate sensor, a sweat sensor, etc. The non-analyte sensors 206 may also include monitors such as a heart rate monitor, an ECG monitor, a blood pressure monitor, a pulse oximeter, calorie intake, and a drug delivery device. One or more of these non-analyte sensors 206 may provide data to the decision-making support engine 114, which is further described below. In some aspects, the user may be able to manually provide a portion of the data for processing by the training server system 140 and / or the decision-making support engine 114 of FIG. 1.

[0090] In certain embodiments, the non-analyte sensor 206 can be combined in any other configuration, such as, for example, in combination with one or more continuous analyte sensors 202. As an illustrative example, a non-analyte sensor, such as a heart rate sensor, can be combined with a continuous analyte sensor 202 configured to measure glucose to form a glucose / heart rate sensor used to transmit sensor data to the sensor electronics module 204 using a common communication circuit. As another illustrative example, a non-analyte sensor, such as a heart rate sensor and / or an ECG sensor, can be combined with a multi-analyte sensor 202 configured to measure glucose and potassium to form a glucose / potassium / heart rate sensor used to transmit sensor data to the sensor electronics module 204 using a common communication circuit.

[0091] In certain embodiments, a wireless access point (WAP) can be used to couple one or more of the continuous analyte monitoring system 104, the plurality of display devices, the medical device 208, and / or the non-analyte sensor 206 to each other. For example, the WAP 138 can provide Wi-Fi and / or cellular connectivity among these devices. Near Field Communication (NFC) and / or Bluetooth can also be used among the devices depicted in the diagram 200 of FIG. 2.

[0092] FIG. 3 illustrates exemplary inputs for use with the decision support system of FIG. 1 and exemplary metrics calculated based on the inputs, according to some embodiments disclosed herein. In particular, FIG. 3 provides a more detailed illustration of the exemplary inputs and exemplary metrics introduced in FIG. 1.

[0093] FIG. 3 illustrates an exemplary input 128 on the left, a decision support engine 114 including an application 106 and a DAM 116 in the center, and metrics 130 on the right. In certain embodiments, each of the metrics 130 may correspond to one or more values, such as discrete numerical values, ranges, or qualitative values (such as high / medium / low, or stable / unstable). The application 106 obtains the input 128 via one or more channels (e.g., manual user input, sensors / monitors, other applications running on the display device 107, EMR, etc.). As described above, in certain embodiments, the input 128 may be processed by the DAM 116 and / or the decision support engine 114 to output the metrics 130. The input and the metrics 130 may be used by the decision support engine 114 to provide decision support to the user. For example, the input 128 and the metrics 130 may be used by a training server system 140 to train and deploy one or more machine learning models used by the decision support engine 114 to provide decision support regarding diabetes to a patient with kidney disease.

[0094] In certain embodiments, starting from the input 128, the food consumption information may include information about one or more of meals, snacks, and / or beverages, such as one or more of quantity, content (milligrams (mg) of potassium, glucose, lactate, carbohydrates, fat, protein, etc.), order of consumption, and time of consumption. In certain embodiments, food consumption may be provided by the user through manual input by providing a photograph through an application configured to recognize the type and quantity of food (e.g., potassium and glucose / carbohydrate content of the food), and / or by scanning a barcode or menu. In various examples, the meal size may be manually entered as one or more of calories, quantity (e.g., "3 cookies"), menu item (e.g., "Royale with Cheese"), and / or food exchange (1 fruit, 1 dairy product). In some examples, the meal information may be received via a convenient user interface provided by the application 106.

[0095] In certain embodiments, the food consumption information entered by the user may be related to the glucose consumed by the user. Glucose for consumption may include any natural or artificial food or beverage containing glucose, dextrose, or carbohydrates (e.g., glucose tablets, bananas, or bread, etc.).

[0096] In certain embodiments, the motion information is also provided as an input. The motion information can be any information surrounding an activity that requires physical exertion by the user, such as an activity. For example, the motion information can include information associated with low-intensity (e.g., taking a few steps of walking) and high-intensity (e.g., running 5 miles) physical exertion. In certain embodiments, the motion information can be provided by an accelerometer sensor on a wearable device such as, for example, a watch, a fitness tracker, and / or a patch. In certain embodiments, the motion information can also be provided through manual user input and / or through a surrogate sensor and a prediction algorithm that measures changes in heart rate (or other cardiac metrics). When the user predicts that they are exercising based on their sensor data, the user can be asked to confirm whether exercise is being performed, what type of exercise it is, and / or the level of intense exercise being used during the exercise over a specific period. This data can be used to train the system 100 to learn about the user's exercise patterns as time progresses and to reduce the need for confirmation questions.

[0097] In certain embodiments, user statistics such as one or more of age, height, weight, BMI, body composition (e.g., body fat percentage), height, build, or other information can also be provided as an input. In certain embodiments, the user statistics can be provided through a user interface, by interfacing with an electronic information source such as an electronic medical record, and / or from a measurement device. In certain embodiments, the measurement device can include one or more of a wireless, e.g., Bluetooth-enabled, scale, and / or a camera that can communicate, for example, with the display device 107 to provide user data.

[0098] In certain embodiments, treatment / medical information is also provided as an input. The medical information may include information regarding the type, dosage, and / or timing when one or more medications are taken by the user. As referred to herein, the medical information may include information regarding one or more antihyperglycemic medications, one or more drugs known to damage the kidneys, one or more drugs known to control complications of renal diseases prescribed to the user, and / or one or more medications for treating one or more symptoms of renal diseases, hyperkalemia, hypokalemia, diabetes, and / or other conditions and diseases the user may have. The treatment information may include information regarding different lifestyle habits, surgical procedures, and / or other non-invasive procedures recommended by the user's physician. For example, the user's physician may recommend to the user to increase / decrease their glucose intake, exercise for at least 30 minutes a day, and / or increase insulin dosage or other medications to maintain and / or improve kidney health, glucose homeostasis, general health, etc. In certain embodiments, the treatment / medical information may be provided through manual user input.

[0099] In certain embodiments, analyte sensor data can also be provided as input, for example, through a continuous analyte monitoring system 104. In certain embodiments, the analyte sensor data can include glucose data (e.g., a user's blood glucose value) measured by at least one continuous glucose sensor (or a multi-analyte sensor configured to measure at least glucose) that is part of the continuous analyte monitoring system 104. In certain embodiments, the analyte sensor data can include potassium data measured by at least one potassium sensor (or a multi-analyte sensor configured to measure at least potassium) that is part of the continuous analyte monitoring system 104. In certain embodiments, the analyte sensor data can include lactate data measured by at least one lactate sensor (or a multi-analyte sensor configured to measure at least lactate) that is part of the continuous analyte monitoring system 104. In certain embodiments, the analyte sensor data can include insulin data measured by at least one insulin sensor (or a multi-analyte sensor configured to measure at least insulin) that is part of the continuous analyte monitoring system 104. In certain embodiments, the analyte sensor data can include pyruvate data measured by at least one pyruvate sensor (or a multi-analyte sensor configured to measure at least pyruvate) that is part of the continuous analyte monitoring system 104. In certain embodiments, the analyte sensor data can include ketone data measured by at least one ketone sensor (or a multi-analyte sensor configured to measure at least ketones) that is part of the continuous analyte monitoring system 104.

[0100] In certain embodiments, the input may also be received from one or more non-analyte sensors, such as non-analyte sensor 206 described with respect to FIG. 2. Inputs from such non-analyte sensors 206 may include information associated with the user's heart rate, heart rate variability (e.g., differences in time between heart beats), ECG data, respiratory rate, oxygen saturation, blood pressure, or body temperature (e.g., for detecting illness, physical activity, etc.). In certain embodiments, the electromagnetic sensor may also detect a low-power radio frequency (RF) field emitted from an object or a tool in contact with or in proximity to the object that may provide information about the user's activity or location.

[0101] In certain embodiments, the input received from the non-analyte sensor may include input related to the user's insulin delivery. In particular, input related to the user's insulin delivery may be received via a wireless connection on a smart insulin pen, via user input, and / or from an insulin pump. The insulin delivery information may include one or more of insulin volume, delivery time, etc. Other parameters such as insulin action time, insulin activity rate, or duration of insulin action may also be received as input.

[0102] In certain embodiments, the time may also be provided as an input, such as the time of day or the time from a real-time clock. For example, in certain embodiments, the input analyte data may be timestamped to indicate the date and time when the analyte measurement was taken for the user.

[0103] Any user input of the inputs 128 referred to above may be through a user interface, such as the user interface of display device 107 in FIG. 1.

[0104] As described above, in certain embodiments, the DAM 116 and / or the decision support engine (e.g., using one or more trained models) determine or calculate the user's metric 130 based on the input 128. An exemplary list of metrics 130 is shown in FIG. 3.

[0105] In certain embodiments, the glucose level can be determined from sensor data (e.g., the glucose level obtained from a continuous glucose sensor of the continuous analyte monitoring system 104). For example, the glucose level refers to a timestamped glucose level or value that is continuously generated and stored over time. In certain embodiments, the glucose metric refers to one or more timestamped glucose levels or values. The glucose metric may refer to, include, or indicate a change between one or more timestamped glucose levels or values. For example, the glucose metric may include or indicate the difference between the glucose level at a first timestamp and the glucose level at a subsequent timestamp. The glucose metric may further include or indicate a rate of change of glucose, which refers to the rate of change of one or more timestamped glucose levels or values with respect to one or more other timestamped glucose levels or values. For example, the glucose metric may include or indicate the derivative of the change between the glucose level at a first timestamp and the glucose levels at one or more subsequent timestamps. The glucose metric may be the average glucose level at a particular time of day or after an event such as a meal or physical activity. The particular time of day may include a time range over which multiple analyte measurements are taken over one or more periods (e.g., time ranges). By comparing the average glucose level at a particular time of day to subsequent days, the change in glucose level over a particular period of time can be determined, which can demonstrate the progression of a disease. Additionally, the glucose metric may include the frequency of deviation from an expected average glucose threshold during the day.

[0106] In certain embodiments, the glucose metric may include glucose level, glucose rate of change, average glucose, GMI, glucose variability, TIR, and / or glucose clearance rate. Further, in certain embodiments, the glucose metric may include changes in one or more glucose metrics, such as absolute change, rate of change, etc. In certain embodiments, DAM116 can continuously calculate the glucose metric, timestamp the calculated glucose metric, and store the corresponding information in the user's profile 118.

[0107] In certain other embodiments, DAM116 can exclude glucose levels measured over a period of time if there are external conditions that affect the user's glucose homeostasis (e.g., the user is engaged in exercise, consuming glucose, or in other similar situations for at least part of the period). In such embodiments, in some examples, DAM116 can first identify which measured glucose values should not be used to calculate the glucose metric by identifying glucose values affected by external events such as food consumption, exercise, medications, or other perturbations that interfere with the capture of the glucose metric. Then, DAM116 can exclude such measurements when calculating the user's glucose metric.

[0108] In certain embodiments, the absolute maximum glucose level may be determined from sensor data (e.g., glucose levels obtained from a continuous glucose sensor of the continuous analyte monitoring system 104), health / illness metrics (e.g., described in more detail below), and / or disease stage metrics (e.g., described in more detail below). The absolute maximum glucose level represents the maximum glucose level of a user determined to be unsafe over a period of time (e.g., hourly, weekly, daily, etc.). In certain embodiments, the absolute maximum glucose level may be consistent across all users (e.g., set to 200 mg / dL based on current medical guidelines). In certain other embodiments, each patient may have a different absolute maximum glucose level. For example, the absolute maximum glucose level may be lower for a user diagnosed with stage 1 CKD (e.g., normal or high GFR (GFR > 90 mL / min)) than for a user diagnosed with stage 5 end-stage CKD (e.g., GFR < 15 mL / min). In certain embodiments, the absolute maximum glucose level per patient may change over time. For example, a user may initially be assigned an absolute maximum glucose level based on clinical input. This assigned absolute maximum glucose level may be adjusted over time based on other sensor data, disease stage, co-morbidities, etc. for the user.

[0109] For example, a user's absolute maximum glucose level may change over time as the user's renal function, kidney disease, and / or one or more other diseases progress and / or improve. Certain embodiments may determine a first absolute maximum glucose level for a period when there are no external conditions affecting the glucose level, and a second absolute maximum glucose level for a period when external conditions affecting the glucose level are present (e.g., periods when the user is consuming glucose, exercising, taking medications that affect the glucose level, etc.).

[0110] In certain embodiments, the absolute minimum glucose level may be determined from sensor data (e.g., glucose levels obtained from the continuous glucose sensor of the continuous analyte monitoring system 104), health / illness metrics (e.g., described in more detail below), and / or disease stage metrics (e.g., described in more detail below). The absolute cellular glucose level represents the minimum glucose level of a user determined to be unsafe over a period of time (e.g., hourly, weekly, daily, etc.). In certain embodiments, the absolute minimum glucose level may be consistent across all users (e.g., set based on current medical guidelines). In certain other embodiments, each user may have a different absolute minimum glucose level. For example, the absolute minimum glucose level may be lower for a user diagnosed with stage 1 CKD (e.g., normal or high GFR (GFR > 90 mL / min)) than for a user diagnosed with stage 5 end-stage CKD (e.g., GFR < 15 mL / min). In certain embodiments, the absolute minimum glucose level per patient may change over time. For example, a user may initially be assigned an absolute minimum glucose level based on clinical input. This assigned absolute minimum glucose level may be adjusted over time based on other sensor data, disease stage, comorbidities, etc. for the user.

[0111] For example, a user's absolute minimum glucose level may change over time as the user's renal function, kidney disease, and / or one or more other diseases progress and / or improve. Certain embodiments may determine a first absolute minimum glucose level for a period in which there are no external conditions affecting the glucose level, and a second absolute minimum glucose level for a period in which external conditions affecting the glucose level are present (e.g., a period in which the user is consuming glucose, exercising, taking a medication that affects the glucose level, etc.). In certain embodiments, the absolute minimum glucose level may be determined for a period in which the user is sleeping.

[0112] In certain embodiments, the glucose level change rate can be determined from sensor data (e.g., blood glucose measurements obtained from the continuous glucose sensor of the continuous analyte monitoring system 104). For example, the glucose level change rate refers to the rate at which one or more timestamped glucose levels or values change relative to one or more other timestamped glucose levels or values. The change rate of the glucose level can be determined over a period of one second or more, several minutes, several hours, several days, etc.

[0113] In certain embodiments, the determined glucose level change rate can be marked as "rapidly increasing" or "rapidly decreasing". As used herein, "rapidly" describes a glucose level change rate that is clinically significant and indicative of a patient's glucose level trend that is likely to breach the absolute maximum glucose level or the absolute minimum glucose level within a defined subsequent period. In other words, the predicted trend (e.g., generated by the decision support engine 114 using one or more trained models) can, in some cases, indicate, based on the determined glucose level change rate, that the patient is likely to reach the absolute maximum glucose level within a specified period (e.g., one hour or two hours). Thus, such a glucose level change rate can be marked as "rapidly increasing". Similarly, the predicted trend (e.g., generated by the decision support engine 114 using one or more trained models) can, in some cases, indicate, based on the determined glucose level change rate, that the patient is likely to reach the absolute minimum glucose level within a specified period (e.g., one hour or two hours). Thus, such a glucose level change rate can be marked as "rapidly decreasing".

[0114] In certain embodiments, the glucose metric change rate may be determined from glucose metrics determined for a user over time. For example, the glucose metric change rate refers to a rate indicating how one or more timestamped glucose metrics for a user change with respect to one or more other timestamped glucose metrics for the same user. The glucose metric change rate can be determined over a period of one second or more, minutes, hours, days, etc.

[0115] In certain embodiments, the glucose clearance rate can be determined from sensor data (e.g., glucose levels obtained from a continuous glucose sensor of the continuous analyte monitoring system 104) following the consumption of a known or estimated amount of glucose. The glucose clearance rate analyzed over time can indicate glucose homeostasis. In particular, the slope of the glucose clearance curve during a first period (e.g., after consuming a known amount of glucose) can indicate the ability of the kidneys, more specifically, to maintain glucose homeostasis, as compared to the slope of the glucose clearance curve during a second period (e.g., after consuming the same amount of glucose) (e.g., the glucose clearance rate can be slower when the user's kidneys are impaired than when the user's kidneys are healthy).

[0116] In certain embodiments, the glucose clearance rate can be determined by calculating the gradient between an initial high glucose value (e.g., the highest glucose level during a 20 - 30 minute period after glucose consumption) and a subsequent low glucose value. The low glucose value (GL) can be determined based on the user's initial high glucose value (GH) before glucose consumption and the baseline glucose value (GB). In certain embodiments, GL is a glucose value between GH and GB, e.g., GL = GB + K *It can be (GH-GB) / 2, where K can be a percentage representing how much the user's glucose level has returned to the user's baseline value. When K is equal to 0, the low glucose value is equal to the baseline glucose value. When K is equal to 0.5, the low glucose value is equal to the average glucose value between the initial glucose value and the baseline glucose value. In certain embodiments, the glucose clearance rate can be determined over one or more periods after glucose consumption. The glucose clearance rate can be calculated for each period to represent the dynamics of the glucose clearance rate after glucose consumption. These glucose clearance rates calculated over time can be timestamped and stored in the user's profile 118. Specific metrics can be derived from the timestamped glucose clearance rates, such as average, median, standard deviation, percentile, etc. In certain embodiments, a user with CKD may have impaired kidney function to metabolize insulin, and the time elapsed from the initial high glucose value to the low glucose value can indicate the ability of the kidneys to function. The time elapsed from the initial high glucose value to the low glucose value and the glucose clearance rate can be timestamped and stored in the user's profile 118.

[0117] In certain embodiments, the glucose trend can be determined based on glucose levels over a specific period. In certain embodiments, the glucose trend can be determined based on glucose metrics over a specific period. In certain embodiments, the glucose trend can be determined based on the absolute glucose level minimum over a specific period. In certain embodiments, the glucose trend can be determined based on the absolute maximum glucose level over a specific period. In certain embodiments, the glucose trend can be determined based on the rate of change of glucose levels over a specific period. In certain embodiments, the glucose trend can be determined based on the rate of change of glucose metrics (e.g., glucose level, average glucose, GMI, glucose variability, TIR, glucose clearance, etc.) over a specific period. In certain embodiments, the glucose trend can be determined based on the calculated glucose clearance rate over a specific period.

[0118] In certain embodiments, insulin sensitivity can be determined using historical data, real-time data, or a combination thereof, based on, for example, one or more of the following inputs 128 such as food consumption information, continuous analyte sensor data, non-analyte sensor data (e.g., insulin delivery information from an insulin device), etc. Insulin sensitivity refers to how a user's cells respond to insulin. Improving insulin sensitivity for a user can help reduce insulin resistance in the user.

[0119] In certain embodiments, the insulin on board can be determined using a known or learned (e.g., from user data) insulin time-action profile that can take into account both non-analyte sensor data inputs (e.g., insulin delivery information) and / or basal metabolic rate (e.g., insulin uptake to maintain body movement) and insulin use driven by activity or food consumption.

[0120] In certain embodiments, the potassium level may be determined from sensor data (e.g., potassium measurements obtained from the continuous analyte monitoring system 104).

[0121] In certain embodiments, the absolute maximum potassium level may be determined from sensor data (e.g., potassium measurements obtained from the continuous potassium sensor of the continuous analyte monitoring system 104), health / illness metrics (e.g., described in more detail below), and / or disease stage metrics (e.g., described in more detail below). The absolute maximum potassium level represents the maximum potassium level of a user determined to be unsafe over a period of time (e.g., hourly, weekly, daily, etc.). Each user may have a different absolute maximum potassium level. A user's absolute maximum potassium level may change over time as the user's kidney function, kidney disease, and / or one or more other diseases progress and / or improve.

