Devices, systems, and methods for administering physiologic insulin resensitization

A web-based platform with digital flowsheets and machine-learning support addresses the challenge of administering PIR therapy, enhancing accuracy and safety while reducing manual workload.

WO2026241561A1PCT designated stage Publication Date: 2026-11-19WELL CELL GLOBAL LLC
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
PCT/US2026/028244
Authority / Receiving Office
WO · WO
Patent Type
Applications
Current Assignee / Owner
Priority Date
2025-05-15
Filing Date
2026-05-14
Publication Date
2026-11-19

AI Technical Summary

Technical Problem

Conventional diabetes management strategies fail to replicate physiologic pulsatile insulin secretion, leading to insulin resistance and complications, and lack integrated, intelligent support for administering physiologic insulin resensitization (PIR) therapy.

Method used

A secure web-based platform with digital flowsheets, prescriptive analytics, machine-learning refinement, and clinician override capabilities to guide and optimize PIR therapy, providing real-time decision support and data management.

Benefits of technology

Improves accuracy and consistency of PIR dosing, reduces clinician workload, enhances safety, and automates treatment protocol refinement.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure US2026028244_19112026_PF_FP_ABST
    Figure US2026028244_19112026_PF_FP_ABST
Patent Text Reader

Abstract

Systems, devices, and methods are disclosed for administering physiologic insulin resensitization therapy. A computing platform may receive patient metabolic metrics, determine eligibility using an Eligibility Qualification Verification Score (EQVS), and generate a PIR Care Plan specifying a recommended insulin concentration (e.g., ORANGE, GREEN, BLUE, WHITE, CLEAR, ULTRA) and a target therapeutic glucose range. The platform may guide treatment sessions through a digital flowsheet that records glucose values, vital signs, IV-placement details, insulin parameters, and oral dextrose dosing. It may generate real-time or near-real-time prescriptive recommendations, evaluate clinician decisions at individual decision points, classify sessions with granular scoring, and refine future recommendations using rule-based logic, machine-learning models, or both. The platform may incorporate laboratory data, validated questionnaires, and clinician overrides, and may provide alerts, field validation, audit trails, treatment summaries, performance metrics, population-level analytics, reporting tools, anonymization, and security features. In some embodiments, it excludes automated insulin delivery and operates solely as clinician-directed decision support.
Need to check novelty before this filing date? Find Prior Art

Description

DEVICES, SYSTEMS, AND METHODS FOR ADMINISTERING PHYSIOLOGIC INSULIN RESENSITIZATIONREFERENCE TO RELATED APPLICATIONS

[0001] This is a Patent Cooperation Treaty application, which claims priority to United States provisional patent application serial number 63 / 806,392 filed on May 15, 2025 that is hereby incorporated by this reference in its entirety.FIELD

[0002] This disclosure relates generally to medical treatment technologies. More particularly, this disclosure relates to devices, systems, and methods for administering, guiding, monitoring, optimizing, and managing physiologic insulin resensitization (“PIR”) therapy. In various embodiments, the disclosure relates to software-based medical devices, systems, and methods implemented on and / or associated with one or more computing systems, optionally networked, for improving insulin sensitivity, reducing insulin resistance, and supporting treatment of metabolic disorders. In various embodiments, the disclosure relates to a secure, web-based platform that assists healthcare providers, office personnel, insurance companies, and / or others in planning, administering, monitoring, documenting, and optimizing PIR therapy for patients with metabolic disorders.BACKGROUND

[0003] Diabetes mellitus is a chronic metabolic disorder affecting hundreds of millions of individuals worldwide. Insulin resistance (“IR”) is a central pathological feature of Type-2 diabetes and contributes to microvascular and macrovascular complications, including neuropathy, nephropathy, retinopathy, and cardiovascular disease. In healthy individuals, pancreatic P-cells secrete insulin in rhythmic pulses, including small oscillatory bursts every several minutes. These pulses maintain receptor sensitivity and support glucose homeostasis. In diabetic patients, this pulsatility may be diminished or absent, contributing to IR and downstream metabolic dysfunction.

[0004] Conventional diabetes management strategies, including basal-bolus insulin therapy, oral hypoglycemics, GLP-1 receptor agonists, and SGLT2 inhibitors, may improveglycemic control but do not replicate physiologic pulsatile insulin secretion. These approaches often fail to restore insulin sensitivity or reverse IR. Physiologic insulin resensitization (“PIR”) therapy seeks to address this gap by delivering exogenous insulin intravenously in rhythmic boluses, typically every four to eight minutes, while maintaining blood glucose within a therapeutic range through controlled oral glucose (dextrose) administration. PIR has shown promise in improving insulin sensitivity and reducing IR-related complications.

[0005] Despite PIR’s therapeutic potential, its administration is complex. Clinicians must monitor blood glucose frequently, adjust insulin bolus size and frequency, titrate oral glucose (dextrose), maintain patient safety, document treatment decisions, evaluate treatment quality, and refine protocols over time. Existing tools do not provide integrated, intelligent, real-time support for PIR administration. Manual workflows are error-prone, inconsistent, and difficult to standardize. There is a need for a system that guides clinicians, reduces cognitive load, improves safety, captures treatment data, analyzes outcomes, and continuously improves through machine learning.

[0006] The closest prior art includes electronic medical record systems and methods, infusion pump interfaces, and diabetes management software. These may collect data or provide limited decision support but do not provide integrated, prescriptive analytics tailored to PIR therapy. The objective technical problem addressed by this disclosure is how to provide a computing-implemented system that assists clinicians in administering PIR therapy safely, consistently, and effectively, while reducing manual workload and improving treatment outcomes.