[0122] In certain embodiments, the rate of change of potassium level may be determined from sensor data (e.g., potassium measurements obtained from the potassium sensor of the continuous analyte monitoring system 104). For example, the rate of change of potassium level refers to the rate at which one or more timestamped potassium measurements or values change with respect to one or more other timestamped potassium measurements or values. The rate of change of potassium level may be determined over a period of one second or more, minutes, hours, days, etc. In certain embodiments, an average potassium level may be calculated to determine the rate of change of the user's calculated average potassium level.

[0123] In certain embodiments, the potassium trend may be determined based on potassium levels over a specific period of time. In certain embodiments, information regarding the time in range (TIR) of potassium may also be determined based on potassium levels over time.

[0124] In certain embodiments, lactate levels can be determined from sensor data (e.g., lactate measurements obtained from the continuous analyte monitoring system 104). In certain embodiments, lactate trends can be determined based on lactate levels over a particular period of time. In certain embodiments, information regarding time in range (TIR) of lactate can also be determined based on lactate levels over time.

[0125] In certain embodiments, pyruvate levels can be determined from sensor data (e.g., pyruvate measurements obtained from the continuous analyte monitoring system 104). In certain embodiments, pyruvate trends can be determined based on pyruvate levels over a particular period of time. In certain embodiments, information regarding time in range (TIR) of pyruvate can also be determined based on pyruvate levels over time. In certain embodiments, pyruvate measurements can help to understand the rate of glucose production and / or consumption.

[0126] In certain embodiments, ketone levels can be determined from sensor data (e.g., ketone measurements obtained from the continuous analyte monitoring system 104). In certain embodiments, ketone trends can be determined based on ketone levels over a particular period of time. In certain embodiments, information regarding time in range (TIR) of ketones can also be determined based on ketone levels over time.

[0127] In certain embodiments, health and illness metrics can be determined from physiological sensors (e.g., body temperature), activity sensors, or combinations thereof, based on, for example, one or more of user input (e.g., pregnancy information or known illness or disease information). In certain embodiments, based on the values of the health and illness metrics, for example, the user's state can be defined as one or more of healthy, ill, resting, or fatigued.

[0128] In certain embodiments, a disease stage metric, such as for a kidney disease, can be determined based on one or more of the user inputs or outputs provided, for example, by the decision support engine 114 illustrated in FIG. 1. In certain embodiments, exemplary stages of a kidney disease can include AKI, stage 1 CKD with normal or high GFR (e.g., GFR > 90 mL / min), stage 2 mild CKD (e.g., GFR = 60 - 89 mL / min), stage 3A moderate CKD (e.g., GFR = 45 - 59 mL / min), stage 3B moderate CKD (e.g., GFR = 30 - 44 mL / min), stage 4 severe CKD (e.g., GFR = 15 - 29 mL / min), and stage 5 end-stage CKD (e.g., GFR < 15 mL / min). In certain embodiments, exemplary disease stages can be represented as GFR values / ranges, severity scores, and the like.

[0129] In certain embodiments, a dietary status metric can indicate the user's status regarding food consumption. For example, the dietary status can indicate whether the user is in a fasting state, pre-meal state, eating state, post-meal reaction state, or stable state. In certain embodiments, the dietary status can also indicate residual nutrition, such as consumed meals, snacks, or beverages, and can be determined from food consumption information, meal time information, and / or digestibility information, which can be correlated, for example, to the type of food (e.g., macronutrient and micronutrient content), amount, and / or order (e.g., which food / beverage was eaten first).

[0130] In certain embodiments, a dietary habit metric is based on the content and timing of the user's meals. For example, if the dietary habit metric is on a scale of 0 - 1, in an example, the more the user eats a better / healthier meal, the higher the user's dietary habit metric is relative to 1. Also, in this example, the more the user's food consumption adheres to a specific time schedule or recommended meals, the closer the user's dietary habit metric is to 1.

[0131] In certain embodiments, medication compliance is measured by one or more metrics indicating how committed the user is to their medication regimen. In certain embodiments, the medication compliance metric is calculated based on one or more of the timing of when the user takes the medication (e.g., whether the user is on time or on schedule), the type of medication (e.g., whether the user is taking the correct type of medication), and the dosage of the medication (e.g., whether the user is taking the correct dosage). In certain embodiments, the user's medication compliance can be determined in a clinical trial where medication consumption and the timing of such medication consumption are monitored through user input and / or based on analytes received from the analyte monitoring system 104.

[0132] In certain embodiments, the activity level metric can indicate the user's activity level. In certain embodiments, the activity level metric is determined based on input from an activity sensor or other physiological sensor, such as, for example, the non-analyte sensor 206. In certain embodiments, the activity level metric can be calculated by the DAM 116 based on one or more of the input 128, such as one or more of exercise information, non-analyte sensor data (e.g., accelerometer data), time, user input, etc. In certain embodiments, the activity level can be represented as the user's step rate. The activity level metric can be timestamped so as to be correlated with the user's analyte level simultaneously.

[0133] In certain embodiments, the exercise regimen metric may indicate one or more of what type of activity the user engages in, the corresponding intensity of such activity, how often the user engages in such activity, and the like. In certain embodiments, the exercise regimen metric may be calculated based on one or more of non-analyte sensor data inputs (e.g., non-analyte sensor data inputs from an accelerometer, a heart rate monitor, an ECG sensor (e.g., for monitoring new features, QRS complexes, abnormalities, premature ventricular contractions (PVCs), premature atrial complexes (PACs), etc.), a temperature sensor, a respiratory rate sensor, etc.), calendar inputs, user inputs, and the like).

[0134] In certain embodiments, the body temperature metric may be calculated by the DAM116 based on input 128, more specifically, non-analyte sensor data from a temperature sensor. In certain embodiments, the heart rate metric (e.g., including heart rate and heart rate variability) may be calculated by the DAM116 based on input 128, more specifically, non-analyte sensor data from a heart rate sensor and / or an ECG sensor. In certain embodiments, the respiration metric may be calculated by the DAM113 based on input 128, more specifically, non-analyte sensor data from a respiratory rate sensor.

[0135] Exemplary methods and systems for providing decision support regarding diabetes and kidney disease FIG. 4 is a flowchart illustrating an exemplary method 400 for providing decision support using a continuous analyte monitoring sensor including at least a continuous glucose sensor, according to certain exemplary aspects of the present disclosure. For example, method 400 may be executed to provide decision support to a user using a continuous analyte monitoring system 104 including at least a continuous glucose sensor 202, as illustrated in FIGS. 1 and 2.

[0136] Method 400 is executed by decision support system 100 to (1) automatically detect patterns associated with glucose metrics, (2) evaluate the presence and severity of diabetes, (3) risk-stratify patients to identify patients at high risk of diabetes, (4) identify risks associated with current diabetes diagnoses (e.g., risk of death, major cardiac events, hyperglycemia, hypoglycemia, etc.), and (5) collect / generate data such as input 128 and metric 130 including, for example, the above-described analyte data, patient information, and non-analyte sensor data to provide patient-specific treatment recommendations for diabetes diagnosis and management. For example, decision support system 100 can execute Method 400 by monitoring one or more analytes of a renal disease patient over multiple periods, where the one or more analytes include at least glucose. Decision support system 100 then determines one or more glucose metrics of the renal disease patient and uses a trained machine learning model to generate a diabetes diagnosis based on the one or more glucose metrics and then generate one or more recommendations for treatment based on the diabetes diagnosis.

[0137] In certain embodiments, decision support system 100 presented herein can be configured to identify that a patient is experiencing diabetes even if the patient is unaware of their condition or if existing diagnostic methods, techniques, and tests do not indicate a diabetes diagnosis. In particular, in patients with renal disease, diabetes and a decline in glucose homeostasis can be masked by the renal disease, such that existing diagnostic methods, techniques, and tests may not diagnose patients having diabetes and a decline in glucose homeostasis. Thus, by identifying or predicting the presence and / or severity of diabetes in patients with renal dysfunction based on sensor data (e.g., generated by at least continuous glucose sensor 202), decision support system 100 presented herein provides a diagnosis that can be important for the detection and management of diabetes. Method 400 is described below with reference to FIGS. 1 and 2 and their components.

[0138] In certain embodiments, the decision-making support engine 114 of the decision-making support system 100 can use various algorithms or artificial intelligence (AI) models, such as machine learning models trained based on patient-specific data and / or population data, to provide a diabetes diagnosis. The algorithms and / or machine learning models can take into account one or more inputs 128 and / or metrics 130 described with respect to FIG. 3 for a patient when predicting whether the patient is experiencing or is likely to experience diabetes.

[0139] Certain embodiments may include a multiple input, single output (MISO) model in which one or more machine learning models are trained to predict the risk or likelihood that a patient will experience diabetes, type, or severity. For example, one model may be trained to predict the likelihood of current or future onset of diabetes, another model may be trained to predict the likelihood of current or future onset of type II diabetes, and yet another model may be trained to predict the likelihood of current or future onset of type II insulin dependence.

[0140] As an illustrative example, a MISO model can be trained to predict the type of diabetes of a patient. In certain embodiments, such a MISO model may be trained to output a vector having a plurality of values, each value corresponding to the likelihood of the type of diabetes that the patient may currently or in the future experience. For example, the vector output by the MISO model can include two values, each value corresponding to a different type of diabetes. For example, a first value indicating 90% may indicate that there is a 90% likelihood that the patient is experiencing type II diabetes, or a 90% risk or likelihood that the patient will experience type II diabetes within a given period. On the other hand, a second value indicating 0% may indicate that there is a 0% likelihood that the patient is experiencing type I diabetes, or a 0% risk or likelihood that the patient will experience type I diabetes within a given time.

[0141] Certain embodiments may include a multi-input multi-output (MIMO) model in which one or more machine learning models are trained to predict the risk or likelihood of the presence, type, and severity of diabetes in a patient. For example, a single model can be trained to predict the likelihood of current or future occurrence of diabetes, its type, and its severity.

[0142] One or more machine learning models described herein for making such predictions can first be trained using population data. A method for training one or more machine learning models can be described in more detail below with respect to FIG. 6.

[0143] In certain embodiments, instead of using a machine learning model, the decision support engine 114 can use a rule-based model to predict the risk or likelihood that a patient has diabetes, its type, and its severity. A rule-based model involves using a set of rules for analyzing data. These rules also tend to follow the "If X happens then do or conclude Y" guideline and may be referred to as "If statements". In particular, the decision support engine 114 can apply rule statements (e.g., if-then statements) to predict the presence of diabetes.

[0144] Such rules can be defined and maintained by the decision support engine 114 in a reference library. For example, the reference library can maintain a range of metrics that can be mapped to different risks of diabetes. In certain embodiments, such rules can be determined based on an analysis of historical patient records, such as records stored in an empirical study or a historical record database 112. In some cases, the reference library can be very granular. For example, other factors can be used in the reference library to create such "rules". Other factors can include gender, age, diet, medical history, family medical history, body mass index (BMI), and the like. An increase in granularity can provide a more accurate output.

[0145] In block 402, method 400 begins by continuously monitoring one or more analytes of a kidney disease patient, such as user 102 illustrated in FIG. 1, during a first period to obtain analyte data. The one or more analytes to be monitored may include at least glucose. Block 402 may be performed, in certain embodiments, by the continuous analyte monitoring system 104 illustrated in FIGS. 1 and 2, and more specifically, by the continuous analyte sensor 202 illustrated in FIG. 2. For example, the continuous analyte monitoring system 104 may include a continuous glucose sensor 202 configured to measure a patient's glucose level over multiple periods.

[0146] The primary analyte for the measurements described herein is glucose, although in certain embodiments, other analytes may be considered. In particular, combining glucose data with additional analyte data can help further inform the analysis regarding the diagnosis of diabetes. For example, monitoring additional types of analytes such as potassium, lactate, pyruvate, insulin, and / or ketones measured by the continuous analyte monitoring system 104 can provide additional insight into the generation of a diabetes diagnosis.

[0147] The additional insight obtained from using combinations of analytes, in addition to glucose, can increase the accuracy of a diabetes diagnosis. For example, the probability of accurately generating a diabetes diagnosis can be a function of the number of analytes measured for a patient. For example, the probability of accurately diagnosing diabetes using only glucose data may be lower than the probability of accurately diagnosing diabetes using both glucose and potassium data, and may also be lower than the probability of accurately diagnosing diabetes using glucose, potassium, and lactate data for the analysis.

[0148] For example, in certain embodiments, at block 402, the continuous analyte monitoring system 104 may continuously monitor a patient's glucose level and potassium level over multiple periods. In certain embodiments, the measured potassium concentration may be used in conjunction with the glucose level to generate a diabetes diagnosis. In certain embodiments, the potassium data may indicate the presence of a kidney disease. The potassium data may also, in certain embodiments, indicate a change in the severity of a kidney disease. The presence and / or change in severity of a kidney disease may indicate a change (e.g., a higher) in the risk of diabetes for the user, based on the user's glucose metrics. For example, the potassium data may indicate that the user is worsening a kidney disease, and as a result, the user's glucose metrics may be mapped to a higher risk of diabetes.

[0149] In addition to continuously monitoring one or more analytes of a patient over multiple periods to obtain analyte data at block 402, optionally, in certain embodiments, method 400 may also include monitoring other sensor data (e.g., non-analyte data) over multiple periods using one or more other non-analyte sensors or devices (e.g., non-analyte sensor 206 and / or medical device 208 of FIG. 2, etc.).

[0150] As described above, non-analyte sensors and devices can include, but are not limited to, one or more of an insulin pump, a tactile sensor, an ECG sensor and / or a heart rate monitor, a blood pressure sensor, a respiratory sensor, a peritoneal dialysis machine, a hemodialysis machine, a sensor or device provided by the display device 107 (e.g., an accelerometer, a camera, a global positioning system (GPS), a heart rate monitor, an ECG sensor, etc.) or other user accessories (e.g., a smartwatch), or any other sensor or device that provides relevant information about the user. Metrics such as the metric 130 illustrated in FIG. 3 can be calculated using measurement data from one or more of these additional sensors. As illustrated in FIG. 3, metrics 130 calculated from non-analyte sensor or device data can include heart rate (including heart rate variability), respiratory rate, and the like. In certain embodiments described in more detail below, metrics 130 calculated from non-analyte sensor or device data can be used to further inform an analysis related to the generation of a diabetes diagnosis.

[0151] In certain embodiments, one or more of the non-analyte sensors and / or devices can be worn by the user to help detect periods of increased physical exertion by the user. Such non-analyte sensors and / or devices can include an accelerometer, an ECG sensor, a blood pressure sensor, a heart rate monitor, an impedance sensor, a dialysis machine, and the like. Glucose levels can vary during exercise and can vary based on the type and intensity of the exercise. Due to these extrinsic variations in glucose levels due to exercise, in certain embodiments, the glucose metrics obtained for the user can exclude glucose levels during exercise. In particular, glucose data measured and collected from periods of increased body movement can be excluded from the calculation of one or more glucose metrics used to predict diabetes. For example, glucose data collected during exercise can be excluded from the calculation of TIR, allowing a diabetes diagnosis to be based on the TIR when the user is not exercising.

[0152] In certain embodiments, one or more of the non-analyte sensors and / or devices may be worn by a user to assist in detecting the duration of treatment. Such non-analyte sensors and / or devices may include accelerometers, ECG sensors, blood pressure sensors, heart rate monitors, impedance sensors, dialysis machines, and the like. As part of dialysis treatment, a large amount of glucose may enter the patient's system. This additional glucose can confound the glucose metrics used to diagnose diabetes. Thus, in certain embodiments, glucose data measured and collected during the treatment period may be excluded from the calculation of one or more glucose metrics used to predict diabetes. For example, glucose data collected during dialysis may be excluded from the calculation of a measure of glucose variability, and a diabetes diagnosis may be based on this measure of glucose variability without dialysis.

[0153] At block 404, method 400 continues by processing the analyte data to determine one or more glucose metrics for the renal disease patient. Block 404 may be performed by the decision support engine 114 in certain embodiments.

[0154] In certain embodiments, a glucose metric refers to a timestamped glucose measurement or value. A glucose metric may refer to a change between one or more timestamped glucose levels or values. For example, a glucose metric may be the difference between a glucose level at a first timestamp and a glucose level at a subsequent timestamp. A glucose metric may be a rate of change of glucose that indicates a change of one or more timestamped glucose levels relative to one or more other timestamped glucose levels. For example, a glucose metric may be a derivative of the change between a glucose level at a first timestamp and glucose levels at one or more subsequent timestamps.

[0155] Additionally, or alternatively, the timestamped glucose levels may reflect the time at which the levels were recorded and may be classified as daytime glucose levels (e.g., between 6:00 a.m. and 10:00 p.m.) or nighttime glucose values (e.g., between 10:00 p.m. and 6:00 a.m.). The daytime glucose values can be compared to the nighttime glucose values to determine how the daytime glucose levels affect the nighttime glucose levels.

[0156] In certain embodiments, the glucose metric may indicate an average glucose level, which may be an average of two or more timestamped glucose levels. In certain embodiments, the average glucose can be calculated based on the glucose levels as well as other inputs 128 such as food consumption information, thereby determining the average glucose using the corresponding glucose levels and food consumption information (e.g., having overlapping timestamps). The average glucose is calculated over a period (e.g., one day) and can be compared to the average glucose on a later day.

[0157] The average glucose is the average of one or more glucose levels over a given period. The average glucose can be used to calculate a glucose management indicator (GMI). The average glucose and the GMI have a known relationship:

[0158]

Number

[0159] The Glucose Management Indicator (GMI) is a measure derived from the average glucose value (i.e., the average glucose level), and is intended to convey the current state of glucose management. The relationship between average glucose and GMI is derived from the relationship between continuous glucose sensor measured glucose levels and clinical A1C. In certain embodiments, the glucose metric refers to the Glucose Management Indicator (GMI). In certain embodiments, the GMI is calculated based on the relationship between the GMI and average glucose, and in particular, the average glucose is determined based on glucose levels over a period of time. For example, the GMI can be calculated based on the average glucose level over a two-week or one-month period. In certain embodiments, the GMI can be calculated based on glucose levels and other inputs 128 such as food consumption information, and thereby, the corresponding glucose levels and food consumption information (e.g., having overlapping timestamps) can be used to determine the GMI.

[0160] In certain embodiments, the glucose metric can indicate the variability of blood glucose, which can be the standard deviation of the average glucose (e.g., the standard deviation of the average of one or more timestamped glucose levels relative to the average of one or more other timestamped glucose levels). For example, the standard deviation of the average glucose may be based on the average of two or more timestamped glucose levels. In another example, the standard deviation of the average glucose may be based on the average of glucose levels over two or more periods. In yet another example, the standard deviation of the average glucose may be based on the average of glucose levels over a specific period. In certain embodiments, the variability of blood glucose may be calculated based on glucose levels and other inputs 128 such as food consumption information, and thereby, the glucose levels and food consumption levels having overlapping timestamps can be used to determine the variability of blood glucose.

[0161] In some embodiments, the glucose metric indicating glucose variability may be a setpoint metric. For example, the decision support engine 114 may determine a setpoint based on an estimate of the "mode" of the patient's glucose values (e.g., the glucose value that appears most frequently in a set of glucose values). The setpoint may be determined, for example, based on historical population data and / or the patient's historical glucose data. Based on the calculated setpoint, the glucose metric can further indicate the time within a range of glucose levels within the range of the setpoint value.

[0162] In certain embodiments, the glucose metric may be a minimum and / or maximum glucose level. For example, the minimum and / or maximum glucose levels may be based on, for example, the minimum and maximum glucose levels over a day or a week.

[0163] In certain embodiments, the glucose metric indicates the period during which the glucose level is within a certain range (e.g., the glucose level meets, exceeds, or is between the threshold levels). The periods during which glucose is within a particular range can be aggregated to determine the time in range (TIR) of glucose for that particular range. For example, all of the periods during which the glucose level is within a healthy glucose range can be aggregated over the monitoring period during which the glucose data was collected so that the TIR for the healthy range can be determined. In another example, all of the periods during which the glucose level is below a healthy glucose range (e.g., a low glucose range) over a monitoring period (e.g., two weeks, four weeks, eight weeks) can be aggregated to determine the low-range glucose TIR. In yet another example, all of the periods during which the glucose level exceeds a healthy glucose range (e.g., a high glucose range) over a monitoring period (e.g., two weeks, four weeks, eight weeks) can be aggregated to determine the high-range glucose TIR. In certain embodiments, glucose ranges such as a healthy glucose range, a low glucose range, and / or a high glucose range may be provided through user input or obtained from a particular data storage device (e.g., an electronic medical record (EMR), historical data, etc.).

[0164] In certain embodiments, the glucose metric may demonstrate a pattern or trend of glucose levels obtained at different times (e.g., 5-minute intervals, 10-minute intervals, etc.). The glucose metric may be an autocorrelation feature that demonstrates the similarity in patterns and trends between glucose levels obtained at different times. The autocorrelation feature may be a numerical value from 1.0 to 0.0, where 1.0 demonstrates that the time series glucose levels are correlated (e.g., the patterns of glucose levels obtained at different times are very similar and / or the same), and 0.0 demonstrates that the time series glucose levels are not correlated (e.g., the patterns of glucose levels obtained at different times are not similar and / or the same).

[0165] In certain embodiments, the machine learning model described herein used to provide diabetes prediction may include one or more features related not only to the patient's glucose levels but also to the patient's glucose metrics described above. For example, an exemplary machine learning model may include weights applied to features associated with one or more glucose metrics of a patient. Thus, in certain embodiments, one or more glucose metrics of a patient can be calculated for input to the model before using the machine learning model.

[0166] Furthermore, in certain embodiments, the rule - based model described herein used to provide diabetes prediction may include one or more rules related to not only the patient's glucose level but also the patient's glucose metrics. For example, a reference library used to define one or more rules for the rule - based model may maintain ranges of glucose levels and ranges of glucose metrics that can be mapped to different risk levels for diabetes. Thus, prior to using the rule - based model, one or more glucose metrics of the patient can be calculated for input to the model. Additional features added to the model, such as additional glucose metrics, may, in some cases, enable a more accurate prediction of the patient's diabetes disease.

[0167] In block 406, method 400 continues by generating a disease prediction using analyte data associated with one or more analytes and at least one or more glucose metrics. Block 406 may be performed, in certain embodiments, by the decision support engine 114 illustrated in FIG. 1. As discussed above, for patients with kidney disease, A1C is not a reliable measure for diabetes diagnosis because the calculation of A1C depends on assumptions about the red blood cell half - life, which are likely inaccurate. Thus, the glucose metrics determined in block 404 can be used by the decision support system 100 to provide an accurate diabetes diagnosis for patients with kidney disease. Based on the diabetes diagnosis, cases of diagnosing patients with diabetes can be provided to the HCP.

[0168] As described above, different methods for generating disease predictions can be used by the decision support engine 114. In particular, in certain embodiments, the decision support engine 114 can use a rule-based model to provide real-time diabetes predictions, including the provision of diabetes diagnosis and staging. In particular, the decision support engine 114 can apply rules to assess the presence and severity of diabetes in a patient and / or to identify risks associated with the patient's current diabetes diagnosis (e.g., risk of death, etc.).

[0169] For example, one rule may relate to the current patient's glucose level or be based on changes in glucose level over time. Another rule may be based on a glucose metric or changes in a glucose metric over time. Another rule may be based on TIR or changes in the patient's TIR over time. Another rule may be based on average glucose or changes in the patient's average glucose over time. Another rule may be based on GMI or changes in the patient's GMI over time. Another rule may be based on the variability of blood glucose or changes in the variability of the patient's blood glucose over time. Another rule may relate to the rate of change of the patient's glucose level, such as a "rapidly increasing" or "rapidly decreasing" rate of change (as described with respect to FIG. 3), or be based on changes in the rate of change of glucose level over time. Another rule may relate to the patient's glucose response or lack thereof to biochemical hypoglycemia (e.g., below 70 mg / dL), regardless of the level of circulating insulin. Another rule may relate to the patient's glucose response or lack thereof to biochemical hyperglycemia (e.g., above 150 mg / dL), regardless of the level of circulating insulin. Another rule may relate to an insulin metric, a potassium metric, a lactate metric, a pyruvate metric, and / or a ketone metric, as described with respect to FIG. 3, or be based on changes in such metrics over time. It is also contemplated that any of the rules may be used alone or in combination with any other rules by the decision support system 100 to determine a diabetes diagnosis based on the patient's glucose metric.

[0170] Such rules may be defined and maintained by the decision support engine 114 in a reference library. For example, the reference library may maintain ranges of glucose levels and / or glucose metrics that may be mapped to different severities, types, and / or stages of diabetes. In certain embodiments, such rules may be determined based on empirical studies and on analyzing historical patient records from the historical record database 112.

[0171] In certain embodiments, as an alternative to using a rule-based model, an AI model, such as a machine learning model, can be used to provide real-time decision-making disease predictions. For example, the decision support engine 114 can deploy one or more of these machine learning models to perform the diagnosis and staging of diabetes in patients with kidney disease. In another example, one or more machine learning models can stratify patients to identify the health risks associated with the current diagnosis of diabetes in patients with kidney disease. Risk stratification can refer to the process of assigning health risks (e.g., the risks of hyperglycemia and / or hypoglycemia) to patients with diabetes. Identification of patients with a high risk of hyperglycemia and / or hypoglycemia can be used to generate treatment recommendations.

[0172] In particular, the decision support engine 114 can obtain information from the user profile 118 associated with the patient stored in the user database 110, characterize the information about the patient stored in the user profile 118 into one or more features, and use these features as inputs to such a model. Alternatively, the information provided by the user profile 118 can be characterized by another entity (e.g., another server or computing device), and then the features can be provided to the decision support engine 114 to be used as inputs to the ML model. In certain embodiments, the features associated with the patient can be used as inputs to one or more of the models to assess the presence and severity of diabetes in the patient.

[0173] In certain embodiments, the features associated with a patient can be used as inputs to one or more of the models for providing disease predictions, which may involve identifying whether the patient is at high or low risk of developing diabetes. In certain embodiments, the features associated with a patient can be used as inputs to one or more of the models to identify the risks associated with the patient's current diabetes diagnosis (e.g., risk of death, etc.). In certain embodiments, the features associated with a patient can be used as inputs to one or more of the models to perform any combination of the above functions. Details related to how one or more machine learning models can be trained to provide real-time decision support for diabetes diagnosis and staging are further considered in connection with FIG. 6.

[0174] As mentioned, in certain embodiments, in addition to glucose, other analyte data can be used by the decision support engine 114 in block 406 to generate disease predictions for a patient. Analyte data including potassium data, lactate data, pyruvate data, insulin data, cystatin C data, and / or ketone data (e.g., from measurements by the continuous analyte monitoring system 104) can be used as inputs to such machine learning models and / or rule-based models to predict the presence and severity of diabetes in a user.

[0175] The decision support engine 114 may use a machine learning model and / or a rule-based model to generate disease predictions based on the continuous analysis of data about patients (e.g., analyte data and in some cases non-analyte data) collected over various periods. The analysis of data collected about patients over various periods may provide insights into whether the patient's health and / or disease is improving or deteriorating. For example, using the models discussed herein, patients previously diagnosed with diabetes may be continuously monitored (e.g., continuously collected for the patient) to determine, among other things, whether the disease is worsening or improving. As an example, a comparison of glucose levels, timestamped glucose levels, glucose baselines, absolute maximum glucose levels, absolute minimum glucose levels, glucose rate of change, glucose metrics (e.g., glucose setpoint metrics, glucose autocorrelation features, etc.), TIR, average glucose, GMI, and / or glucose variability for a patient over multiple periods (e.g., days, weeks, months, etc.) may indicate the progression of the patient's disease. For example, if the decision support engine 114 determines that the patient's absolute minimum glucose level begins to decrease over time while the patient is sleeping, it may determine that the patient's kidney disease is progressing.

[0176] In certain embodiments, the decision support engine 114 may provide disease predictions based on the patient's timestamped glucose values. For example, a patient with kidney disease may have, over time, higher daytime glucose levels, more daytime hyperglycemic events, lower nighttime glucose levels, and / or more nighttime hypoglycemic events, which may indicate a worsening of glycemic control and thus a worsening of kidney disease. Additionally, a patient with kidney disease may experience hyperglycemic spikes after dinner and / or when the patient goes to bed. If the magnitude of the patient's hyperglycemic spikes after dinner and / or when the user goes to bed increases over time, the decision support engine 114 may determine that the patient is experiencing worsening kidney disease.

[0177] In certain embodiments, lower autocorrelation features (e.g., less than 0.5) may indicate a worsening of kidney disease. In certain embodiments, higher maximum glucose levels and lower minimum glucose levels may indicate a worsening of kidney disease. In certain embodiments, the decision support engine 114 may use setpoint metrics to determine a patient's kidney disease stage. For example, as kidney disease worsens, the variability of glucose measurements increases, and as a result, the time within the range of glucose levels (specifically, within the range of the setpoint) may decrease. As the frequency of glucose levels within the setpoint range decreases, the decision support engine 114 may determine that the user's kidney disease is progressing.

[0178] In certain embodiments, the rates of change and directional movement of glucose and potassium that coincide and oppose each other over time, and / or the time delay in which one moves and affects the other, can be used to (1) identify improvements in the user's diabetes and / or kidney disease, (2) serve as a surrogate for when insulin is activated to push potassium and glucose into cells, but also helps to rule out certain other conditions. Similarly, the difference in the rates of change between potassium and glucose can be used to derive insights because under normal conditions (e.g., when the patient is not exercising and / or is in a disease or injury state), potassium appears to change at a slower rate than glucose. That time delay can be a predictor of potassium change, or the length of time from glucose change to potassium change can be an indicator of insulin resistance or another factor. For example, a shorter time delay versus a longer time delay may indicate that the reaction kinetics of uptake has changed from one intracellular pathway to another. Thus, a change in the time delay can indicate a change in the user's diabetes and / or kidney disease state.

[0179] In some cases, method 400 continues at block 408, where the decision support engine 114 generates one or more recommendations for management and / or treatment, at least partially based on the disease prediction generated at block 406. In particular, the decision support engine 114 may provide recommendations for the management and / or treatment of diabetes, such as lifestyle recommendations, pharmaceutical recommendations, medical intervention recommendations, or other recommendations for managing diabetes and glucose homeostasis. The decision support engine 114 may output such recommendations for treatment to the user (e.g., through application 106). In certain other cases, method 400 ends at block 406 with the patient's disease prediction (e.g., block 408 may be optional).

[0180] In certain embodiments, the decision support engine 114 can use one or more other machine learning models trained based on patient-specific data and / or population data to provide management and / or treatment recommendations to the user. The machine learning model may consider one or more inputs 128 and / or metrics 130 (e.g., including glucose metrics) described with respect to FIG. 3 to determine optimal recommendations for the management and / or treatment of the patient's diabetes. In certain embodiments, the model can look at different patterns of analyte measurements collected for the patient to guide the patient in the management and / or treatment of the disease.

[0181] In certain embodiments, instead of using a machine learning model, the decision support engine 114 can use one or more decision trees to provide management and / or treatment recommendations. The decision trees may be rule-based and may provide recommendations regarding the management and / or treatment of the patient's diabetes based on one or more rules.

[0182] Recommendations for the management and / or treatment of diabetes may, in some cases, be based on determined glucose metrics (e.g., glucose level, rate of change of glucose level, TIR, average glucose, GMI, glucose variability, and / or glucose clearance rate), and / or on changes in a patient's glucose metrics. For example, recommendations for the management and / or treatment of diabetes may be based on a low TIR, a high GMI, and / or high glucose variability. Recommendations regarding the management and treatment of diabetes may include diet, exercise, lifestyle, treatment, insulin administration, and / or pharmaceutical recommendations, as discussed in more detail with respect to FIG. 5.

[0183] In certain embodiments, diabetes management and / or treatment intervention advisories can include recommendations for a patient to seek medical assistance. For example, in certain embodiments, diabetes management and / or treatment advisories may indicate to the patient that the patient needs to go to the emergency room immediately and / or needs to contact a physician. In certain other embodiments, the management / or treatment recommendations may automatically alert the patient's HCP regarding the patient's condition for intervention by the HCP. In certain other embodiments, the diabetes management and / or treatment recommendations can alert medical personnel to activate an ambulance service or emergency medical service to provide emergency pre-hospital care and stabilization to the patient and / or to transport the patient to definitive care. Certain embodiments are such that the decision support engine 114 can make management and / or treatment recommendations based on the patient's ability to seek medical assistance and / or the patient's access to medical assistance.

[0184] Method 400 is described in relation to providing diabetes determination support for patients with kidney disease. However, it should be noted that in certain embodiments, method 400 can be similarly performed on patients without kidney disease. The average glucose is expected to be similar between patients with and without kidney disease (e.g., CKD). However, the glucose variability may be higher in patients with CKD, particularly those with stage 3 or 4 CKD (progressive CKD) and / or patients on dialysis, compared to patients without CKD. In addition, A1C may be artificially suppressed low in such patients, and thus, an "average shift" is observed and may be highly variable depending on the therapy administered to such patients. In humans with diabetes and without CKD, the relationship between the average glucose measured by a continuous glucose monitor and A1C is consistent. However, many of the treatments for CKD (e.g., erythropoietin (EPO) administration, anemia, dialysis) can affect red blood cell metabolism. Thus, for CKD patients, higher glucose variability, similar average glucose values, and a very slightly consistent relationship over time for A1C can be expected.

[0185] FIG. 5 is a flow diagram illustrating an exemplary method 500 for providing decision support using a continuous analyte sensor that includes at least a continuous glucose sensor 202, according to certain exemplary aspects of the present disclosure. For example, method 500 can be performed to provide decision support to a user using a continuous analyte monitoring system 104 that includes at least a continuous glucose sensor 202, as illustrated in FIGS. 1 and 2.

[0186] Method 500 may be performed by decision support system 100 to, for example, collect / generate data such as input 128 and metric 130 including the above-described analyte data, patient information, and non-analyte sensor data, and (1) automatically detect and determine glucose metrics, including, for example, glucose clearance rate; (2) assess the presence and severity of an atypical glucose trend; (3) risk stratify patients to identify patients at high risk of hypertension and / or hypotension; (4) identify risks associated with current glucose metrics (e.g., risk of death, risk of significant cardiac events, etc.); (5) make patient-specific treatment decisions or recommendations for glucose, diabetes, and kidney disease management; and (6) provide information regarding the effect of an intervention (e.g., the effect of a patient's lifestyle modification, the effect of a surgical procedure, the effect of a patient taking a new medication, etc.). In other words, the decision support system presented herein may provide information useful for directing and improving care for patients having or at risk of having diabetes and kidney disease.

[0187] Patient care can be improved by providing recommendations tailored to manage and treat diabetes in patients with kidney disease. As contemplated, kidney disease affects glucose homeostasis by affecting glucose metabolism. A user's glucose clearance rate and / or other metrics may be used to approximate glucose metabolism in the user's current renal function and to improve management of glucose homeostasis. Specifically, as described below in connection with method 500, a user's glucose clearance rate may be used in conjunction with other glucose metrics, including current glucose levels and rates of change, to provide context to that user's glucose trends. The glucose clearance rate can reveal when glucose levels are trending towards adverse events (e.g., hyperglycemia and / or hypoglycemia) for the user and their current renal function. For example, a rapidly increasing glucose trend and a slow glucose clearance rate may indicate that the user is at risk of hyperglycemia. Alternatively, a user with kidney disease may experience a rapid decrease in glucose as a result of low insulin clearance, decreased renal glucose consumption, and decreased renal gluconeogenesis, which may indicate that the user is at risk of hypoglycemia. Then, as described in connection with method 500, treatment recommendations can be provided to the user to manage the adverse events. Method 500 is described below with reference to FIGS. 1 and 2, and their components.

[0188] At block 502, method 500 begins by continuously monitoring one or more analytes of a patient, such as patient 102 illustrated in FIG. 1, during a first period, to obtain analyte data. The one or more analytes to be monitored can include at least glucose. For example, block 502 can be performed by the continuous analyte monitoring system 104 illustrated in FIGS. 1 and 2, and more specifically, by the continuous analyte sensor 202 illustrated in FIG. 2, in certain embodiments. The continuous analyte monitoring system 104 can comprise a continuous glucose sensor 202 configured to measure the patient's glucose level.

[0189] Similar to method 400 of FIG. 4, the main analyte for the measurements described herein is glucose, although in certain embodiments, other analytes may be considered. In particular, combining glucose levels with additional analyte data can help further inform the analysis regarding the management of glucose homeostasis. For example, monitoring additional types of analytes such as potassium, lactate, pyruvate, insulin, and / or ketones measured by continuous analyte monitoring system 104 can provide additional insights into supplementary information used to determine glucose homeostasis and / or optimal treatment to maintain glucose homeostasis. The additional insights obtained from using combinations of analytes, in addition to glucose, can increase the accuracy of glucose management, as described with respect to method 400.

[0190] As described above, renal dysfunction affects at least glucose, insulin, and potassium metabolism. Thus, the additional insights obtained from potassium levels can improve the analysis of glucose clearance rate in block 504 and atypical glucose trends in block 506, as glucose homeostasis and the risk of adverse events are affected by the level and changes in renal function. For example, potassium levels may indicate a worsening of renal dysfunction, which can negatively impact glucose homeostasis and increase the risk of harmful hypoglycemia / hyperglycemia events. Alternatively, glucose metrics and corresponding trends derived from glucose metrics may indicate a worsening of renal function, which can negatively impact potassium homeostasis and increase the risk of hypokalemia / hyperkalemia events. Additionally, changes in renal function can affect treatment recommendations, as the therapeutic effect may be dependent on renal function.

[0191] In another example, lactate levels may be associated with glucose, insulin, and potassium metabolism. Lactate levels may be used to detect food consumption, exercise, rest, and / or stress. Thus, additional insights obtained from lactate levels may improve the analysis of glucose clearance rate at block 504 and the analysis of the atypical glucose trend at block 506 by associating glucose metrics with different body states (e.g., consumed food, exercise, rest, stress, etc.).

[0192] Additionally, pyruvate is used for gluconeogenesis to produce glucose. In renal dysfunction, renal gluconeogenesis is impaired, thereby reducing glucose homeostasis. Thus, pyruvate levels may provide insights into gluconeogenesis and the resulting increase in glucose levels. For example, an increase in pyruvate levels may indicate an expected increase in glucose levels, which may improve the determination of glucose metrics. Thus, considering the interactions of the comorbidities described above, the decision support algorithm and / or the parameters, thresholds, and / or rules of the model may be changed based on the number of analytes being measured for input to reflect the knowledge obtained from each of the other measured analytes and / or the prevalence associated with the additional analytes being measured.

[0193] In addition to continuously monitoring one or more analytes of a patient over a plurality of periods to obtain analyte data at block 502, optionally, in certain embodiments, method 500 may also include monitoring other sensor data (e.g., non-analyte data) over a plurality of periods using one or more other non-analyte sensors or devices (e.g., non-analyte sensor 206 and / or medical device 208 of FIG. 2, etc.).

[0194] As described above, the non-analyte sensors and devices can include, but are not limited to, one or more of an insulin pump, a tactile sensor, an ECG sensor and / or a heart rate monitor, a blood pressure sensor, a publication sensor, a respiratory sensor, a thermometer, a peritoneal dialysis machine, a hemodialysis machine, a sensor or device provided by the display device 107 (e.g., an accelerometer, a camera, a global positioning system (GPS), a heart rate monitor, etc.) or other user accessories (e.g., a smartwatch), or any other sensor or device that provides related information about the user.