[0007] The disclosed devices, systems, and methods solve this problem by providing a secure web-based platform, digital flowsheets for real-time documentation, prescriptive analytics for insulin and glucose (dextrose) dosing, machine-learning refinement of treatment protocols, treatment scoring and classification, laboratory integration, questionnaire administration, field validation and data integrity features, and clinician override capabilities. The technical effects include improved accuracy of PIR dosing decisions, reduced clinician workload, improved consistency of treatment protocols, enhanced safety through alerts and validation, improved long-term patient outcomes, and automated refinement of treatment recommendations .

[0008] In various embodiments, the disclosed devices, systems, and methods may include any combination of the features described herein. In some embodiments, the devices, systems, and methods may exclude one or more features described herein. In some embodiments, the devices, systems, and methods may require inclusion of one or more features described herein.

[0009] The embodiments described herein are not limiting. A person skilled in the art may implement the disclosed concepts using alternative architectures, data structures, algorithms, and hardware configurations without departing from the scope of the invention. The disclosed devices, systems, and methods may be implemented using cloud computing, local servers, distributed systems, or standalone devices. The devices, systems, and methods may be implemented using any programming language, database structure, or communication protocol.SUMMARY OF THE DISCLOSURE

[0010] In various embodiments, disclosed are devices, systems, and methods for administering PIR therapy. The systems may include a secure web-based platform, digital flowsheets for capturing and storing treatment session data, field validation and reminder functionality for complete and accurate data collection, prescriptive analytics for recommending insulin and glucose (dextrose) dosing, machine-learning components for refining recommendations, laboratory integration modules for receiving and processing patient lab results, questionnaire administration modules for validated medical questionnaires, treatment scoring and classification capabilities, dose calculation tools for initial and intra-session dosing modifications, and clinician override capabilities. Referring generally to figures and tables, the disclosed platform may serve as a central, web-based repository for PIR infusion data and may assist healthcare providers in making evidencebased treatment decisions.

[0011] In other various embodiments, devices, systems, and methods may generate a PIR Care Plan based on patient metrics, including HbAlc, fasting blood glucose, fasting insulin level, HOMA-IR score, creatinine, BUN, eGFR, UACR, C-peptide, hs-CRP, DNA methylation, blood pressure, lipid profile, weight, and validated questionnaires including provider-facing questionnaires and patient-facing questionnaires such as RAND-SF-36 and Goldberg Depression scales. In various embodiments, devices, systems, and methods maydetermine patient eligibility using an Eligibility Qualification Verification Score (“EQVS”) and may recommend insulin concentrations corresponding to reference names such as ORANGE, GREEN, BLUE, WHITE, CLEAR, and ULTRA, each associated with a specific insulin concentration or range, e.g., insulin-to-saline ratio(s) or other solvent(s) than saline and optionally other solutes.

[0012] In still other various embodiments, devices, systems, and methods may provide real-time or near-real-time recommendations during PIR treatment sessions, including insulin bolus adjustments, oral glucose (dextrose) dosing, insulin infusion pump settings (which may be automatic, manual, or suggested by the system), titration adjustments during treatment, treatment session duration and frequency recommendations, and safety alerts. In various embodiments, devices, systems, and methods may score treatment decisions at individual decision points, classify sessions as “good” or “bad” or with greater granularity, and incorporate successful treatment patterns into machine-learning models while excluding unsuccessful patterns. In various embodiments, devices, systems, and methods may be implemented on a generalized computing system comprising one or more processors, memory, storage, network interfaces, and peripherals.

[0013] In various embodiments, the PIR Care Plan may further recommend a target therapeutic blood glucose range, an initial insulin concentration, an initial insulin dosing protocol, an oral dextrose administration protocol, treatment session frequency and duration, medication monitoring and adjustment recommendations, and diet and exercise instractions). The Care Plan may encompass induction phases, maintenance phases, and tapering phases of treatment.

[0014] In various embodiments, the disclosed platform may provide digital flowsheet management for capturing and storing PIR treatment session data in real time. The platform may integrate with laboratory information systems to receive and process patient lab results, track laboratory parameters overtime, alert providers to significant changes in lab values, and generate treatment-specific laboratory reports. The platform may perform dosing calculations for initial dose determination and intra-infusion dosing modifications.

[0015] In various embodiments, the platform may include reporting capabilities for a-priori delivery of patient treatment summaries, longitudinal tracking of key metabolic parameters, provider performance metrics, population-level outcomes analysis, and customreport generation. The platform may further include data management features such as secure structured storage, backup and disaster recovery, data export capabilities for research and analysis, and anonymization tools for de-identified data sharing.

[0016] In various embodiments, the devices, systems, and methods may explicitly exclude automated insulin delivery, automated glucose delivery, autonomous medical decision-making, direct patient treatment actions, robotic infusion systems, or closed-loop insulin pumps. These exclusions ensure that the devices, systems, and methods may operate as a decision-support tool and that all treatment actions are performed by a clinician.BRIEF DESCRIPTION OF THE DRAWINGS

[0017] The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate various embodiments of the disclosed devices, systems, and methods and, together with the description, serve to explain the principles of the invention. The drawings are not necessarily to scale, and the dimensions and proportions of various elements may be exaggerated for clarity of illustration.

[0018] FIG. 1 is a snapshot from a visual display, optionally having audio and in communication with enabling logic reduced to hardware and / or software with computer system(s), illustrating an example portal appointments webpage for use in / with and enabled by the disclosed devices, systems, and methods.

[0019] FIG. 2 is a snapshot from a visual display, optionally having audio and in communication with enabling logic reduced to hardware and / or software with computer system(s), illustrating an example infusion flowsheet for use in / with and enabled by the disclosed devices, systems, and methods.

[0020] FIG. 3 is a snapshot from a visual display, optionally having audio and in communication with enabling logic reduced to hardware and / or software with computer system(s), illustrating an example infusion flowsheet showing data input, some or all of which may be required prior to starting infusion for use in / with and enabled by the disclosed devices, systems, and methods.