[0195] In block 504, method 500 continues by processing the analyte data to determine at least one glucose clearance rate of the patient. Block 504 can be performed by the decision support engine 114 in certain embodiments. Block 504 in method 500 of FIG. 5 can be similar to block 404 in workflow 400 of FIG. 4.

[0196] In certain embodiments, the glucose clearance rate can be determined from sensor data (e.g., glucose levels obtained from the continuous glucose sensor 202 of the continuous analyte monitoring system 104). For example, the glucose clearance rate can be determined by a glucose challenge that can test the user's metabolic function to determine the user's response (e.g., glucose clearance rate) to the consumption of a known amount of glucose. The glucose challenge introduces a known or estimated amount of glucose into the body for the purpose of measuring an increase and / or decrease in the rate of change of the resulting glucose level or other analyte level. The glucose clearance rate can indicate the ability of the kidneys to metabolize glucose, including the ability to respond to an increase and / or decrease in glucose levels. Thus, the glucose clearance rate of a user with kidney disease may include a more rapid increase in glucose levels and a higher peak glucose level in response to the introduction of glucose compared to a healthy user. A user with kidney disease may have a faster and more pronounced glucose peak after the introduction of glucose due to a decrease in the kidney's glucose filtration ability and a decrease in glucose consumption.

[0197] A user's response to a challenge can be obtained and recorded in the form of absolute glucose levels, changes in glucose levels, rates of change of glucose levels, average glucose, GMI, TIR, glucose variability, and / or any other glucose metric. Glucose levels may be determined at specified intervals (e.g., 5 minutes, 10 minutes, 30 minutes, 1 hour, 2 hours, etc. after glucose consumption), over a specified period of time (e.g., 2 hours after glucose consumption), at a specified time of day (e.g., morning, evening, night), etc. In certain embodiments, the glucose clearance rate may be determined by calculating the gradient between an initial high glucose value (e.g., an increase resulting as a consequence associated with the consumption of a known or estimated amount of glucose) and a subsequent low glucose value. In certain embodiments, the glucose clearance rate after administration of a particular type of insulin, such as rapid-acting insulin, can be obtained, acquired, and recorded. In the advanced stages of CKD, insulin clearance disorders are expected to manifest as longer glucose clearance times and perhaps higher clearance rates.

[0198] In certain embodiments, the consumption of a known amount of glucose can be based on a recommendation provided by the decision support engine 114. The decision support engine 114 can provide a recommendation for a glucose challenge by recommending the consumption of a known amount of glucose to increase the glucose level. For example, the glucose challenge recommendation can be to consume 25 grams of glucose. In certain embodiments, the glucose challenge recommendation can be to consume a particular food item. For example, the recommendation can be to consume an apple. In certain embodiments, the glucose challenge recommendation can include a recommendation for the patient to consume a meal having a particular dietary composition. For example, the glucose challenge recommendation can be to consume a high-carbohydrate meal, such as 60% total carbohydrates.

[0199] Furthermore, in certain embodiments, the glucose challenge recommendation may include a recommendation for the user to enter food consumption information manually, through an application, by scanning a barcode or menu, or by other input methods. For example, food consumption information regarding food (e.g., carbohydrates) that the user has already consumed may provide a known amount of glucose introduced for the glucose challenge. In certain embodiments, the food consumption information may be provided as part of input 128 and may include information regarding one or more of meals, snacks, and / or beverages, such as one or more of dimensions, content (e.g., milligrams (mg) of glucose, carbohydrates, fat, protein, etc.), order of consumption, and time of consumption, as described above in relation to input 128. In certain embodiments, the decision support engine 114 may prompt the user to confirm glucose consumption. For example, the decision support engine 114 may ask the user to confirm a meal (e.g., "Did you have a meal?").

[0200] In certain embodiments, glucose consumption information may not be available or may not be provided. In such embodiments, sensor data can be used to determine the amount of glucose consumed. The sensor data may indicate the user's glucose level, the change in the user's glucose level, the rate of change of the glucose level, and / or any other glucose metric. In certain embodiments, the sensor data may be combined with glucose data to help determine the amount of glucose consumed, the composition of the meal / food consumed, notify food consumption information, and / or otherwise indicate current or impending glucose metrics, and may include other analyte data. Additional analytes may include potassium, lactate, ketones, pyruvate, dextrose, insulin, and / or any other analyte. For example, the sensor data may include glucose levels and lactate levels that may indicate or otherwise represent food consumption information and may enable the determination of a glucose clearance rate based on the indicated consumed glucose and the resulting rate of change of glucose.

[0201] In certain embodiments, when glucose consumption information is unknown, historical data may be used to determine the amount of glucose consumed. The user's own historical data may include previous sensor data, food consumption information, exercise information, time information, location information, and the like. In certain embodiments, the historical data includes known and / or unknown amounts of glucose, the composition of the consumed meals, and / or previous glucose clearance rates that are used to determine whether the user consumed glucose and / or the amount or composition of glucose consumed when starting a glucose challenge, and may indicate the timing of glucose consumption. For example, the user may consume meals at the same time every day. In another example, the user may consume a similar meal composition every day. In yet another example, the user may consume a meal of a similar composition at a particular location. As another example, the user may consume a similar composition after an activity (e.g., the user consumes a sports drink after exercise). Thus, the historical information can be used to predict or estimate the amount of glucose consumption used to determine the glucose clearance rate. For example, if the user's historical data indicates that a meal typically corresponding to a 20 mg / dL increase in glucose level is consumed by the user every day for lunch, when calculating the user's glucose clearance rate on a particular day, the decision support engine 114 may predict that a similar amount of glucose will be consumed for lunch on that particular day.

[0202] In certain embodiments, a glucose challenge notification is provided by the decision support engine 114 to confirm glucose consumption based on historical data. For example, the glucose challenge notification may ask the user to confirm a meal (e.g., "Did you have a meal?"). Further, in certain embodiments, the glucose challenge notification may include a recommendation for the user to enter food consumption information manually, through an application, by scanning a barcode or menu, or by other input methods.

[0203] In certain embodiments, the amount of glucose consumed (e.g., as determined by any of the methods described above) is used to calculate the initial increase and subsequent decrease in glucose levels when determining the glucose clearance rate. For example, an amount of glucose is consumed and the glucose clearance rate is determined based on the resulting increase associated with the amount of glucose consumed and then the rate of decrease in glucose levels.

[0204] In certain embodiments, the glucose clearance rate can be determined based on not only the glucose data obtained from the continuous glucose sensor 202 but also insulin administration information. For example, some users with known diabetes may be taking exogenous insulin to reduce the risk of hyperglycemia. In the case of users taking exogenous insulin, the glucose challenge may include administration of insulin to prevent hyperglycemia during the challenge. The glucose clearance rate of a user determined by the glucose challenge can thus be based on both the amount of glucose consumed and the dosage of insulin administered.

[0205] In the case of users with renal impairment who are receiving exogenous insulin, insulin administration can have unexpected, atypical, or other different effects on the glucose clearance rate, particularly when compared to users without renal impairment. Also, as renal impairment progresses or regresses, insulin administration can have different effects on the glucose clearance rate compared to the previous glucose clearance rate associated with the user's previous state of renal impairment. Thus, for users with renal impairment, the glucose clearance rate determined based on a glucose challenge with insulin administration can provide further insight into glucose homeostasis and the risk of adverse events. The glucose clearance rate for such a glucose challenge is calculated based on both the amount of glucose consumed and both the dosage and type of insulin administered.

[0206] In certain embodiments, glucose consumption and insulin administration include consuming a known amount of glucose and administering a known dosage of insulin. The known amount of glucose can be determined based on any of the methods described above. Information regarding the administration of the known dosage of insulin (the "insulin information") may include information regarding the type, dosage, rate of activity, duration of activity, and / or timing of the insulin administered as part of the challenge, and may be provided as pharmaceutical information provided as part of input 128 of FIG. 1.

[0207] In certain embodiments, the type of insulin administered for a glucose challenge is rapid-acting insulin. In patients without renal impairment, the onset (e.g., the length of time it takes for insulin to begin to act on blood glucose) is 5 - 15 minutes, the peak of activity is 30 - 90 minutes, and the total duration of activity is 3 hours. For example, a user can consume a known amount of glucose and administer rapid-acting insulin as part of the glucose challenge, thereby determining the user's glucose clearance rate for rapid-acting insulin and correlating it with the onset, peak, and duration of insulin activity. For users with renal impairment due to insulin resistance and insulin clearance disorders, the onset and peak of activity of rapid-acting insulin can be more rapid and the duration of activity can be extended. Thus, determining the glucose clearance rate associated with exogenous insulin in users with renal impairment can improve decision-making support by explaining the altered effect of insulin on glucose levels when determining glucose trends and treatment recommendations.

[0208] In certain embodiments, insulin administration may be based on recommendations provided by the decision support engine 114. The decision support engine 114 may provide recommendations regarding insulin-dependent glucose challenges by recommending insulin administration. The recommendations may indicate the type, dosage, and / or timing of insulin to be administered as part of the challenge. For example, the recommendation may be to administer 10 units of rapid-acting insulin after a meal.

[0209] In certain embodiments, insulin administration may be determined using historical medical information, real-time medical information, or a combination thereof. For example, the medical information may indicate that the user typically administers insulin at a particular time, in conjunction with a particular activity, and / or at a particular glucose level. The medical information provided as part of input 128 in FIG. 1 may indicate that the user administered insulin as part of a glucose challenge to determine the glucose clearance rate. In certain embodiments, a notification can be provided by the decision support engine 114 to confirm insulin administration based on the medical information. For example, the notification may ask the user to confirm the administration (e.g., "Did you administer rapid-acting insulin?"). The confirmed insulin administration may then form part of a glucose challenge to determine the glucose clearance rate.

[0210] In certain embodiments, the medical information may be provided by the user through manual entry, by providing a photograph through an application configured to recognize the type and amount of insulin, and / or by scanning a barcode. In some examples, the insulin information may be received via a convenient user interface provided by the application 106. In certain embodiments, the insulin information may be received via a wireless connection on a smart pen and / or from an insulin pump.

[0211] In certain embodiments, insulin administration can be determined using sensor data. The sensor data can include the user's glucose level, the change in the user's glucose level, the rate of change of the glucose level, and / or any other glucose metric. Certain embodiments may include other analyte data associated with one or more analytes other than glucose in the sensor data. Additional analytes can include potassium, lactate, ketones, pyruvate, dextrose, insulin, which may vary as a result of insulin administration. Changes in one or more additional analyte levels combined with changes in glucose levels can indicate insulin administration. For example, a decrease in glucose level and potassium level can indicate that insulin has been administered. As a result, glucose data and additional analyte data can be used to indicate that insulin has been administered and the type and / or dosage of insulin used as medical information to determine the glucose clearance rate.

[0212] In certain embodiments, the glucose challenge can be repeated periodically one or more times to determine the user's glucose clearance rate at different times. By way of example, the user's glucose clearance rate can vary for different activities (e.g., activity, exercise, sleep, diet, medications, etc.), and thus, in certain embodiments, multiple glucose clearance rates can be determined for a particular user following different user activities to obtain an average glucose clearance rate. The different activities may be automatically determined based on sensor data including additional analyte data and non-analyte data. For example, lactate data may be used in certain embodiments to determine that the user is exercising, consuming a meal, at rest and / or under stress. In certain embodiments, when determining the glucose clearance rate for various activities, an activity profile can be created for each of those activities, and each activity profile stores the corresponding glucose clearance rate. For example, the user's glucose clearance rate for exercise can be determined and used to generate an exercise activity profile. The decision support provided to the user in connection with the user's exercise can then be determined based on the glucose clearance rate determined for the user's exercise profile.

[0213] In certain embodiments, a user's glucose clearance rate can be determined periodically. In another example, a user's glucose clearance rate can be determined at regular intervals (e.g., weekly, monthly, bi - monthly, annually). In certain embodiments, a user's glucose clearance rate can be determined upon a change to user input 128 or metric 130. For example, a user's glucose clearance rate can be determined following a change in disease state (e.g., worsening or improvement of renal dysfunction). As described above, in certain embodiments, an additional analyte (e.g., potassium) can indicate a change in disease state and, based thereon, a glucose clearance rate associated with the change in disease state can be determined. In another example of a change to user input 128 or metric 130, in certain embodiments, a user's glucose clearance rate can be determined following a change in medication or treatment information (e.g., start or stop of a medication, change in medication type, dosage, timing, etc.). Additionally, in certain embodiments, a user's glucose clearance rate can be determined after a delay following the change (e.g., two weeks after the start of a new exercise program). In certain embodiments, a user's glucose clearance rate can be determined upon user request. For example, a user can request a glucose clearance rate determination via a convenient user interface provided by application 106.

[0214] In certain embodiments, a warning can be provided by decision - making support engine 114 in response to a change in a user's glucose clearance rate. For example, the decision - making support notification can be that the user's glucose clearance rate has increased or decreased since the previous glucose clearance rate determination.

[0215] In certain embodiments, the glucose clearance rate of a user can be determined using any of the above - described methods for determining a glucose clearance rate, including any combination of the above - described methods.

[0216] In block 506, method 500 continues by determining the likelihood of at least one non - typical glucose trend associated with the glucose clearance rate of a patient with renal dysfunction. As discussed, users with renal disease may experience different (i.e., non - typical) glucose trends compared to users without renal disease. Additionally, glucose trends may have different (i.e., non - typical) effects on users with renal disease compared to those without renal disease. For example, in the case of a user with renal disease, an increasing glucose trend may indicate a more significant risk of hyperglycemia based on the glucose clearance rate associated with the user's current renal dysfunction. In another example, a decreasing glucose trend may indicate a more significant risk of hypoglycemia based on the glucose clearance rate associated with the user's current renal dysfunction. Block 506 may, in certain embodiments, be performed by the decision support engine 114 illustrated in FIG. 1.

[0217] As mentioned, patients with renal disease may have a decreased glucose homeostasis, diabetes, and / or a higher risk of hyperglycemia and / or hypoglycemia as compared to patients without renal disease. Additionally, the treatment efficacy and / or treatment eligibility of patients with renal disease may be subject to change. Due to these effects of renal disease, the success of diabetes management in patients with renal disease may depend on the recognition of the presence and risk of non - typical glucose trends and effective treatment options.

[0218] Different methods for determining the likelihood of an atypical glucose trend associated with the glucose clearance rate can be used by the decision support engine 114. In certain embodiments, the decision support engine 114 can use a rule-based model to provide real-time decision support based on the determined likelihood of an atypical glucose trend. For example, the decision support engine 114 can apply rule statements to evaluate the presence of an atypical glucose trend in a patient, perform risk stratification on the patient (e.g., the patient has a high or low risk of hyperglycemia and / or hypoglycemia based on the glucose clearance rate), and / or identify risks associated with the patient's current atypical glucose trend. Any rule can be used by the decision support engine 114 alone and / or in combination with any other rule.

[0219] For example, one rule can be related to the patient's glucose clearance rate or a change in the patient's glucose clearance rate. Another rule can be based on the patient's glucose clearance rate and the current glucose level. For example, an extended glucose clearance rate and a low glucose level can be an atypical glucose trend that results in hypoglycemia. Another rule can be based on the patient's glucose clearance rate and the rate of change of the glucose level. For example, an extended glucose clearance rate and a rapid decrease in the glucose level can be an atypical glucose trend that results in hypoglycemia. Another rule can be based on the patient's glucose clearance rate and glucose metrics, TIR, average glucose, GMI, and / or glucose variability. Various other rules based on combinations of the user's glucose clearance rate with other glucose metrics such as TIR, average glucose, GMI, glucose variability, absolute maximum glucose level, absolute minimum glucose level, rate of change of the glucose level, and changes in such glucose metrics can also be used.

[0220] Specific rules may also be based on insulin administration history, whether the glucose clearance rate is exogenous insulin-dependent, and / or insulin administration information such as insulin on board. For example, insulin administration and the corresponding exogenous insulin-glucose clearance rate may exhibit an atypical glucose trend that results in hypoglycemia. In another example, a change in insulin administration (e.g., skipping an insulin dose) and the corresponding exogenous insulin-glucose clearance rate may exhibit an atypical glucose trend that results in hyperglycemia.

[0221] Specific rules may also utilize the user's glucose clearance rate as indicated by the user's activity profile, based on the glucose clearance rate for the corresponding activity. For example, when the user is exercising, the user's glucose clearance rate as indicated by the user's exercise activity profile is used to determine the likelihood of an atypical glucose trend. Specific rules may also include potassium metrics, lactate metrics, pyruvate metrics, and / or any sensor data as described with respect to Figure 3.

[0222] Rules can be defined and maintained by the decision support engine 114 in the reference library. For example, the reference library may maintain ranges of glucose levels, ranges of glucose level change rates, ranges of glucose clearance rates, etc. that can be mapped to glucose levels. In certain embodiments, such rules can be determined based on empirical studies and on analyzing historical patient records from the historical record database 112.

[0223] In certain embodiments, as an alternative to using a rule-based model, an AI model, such as a machine learning model, may be used to provide real-time decision-making support for the diagnosis and staging of kidney diseases. In certain embodiments, the decision support engine 114 may deploy one or more of these machine learning models to perform a determination of the likelihood of an atypical glucose trend. In particular, the decision support engine 114 obtains information from the user profile 118 associated with the user stored in the user database 110, characterizes the information about the user stored in the user profile 118 in terms of one or more features, and may use these features as inputs to such a model. Alternatively, the information provided by the user's profile 118 may be characterized by another entity and then the features may be provided to the decision support engine 114 to be used as inputs to the ML model. In certain embodiments, features associated with a patient may be used as inputs to one or more of the models to assess the likelihood of an atypical glucose trend in the patient.

[0224] In certain embodiments, features associated with a patient can be used as inputs to one or more models to identify whether the patient has a high or low risk of developing an atypical glucose trend within a particular period. In certain embodiments, features associated with a patient may be used as inputs to one or more of the models to identify risks associated with the patient's current atypical glucose trend (e.g., hyperglycemia and / or hypoglycemia). Details related to how one or more machine learning models can be trained to provide real-time decision-making support for atypical glucose trends are further considered in connection with FIG. 6.

[0225] As mentioned, in certain embodiments, in addition to glucose, other analyte data can be used by the decision support engine 114 in block 506 to generate an outlier glucose trend for a patient. Analyte data including glucose and potassium data, lactate data, pyruvate data, insulin data, cystatin C data, or ketone data (e.g., from measurements by the continuous analyte monitoring system 104) can be used as input to such machine learning models and / or rule-based models to predict the user's outlier glucose trend.

[0226] The decision support engine 114 can use machine learning models and / or rule-based models to generate an outlier glucose trend based on a continuous analysis of data for a patient collected over various periods (e.g., analyte data and in some cases non-analyte data). Analysis of data collected for a patient over various periods can provide insights into the direction of the outlier glucose trend, such as whether the patient's health is improving or deteriorating. For example, a patient with an outlier glucose trend may continue to be monitored (e.g., by continuously collecting the patient's data) to determine whether the outlier glucose trend is improving or deteriorating (e.g., whether the glucose level is returning to a healthy range). As an example, a comparison of outlier glucose trends including treatment for the outlier glucose trend may be monitored over a period (e.g., minutes, hours, days, weeks, months) and may indicate a reduction in the outlier glucose trend and / or an increase in the effectiveness of treatment (e.g., improvement in glucose homeostasis and / or diabetes and kidney disease management).

[0227] In block 508, method 500 continues by generating a decision support output based on the determined likelihood of at least one outlier glucose trend. Block 508 can be performed by the decision support engine 114 illustrated in FIG. 1 in certain embodiments.

[0228] In certain embodiments, the decision support output may include warnings of harmful glycemic events such as glucose levels exceeding a threshold, a rate of change exceeding a threshold, hyperglycemia, and / or hypoglycemia. The decision support engine 114 may output such warnings to the user (e.g., through the application 106). In certain embodiments, the decision support engine 114 can use one or more machine learning models trained based on patient-specific data and / or population data to provide warnings about harmful glycemic events. The algorithms and / or machine learning models can take into account at least one determined likelihood of an atypical glucose trend determined at block 508, as well as one or more inputs 128 and / or metrics 130 (e.g., including glucose metrics) described with respect to FIG. 3 for the patient to determine warnings about harmful glycemic events.

[0229] In certain embodiments, the decision support output may include one or more recommendations for treatment based on the determined likelihood of at least one atypical glucose trend. In particular, the decision support engine 114 can provide recommendations for the treatment or prevention of harmful glycemic events, such as pharmaceutical recommendations, dietary recommendations, lifestyle recommendations, medical intervention recommendations, or other recommendations for managing atypical glucose trends. The decision support engine 114 may output such recommendations for treatment to the user (e.g., through the application 106).

[0230] In certain embodiments, the decision support engine 114 may use one or more machine learning models trained based on patient-specific data and / or population data to provide recommendations for the treatment and / or prevention of atypical glucose trends. The algorithms and / or machine learning models can take into account at least one determined likelihood of an atypical glucose trend determined at block 508, as well as one or more inputs 128 and / or metrics 130 (e.g., glucose metrics) described with respect to FIG. 3 for the patient to determine optimal recommendations for the prevention and / or management of the patient's atypical glucose trend.

[0231] In certain embodiments, instead of using a machine learning model, the decision support engine 114 can use one or more decision trees to provide recommendations for the treatment or prevention of adverse glycemic events. The decision trees may be rule-based and may include one or more rules used to provide recommendations regarding the management of a patient's atypical glucose trends.

[0232] Recommendations for the treatment or prevention of atypical glucose trends may, in some cases, be based on the optimal glucose and insulin levels determined for the patient. In particular, in certain embodiments, one or more algorithms may be used to determine the optimal balance of the patient's glucose and insulin levels, which may then be used to form one or more recommendations for the patient regarding treatment (e.g., insulin and / or medications), diet, and / or lifestyle. Further, recommendations for the treatment and / or prevention of atypical glucose trends may, in some cases, be based on the management of kidney disease. Specifically, in certain embodiments, the optimal balance of potassium, glucose, and insulin levels for the patient may be determined and used to provide recommendations to the user for the prevention of atypical glucose trends. As an example, pharmaceutical recommendations for insulin administration can be based on both glucose levels and potassium levels, such as for managing glucose homeostasis and kidney disease.

[0233] Generally, in the case of patients with CKD, renal dysfunction affects insulin metabolism by impairing insulin clearance. Metabolic syndrome and / or diabetes, which often co-occur with renal dysfunction, can increase or induce insulin resistance, so higher doses of insulin may be required to lower glucose levels to a desired level (e.g., within a healthy range). However, renal dysfunction also impairs insulin clearance, which often results in insulin excess (i.e., long-term and significant insulin), so insulin administration may need to be adjusted to avoid hypoglycemia. Thus, in certain embodiments, the medical recommendation can include insulin administration recommendations that include the type, dosage, amount, activity rate, duration of activity, and / or timing of insulin.

[0234] In certain embodiments, the insulin administration recommendations can be for changes in the type, dosage, amount, activity rate, and / or duration of activity of insulin, and / or the timing of insulin administration, such as to avoid both short-term hyperglycemia and long-term hypoglycemia. Further, the insulin administration recommendations can be individualized for each patient to account for differences in insulin resistance across patients based on the patient's glucose clearance rate. As described above, since insulin resistance changes the ability of insulin to lower glucose levels, the glucose clearance rate can indicate the level of a patient's insulin resistance. Patients with higher insulin resistance may require different insulin dosages compared to another patient with a different level of insulin resistance to remove the same amount of glucose. In certain aspects, the individualized insulin dosage recommendations may be modified over time as insulin resistance increases or decreases in a patient. For example, the algorithm used by the decision support engine 114 can calculate the amount of insulin required to decrease glucose levels by a value X (e.g., X is a value greater than 0) based on the user's glucose clearance rate.

[0235] In certain embodiments, insulin administration recommendations may be for rapid-acting insulin and may be determined based on the progression of the user's kidney disease. For example, in patients with early kidney disease (e.g., GFR > 50 ml / min), normal doses of insulin without any adjustment may be recommended. In another example, in patients with moderate kidney disease (e.g., 50 > GFR > 10 ml / min), it may be determined and may be advisable that a dose reduction (e.g., 25% reduction) is optimal, and thus may be recommended. In yet another example, in patients with ESRD (e.g., 10 ml / min > GFR), a larger dose reduction (e.g., 50% reduction) may be optimal and may be determined to be recommended.

[0236] In certain embodiments, insulin administration recommendations may be for long-acting insulin. In certain cases, long-acting insulin may not necessarily be affected by renal dysfunction, at least to the extent that insulin administration adjustment is required, while in other cases, long-acting insulin may require insulin administration adjustment. For example, some long-acting insulins (e.g., glargine and levemir) may require insulin administration adjustment (e.g., 30% reduction) based on a specific progression of renal dysfunction (60 ml / min > GFR), and thus a reduced dose of insulin (e.g., 70% of the normal dose) may be recommended.

[0237] Furthermore, in certain embodiments, insulin administration recommendations may include recommendations regarding specific insulin dosages as well as dietary consumption, which may be referred to herein as combination dosages. For CKD patients, potassium regulation is important for reducing the risks of hyperkalemia and hypokalemia. Since potassium levels are decreased by insulin, the interaction between potassium, insulin, and glucose can be considered when providing insulin administration recommendations. Thus, in patients with renal dysfunction who are receiving insulin and have increased potassium levels, the algorithm can be used for insulin administration to avoid or treat hyperkalemia while also preventing hypoglycemia. For example, the algorithm can calculate the amount of insulin necessary to lower the potassium level by a value X (e.g., X is a value greater than 0), as well as the amount of glucose and the timing of glucose consumption to prevent hypoglycemia. In another example, combination dosages may be recommended to prevent hypokalemia and hypoglycemia.

[0238] In certain embodiments, pharmaceutical recommendations may include adjustments to other antihyperglycemic medications to reduce the likelihood of atypical glucose trends. Generally, renal dysfunction impairs the metabolism of other antihyperglycemic medications in addition to insulin metabolism. Typically, these other antihyperglycemic medications are more effective in reducing glucose levels in renal dysfunction and preventing hyperglycemia in patients with renal dysfunction compared to patients without renal dysfunction. Thus, for patients with renal dysfunction, adjustments to these antihyperglycemic medications may also be required to avoid hypoglycemia.

[0239] In certain embodiments, the medication recommendation may include an adjustment to the type, timing, and / or amount of the blood glucose regulating medication. For example, the recommendation may be to take a lower dosage of the blood glucose regulating medication. In another example, the recommendation may be to adjust the administration timing (e.g., morning administration, before meals, etc.). In yet another example, the recommendation may be to discontinue the use of the blood glucose regulating medication. Additionally, alone or in combination with other medication recommendations, the recommendation may be to consult with the HCP regarding the adjustment to the blood glucose regulating medication.

[0240] Some blood glucose regulating medications (e.g., metformin) are associated with the risk of lactic acidosis. Although rare, lactic acidosis is associated with high mortality. Additionally, an increased risk of lactic acidosis is associated with renal dysfunction. Due to this increased risk, dose reduction is recommended for mild renal dysfunction (e.g., 45 mL / min / 1.73 m2 > GFR). Also, metformin was contraindicated for patients with severe renal dysfunction (30 mL / min / 1.73 m2 > GFR). However, metformin has known renal protective properties. Therefore, when the risk of lactic acidosis is reduced, it is desirable to continue its use in patients with renal dysfunction. In certain embodiments, based on the input medication information and lactate level, the decision support engine 114 can determine that the risk of lactic acidosis is increased. In certain embodiments, the medication recommendation may be an adjustment to the dosage of metformin. For example, the recommendation may be to decrease the metformin dosage. In another example, the recommendation may be to discontinue metformin administration. In yet another example, the recommendation may be to consult with the HCP regarding the reduction or discontinuation of metformin administration.

[0241] In certain embodiments, the treatment recommendation may include a recommendation for dialysis for the patient. Dialysis is a treatment for kidney failure that removes unwanted toxins, waste products, and excess fluid from the body by filtering the patient's blood. Dialysis helps to maintain the balance of potassium, glucose, insulin, phosphorus, and sodium levels in the patient's body. In certain embodiments, dialysis may be recommended based on analyte levels, glucose metrics, glucose clearance rate, and / or any other analyte data.

[0242] In certain embodiments, the dietary recommendation may include recommending to the patient to consume a fixed amount of glucose at specific intervals (e.g., daily, weekly, etc.). For example, the patient may be recommended to ingest a certain amount of glucose per day based on the level of glucose that the patient's kidneys are clearing in real time. Monitoring glucose consumption can help to reduce the risk of atypical glucose trends while also ensuring that the patient is consuming enough glucose to maintain glucose homeostasis.

[0243] In certain embodiments, the dietary recommendations may include recommendations for the patient to increase the patient's glucose consumption to improve glucose homeostasis. For example, a patient may be recommended to increase glucose consumption if excessive insulin or decreased insulin clearance may result in an atypical glucose tendency of hypoglycemia. In certain embodiments, the dietary recommendations may relate to the timing and / or amount of glucose consumed. For example, a patient may be recommended to consume additional glucose following insulin administration to reduce the risk of hypoglycemia. In certain embodiments, the dietary recommendations may include recommendations for the patient to reduce the patient's glucose consumption to maintain glucose homeostasis. For example, a patient may be recommended to decrease glucose consumption if inappropriate insulin or insulin resistance may result in an atypical glucose tendency of hyperglycemia. Certain embodiments may be that the dietary recommendations may be made based on, or in conjunction with, pharmaceutical recommendations. For example, the recommendation may be to consume glucose and administer insulin. Certain embodiments may be that the dietary recommendations may be based on, or in conjunction with, lifestyle recommendations. For example, the recommendation may be to consume glucose and increase exercise.

[0244] In certain embodiments, the dietary recommendations may include recommendations for managing renal disease and glucose homeostasis by recommending that the patient consume or avoid consuming potassium and / or glucose. For example, a patient may comply with a low-potassium diet, and a recommendation to increase glucose consumption may include avoiding potassium consumption. In certain embodiments, the dietary recommendations may be based on, or in conjunction with, pharmaceutical recommendations. For example, a patient may be recommended to administer insulin to reduce the risk of hyperkalemia and, at the same time, reduce the risk of hyperglycemia, and be recommended to increase potassium consumption rather than glucose consumption.

[0245] In certain embodiments, the dietary recommendations may include recommendations for a patient based on the patient's adherence to a meal plan to maintain glucose homeostasis. For example, the recommendation may be to reduce glucose consumption and maintain glucose consumption within a set threshold. In certain embodiments, the dietary recommendations may include recommendations for a patient based on the type, timing, and / or composition of the meal. For example, the recommendation may be for the user to have a meal now. In another example, the recommendation may be for the user to have their last meal earlier in the evening (e.g., have dinner at 5:00 PM instead of 7:00 PM). In another example, the recommendation may be for the user to have a meal with less glucose. In another example, the recommendation may be for the user to consume a meal that includes complex carbohydrates to maintain glucose levels throughout the night. In yet another example, the recommendation may be for the user to have a smaller meal or skip a meal.

[0246] In certain embodiments, lifestyle recommendations (including, for example, exercise recommendations) may include recommendations to increase a patient's physical activity to maintain glucose homeostasis. Increasing physical activity is one way to lower glucose levels. Certain embodiments may be that the exercise recommendation may be to engage in physical activity. In certain embodiments, the exercise recommendation may be provided along with other recommendations (such as dietary and / or treatment recommendations). In certain embodiments, one or more models may determine when such a patient should engage in physical activity based on the patient's glucose level, renal function, and / or insulin level. The recommendations may include a modified physical activity schedule and / or additional breaks. In certain embodiments, other analyte data or non-analyte data, such as heart rate or respiratory rate data, may be used in combination with glucose data to provide such exercise recommendations. For example, glucose data, potassium data, lactate data, and heart rate may indicate that the user should take a break to maintain glucose homeostasis and manage renal disease. In another example, these analyte or non-analyte data may be used to guide specific exercise recommendations, such as the intensity and duration of exercise, to optimize glucose levels.

[0247] In certain embodiments, medical intervention recommendations may include recommendations for a patient to seek medical assistance. For example, in certain embodiments, the medical intervention recommendation may indicate to the patient that the patient needs to go to the emergency room immediately and / or needs to contact a physician. In certain other embodiments, the medical intervention recommendation can automatically alert the patient's healthcare provider (HCP) regarding the patient's condition for intervention by the HCP. In certain other embodiments, the medical intervention recommendation can alert healthcare providers to activate an ambulance service or emergency medical service to provide the patient with emergency pre-hospital care and stabilization and / or to transport the patient to definitive care. In certain embodiments, the decision support engine 114 may make medical intervention recommendations based on the patient's ability to seek medical assistance and / or the patient's access to medical assistance.

[0248] In certain embodiments, medical intervention recommendations may be based on, for example, blood pressure levels. Blood pressure monitoring over time can be used to determine whether a patient has developed hypertension. For example, if a patient's average blood pressure increases over a period of time (e.g., one month or two months), the medical intervention recommendation may be to seek medical treatment and / or contact a physician. In certain embodiments, the decision support engine 114 and / or the patient's physician may recommend that the patient initiate RAASi dosing and / or complete light exercise after meals to lower glucose levels and prevent the onset and / or progression of hypertension.

[0249] In certain embodiments, the atypical glucose trend generated at block 506 can be used to determine the risk of a patient's hospitalization / readmission. In certain embodiments, the atypical glucose trend generated at block 506 can be used to better understand the post-discharge stability of a patient when determining whether to discharge the patient. In certain embodiments, the atypical glucose trend generated at block 506 can be used to determine the level of care a patient should receive upon admission. For example, a patient admitted to a hospital can be one of many patients in the hospital. Thus, in certain aspects, the atypical glucose trend generated for a patient is compared to the atypical glucose trends generated for other patients in the hospital to better inform medical staff in the hospital where the patient ranks among other patients with respect to the level of care required and the urgency of attention required by the patient among other patients. This can be particularly important if a patient is at high risk of experiencing an acute life-threatening event within a short time after admission (e.g., as opposed to waiting more than four hours before receiving support or care by medical staff).

[0250] As described herein, the machine learning models deployed by the decision support engine 114 include one or more models trained by a training server system 140 as illustrated in FIG. 1 to provide various types of predictions as described in connection with FIGS. 4 and 5. FIG. 6 further details a technique for training one or more machine learning models to generate predictions associated with diabetes in patients with kidney disease, according to a particular embodiment of the present disclosure. Predictions associated with diabetes can include (1) predictions regarding the presence and / or severity of diabetes in patients with kidney disease (e.g., the user illustrated in FIG. 1), (2) identification of the risk of hyperglycemia and / or hypoglycemia, and / or (3) predictions regarding optimal treatment for patients with diabetes and kidney disease.

[0251] Method 600 begins, at block 602, by retrieving data from a historical record database, such as historical record database 112 illustrated in FIG. 1, by a training server system, such as training server system 140 illustrated in FIG. 1. As referred to herein, historical record database 112 can provide up-to-date and historical information for users of a continuous analyte monitoring system and connected mobile health applications, such as users of continuous analyte monitoring system 104 and application 106 illustrated in FIG. 1, as well as a repository of data for one or more patients who are not, or were not previously, users of continuous analyte monitoring system 104 and / or application 106. In certain embodiments, historical record database 112 can include one or more data sets of historical patients who (1) have neither diabetes nor kidney disease, (2) have neither diabetes nor various stages of kidney disease, (3) have various stages of diabetes but no kidney disease, or (4) have various stages of both diabetes and various stages of kidney disease.

[0252] The retrieval of data from the history record database 112 by the training server system 140 may include, at block 602, the retrieval of all or any subset of the information maintained by the history record database 112. For example, if the history record database 112 stores information about 100,000 patients (e.g., non-users and users of the continuous analyte monitoring system 104 and the application 106), the data retrieved by the training server system 140 to train one or more machine learning models may include information about all 100,000 patients, or only a subset of the data for those patients, e.g., data associated with only 50,000 patients, or only data from the past 10 years.

[0253] As an illustrative example, integrating with an in-house or cloud-based medical record database through, for example, Fast Healthcare Interoperability Resources (FHIR), a web application programming interface (API), Health Level 7 (HL7), and / or other computer interface languages may enable the aggregation of healthcare history records for baseline assessment in addition to the aggregation of anonymized patient data from a cloud-based repository.

[0254] As an illustrative example, at block 602, the training server system 140 may retrieve information about 100,000 patients having various stages of diabetes and / or various stages of kidney disease stored in the history record database 112 to train a model to predict the risk, presence, and / or severity of liver disease in a user. Each of the 100,000 patients may have a corresponding data record (e.g., based on their corresponding user profile) stored in the history record database 112. Each user profile 118 may include information such as that discussed with respect to FIG. 3.

[0255] Next, the training server system 140 uses the information in each of the records to train an artificial intelligence or ML model (referred to herein as an "ML model" for simplicity). Examples of the types of information included in a patient's user profile were provided above. The information in each of these records can be characterized (e.g., manually or by the training server system 140) to yield features that can be used as input features for training the ML model. For example, patient records can include or be used to generate features associated with a patient's age, patient's gender, patient's occupation, a patient's analyte levels over time, the rate of change and / or trend of a patient's analyte levels over time, physiological parameters associated with a patient's different kidney diseases and / or diabetes stages over time, and / or any information provided by input 128 and / or metric 130. The features used to train the machine learning model can vary in different embodiments.

[0256] In certain embodiments, each historical patient record retrieved from the historical record database 112 is further associated with a label indicating whether the patient was healthy or experienced some form of kidney disease and / or diabetes, a diagnosis of kidney disease and / or diabetes previously determined for the patient, and / or the stage of chronic kidney disease (CKD), a kidney disease risk assessment, a diabetes risk assessment, treatment, and / or similar metrics. What the record is labeled with depends on what the model is being trained to predict.

[0257] In block 604, method 600 continues by the training server system 140 training one or more machine learning models based on features and labels associated with historical patient records. In some embodiments, the training server does this by providing the features as inputs to the model. This model can be a new model initialized with random weights and parameters, or can be partially or fully pre-trained (e.g., based on a previous training round). Based on the input features, the model-in-training generates some output. In certain embodiments, the output can indicate a diagnosis of a kidney disease, a diagnosis of diabetes, the severity of diabetes, the type of diabetes, a risk assessment associated with the current diabetes diagnosis, a decrease in glucose homeostasis, and / or the level of risk that the patient has experienced hyperglycemia / hypoglycemia, the likelihood of an atypical glucose trend, decision support recommendations (e.g., treatment recommendations, diet recommendations, exercise recommendations, etc.). In certain embodiments, the output may indicate whether the recommended treatment is effective in reducing the risk (e.g., associated with an atypical glucose trend).

[0258] In certain embodiments, the training server system 140 compares this generated output to the actual labels associated with the corresponding historical patient records in order to calculate a loss based on the difference between the actual and the generated results. This loss is then used to refine one or more internal weights and parameters of the model (e.g., via backpropagation) so that the model learns to more accurately predict the presence, type, and / or severity of diabetes (or its recommended treatment).

[0259] One of various machine learning algorithms can be used to train the model described above. For example, one of a supervised learning algorithm, a neural network algorithm, a deep neural network algorithm, a deep learning algorithm, etc. can be used.