[0021] FIG. 4 is a snapshot from a visual display, optionally having audio and in communication with enabling logic reduced to hardware and / or software with computersystem(s), illustrating an example infusion flowsheet showing an out-of-range blood glucose level for use in / with and enabled by the disclosed devices, systems, and methods.

[0022] FIG. 5 is a snapshot from a visual display, optionally having audio and in communication with enabling logic reduced to hardware and / or software with computer system(s), illustrating an example infusion flowsheet showing first-hour-data capture for blood glucose level (B / S), infusion pump settings (Q and V for interval duration and bolus volume settings, respectively), glucose (oral dextrose taken), and pump (identification number of infusion pump used) for use in / with and enabled by the disclosed devices, systems, and methods.

[0023] FIG. 6 is a snapshot from a visual display, optionally having audio and in communication with enabling logic reduced to hardware and / or software with computer system(s), illustrating an example infusion flowsheet showing paused infusion pump for use in / with and enabled by the disclosed devices, systems, and methods.

[0024] FIG. 7 is a snapshot from a visual display, optionally having audio and in communication with enabling logic reduced to hardware and / or software with computer system(s), illustrating an example infusion flowsheet showing pop-up session comments box for use in / with and enabled by the disclosed devices, systems, and methods.

[0025] FIG. 8 is a snapshot from a visual display, optionally having audio and in communication with enabling logic reduced to hardware and / or software with computer system(s), illustrating an example infusion flowsheet showing session comments, session medications given and comments for next session for use in / with and enabled by the disclosed devices, systems, and methods.

[0026] FIG. 9 is a snapshot from a visual display, optionally having audio and in communication with enabling logic reduced to hardware and / or software with computer system(s), illustrating an example end-of-infusion flowsheet for use in / with and enabled by the disclosed devices, systems, and methods.

[0027] FIG. 10 depicts a computer system capable of being a networked computer system or computer system connected to the networked computer system as discussed in this disclosure, when provided with the appropriate messaging service software, for use in / with and enabled by the disclosed devices, systems, and methods.DETAILED DESCRIPTION

[0028] The following detailed description sets forth various embodiments, variations, and implementations of the disclosed devices, systems, and methods for administering PIR therapy. The embodiments described herein are provided to illustrate the breadth and depth of the invention and are not intended to limit the scope of the claims. A person of ordinary skill in the art will appreciate that the disclosed concepts may be implemented in numerous ways, including through alternative system architectures, data structures, algorithms, and hardware configurations, without departing from the spirit of the invention. As described herein and also with reference to the figures and tables, the disclosed platform may combine digital flowsheet management, data validation, medical questionnaire administration, laboratory integration, prescriptive analytics, treatment scoring and evaluation, machine-learning refinement, and clinician override capabilities into an integrated system for PIR therapy administration.

[0029] As used herein, the term “oral glucose” and the term “dextrose” may be used interchangeably to refer to the glucose administered orally to a patient during PIR therapy to maintain blood glucose within a target therapeutic range. The term “landing blood glucose” refers to the target blood glucose level that is sought to be achieved at the conclusion of a PIR treatment session. The term “Eligibility Qualification Verification Score” or “EQVS” refers to a computed or manually entered composite score used to determine a patient’s eligibility for PIR therapy.System Architecture

[0030] In various embodiments, the disclosed devices, systems, and methods may be implemented using a multi-tier computing architecture. A front-end interface may be provided to clinicians through a web-based portal, a mobile application, a tablet interface, a desktop application, or any combination thereof. The front-end interface may communicate with a back-end server environment that may be cloud-hosted, on-premises, hybrid, or distributed across multiple computing nodes. A data storage layer may be implemented using relational databases, NoSQL databases, encrypted file systems, or combinations thereof. An analytics engine may be implemented using prescriptive analytics, machine-learning models, rule-based logic, or combinations thereof. Integration modules may facilitate communication with electronic medical record systems, laboratory information systems, infusion pumps, continuous glucose monitors, and other medical devices. Security and authentication modulesmay enforce role-based access control, policy-based access control, multifactor authentication, and secure data transmission.

[0031] In some embodiments, the devices, systems, and methods may exclude one or more of the foregoing components. For example, a minimal configuration may include only a front-end interface and a local analytics engine operating on a standalone computing device. In other embodiments, the devices, systems, and methods may require inclusion of at least a front-end interface, a processor, and memory storing executable instructions. The devices, systems, and methods may be implemented using any programming language, framework, or database technology, and the disclosure is not limited to any particular implementation.

[0032] In various embodiments, the front-end interface may provide one or more digital flowsheets that capture PIR treatment session data in real time. The digital flowsheets may include fields for patient demographics, vitals, blood glucose readings, insulin administration parameters, oral glucose (dextrose) administration amounts, clinician notes, and session summaries. The back-end server environment may receive, validate, process, and store data from the front-end interface. The analytics engine may analyze session data and generate prescriptive recommendations for insulin and glucose (dextrose) dosing, treatment frequency, treatment duration, and medication adjustments. The integration modules may receive laboratory results, import electronic medical record data, communicate with infusion pump interfaces, and exchange data with continuous glucose monitoring devices.Computing System