[0260] In block 606, the training server system 140 deploys the trained model to make predictions associated with diabetes at runtime. In some embodiments, this includes transmitting some indication (e.g., a weight vector) of the trained model that can be used to instantiate the model on another device. For example, the training server system 140 may transmit the weights of the trained model to the decision support engine 114. The model can then be used with the application 106 to assess the presence and / or severity of the user's diabetes in real time, provide treatment recommendations, and / or make other types of predictions discussed above. In certain embodiments, the training server system 140 may continue to train the model in an “online” fashion by using input features and labels associated with new patient records.

[0261] Furthermore, a similar method for training illustrated in FIG. 6 using historical patient records can also be used to train the model using patient-specific records to create a more personalized model for making predictions related to the non-standard glucose trends for patients with renal dysfunction. In particular, a model trained using historical patient records deployed for a particular user can be further retrained after deployment. For example, the model can be retrained after the model has been deployed for a particular patient to create a more personalized model for that patient. The more individualized model may be able to more accurately make non-standard glucose trend predictions for the patient based on the patient's own data (as opposed to only historical patient record data) including the patient's own glucose metrics.

[0262] FIG. 7 is a block diagram depicting a computing device 700 configured for (1) prediction regarding the presence and / or severity of diabetes in a patient having a kidney disease (e.g., the user illustrated in FIG. 1), (2) identification of the risk of hyperglycemia and / or hypoglycemia, and / or (3) prediction regarding optimal treatment for a patient. Although depicted as a single physical device, in embodiments, the computing device 700 may be implemented using virtual devices and / or across several devices such as a cloud environment. As illustrated, the computing device 700 includes a processor 705, a memory 710, a storage 715, a network interface 725, and one or more I / O interfaces 720. In the illustrated embodiment, the processor 705 fetches and executes programming instructions stored in the memory 710 and stores and fetches application data present in the storage 715. The processor 705 generally represents a single CPU and / or GPU, multiple CPUs and / or GPUs, a single CPU and / or GPU having multiple processing cores, etc. The memory 710 is generally included to represent random access memory (RAM). The storage 715 can be any combination of a disk drive, flash-based storage device, etc., and can include fixed and / or removable storage devices such as a fixed disk drive, removable memory card, cache, optical storage, network attached storage (NAS), or storage area network (SAN).

[0263] In some embodiments, input and output (I / O) devices 735 (such as keyboards, monitors, etc.) may be connected via an I / O interface 720. Further, via a network interface 725, computing device 700 can be communicatively coupled to one or more other devices and components, such as user database 110 and / or history record database 112. In certain embodiments, computing device 700 is communicatively coupled to other devices via a network that may include the Internet, a local network, etc. The network may include a wired connection, a wireless connection, or a combination of wired and wireless connections. As illustrated, processor 705, memory 710, storage 715, network interface 725, and I / O interface 720 are communicatively coupled by one or more interconnects 730. In certain embodiments, computing device 700 represents a display device 107 associated with a user. In certain embodiments, as discussed above, display device 107 can include a user's laptop, computer, smartphone, etc. In another embodiment, computing device 700 is a server running in a cloud environment.

[0264] In the illustrated embodiment, storage 715 includes user profile 118. Memory 710 includes decision support engine 114 which itself includes DAM 116. Decision support engine 114 is executed by computing device 700 to perform the operations of workflow 400 of FIG. 4, the operations of method 500 of FIG. 5, and / or the operations of method 600 of FIG. 6.

[0265] As described above, the continuous analyte monitoring system 104 described in connection with FIG. 1 can be a multi-analyte sensor system that includes a multi-analyte sensor. FIGS. 8-12 depict exemplary multi-analyte sensors used to measure multiple analytes.

[0266] As used herein, the terms "analyte measurement device", "analyte monitoring device", "analyte sensing device", and / or "multianalyte sensor device" are broad terms and are given their ordinary and customary meaning to one of ordinary skill in the art (and are not limited to a special or customized meaning), and refer to, but are not limited to, devices and / or systems that are responsible for the detection of a particular analyte or combination of analytes or the conversion of a signal associated therewith. For example, these terms can refer to, but are not limited to, devices that are responsible for the detection of a particular analyte or combination of analytes. In one embodiment, the device includes a sensor coupled to a circuit arranged within a housing and configured to process a signal associated with an analyte concentration into information. In one embodiment, such a device and / or system can use a biological recognition element combined with a conversion (detection) element to provide specific quantitative, semi-quantitative, qualitative, and / or semi-qualitative analysis information.

[0267] As used herein, the terms "biosensor" and / or "sensor" are broad terms and are given their ordinary and customary meaning to one of ordinary skill in the art (and are not limited to a special or customized meaning), and refer to, but are not limited to, a portion of an analyte measurement device, analyte monitoring device, analyte sensing device, and / or multianalyte sensing device that is responsible for the detection of a particular analyte or combination of analytes or the conversion of a signal associated therewith. In one embodiment, a biosensor or sensor generally comprises a body and a working electrode, a reference electrode, and / or a counter electrode coupled to the body and forming a surface configured to provide a signal during an electrochemical reaction. One or more membranes can be affixed to the body and can cover the electrochemically reactive surface. In one embodiment, such a biosensor and / or sensor can use a biological recognition element combined with a conversion (detection) element to provide specific quantitative, semi-quantitative, qualitative, semi-qualitative analysis signals.

[0268] As used herein, the terms "sensing portion", "sensing membrane", and "sensing mechanism" are broad terms and are given their ordinary and customary meaning to those of ordinary skill in the art (and are not limited to special or customized meanings), and refer to, but are not limited to, a biosensor and / or a part of a sensor that is responsible for detecting a specific analyte or combination of analytes or converting a signal associated therewith. In one embodiment, the sensing portion, sensing membrane, and / or sensing mechanism generally comprises an electrode configured to provide a signal during an electrochemical reaction with one or more membranes covering an electrochemically reactive surface. In one embodiment, such sensing portion, sensing membrane, and / or sensing mechanism can use a biological recognition element combined with a conversion (detection) element to provide specific quantitative, semi-quantitative, qualitative, semi-qualitative analysis information.

[0269] As used interchangeably herein, the terms "biomimetic membrane" and "biomimetic layer" are broad terms and are given their ordinary and customary meaning to those of ordinary skill in the art (and are not limited to special or customized meanings), and refer to, but are not limited to, a permeable membrane (which can include multiple domains) or layer that functions as a biodefense interface between a host tissue and an implantable device. The terms "biomimetic interface" and "biodefense" are used interchangeably herein.

[0270] As used herein, the term "cofactor" is a broad term and is given its ordinary and customary meaning to those of ordinary skill in the art (and is not limited to special or customized meanings), and non-limitingly refers to one or more substances whose presence contributes to or is required for the analyte-related activity of an enzyme, but is not limited thereto. Analyte-related activity can include, but is not limited to, any one or a combination of binding, electron transfer, and chemical conversion. Cofactors include coenzymes, non-protein chemical compounds, metal ions and / or metal-organic complexes. Coenzymes include prosthetic molecular groups and co-substrates.

[0271] As used herein, the term "continuous" is a broad term and is given its ordinary and customary meaning to one of ordinary skill in the art (and is not limited to a special or customized meaning), and refers to a portion, domain, coating, or layer that is uninterrupted or seamless, but is not limited thereto.

[0272] As used herein, the phrases "continuous analyte sensing" and "continuous multi - analyte sensing" are broad phrases and are given their ordinary and customary meaning to one of ordinary skill in the art (and are not limited to a special or customized meaning), and refer to a period during which monitoring of analyte concentration is carried out continuously, intermittently, and / or periodically (but regularly), for example, every period from about less than 1 second to about more than 1 week, but is not limited thereto. In further embodiments, the monitoring of analyte concentration is carried out every about 2, 3, 5, 7, 10, 15, 20, 25, 30, 35, 40, 45, 50, 55, or 60 seconds to about 1.25, 1.50, 1.75, 2.00, 2.25, 2.50, 2.75, 3.00, 3.25, 3.50, 3.75, 4.00, 4.25, 4.50, 4.75, 5.00, 5.25, 5.50, 5.75, 6.00, 6.25, 6.50, 6.75, 7.00, 7.25, 7.50, 7.75, 8.00, 8.25, 8.50, 8.75, 9.00, 9.25, 9.50, or 9.75 minutes. In further embodiments, the monitoring of analyte concentration is carried out every about 10, 20, 30, 40, or 50 minutes to about 1, 2, 3, 4, 5, 6, 7, or 8 hours. In further embodiments, the monitoring of analyte concentration is carried out every about 8 hours to about 12, 16, 20, or 24 hours. In further embodiments, the monitoring of analyte concentration is carried out every about 1 day to about 1.5, 2, 3, 4, 5, 6, or 7 days. In further embodiments, the monitoring of analyte concentration is carried out every about 1 week to about 1.5, 2, 3 weeks or more.

[0273] As used herein, the term "coaxial" should be broadly interpreted to include a sensor architecture having elements aligned along a common axis around a core that can be configured to have a circular, elliptical, triangular, polygonal, or other cross-section, such elements can include electrodes, insulating layers, or other elements that can be circumferentially located around a core layer such as a core electrode or a core polymer wire.

[0274] As used herein, the term "coupled" is a broad term and is given its ordinary and customary meaning to one of ordinary skill in the art (and is not limited to a special or customized meaning), and refers to two or more system elements or components that are configured to be attached to at least one of electrically, mechanically, thermally, operably, chemically, or otherwise, but is not limited thereto. For example, an element is "coupled" if it is covalently, communicatively, electrostatically, thermally, mechanically, magnetically, or ionically associated with, or physically captured, adsorbed, or absorbed by, another element. Similarly, as used herein, the phrases "operably connected," "operably linked," and "operably coupled" can refer to one or more components that are linked to another component in a manner that facilitates the transmission of at least one signal between the components. In some embodiments, the components are part of the same structure and / or integrated with each other as if they were covalently, electrostatically, mechanically, thermally, magnetically, ionically associated with, or physically captured or absorbed (i.e., "directly coupled" as if there were no intervening elements). In other embodiments, the components are connected via remote means. For example, one or more electrodes can be used to detect an analyte in a sample and convert that information into a signal, which can then be transmitted to an electronic circuit. In this example, the electrode is "operably linked" to the electronic circuit. As used herein, the phrase "removably coupled" can refer to two or more system elements or components that are configured to be attached and removed electrically, mechanically, thermally, operably, chemically, or otherwise, without damaging any of the coupled elements or components, or are so configured.As used herein, the phrase "permanently coupled" refers to two or more system elements or components that are configured or attached to be electrically, mechanically, thermally, operably, chemically, or otherwise attached, or are attached, and cannot be separated without damaging at least one of the coupled elements or components, and can be covalently, electrostatically, ionically associated, or physically trapped or absorbed.

[0275] As used herein, the term "discontinuous" is a broad term and is given its ordinary and customary meaning to those of ordinary skill in the art (and is not limited to a special or customized meaning), and refers to, but is not limited to, a cut, interrupted, or separated portion, layer, coating, or domain.

[0276] As used herein, the term "distal" is a broad term and is given its ordinary and customary meaning to those of ordinary skill in the art (and is not limited to a special or customized meaning), and refers to, but is not limited to, a region spaced relatively far from a reference point such as an origin or an attachment point.

[0277] As used herein, the term "domain" is a broad term and is given its ordinary and customary meaning to those of ordinary skill in the art (and is not limited to a special or customized meaning), and refers to, but is not limited to, a region of a membrane system that can be a layer, a homogeneous or inhomogeneous gradient (e.g., an anisotropic region of a membrane), or a portion of a membrane capable of sensing one, two, or more analytes. The domains discussed herein can be formed as a single layer, as two or more layers, as a bilayer pair, or as a combination thereof.

[0278] As used herein, the term "electrochemically reactive surface" is a broad term and is given its ordinary and customary meaning to one of ordinary skill in the art (and is not limited to a special or customized meaning), and refers to, but is not limited to, the surface of an electrode at which an electrochemical reaction occurs. In one example, this reaction is Faradaic and results in charge transfer between the surface and its environment. In one example, hydrogen peroxide produced by an enzyme-catalyzed reaction of an analyte oxidized on the surface results in a measurable electron current. For example, in the detection of glucose, glucose oxidase produces hydrogen peroxide (H2O2) as a byproduct. H2O2 reacts with the surface of the working electrode to produce two protons (2H + ), two electrons (2e - ), and one oxygen molecule (O2), which produces the electron current to be detected. At the counter electrode, a reducible species, such as O2, is reduced at the electrode surface to balance the current generated by the working electrode.

[0279] As used herein, the term "electrolysis" is a broad term and is given its ordinary and customary meaning to one of ordinary skill in the art (and is not limited to a special or customized meaning), and non-limitingly refers to, but is not limited to, the electrochemical oxidation or reduction (collectively "redox") of a compound, either directly or indirectly, by one or more enzymes, cofactors, or mediators.

[0280] As used herein, the terms "implanted," "being implanted," "embedded," or "implantable" are broad terms and are given their ordinary and customary meaning to one of ordinary skill in the art (and are not limited to a special or customized meaning), and refer to an object (e.g., a sensor) that is inserted or configured to be inserted subcutaneously (i.e., within the fat layer between the skin and muscle), intradermally (i.e., penetrating the stratum corneum and located within the epidermal or dermal layer of the skin), or transdermally (i.e., penetrating, entering, or passing through intact skin), which can result in a sensor having an in vivo portion and an ex vivo portion, but is not limited thereto. The term "implanted" also encompasses an object configured to be inserted subcutaneously, intradermally, or transdermally, whether or not it is itself inserted.

[0281] As used herein, the terms "interferent" and "interfering species" are broad terms and are given their ordinary and customary meaning to one of ordinary skill in the art (and are not limited to a special or customized meaning), and refer to an effect and / or species that interferes with the measurement of an analyte of interest in a sensor and produces a signal that does not accurately represent the analyte measurement, but is not limited thereto. In one example of an electrochemical sensor, an interfering species is a compound that produces a signal that is not analyte-specific due to a reaction on the electrochemically active surface.

[0282] As used herein, the term "in vivo" is a broad term and is given its ordinary and customary meaning to one of ordinary skill in the art (and are not limited to a special or customized meaning), and includes, but is not limited to, a portion of a device (e.g., a sensor) that is adapted for insertion into and / or presence within the body of a host.

[0283] As used herein, the term "ex vivo" is a broad term and is given its ordinary and customary meaning to one of ordinary skill in the art (and is not limited to a special or customized meaning), and includes, but is not limited to, a portion of a device (e.g., a sensor) adapted to remain and / or exist outside of the host organism.

[0284] As used herein, the terms and phrases "mediator" and "redox mediator" are broad terms and phrases and are given their ordinary and customary meaning to one of ordinary skill in the art (and are not limited to a special or customized meaning), and refer to any chemical compound or aggregate of compounds capable of directly or indirectly enabling electron transfer between an analyte, an analyte precursor, an analyte surrogate, an analyte-reducing enzyme or an analyte-oxidizing enzyme, or a cofactor, and an electrode surface maintained at a potential, but are not limited thereto. In one example, the mediator receives electrons from or transfers electrons to one or more enzymes or cofactors and / or exchanges electrons with the sensor system electrode. In one example, the mediator is a transition metal coordination organic molecule capable of reversible oxidation and reduction reactions. In other examples, the mediator can be an organic molecule or a metal capable of reversible oxidation and reduction reactions.

[0285] As used herein, the term "membrane" is a broad term and is given its ordinary and customary meaning to one of ordinary skill in the art (and is not limited to a special or customized meaning), and serves as a protection for the exposed electrode surface from the biological environment, a diffusion resistance (limitation) for the analyte, a matrix for a catalyst (e.g., one or more enzymes) to enable an enzymatic reaction, a limitation or blocking of interfering species, a provision of hydrophilicity at the electrochemically reactive surface of the sensor interface, a service as an interface between the host tissue and the implanted device, a regulation of the host tissue response via drug (or other substance) release, and combinations thereof, but is not limited thereto. Structures configured to perform functions including, but not limited to, those described above are referred to herein. When used herein, the terms "membrane" and "matrix" are meant to be interchangeable.

[0286] As used herein, the phrase "membrane system" is a broad phrase and is given its ordinary and customary meaning to one of ordinary skill in the art (and is not limited to a special or customized meaning), and can be composed of two or more domains, layers, or layers within a domain, and typically is composed of a material having a thickness of several microns or more, is permeable to oxygen, and optionally is permeable to, for example, glucose or another analyte, but is not limited thereto. In one embodiment, the membrane system includes an enzyme that enables an analyte reaction to occur, whereby the concentration of the analyte can be measured.

[0287] As used herein, the term "plane" should be broadly construed to describe a sensor architecture having a substrate that includes at least a first surface and an opposing second surface, and includes a plurality of elements disposed on one or more surfaces or edges of the substrate. The plurality of elements can include a conductive layer, an insulating layer, or elements configured to operate as a circuit. The plurality of elements may or may not be electrically or otherwise coupled. In one embodiment, the plane includes one or more edges that separate the opposing surfaces.

[0288] As used herein, the term "proximal" is a broad term and is given its ordinary and customary meaning to one of ordinary skill in the art (and is not limited to a special or customized meaning), and refers to, but is not limited to, the spatial relationship between various elements compared to a particular reference point. For example, some embodiments of a device include a membrane system having a biointerface layer and an enzyme domain or enzyme layer. If the sensor is considered the reference point and the enzyme domain is located closer to the sensor than the biointerface layer, the enzyme domain is more proximal to the sensor than the biointerface layer.

[0289] As used herein, the phrases "sensing portion", "sensing membrane", and "sensing mechanism" are broad phrases and are given their ordinary and customary meaning to one of ordinary skill in the art (and are not limited to a special or customized meaning), and refer to, but are not limited to, a biosensor and / or a part of a sensor that is responsible for the detection of a particular analyte or combination of analytes or the conversion of a signal associated therewith. In one embodiment, the sensing portion, sensing membrane, and / or sensing mechanism generally comprises an electrode configured to provide a signal during an electrochemical reaction with one or more membranes covering an electrochemically reactive surface. In one embodiment, such a sensing portion, sensing membrane, and / or sensing mechanism can use a biological recognition element combined with a conversion (detection) element to provide specific quantitative, semi-quantitative, qualitative, semi-qualitative analysis information.

[0290] During the general operation of an analyte measurement device, biosensor, sensor, sensing region, sensing part, or sensing mechanism, a biological sample, such as blood or interstitial fluid, or components thereof, directly or after passing through one or more membranes, comes into contact with an enzyme, such as glucose oxidase, DNA, RNA, or a protein or aptamer, such as one or more periplasmic binding proteins (PBPs) having one or more analyte binding regions or variants or fusion proteins thereof, and each region is capable of specifically or reversibly binding and / or reacting with at least one analyte. The interaction between the biological sample or its components and the analyte measurement device, biosensor, sensor, sensing region, sensing part, or sensing mechanism results in the conversion of a signal that enables qualitative, semi-qualitative, quantitative, or semi-quantitative determination of analyte levels in the biological sample, such as glucose, ketones, lactate, potassium, etc.

[0291] In one embodiment, the sensing region or sensing part can comprise at least a portion of a conductive substrate or at least a portion of a conductive surface, such as a wire (coaxial) or conductive trace, or a substantially planar substrate comprising a substantially planar trace, and a membrane. In one embodiment, the sensing region or sensing part can comprise a non-conductive body, a working electrode, a reference electrode, and an optional counter electrode that form an electrochemically reactive surface at one location on the body and an electrical connection at another location on the body, and a sensing membrane attached to the body and covering the electrochemically reactive surface. In some embodiments, the sensing membrane further comprises an enzyme domain, such as an enzyme domain, and an electrolyte phase, such as a free-flowing liquid phase comprising an electrolyte-containing fluid further described below. These terms are broad enough to include the entire device or only the sensing part thereof (or something in between).

[0292] In another embodiment, the sensing region can include one or more peripheral membrane-binding proteins (PBPs) that include the variant or fusion protein, or an aptamer having one or more analyte-binding regions, each region being capable of specifically and reversibly binding to at least one analyte. Changes in the aptamer or mutations in the PBP can contribute to or alter the long-term stability of the protein, including one or more binding constants, thermal stability, to bind the protein to a particular encapsulation matrix, membrane or polymer, or to attach a detectable reporter group or "label" to indicate a change in the binding region, or to convert a signal corresponding to one or more analytes present in a biological fluid. Specific examples of changes in the binding region include, but are not limited to, changes in hydrophobic / hydrophilic environmental conditions, three-dimensional conformational changes, changes in the orientation of amino / nucleic acid side chains in the binding region of the protein, and redox states of the binding region. Such changes to the binding region provide for the conversion of a detectable signal corresponding to one or more analytes present in a biological fluid.

[0293] In one embodiment, the sensing region determines the selectivity between one or more analytes such that only the analyte to be measured results in (converts) a detectable signal. This selection can be based on any chemical or physical recognition of the analyte by the sensing region, if the chemical composition of the analyte does not change, or if the sensing region causes or catalyzes an analyte reaction that changes the chemical composition of the analyte.

[0294] As used herein, the term "sensitivity" is a broad term and is given its ordinary and customary meaning to those of ordinary skill in the art (and is not limited to a special or customized meaning), and refers to the amount of signal (e.g., in the form of current and / or voltage) produced by a given quantity (unit) of the measured analyte, but is not limited thereto. For example, in one embodiment, the sensor has a sensitivity (or slope) of about 1 to about 100 picoamperes of current per mg / dL of analyte.

[0295] The terms "signal medium" or "transmission medium" should be construed to include any form such as a modulated data signal, a carrier wave, etc. The term "modulated data signal" means a signal in which one or more of its characteristics are set or changed so as to encode information in the signal.

[0296] As used herein, the terms "convert" or "conversion" and their grammatical equivalents are broad terms and are given their ordinary and customary meaning to one of ordinary skill in the art (and are not limited to a special or customized meaning), and refer to, but are not limited to, optical, electrical, electro-chemical, acoustic / mechanical, or colorimetric techniques and methods. Electrochemical properties include current and / or voltage, inductance, capacitance, impedance, and potential. Optical properties include absorbance, fluorescence / phosphorescence, fluorescence / phosphorescence decay rate, wavelength shift, dual wave phase modulation, bioluminescence / chemiluminescence, reflectivity, light scattering, and refractive index. For example, the sensing region converts the recognition of an analyte into a semi-quantitative or quantitative signal.

[0297] As used herein, the term "transducing element" is a broad term when used herein and is given its ordinary and customary meaning to one of ordinary skill in the art (and is not limited to a special or customized meaning), and refers to, but is not limited to, an analyte recognition moiety capable of directly or indirectly facilitating detectable signal transduction corresponding to the presence and / or concentration of a recognized analyte. In one embodiment, the transducing element is one or more enzymes, one or more aptamers, one or more ionophores, one or more capture antibodies, one or more proteins, one or more living cells, one or more oligonucleotides, and / or one or more DNA or RNA moieties. The transdermal continuous multi-analyte sensor can be used in vivo over various lengths of time. The continuous multi-analyte sensor system contemplated herein can be a transdermal device in that a portion of the device can be inserted into the soft tissue below the host's skin while a portion of the device remains on the surface of the host's skin. In one aspect, to overcome short-term noise or problems associated with other sensor functions, one embodiment uses a material that promotes the formation of a fluid pocket around the sensor, such as a porous biocompatible interfacial membrane or matrix that creates a space between the sensor and the surrounding tissue. In some embodiments, the sensor comprises a spacer adapted to provide a fluid pocket between the sensor and the host tissue. This spacer, such as a biocompatible material, matrix, structure, etc., described in more detail elsewhere herein, is thought to provide oxygen and / or glucose transport to the sensor.

[0298] Membrane system The membrane systems disclosed herein are suitable for use in implantable devices that contact biological fluids. For example, the membrane systems can be utilized with implantable devices for monitoring and determining analyte levels in biological fluids, such as devices for monitoring glucose levels in individuals with diabetes. In some embodiments, the analyte measurement device is a continuous device. The analyte measurement device can use any suitable sensing element for providing a biological signal, including but not limited to those involving elements such as enzymes, chemistry, physics, electrochemistry, spectrophotometry, amperometry, potentiometry, polarimetry, calorimetry, radiometry, immunochemistry, etc.

[0299] Suitable membrane systems for the multi - analyte systems and devices described above can include, for example, the membrane systems disclosed in U.S. Patent No. 6,015,572, U.S. Patent No. 5,964,745, and U.S. Patent No. 6,083,523, the teachings of which are hereby incorporated herein by reference in their entirety.

[0300] Generally, the membrane system includes a plurality of domains, such as an electrode domain, an interference domain, an enzyme domain, a resistance domain, and a bio - interface domain. The membrane system can be deposited on an exposed electroactive surface using known thin - film techniques (e.g., vapor deposition, spraying, electrodeposition, dipping, brush coating, film coating, droplet coating, etc.). Additional steps can be applied, such as drying, annealing, and curing (e.g., UV curing, thermal curing, moisture curing, radiation curing, etc.) to improve specific properties such as mechanical properties, signal stability, and selectivity, following the deposition of the membrane material. In a typical process, when depositing the resistance domain membrane, a bio - interface / drug - release layer with a “dry - film” thickness of about 0.05 microns (μm) or less to about 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, or 16 μm is formed. The “dry - film” thickness refers to the thickness of a cured film cast from a coating formulation by standard coating techniques.

[0301] In certain embodiments, the biointerface / drug release layer is formed from a biointerface polymer, which includes one or more membrane domains comprising polyurethane and / or polyurea segments and one or more zwitterionic repeating units. In some embodiments, the biointerface / drug release layer coating is formed from a polyurethane urea having carboxybetaine groups and nonionic hydrophilic polyethylene oxide segments incorporated into the polymer, and the polyurethane urea polymer is dissolved in an organic solvent system or a non-organic solvent system according to a predetermined coating formulation, crosslinked with an isocyanate crosslinking agent, and cured at a medium temperature of about 50 °C. The solvent system can be a single solvent or a mixture of solvents useful for dissolving or dispersing the polymer. The solvent can be one selected as the polymerization medium or added after polymerization is complete. The solvent is selected from those having a lower boiling point to facilitate drying and be less toxic for implant applications. Examples of these solvents include aliphatic ketones, esters, ethers, alcohols, hydrocarbons, etc. Depending on the final thickness of the biointerface / drug release layer and the solution viscosity (related to the percentage of polymer solids), the coating can be applied in a single step or multiple iterative steps of a selected process such as dipping to build the desired thickness. In still other embodiments, the bioprotective polymer is formed from a polyurethane urea having carboxylic acid groups and carboxybetaine groups incorporated into the polymer and a nonionic hydrophilic polyethylene oxide segment, and the polyurethane urea polymer is dissolved in an organic solvent system or a non-organic solvent system in the coating formulation, crosslinked with a carbodiimide (e.g., 1-ethyl-3-(3-dimethylaminopropyl)carbodiimide (EDC)), and cured at a medium temperature of about 50 °C.

[0302] In other embodiments, the biointerface / drug release layer coating is formed from a polyurethane urea having sulfobetaine groups and nonionic hydrophilic polyethylene oxide segments incorporated into the polymer. The polyurethane urea polymer is dissolved in an organic or non-organic solvent system according to a predetermined coating formulation, crosslinked with an isocyanate crosslinking agent, and cured at a moderate temperature of about 50°C. The solvent system can be a single solvent or a mixture of solvents useful for dissolving or dispersing the polymer. The solvent can be one selected as the polymerization medium or added after polymerization is complete. The solvent is selected from those having a lower boiling point to facilitate drying and be less toxic for implant applications. Examples of these solvents include aliphatic ketones, esters, ethers, alcohols, hydrocarbons, etc. Depending on the final thickness of the biointerface / drug release layer and the solution viscosity (related to the percentage of polymer solids), the coating can be applied in a single step or multiple iterative steps of a selected process such as dipping to build the desired thickness. In still other embodiments, the biointerface polymer is formed from a polyurethane urea having unsaturated hydrocarbon groups and sulfobetaine groups incorporated into the polymer, as well as nonionic hydrophilic polyethylene oxide segments. The polyurethane urea polymer is dissolved in an organic or non-organic solvent system in the coating formulation and crosslinked in the presence of an initiator by irradiation including heat or UV, LED light, electron beam, etc., and cured at a moderate temperature of about 50°C. Examples of unsaturated hydrocarbons include allyl groups, vinyl groups, acrylates, methacrylates, alkenes, alkynes, etc.

[0303] In some embodiments, a tether is used. The tether is a polymer or chemical moiety that does not participate in the (electro)chemical reaction involved in sensing but forms a chemical bond with the (electro)chemical active component of the membrane. In some embodiments, these bonds are covalent bonds. In one embodiment, the tether can be formed in solution before one or more intermediate layers of the membrane are formed, and the tether either directly binds two (electro)chemical active components to each other or, alternatively, the tether binds the (electro)chemical active component to a polymer backbone structure. In another embodiment, the (electro)chemical active components are mixed with a crosslinking agent (and optionally a polymer) having an adjustable length, and the tethering reaction occurs as an in situ crosslinking. Tethering can be used to maintain a predetermined degree of freedom of NAD(P)H for effective enzyme catalysis, and "effective" enzyme catalysis allows the analyte sensor to continuously monitor one or more analytes over a period of about 5 days to about 15 days or more.

[0304] Membrane Manufacturing Polymers can be processed by solution-based techniques such as, for example, spraying, dipping, casting, electrospinning, vapor deposition, spin coating, coating, etc. Aqueous polymer emulsions can be manufactured to form membranes in a manner similar to that used for solvent-based materials. In both cases, the evaporation of the volatile liquid (e.g., an organic solvent or water) leaves behind a film of the polymer. Crosslinking of the deposited film or layer can be carried out through the use of multifunctional reactive components in many ways. The liquid system can be cured by heat, moisture, high-energy radiation, ultraviolet light, or by completing the reaction, which produces the final polymer either in a mold or on a substrate being coated.

[0305] In some embodiments, the wetting characteristics of the membrane (and thus the degree of sensor drift indicated by the sensor) can be adjusted and / or controlled by creating covalent cross-links between a surface-active group-containing polymer, a functional-group-containing polymer, a polymer having zwitterionic groups (or a precursor or derivative thereof), and combinations thereof. The cross-linking can have a substantial effect on the film structure, which can in turn affect the surface wetting characteristics of the film. The cross-linking can also affect the tensile strength, mechanical strength, water absorption rate, and other properties of the film.

[0306] Cross-linked polymers can have different cross-linking densities. In certain embodiments, cross-linking agents are used to facilitate cross-linking between layers. In other embodiments, heat is used to form cross-links instead of (or in addition to) the cross-linking techniques described above. For example, in some embodiments, imide bonds and amide bonds can be formed between two polymers as a result of high temperature. In some embodiments, photo-cross-linking is performed to form covalent bonds between a polycationic layer and a polyanionic layer. One major advantage of photo-cross-linking is that it provides the possibility of patterning. In certain embodiments, patterning using photo-cross-linking is performed to modify the film structure and thus to adjust the wetting characteristics of the membrane and membrane systems, as discussed herein.

[0307] Polymers having domains or segments functionalized to enable crosslinking can be made by at least the methods contemplated herein. For example, a polyurethane urea polymer having an aromatic segment or an aliphatic segment with an electrophilic functional group (e.g., a carbonyl group, an aldehyde group, an anhydride group, an ester group, an amide group, an isocyano group, an epoxy group, an allyl group, or a halo group) can be crosslinked with a crosslinking agent having a plurality of nucleophilic groups (e.g., a hydroxyl group, an amine group, a urea group, a urethane group, or a thiol group). In a further embodiment, a polyurethane urea polymer having an aromatic segment or an aliphatic segment with a nucleophilic functional group can be crosslinked with a crosslinking agent having a plurality of electrophilic groups. Further, a polyurethane urea polymer having a hydrophilic segment with a nucleophilic functional group or an electrophilic functional group can be crosslinked with a crosslinking agent having a plurality of electrophilic groups or nucleophilic groups. Unsaturated functional groups on the polyurethane urea can also be used for crosslinking by reacting with a polyvalent free radical agent. Non-limiting examples of suitable crosslinking agents include isocyanates, carbodiimides, glutaraldehyde, aziridines, silanes, or other aldehydes, epoxies, acrylates, free radical-based agents, ethylene glycol diglycidyl ether (EGDE), poly(ethylene glycol) diglycidyl ether (PEGDE), or dicumyl peroxide (DCP). In one embodiment, about 0.1% to about 15% w / w of the crosslinking agent is added relative to the total dry weight of the crosslinking agent and polymer added when the components are blended. In another embodiment, about 1% to about 10% w / w of the crosslinking agent is added relative to the total dry weight of the crosslinking agent and polymer added when the components are blended. In yet another embodiment, about 5% to about 15% w / w of the crosslinking agent is added relative to the total dry weight of the crosslinking agent and polymer added when the components are blended. During the curing process, it is believed that substantially all of the crosslinking agent reacts to leave substantially no detectable unreacted crosslinking agent in the final film.

[0308] The polymers disclosed in this specification can be formulated into a mixture that can be stretched into a film or applied to a surface using methods such as self-assembling monolayer (SAM), spraying, coating, dip coating, vapor deposition, molding, 3D printing, lithography techniques (e.g., photolithography), micro- and nanopipetting printing techniques, silk screen printing, and so on. The mixture can then be cured at a high temperature (e.g., about 30 °C to about 150 °C). Other suitable curing methods can include, for example, ultraviolet light, electron beam, or gamma rays.

[0309] In some situations, using a continuous multi-analyte monitoring system that includes sensors composed of a bioprotective membrane and / or a drug release membrane, it is considered that the foreign body reaction can be managed or manipulated to be a major event surrounding the extended embedding of the implanted device and to support rather than interfere with or block analyte transport. In another aspect, to extend the lifespan of the sensor, one example uses a material that promotes intussusceptive angiogenesis within, for example, a porous biocompatible membrane. For example, intussusceptive angiogenesis within the porous biocompatible material surrounding the sensor can promote sensor function over a long period (e.g., weeks, months, or years). It has been observed that intussusceptive angiogenesis and formation of the tissue bed can take up to 3 weeks. Intussusceptive angiogenesis and tissue bed formation are considered to be part of the foreign body reaction. As discussed herein, the foreign body reaction can be manipulated by the use of a porous bioprotective material that surrounds the sensor and promotes intussusceptive angiogenesis of the tissue and microvasculature over time.

[0310] Therefore, a sensor as contemplated in the examples of this specification can include a biocompatible layer. The biocompatible layer can include, but is not limited to, a porous biocompatible material that includes, for example, a solid portion and interconnected cavities, all of which are described in more detail elsewhere in this specification. The biocompatible layer can be used to improve sensor function over the long term (e.g., after intussusceptive angiogenesis).

[0311] Accordingly, a sensor as contemplated in the examples herein can include a drug release membrane that functions at least in part as, or in combination with, a biological interface membrane. The drug release membrane can include, for example, materials that include hard-soft segment polymers having hydrophilic domains and optionally hydrophobic domains, all of which are described in more detail elsewhere herein and can be used to improve sensor functionality over a long period (e.g., after in-growth within tissue). In one example, a material that includes a hard-soft segment polymer having hydrophilic domains and optionally hydrophobic domains is configured to release a combination of dexamethasone or a derivative form of dexamethasone acetate and dexamethasone such that one or more different release rates of an anti-inflammatory drug are achieved and the service life of the sensor is extended. Other suitable drug release membranes of the present disclosure include silicone polymers, polytetrafluoroethylene, expanded polytetrafluoroethylene, ethylene tetrafluoroethylene copolymer, polyolefin, polyester, polycarbonate, biostable polytetrafluoroethylene, homopolymers, copolymers, terpolymers of polyurethane, polypropylene (PP), polyvinyl chloride (PVC), polyvinylidene fluoride (PVDF), polyvinyl alcohol (PVA), ethylene vinyl acetate (EVA), polybutylene terephthalate (PBT), polymethylmethacrylate (PMMA), polyether ether ketone ether (polyether etherketone, PEEK), polyamide, polyurethane and copolymers and blends thereof, polyurethane urea polymers and copolymers and blends thereof, cellulose polymers and copolymers and blends thereof, poly(ethylene oxide) and copolymers and blends thereof, poly(propylene oxide) and copolymers and blends thereof, polysulfone and block copolymers (including, for example, diblock, triblock, alternating, random and graft copolymers cellulose), hydrogel polymers, poly(2-hydroxyethyl methacrylate) (pHEMA) and copolymers and blends thereof, hydroxyethyl methacrylate (HEMA) and copolymers and blends thereof, polyacrylonitrile-polyvinyl chloride (PAN-PVC) and copolymers and blends thereof, acrylic copolymers and copolymers and blends thereof, nylon and copolymers and blends thereof, polyvinylidene fluoride, polyanhydrides, poly(l-lysine), poly(L-lactic acid), pans and copolymers and blends thereof, hydroxyapatite and copolymers and blends thereof.

[0312] Exemplary multi-analyte sensor membrane configurations Continuous multi-analyte sensors are provided that have various membrane configurations suitable for facilitating signal conversion corresponding to analyte concentration simultaneously, intermittently, and / or sequentially. In one embodiment, such a sensor can be configured using a signal transducer comprising one or more transducing elements ("TL"). Such continuous multi-analyte sensors can use various transduction means, such as, among other techniques, amperometric measurements, coulometric measurements, potentiometric measurements, and impedance measurement methods.

[0313] In one embodiment, the conversion element includes one or more membranes that can include one or more layers and / or domains, and each of the one or more layers and / or domains can independently include one or more signal converters, such as enzymes, RNA, DNA, aptamers, binding proteins, and the like. As used herein, the conversion element includes enzymes, ionophores, RNA, DNA, aptamers, binding proteins, and is used interchangeably.

[0314] In one embodiment, the conversion element is present within one or more membranes, layers, or domains formed over the sensing region. In one embodiment, such a sensor can be constructed using a membrane domain that includes one or more enzyme domains, such as an enzyme domain also referred to as an EZ layer (“EZL”), and each enzyme domain can include one or more enzymes. References to the following “enzyme layer” are intended to include all or part of the enzyme domain, any of which can be all or part of a membrane system as contemplated herein, for example, as a single layer, as two or more layers, as a pair of bilayers, or as a combination thereof.

[0315] In one embodiment, the continuous multi - analyte sensor uses one or more of the following analyte - substrate / enzyme pairs, for example, sarcosine oxidase combined with creatinine amidohydrolase, the creatinine amidohydrolase used for the sensing of creatinine. Other examples of analyte / oxidase enzyme combinations that can be used in the sensing region are, for example, alcohol / alcohol oxidase, cholesterol / cholesterol oxidase, galactose:galactose / galactose oxidase, choline / choline oxidase, glutamate / glutamate oxidase, glycerol / glycerol - 3 - phosphate oxidase (or glycerol oxidase), bilirubin / bilirubin oxidase, ascorbic acid / ascorbic acid oxidase, uric acid / uric acid oxidase, pyruvate / pyruvate oxidase, hypoxanthine:xanthine / xanthine oxidase, glucose / glucose oxidase, lactate / lactate oxidase, L - amino acid oxidase, and glycine / sarcosine oxidase. Other analyte - substrate / enzyme pairs can be used that include analyte - substrate / enzyme pairs such as those that include recombinant enzymes, immobilized enzymes, mediator - wired enzymes, dimerizing enzymes, and / or fusion enzymes.