[0033] In various embodiments, the disclosed devices, systems, and methods may be implemented on a generalized computing system comprising one or more processors, memory, non-transitory storage, network interfaces, input and output peripherals, display devices, and secure authentication modules. FIG. 10 illustrates such a generalized computing system or information handling system 1001 which is a simplified example of a computer system capable of performing the operations described herein. Computer system 1001 includes processor 1000 which is coupled to host bus 1005. A level two (L2) cache memory 1010 is also coupled to the host bus 1005. Host-to-PCI bridge 1015 is coupled to main memory 1020, includes cache memory and main memory control functions, and provides bus control to handle transfers among PCI bus 1025, processor 1000, L2 cache 1010, main memory 1020, and host bus 1005. PCI bus 1025 provides an interface for a variety of devices including, for example, LAN card 1030. PCI-to-ISA bridge 1035 provides buscontrol to handle transfers between PCI bus 1025 and ISA bus 1040, universal serial bus (USB) functionality 1045, IDE device functionality 1050, power management functionality 1055, and can include other functional elements not shown, such as a real-time clock (RTC), DMA control, interrupt support, and system management bus support. Peripheral devices and input / output (I / O) devices can be attached to various interfaces 1060 (e.g., parallel interface 1062, serial interface 1064, infrared (IR) interface 1066, keyboard interface 1068, mouse interface 1070, fixed disk (HDD) 1072, removable storage device 1074) coupled to ISA bus 1040. Alternatively, many I / O devices can be accommodated by a super I / O controller (not shown) attached to ISA bus 1040.

[0034] BIOS 1080 is coupled to ISA bus 1040, and incorporates the necessary processor executable code for a variety of low-level system functions and system boot functions. BIOS 1080 can be stored in any computer readable medium, including magnetic storage media, optical storage media, flash memory, random access memory, read only memory, and communications media conveying signals encoding the instructions (e.g., signals from a network). In order to attach computer system 1001 to another computer system to copy files over a network, LAN card 1030 is coupled to PCI bus 1025 and to PCI-to-ISA bridge 1035. Similarly, to connect computer system 1001 to an ISP to connect to the Internet using a telephone line connection, modem 1075 is connected to serial port 1064 and PCI-to-ISA Bridge 1035.

[0035] While the computer system described in FIG. 10 is capable of executing the invention described herein, this computer system is simply one example of a computer system. Those skilled in the art will appreciate that many other computer system designs are capable of performing the invention described herein.

[0036] Another embodiment of the invention is implemented as a program product for use with a computer system such as, for example, the disclosed devices, systems, and methods. The program(s) of the program product defines functions of the embodiments (including the methods described herein) and can be contained on a variety of signal-bearing media. Illustrative signal-bearing media include, but are not limited to: (i) information permanently stored on non-writable storage media (e.g., read-only memory devices within a computer such as CD-ROM disks readable by a CD-ROM drive); (ii) alterable information stored on writable storage media (e.g., floppy disks within a diskette drive or hard-disk drive); and (iii) information conveyed to a computer by a communications medium, such asthrough a computer or telephone network, including wireless communications. The latter embodiment specifically includes information downloaded from the Internet and other networks. Such signal-bearing media, when carrying computer-readable instructions that direct the functions of the present invention, represent embodiments of the present invention.

[0037] In general, the routines executed to implement the embodiments of the invention, may be part of an operating system or a specific application, component, program, module, object, or sequence of instractions. The computer program of the present invention typically is comprised of a multitude of instructions that will be translated by the native computer into a machine-readable format and hence executable instructions. Also, programs are comprised of variables and data structures that either reside locally to the program or are found in memory or on storage devices. In addition, various programs described hereinafter may be identified based upon the application for which they are implemented in a specific embodiment of the invention. However, it should be appreciated that any particular program nomenclature that follows is used merely for convenience, and thus the invention should not be limited to use solely in any specific application identified and / or implied by such nomenclature.

[0038] The processors may include central processing units, graphics processing units, application-specific integrated circuits, field-programmable gate arrays, neural processing units, or combinations thereof. The memory may include random-access memory, read-only memory, flash memory, or persistent memory. The non-transitory storage may include solid-state drives, hard-disk drives, cloud-based storage, or distributed storage systems. Network interfaces 240 may include Ethernet, Wi-Fi, cellular, Bluetooth, or other communication technologies. Input and output peripherals may include keyboards, touchscreens, barcode scanners, or other devices. Display devices may include monitors, tablets, or integrated screens. Secure authentication modules may include trusted platform modules, hardware security modules, or secure enclaves.

[0039] The computing system 1001 may execute enabling logic reduced to hardware, software, firmware, or any combination thereof to perform any method described herein. The computing system 1001 may operate in a standalone configuration or in communication with remote servers. In various embodiments, the computing system may include a rales engine for deterministic recommendations, a machine-learning engine for adaptive recommendations, a data-validation engine, a treatment-scoring engine, a session-classification engine, a dose-calculation engine, a field-validation engine, and a safety-alert engine. In some embodiments, the computing system 200 may exclude machine learning entirely and rely solely on rule-based logic. In other embodiments, machine learning may be required.

[0040] In other example embodiment(s), a computing -implemented system 1001 may be provided for devices, systems, and methods for generating, updating, and presenting clinician-directed treatment guidance based on patient-specific metabolic data. The system 1001 may include a front-end interface in communication with a back-end analytics engine configured to receive physiological measurements, laboratory values, questionnaire responses, and treatment-session data, and to process such data to produce treatment recommendations. In some embodiments, the analytics engine may apply rule-based logic, machine-learning models, or combinations thereof to determine one or more recommended treatment parameters, including but not limited to insulin concentration selection, bolus volume, bolus frequency, oral glucose (dextrose) administration amounts, and monitoring intervals. In various embodiments, the devices, methods and systems may further include a data-validation module configured to confirm completeness and physiological plausibility of received data, a safety-alert module configured to detect out-of-range values and generate alerts, and a clinician-override module configured to accept and record clinician-entered modifications to system-generated recommendations. In some embodiments, the devices, methods and systems may compute or receive an EQVS based on patient metrics and may use the EQVS to determine eligibility status and initiate Care Plan generation. In still other embodiments, the devices, methods and systems may incorporate treatment-scoring and session-classification logic that evaluates clinician decisions at individual decision points, classifies treatment sessions based on predefined criteria, and selectively incorporates data from successful sessions into a machine -learning refinement process while excluding data from unsuccessful sessions. The devices, methods and systems may be implemented using any combination of processors, memory, storage, and network interfaces as described herein, and may operate as a secure, web-based platform, a local application, or a hybrid architecture.Patient Assessment and Eligibility Determination

[0041] In various embodiments, the devices, systems, and methods may determine patient eligibility for PIR therapy through a structured assessment workflow. A patient intake flow may receive patient metrics through manual entry, electronic medical record import,laboratory integration, or combinations thereof. The patient metrics may include, without limitation, those set forth in Table 1.