[0316] NAD - based multi - analyte sensor platform Nicotinamide adenine dinucleotide (NAD(P) + / NAD(P)H) is a dinucleotide consisting of two nucleotides linked through their phosphate groups. One nucleotide contains the adenine nucleobase and the other contains nicotinamide. NAD exists in two forms, for example, the oxidized form (NAD(P)+) and the reduced form (NAD(P)H) (H = hydrogen). The reaction between NAD+ and NADH is reversible, and thus the coenzyme can continuously cycle between the NAD(P) + / form and the NAD(P)H form without being essentially consumed.

[0317] In one embodiment, one or more enzyme domains of the sensing region of the continuous multi - analyte sensor device of the present disclosure include an amount of NAD+ or NADH to provide a detectable signal conversion corresponding to the presence or concentration of one or more analytes. In one embodiment, one or more enzyme domains of the sensing region of the continuous multi - analyte sensor device of the present disclosure include an excess amount of NAD+ or NADH to provide an enhanced conversion of a detectable signal corresponding to the presence or concentration of one or more analytes.

[0318] In one embodiment, NAD, NADH, NAD+, NAD(P)+, ATP, flavin adenine dinucleotide (FAD), magnesium (Mg++), pyrroloquinoline quinone (PQQ), and their functionalized derivatives can be used in combination with one or more enzymes in a continuous multi - analyte sensor device. In one embodiment, NAD, NADH, NAD+, NAD(P)+, ATP, flavin adenine dinucleotide (FAD), magnesium (Mg++), pyrroloquinoline quinone (PQQ), and their functionalized derivatives are incorporated into the sensing region. In one embodiment, NAD, NADH, NAD+, NAD(P)+, ATP, flavin adenine dinucleotide (FAD), magnesium (Mg++), pyrroloquinoline quinone (PQQ), and their functionalized derivatives are dispersed or distributed in one or more membranes or domains of the sensing region.

[0319] In one aspect of the present disclosure, the sequential sensing of one or more or two or more analytes using NAD+-dependent enzymes is provided in one or more membranes or domains of the sensing region. In one example, the membrane or domain provides a mechanism for retaining and stably recycling NAD+ and for converting NADH oxidation or NAD+ reduction into a current measurable by amperometry. In one example described below, a sequential sensing of multiple analytes is provided, where one or more of them are reversibly bound or at least one of them is either oxidized or reduced by an NAD+-dependent enzyme, such as ketone (beta-hydroxybutyrate dehydrogenase), glycerol (glycerol dehydrogenase), cortisol (11β-hydroxysteroid dehydrogenase), glucose (glucose dehydrogenase), alcohol (alcohol dehydrogenase), aldehyde (aldehyde dehydrogenase), and lactate (lactate dehydrogenase). In other examples described below, membranes are provided that enable sequential on-body sensing of multiple analytes using FAD-dependent dehydrogenases, such as fatty acid (acyl-CoA dehydrogenase).

[0320] Exemplary configurations of one or more membranes or portions thereof are provided for providing retention and reuse of NAD⁺. Thus, the electrode surface of a conductive wire (coaxial) or a planar conductive surface is coated with at least one layer containing at least one enzyme, as depicted in FIG. 8A. Referring to FIG. 8B, one or more optional layers can be positioned between the electrode surface and one or more enzyme domains. For example, one or more interference domains (also referred to as an “interferent blocking layer”) can be used to reduce or eliminate signal contributions from unwanted species present, or one or more electrodes (not shown) can be used to assist with wetting, system equilibration, and / or startup. As shown in FIGS. 8A and 8B, one or more membranes provide a NAD⁺ reservoir domain that provides a reservoir for NAD⁺. In one embodiment, one or more interferent blocking membranes are used and a potentiostat is utilized to measure H₂O₂ production or O₂ consumption of an enzyme such as or similar to NADH oxidase, and the NAD⁺ reservoir and enzyme domain positions can be switched to facilitate better consumption of excess NAD⁺ and slower unwanted outward diffusion. Exemplary sensor configurations can be found in U.S. Patent Application No. 63 / 321340, “CONTINUOUS ANALYTE MONITORING SENSOR SYSTEMS AND METHODS OF USING THE SAME,” filed Mar. 18, 2022, which is hereby incorporated by reference in its entirety.

[0321] In one embodiment, one or more mediators optimal for NADH oxidation are incorporated into one or more electrode domains or enzyme domains. In one embodiment, organic mediators such as phenanthroline dione or nitrosoaniline are used. In another embodiment, organometallic mediators such as ruthenium-phenanthroline-dione or osmium(bpy)2Cl, polymers containing covalently bonded organic mediators or organometallic coordination mediator polymers, such as polyvinylimidazole-Os(bpy)2Cl, or polyvinylpyridine-organometallic coordination mediators (including ruthenium-phenanthroline dione) are used. Other mediators can be used as further discussed below.

[0322] In humans, the serum levels of beta-hydroxybutyrate (BHB) are normally in the low micromolar range but can increase up to about 6 - 8 mM. The serum levels of BHB can reach 1 - 2 mM after intense exercise or non-conflicting levels above 2 mM are achieved by a ketogenic diet that contains little carbohydrate. Other ketones such as acetoacetate and acetone are present in serum, but most of the dynamic range in ketone levels is in the form of BHB. Thus, monitoring of BHB, e.g., continuous monitoring, is useful for providing health information to the user or healthcare provider.

[0323] Another embodiment of a continuous ketone analyte detection configuration using an electrode-related mediator-bound diaphorase / NAD+ / dehydrogenase is depicted below.

[0324]

Chemical formula

[0325] In one embodiment, the diaphorase is electrically coupled to an electrode having an organometallic coordination mediator polymer. In another embodiment, the diaphorase is covalently coupled to an electrode having an organometallic coordination mediator polymer. Alternatively, multiple enzyme domains can be used in the enzyme layer to essentially separate NADH oxidation from analyte (ketone) oxidation, e.g., to separate the electrode-associated diaphorase (closest to the electrode surface) from more distal adjacent NAD+ or dehydrogenase enzymes. Alternatively, NAD+ can be closer to the electrode surface than an adjacent enzyme domain containing a dehydrogenase enzyme. In one embodiment, NAD+ and / or HBDH are present in the same or different enzyme domains, and both can be immobilized, e.g., using an amine-reactive crosslinking agent (e.g., glutaraldehyde, epoxide, NHS ester, imide ester). In one embodiment, NAD+ is bound to a polymer and is present in the same or different enzyme domain as HBDH. In one embodiment, the molecular weight of NAD+ is increased to prevent or eliminate migration from the sensing region, e.g., NAD+ is dimerized using its C6-terminal amine with any amine-reactive crosslinking agent. In one embodiment, NAD+ can be covalently coupled to an aspect of an enzyme domain having a higher molecular weight than NAD+, which can improve the stability profile of NAD+ and improve the ability to retain and / or immobilize NAD+ in the enzyme domain. For example, dextran-NAD.

[0326] In one embodiment, the sensing region includes one or more NADH, acceptor oxidoreductase, and one or more NAD-dependent dehydrogenases. In one embodiment, the sensing region includes one or more NADH, acceptor oxidoreductase, and one or more NAD(P)-dependent dehydrogenases having NAD(P)+ or NAD(P)H as a cofactor present in the sensing region. In one embodiment, the sensing region includes an amount of diaphorase.

[0327] In one embodiment, a ketone sensing configuration suitable for combination with another analyte sensing configuration is provided. Thus, an EZL layer about 1-20 μm thick is prepared by providing an EZL solution composition in 10 mM HEPES in water having about 20 μL of 500 mg / mL HBDH, about 20 μL of [500 mg / mL NAD(P)H, 200 mg / mL polyethylene glycol-diglycol ether (PEG-DGE) of about 400 MW], about 20 μL of 500 mg / mL diaphorase, and adding about 40 μL of 250 mg / mL polyvinylimidazole-osmium bis(2,2'-bipyridine) chloride (PVI-Os(bpy)2Cl) to a substrate such as a working electrode, and after drying to provide about 15-40 wt% HBDH, about 5-30% diaphorase, about 5-30% NAD(P)H, about 10-50% PVI-Os(bpy)2Cl, and about 1-12% PEG-DGE (400 MW). Substrates contemplated herein that may include the working electrode are gold, platinum, palladium, rhodium, iridium, titanium, tantalum, chromium, and / or their alloys or combinations, or carbon (e.g., graphite, vitreous carbon, carbon nanotubes, graphene, or doped diamond, and combinations thereof).

[0328] The above enzyme domain was contacted with a resistance domain, also referred to as a resistance layer ("RL"). In one embodiment, the RL comprises about 55-100% PVP and about 0.1-45% PEG-DGE. In another embodiment, the RL comprises about 75-100% PVP and about 0.3-25% PEG-DGE. In yet another embodiment, the RL comprises about 85-100% PVP and about 0.5-15% PEG-DGE. In yet another embodiment, the RL consists essentially of 100% PVP.

[0329] The exemplary continuous ketone sensor depicted in FIGS. 8A and 8B, which includes a NAD(P)H reservoir domain, is configured such that NAD(P)H is not rate-limiting in any of the enzyme domains of the sensing region. In one embodiment, the loading of NAD(P)H in the NAD(P)H reservoir domain is greater than about 20%, 30%, 40% or 50% w / w. One or more membranes or portions of one or more membrane domains (hereinafter also referred to as "membrane") may also contain a polymer or protein binder such as zwitterionic polyurethane and / or albumin. Alternatively, in addition to NAD(P)H, the membrane may contain one or more analyte-specific enzymes (e.g., HBDH, glycerol dehydrogenase, etc.), such that, optionally, the NAD(P)H reservoir membrane also provides a catalytic function. In one embodiment, NAD(P)H is dispersed or distributed in or with a polymer (or protein) and crosslinked to such an extent that proper enzyme / cofactor function and / or reduced NAD(P)H flux within the domain is still possible.

[0330] In one embodiment, the NADH oxidase enzyme is used in one or more membranes of the sensing region, either alone or in combination with superoxide dismutase (SOD). In one embodiment, an amount of superoxide dismutase (SOD) capable of scavenging some or most of the one or more free radicals generated by NADH oxidase is used. In one embodiment, the NADH oxidase enzyme is used in combination with a functionalized polymer having NAD(P)H immobilized on the polymer from NAD(P)H and / or the C6-terminus amine in one or more membranes of the sensing region, either alone or in combination with superoxide dismutase (SOD).

[0331] In one embodiment, NAD(P)H is immobilized to the extent that it maintains its NAD(P)H catalytic function. In one embodiment, dimeric NAD(P)H is used to capture NAD(P)H in one or more membranes by crosslinking their respective C6-terminal amines with a suitable amine-reactive crosslinking agent such as glutaraldehyde or PEG-DGE.

[0332] The continuous ketone sensor configuration described above can be adapted for other analytes or used in combination with other sensor configurations. For example, an analyte-dehydrogenase enzyme combination can be used in any of the membranes of the sensing region including glycerol (glycerol dehydrogenase), cortisol (11β-hydroxysteroid dehydrogenase), glucose (glucose dehydrogenase), alcohol (alcohol dehydrogenase), aldehyde (aldehyde dehydrogenase), and lactate (lactate dehydrogenase).

[0333] In one embodiment, the semipermeable membrane is used in the sensing region, adjacent thereto, or adjacent to one or more membranes of the sensing region so as to attenuate the flow of at least one analyte or chemical species. In one embodiment, the semipermeable membrane attenuates the flow of at least one analyte or chemical species so as to provide a linear response from the converted signal. In another embodiment, the semipermeable membrane prevents or precludes the outflow of NAD(P)H from the sensing region or any membrane or domain. In one embodiment, the semipermeable membrane can be an ion-selective membrane selective for ionic analytes of interest such as ammonium ions.

[0334] In another example, a continuous multi - analyte sensor configuration comprising one or more enzymes and / or at least one cofactor was prepared. FIG. 8C depicts this exemplary configuration of an enzyme domain 850 that includes an enzyme having an amount of cofactor positioned proximate to at least a portion of the surface of a working electrode (WE), where the WE includes an electrochemically reactive surface. In one example, a second membrane 851 containing an amount of cofactor is positioned adjacent to the first enzyme domain. The amount of cofactor in the second membrane can provide an excess amount for the enzyme, for example, to extend the sensor lifetime. One or more resistance domains 852 ("RL") are positioned adjacent to the second membrane (or can be between the membranes). The RL can be configured to block the diffusion of cofactor from the second membrane. The electron transfer from the cofactor to the WE converts a signal that directly or indirectly corresponds to the analyte concentration.

[0335] FIG. 8D depicts an alternative enzyme domain configuration that includes a first membrane 851 having an amount of cofactor positioned more proximally to at least a portion of the WE surface. An enzyme domain 850 containing an amount of enzyme is positioned adjacent to the first membrane.

[0336] In the membrane configurations depicted in FIGS. 8C and 8D, the generation of electrochemically active species in the enzyme domain diffuses to the WE surface and converts a signal that directly or indirectly corresponds to the analyte concentration. In some examples, the electrochemically active species includes hydrogen peroxide. In sensor configurations that include cofactors, the cofactor from the first layer can diffuse into the enzyme domain and, for example, extend the sensor lifetime by regenerating the cofactor. In other sensor configurations, cofactors can optionally be included to improve performance attributes such as stability. For example, a continuous ketone sensor can include NAD(P)H and Mg +2It can contain divalent metal cations such as etc. One or more resistance domains RL can be located adjacent to the second film (or can be in the interlayer). RL can be configured to block the diffusion of cofactors and / or interfering substances from the second film from reaching the WE surface. Other components such as electrodes, resistors, biological interfaces, and drug release films, layers, or domains can be used in the aforementioned configurations. In other embodiments, the continuous analyte sensor contains one or more cofactors that contribute to sensor performance.

[0337] Figure 8E depicts another continuous multi - analyte film configuration where β - hydroxybutyrate dehydrogenase BHBDH within the first enzyme domain 853 is located in proximity to the working electrode WE, and a second enzyme domain 854 containing, for example, alcohol dehydrogenase (ADH) and NADH is located adjacent to the first enzyme domain. One or more resistance domains RL852 can be deployed adjacent to the second enzyme domain 854. In this configuration, the presence of a combination of alcohol and ketone in serum serves to collectively provide a converted signal corresponding to at least one of the analyte concentrations, for example, the ketone. Thus, as the NADH present in the more distal second enzyme domain consumes the alcohol present in the serum environment, the NADH is oxidized to NAD(P)H, diffuses into the first membrane layer, and provides the electron transfer of the BHBDH catalyst for acetoacetate ketone and the conversion of the detection signal corresponding to the concentration of the ketone. In one embodiment, the enzyme can be configured for reverse catalysis and can create substrates for the catalysis of another enzyme present in either the same or different layer or domain. Other components such as electrodes, resistors, biological interfaces, and drug release films, layers, or domains can be used in the aforementioned configurations. Thus, the first enzyme domain, which is more distal from the WE than the second enzyme domain, can be configured to generate cofactors or other elements for the second enzyme domain to act as a reactant (and / or reactant substrate) for detecting one or more target analytes.

[0338] Alcohol sensor configuration In one embodiment, a continuous alcohol (e.g., ethanol) sensor device configuration is provided. In one embodiment, one or more enzyme domains containing alcohol oxidase (AOX) are provided, and the presence and / or amount of alcohol is converted by the production of hydrogen peroxide, either alone or in combination with an oxygen consumption or another substrate-oxidase enzyme system, e.g., glucose-glucose oxidase, and hydrogen peroxide and / or oxygen and / or glucose can be detected and / or measured qualitatively or quantitatively using amperometric measurements.

[0339] In one embodiment, the sensing region for the aforementioned enzyme substrate-oxidase enzyme configuration has one or more enzyme domains containing one or more electrodes. In one embodiment, the sensing region for the aforementioned enzyme substrate-oxidase enzyme configuration has one or more enzyme domains with or without one or more electrodes, and further comprises one or more interference blocking membranes (e.g., a selective permeability membrane, a charge exclusion membrane) to attenuate the diffusion of one or more interferents through the membrane to the working electrode. In one embodiment, the sensing region for the aforementioned substrate-oxidase enzyme configuration has one or more enzyme domains with or without one or more electrodes, and further comprises one or more resistance domains with or without one or more interference blocking membranes to attenuate one or more analytes or enzyme substrates. In one embodiment, the sensing region for the aforementioned substrate-oxidase enzyme configuration has one or more enzyme domains with or without one or more electrodes, and one or more resistance domains with or without one or more interference blocking membranes independently further comprise one or more biocompatible membranes and / or drug release membranes to attenuate one or more analytes or enzyme substrates and to attenuate the immune response of the host after insertion.

[0340] In one embodiment, one or more interference blocking films are deposited adjacen...

Claims

1. A monitoring system comprising: a continuous analyte sensor configured to generate an analyte measurement associated with a patient's analyte level; and a sensor electronics module coupled to the continuous analyte sensor and configured to receive and process the analyte measurement.

2. The continuous analyte sensor comprises: a substrate; a working electrode disposed on the substrate; and a reference electrode disposed on the substrate, wherein the analyte measurement generated by the continuous analyte sensor corresponds to an electromotive force that is at least partially based on a potential difference generated between the working electrode and the reference electrode. The monitoring system according to claim 1.

3. The continuous analyte sensor comprises a continuous glucose sensor, and the analyte measurement includes a glucose measurement. The monitoring system according to claim 1.

4. a memory including executable instructions; and one or more processors in data communication with the memory, the one or more processors being configured to: receive glucose data associated with the glucose measurement from the sensor electronics module; process the glucose data to determine at least one glucose clearance rate for the patient based on the glucose data; determine a likelihood of at least one non - typical glucose trend associated with the at least one glucose clearance rate; and execute the executable instructions to generate a decision - making support output based on the determined likelihood. The monitoring system according to claim 3.

5. further comprising one or more non - analyte sensors, the processor being further configured to: receive non - analyte sensor data generated for the patient using the one or more non - analyte sensors, wherein the determined likelihood of at least one non - typical glucose trend is further based on the non - analyte sensor data. The monitoring system according to claim 4.

6. The one or more non - analyte sensors comprise at least one of an insulin pump, a tactile sensor, an ECG sensor, a heart rate monitor, a blood pressure sensor, a respiratory sensor, a peritoneal dialysis machine, or a hemodialysis machine. The monitoring system according to claim 5.

7. The determined likelihood of at least one non - typical glucose trend indicates a risk of hyperglycemia or hypoglycemia. The monitoring system according to claim 4.

8. The decision - making support output is Warning of harmful blood glucose events, treatment recommendations, and at least one of the recommendations for preventing the at least one atypical glucose trend, the monitoring system according to claim 4.

9. A memory including executable instructions, and one or more processors in data communication with the memory, wherein receives glucose data associated with the glucose measurement value from the sensor electronics module, processes the glucose data to determine at least one glucose metric for the patient based on the glucose data, one or more processors configured to execute the executable instructions to generate a diabetes disease prediction based on the at least one glucose metric, the monitoring system according to claim 3.

10. The monitoring system according to claim 9, wherein the processor is further configured to generate one or more treatment recommendations at least partially based on the diabetes disease prediction.

11. further comprising one or more non-analyte sensors, wherein the processor is configured to receive non-analyte sensor data generated for the patient using the one or more non-analyte sensors, and the diabetes disease prediction is further generated based on the non-analyte sensor data, the monitoring system according to claim 9.

12. The monitoring system according to claim 11, wherein the one or more non-analyte sensors comprise at least one of an insulin pump, a tactile sensor, an ECG sensor, a heart rate monitor, a blood pressure sensor, a respiratory sensor, a peritoneal dialysis machine, or a hemodialysis machine.

13. The monitoring system according to claim 9, wherein the diabetes disease prediction indicates the risk of developing diabetes or the current diabetes diagnosis of the patient.

14. The monitoring system according to claim 9, wherein the diabetes disease prediction is generated using a model trained based on population data including records of historical patients having various stages of diabetes.

15. The monitoring system according to claim 10, wherein the treatment recommendations include at least one of lifestyle recommendations, pharmaceutical recommendations, or medical intervention recommendations.

Images