[0042] Table 1 below sets forth patient assessment metrics, wherein each or one or more may be collected at baseline and / or at predetermined intervals, which may include but are not limited to every session, weekly, monthly, quarterly, or at other clinician-determined intervals.Metric Category Collection Frequency HbAlc (glycated Laboratory Baseline; quarterly or as indicated hemoglobin)Fasting blood glucose Laboratory Baseline; each session Fasting insulin level Laboratory Baseline; quarterly or as indicated HOMA-IR score Calculated Baseline; quarterly or as indicated Creatinine Laboratory / Renal Baseline; quarterly or as indicated Blood urea nitrogen Laboratory / Renal Baseline; quarterly or as indicated (BUN)Estimated glomerular Laboratory / Renal Baseline; quarterly or as indicated filtration rate (eGFR)Urine albumin-to- Laboratory / Renal Baseline; quarterly or as indicated creatinine ratio (UACR)C-peptide Laboratory Baseline; quarterly or as indicated High-sensitivity C-reactive Laboratory / Inflammatory Baseline; quarterly or as indicated protein (hs-CRP)DNA methylation markers Epigenetic Baseline; as indicatedBlood pressure Biometric Baseline; each session (systolic / diastolic)Lipid profile (total Laboratory Baseline; quarterly or as indicated cholesterol, LDL, HDL,triglycerides)Weight / Body Mass Index Biometric Baseline; each session or as (BMI) indicatedProvider questionnaire Questionnaire Baseline; as indicated RAND-SF-36 (Short Form Patient Questionnaire Baseline; quarterly or as indicated Health Survey)Goldberg Depression Scale Patient Questionnaire Baseline; quarterly or as indicated Other validated patient Patient Questionnaire As indicatedquestionnairesTable 1

[0043] The devices, systems, and methods may validate the completeness of the patient metrics, confirm that required fields are present, and verify that values fall within acceptable physiological ranges. In various embodiments, the devices, systems, and methods may compute an Eligibility Qualification Verification Score (“EQVS”) based on the collected patient metrics. The EQVS may be computed automatically by the analytics engine, entered manually by a clinician, or determined through a combination of automated computation and manual clinician input. The EQVS may incorporate weighted values for one or more of the metrics set forth in Table 1 and may produce an eligibility status indicating whether the patient is eligible, ineligible, or conditionally eligible for PIR therapy.

[0044] In various embodiments, the eligibility criteria evaluated by the EQVS may include documented diabetes diagnosis, documented chronic kidney disease stage, HbAlc thresholds, fasting glucose values, fasting insulin levels, HOMA-IR score, renal function metrics (including creatinine, BLTN, eGFR, and UACR), inflammatory markers, and exclusion conditions such as pregnancy, active cancer, or unstable psychiatric disorder. In some embodiments, the devices, systems, and methods may require inclusion of at least HbAlc, fasting glucose, and estimated glomerular filtration rate. In other embodiments, the devices, systems, and methods may exclude one or more metrics and rely on clinician judgment. The EQVS may allow for manual entry to accommodate clinician override of automated eligibility determinations.

[0045] In various embodiments, the devices, systems, and methods may present the eligibility status to the clinician along with a rationale and may allow the clinician to override the determination. The devices, systems, and methods may store the eligibility determination,the rationale, and any clinician override in a secure audit trail. The devices, systems, and methods may trigger PIR Care Plan generation logic upon a determination of eligibility.

[0046] In other example embodiments, a computer-implemented method may be provided for guiding a clinician during administration of a metabolic-therapy infusion session. The devices, methods and systems may include receiving, through a digital flowsheet interface, real-time or near-real-time physiological measurements including blood glucose values, vital signs, and treatment-session parameters. The devices, methods and systems may further include processing the received data using an analytics engine configured to generate one or more treatment recommendations, such as recommended insulin bolus adjustments, insulin concentration selection, oral glucose (dextrose) dosing, session duration modifications, or monitoring-interval adjustments. In some embodiments, the devices, methods and systems may include validating received data for completeness and physiological plausibility, generating alerts when values fall outside predetermined ranges, and presenting the alerts and recommendations to a clinician through a user interface. In various embodiments, the devices, methods and systems may further include computing or receiving an EQ VS based on patient metrics and using the EQVS to determine whether a patient is eligible, ineligible, or conditionally eligible for treatment. In some embodiments, the devices, methods and systems may include scoring clinician decisions at individual decision points, classifying treatment sessions based on predefined criteria, and updating one or more machine-learning models using data from sessions classified as successful while excluding data from sessions classified as unsuccessful. In still other embodiments, the devices, methods and systems may include generating a PIR Care Plan comprising a recommended therapeutic glucose range, an initial insulin concentration selected from a plurality of reference concentrations, an initial dosing protocol, and a monitoring schedule, and presenting the Care Plan to a clinician for review and approval.PIR Care Plan Generation

[0047] Referring now to Table 2, in various embodiments, a PIR Care Plan generation flow may analyze validated patient metrics, determine an insulin sensitivity category, and generate a PIR Care Plan for clinician approval. The devices, systems, and methods may determine the insulin sensitivity category based on one or more of the patient metrics set forth in Table 1 and may select a recommended insulin concentration from the set of reference concentrations set forth in Table 2 below.

[0048] Table 2 below sets forth insulin concentration categories, reference names, insulin-to-saline ratios or other solvent(s), and associated insulin sensitivity classifications used in various embodiments of the disclosed devices, systems, and methods.Reference Name Insulin Insulin Saline Ratio (mU / mL)Sensitivity Concentration Volume (mL)Classification (mU)ORANGE Highly 2.0 7.0 2.0 mU / 7.0 mL ResistantGREEN Moderately 1.0 5.0 1.0 mU / 5.0 mL ResistantBLUE Moderately 0.5 5.5 0.5 mU / 5.5 mL SensitiveWHITE Sensitive 0.25 5.75 0.25 mU / 5.75 mL CLEAR Extremely 0.125 5.875 0.125 mU / 5.875Sensitive mLULTRA Reserved / Custom Custom Clinician- Custom determinedTable 2

[0049] In various embodiments, the PIR Care Plan may recommend one or more of: a target therapeutic blood glucose range personalized to the patient’s condition; an initial insulin concentration selected from Table 2; an initial insulin dosing protocol including bolus volume and bolus frequency; an oral dextrose administration protocol; treatment session frequency and duration; medication monitoring and adjustment recommendations; and diet and exercise instructions. The Care Plan may be generated for induction phases, maintenance phases, or tapering phases of treatment, and may be updated over time based on patient response and outcomes data.

[0050] In some embodiments, the devices, systems, and methods may exclude one or more concentration levels from Table 2. In other embodiments, the devices, systems, and methods may require inclusion of at least one concentration level from Table 2. In still other embodiments, additional insulin concentration categories beyond those listed in Table 2 maybe added by a clinician or by the analytics engine based on patient population data. The ULTRA category may represent a reserved or custom concentration determined entirely by the clinician or by a machine-learning model trained on a sufficient body of clinical data.

[0051] In various embodiments, the PIR Care Plan may further include a recommended monitoring schedule specifying the frequency of blood glucose measurements, vital sign assessments, and laboratory evaluations during and between treatment sessions. The Care Plan may include specific guidance for achieving a target “landing” blood glucose at the conclusion of each treatment session, which may be defined as the blood glucose value sought at the termination of intravenous insulin delivery to ensure patient safety and optimize therapeutic effect.

[0052] In various embodiments, the devices, systems, and methods may present the generated PIR Care Plan to the clinician through the front-end interface and may require clinician approval before implementing the Care Plan. The clinician may accept the Care Plan in whole, reject the Care Plan, or modify individual parameters of the Care Plan. All clinician actions with respect to the Care Plan may be recorded in a secure audit trail.Digital Flowsheet Management

[0053] In various embodiments, the devices, systems, and methods may include digital flowsheet management capabilities for capturing and storing PIR treatment session data. The digital flowsheets may serve as a central, web-based repository for PIR infusion data and may assist healthcare providers in capturing all necessary information at the correct times during a treatment session. The digital flowsheets may include structured fields for patient identification, session date and time, clinician identification, vitals (including blood pressure, heart rate, temperature, respiratory rate, and oxygen saturation), blood glucose readings at regular intervals, insulin administration parameters (including concentration, bolus volume, bolus frequency, and infusion pump settings), oral glucose (dextrose) administration amounts and times, clinician notes, adverse event documentation, and session summary information.

[0054] Also in various embodiments, the digital flowsheets may automatically stamp the date and time of each data entry, track cumulative insulin delivered and glucose (dextrose) administered during a session, document patient responses and any adverse events, and generate a session summary with professional notes for future reference. The digitalflowsheets may support data entry through manual keyboard input, touchscreen input, barcode scanning, integration with connected medical devices, or combinations thereof. Field Validation and Data Integrity

[0055] In various embodiments, the devices, systems, and methods may include field validation and data integrity features to ensure complete and accurate data collection before, during, and / or after PIR treatment sessions. These features may include required field indicators that prevent a clinician from progressing through a workflow without entering critical data, time-based reminders for regular monitoring activities (such as blood glucose checks at specified intervals), range validation for physiological parameters with optional alerts when entered values fall outside of acceptable ranges, automated calculations to reduce manual entry errors (such as automatic computation of insulin doses, HOMA-IR scores, or cumulative dosing totals), and audit trails for all data entries and modifications.

[0056] Also in various embodiments, the field validation features may be configurable by a system administrator or clinician to define which fields are required, which ranges trigger alerts, and which intervals trigger reminders. The audit trail may record the identity of the user who made each entry, the timestamp of each entry, any modifications made to prior entries, and the reason for any modification. In some embodiments, the devices, systems, and methods may require completion of all mandatory fields before a treatment session may be finalized. In other embodiments, the devices, systems, and methods may permit session finalization with incomplete data while generating an alert or flag for review.Medical Questionnaire Administration

[0057] In various embodiments, the devices, systems, and methods may include medical questionnaire administration capabilities. The platform may administer provider-facing questionnaires for clinical evaluation of the patient and patient-facing questionnaires for symptoms, quality-of-life assessment, and subjective outcome reporting. Validated patient questionnaires may include, without limitation, the RAND-SF-36 (Short Form Health Survey), the Goldberg Depression Scale, and other validated instruments for assessing physical function, mental health, pain, vitality, social function, and general health perceptions.

[0058] Also in various embodiments, the platform may schedule and track questionnaire completion, score and interpret questionnaire results using validated scoring algorithms, andtrack questionnaire responses longitudinally over time to assess treatment efficacy and changes in patient-reported outcomes. The platform may present questionnaire results to clinicians alongside laboratory and biometric data to provide a comprehensive view of patient status. In some embodiments, questionnaire responses may be incorporated into the analytics engine as input variables for prescriptive recommendations.Laboratory Integration

[0059] In various embodiments, the devices, systems, and methods may include laboratory integration capabilities. The platform may receive patient lab results directly from laboratory information systems through electronic interfaces, application programming interfaces (APIs), or Health Level Seven (HL7) messaging protocols. The platform may incorporate received lab values into treatment algorithms executed by the analytics engine, track laboratory parameters over time to identify trends, alert providers to significant changes in lab values (such as a clinically significant rise in creatinine or decline in eGFR), and generate treatment-specific laboratory reports.

[0060] Also in various embodiments, laboratory reports may be generated in a presenttype format (showing current values) or a past-type format (showing historical values and trends). The laboratory integration module may support both automated import of lab results and manual entry by a clinician. In some embodiments, the devices, systems, and methods may exclude laboratory integration entirely and rely on manual entry. In other embodiments, at least one laboratory integration module may be required.Treatment Session Workflow

[0061] In various embodiments, a treatment session workflow may include three phases: a pre-infusion phase, an active infusion phase, and a post-infusion phase as discussed, for example, in the next six paragraphs. Each phase may involve specific data collection, analytics, documentation, and safety functions as described herein.

[0062] Pre-Infusion Phase. In various embodiments, the pre -infusion phase may begin with the collection of patient vitals (including blood pressure, heart rate, temperature, respiratory rate, and oxygen saturation), an initial blood glucose reading, an initial HbAlc confirmation if applicable, confirmation of pre-treatment preparations (such as fasting status, medication compliance, and patient consent), and IV-placement information including IV site, gauge, access type, and confirmation of patency. The devices, systems, and methodsmay retrieve the patient's current PIR Care Plan and display the recommended initial insulin concentration (from TABLE 1), initial insulin dosing protocol, and initial oral dextrose dosing for the session.

[0063] Active Infusion Phase. In various embodiments, during the active infusion phase, the devices, systems, and methods may display recommended insulin infusion pump settings, which may include bolus volume, bolus frequency (typically every four to eight minutes), and infusion rate. The pump settings may be automatic (where the system communicates directly with a connected infusion pump), manual (where the clinician manually enters pump settings based on system recommendations), or suggested by the system (where the system displays recommended settings and the clinician confirms or modifies them before application). The devices, systems, and methods may display recommended initial oral dextrose dosing and may update dextrose recommendations throughout the session based on blood glucose trends.

[0064] In various embodiments, during the active infusion phase, the devices, systems, and methods may receive blood glucose readings at regular intervals (such as every four, six, eight, or ten minutes), vital sign measurements, insulin administration information (including actual bolus volumes and times), and oral glucose (dextrose) administration information (including amounts and times). The devices, systems, and methods may provide a monitoring schedule indicating when each type of data should be collected. The devices, systems, and methods may generate alerts for out-of-range blood glucose values, abnormal vital signs, or missing documentation entries.

[0065] Intra-Session Prescriptive Analytics. In various embodiments, during the active infusion phase, the devices, systems, and methods may provide intra-session prescriptive analytics comprising near-real-time or real-time recommendations for adjusting insulin infusion parameters (including bolus volume, bolus frequency, and concentration adjustments) based on blood glucose trends. The intra-session analytics may further provide recommendations for oral dextrose administration timing and amount, alerts for potential hypoglycemia (when blood glucose trends downward toward or below a lower threshold) or hyperglycemia (when blood glucose trends upward toward or above an upper threshold), and guidance for achieving the target “landing” blood glucose at session conclusion. The prescriptive analytics may be generated by the analytics engine using rale -based logic, machine-learning models, or combinations thereof.

[0066] Titration Adjustments. In various embodiments, the devices, systems, and methods may recommend titration adjustments to insulin concentration and / or dosing during a treatment session in response to the patient’s measured blood glucose trajectory, insulin response profile, and clinical status. Titration adjustments may include increasing or decreasing the insulin bolus volume, changing the bolus frequency, switching to a different insulin concentration from TABLE 1, adjusting the oral dextrose dosing schedule, or any combination thereof. All titration recommendations may be subject to clinician review and approval before implementation.

[0067] Post-Infusion Phase. In various embodiments, the post-infusion phase may include documentation of the final blood glucose reading (including comparison to the target “landing” blood glucose), documentation of total insulin delivered and total oral glucose (dextrose) administered during the session, documentation of any adverse events, a postsession vital sign assessment, generation of a session summary report, and entry of professional notes for future session reference. The session summary may be automatically generated by the devices, systems, and methods and may include timestamps, cumulative dosing data, blood glucose trajectory, clinician decisions, alerts generated, and overall session classification.Prescriptive Analytics and Recommendations

[0068] In various embodiments, the devices, systems, and methods may include a prescriptive analytics engine within the analytics engine that processes multiple data points to generate treatment recommendations. The prescriptive analytics engine may receive and analyze current and historical blood glucose values, insulin dosing history and patient response to prior doses, oral dextrose administration history and patient response, vital sign data, laboratory values (including those set forth in Table 1), and patient-reported symptoms and outcomes from administered questionnaires. In various embodiments, validating physiological plausibility may further include confirming cross-variable coherence among blood glucose values, insulin administration parameters, and oral glucose (dextrose) amounts, such that the physiological responses of each variable are consistent with the expected effects of the others.

[0069] Validating completeness may include verifying that all required data fields for a treatment cycle are present, that time-dependent entries occur within predefined intervals, and that each insulin bolus event is associated with corresponding glucose measurements.Validating physiological plausibility may include verifying that received glucose, insulin, and oral glucose values fall within human-possible physiological ranges, exhibit biologically realistic rates of change, maintain logical coherence across variables, and occur in a temporally consistent sequence.

[0070] Also in various embodiments, the prescriptive analytics engine may generate recommendations for: initial treatment protocol determination based on patient characteristics and insulin sensitivity category; near-real-time or real-time adjustments to insulin and dextrose dosing during treatment sessions; treatment session frequency and duration based on patient response; medication adjustment suggestions based on improved insulin sensitivity over time; and personalized target ranges for blood glucose based on the individual patient’s condition, comorbidities, and treatment history.

[0071] In still more various embodiments, the prescriptive analytics engine may present recommendations to the clinician through the front-end interface and may record whether the clinician accepts, rejects, or modifies each recommendation. The acceptance or rejection patterns of clinicians may be used as input to the machine-learning refinement process described herein. In some embodiments, the prescriptive analytics engine may generate recommendations only at fixed intervals (such as every four minutes, every six minutes, or every eight minutes). In other embodiments, the prescriptive analytics engine may generate recommendations continuously. In still other embodiments, the prescriptive analytics engine may generate recommendations only at the start of a session or only in response to specific clinical triggers.

[0072] And in yet additional various embodiments, the prescriptive analytics engine may provide dose calculation tools for both initial dose determination and intra-infusion dosing modifications. The dose calculation tools may compute insulin dosages based on patient weight, insulin sensitivity category, current blood glucose, target blood glucose, and historical response patterns. The dose calculation tools may further compute recommended oral dextrose amounts based on current blood glucose, insulin administration rate, and desired blood glucose trajectory.Treatment Scoring, Evaluation, and Classification

[0073] In various embodiments, the devices, systems, and methods may include treatment scoring, evaluation, and classification capabilities. Individual treatment decisions madeduring a PIR session may be scored at each decision point based on one or more criteria, including: alignment of the clinician’s decision with the system’s recommendation; the resulting blood glucose trajectory following the decision; stability of vital signs during and after the decision; absence of adverse events; and achievement of the target “landing” blood glucose at session conclusion.

[0074] In additional various embodiments, the devices, systems, and methods may evaluate each treatment session as a whole by aggregating the individual decision point scores and additional session-level metrics. The session evaluation may classify sessions as “good,” “bad," or with greater granularity using a multi-tier classification system (such as excellent, good, acceptable, suboptimal, poor, or any other scale defined by the system administrator or clinician). In some embodiments, the classification may use a numerical score. In other embodiments, the classification may use categorical labels. In still other embodiments, the classification may use both numerical scores and categorical labels.

[0075]

[0070] In various embodiments, the session classification may be used by the machine-learning refinement component as described herein. Sessions classified as “good” (or above a threshold quality score) may be incorporated into the machine-learning training database. Sessions classified as “bad” (or below a threshold quality score) may be flagged for review but may not be used for algorithm improvement, thereby preventing suboptimal treatment patterns from propagating into future recommendations. In some embodiments, sessions classified at an intermediate level may be subjected to additional review before inclusion or exclusion from training data.

[0076] In various embodiments, the treatment scoring and evaluation process may score decisions made by the healthcare professional and compare those decisions to patient outcomes, thereby providing feedback to the clinician and to the machine-learning system. The treatment scoring may be performed in real time during a session, retrospectively after session completion, or both.Machine- Learning Refinement

[0077] In various embodiments, the devices, systems, and methods may include a machine-learning refinement component within the analytics engine. The machine-learning component may receive session data and treatment decisions, score the treatment decisions using the scoring criteria described herein, classify each session, and incorporate sessionsclassified as “good” into a training dataset. Sessions classified as “bad” may be excluded from the training dataset. The machine-learning component may update model parameters based on the training dataset and may generate improved recommendations for future sessions.

[0078] In various embodiments, the machine-learning component may employ supervised learning, unsupervised learning, reinforcement learning, deep learning, ensemble methods, or combinations thereof. Machine -learning models may include, without limitation, gradient-boosted decision trees, random forests, neural networks, support vector machines, Bayesian models, or any other model architecture suitable for clinical decision support. The machinelearning component may adapt to clinician acceptance or rejection patterns over time, thereby personalizing recommendations to the clinical judgment of individual providers or practice groups.

[0079] In some embodiments, the machine-learning component may be disabled entirely and the devices, systems, and methods may rely solely on rule-based logic for generating recommendations. In other embodiments, the machine-learning component may be required. In still other embodiments, the recommendation algorithms may be optionally generated by artificial intelligence, by a computer system’s enabling logic reduced to hardware and / or software, and / or by manual selection by a clinician or system administrator. The machinelearning component may continuously refine recommendations based on clinical outcomes and may generate periodic reports on model performance, recommendation accuracy, and outcome trends.

[0080] In various embodiments, “good” infusions may be incorporated into a learning database to improve and refine the recommendation algorithms, while “bad” infusions may be flagged for clinician review but may not be used to update or improve the algorithms. This selective training approach may ensure that the machine-learning models are trained only on successful treatment patterns, thereby avoiding the incorporation of suboptimal or harmful patterns into future recommendations.Safety, Overrides, and Clinician Control

[0081] In various embodiments, the devices, systems, and methods may include safety, override, and clinician-control mechanisms. The devices, systems, and methods may include required field validation (preventing progression without critical data), range validation(alerting when entered values fall outside acceptable physiological ranges), alerts for abnormal values, alerts for missing data, alerts for potential hypoglycemia or hyperglycemia, clinician override capabilities, audit trails for all actions, and timestamping of all entries.

[0082] In additional various embodiments, the clinician may override any recommendation generated by the analytics engine, including insulin concentration, bolus volume, bolus frequency, oral glucose (dextrose) dosing, session duration, treatment frequency, eligibility determination, and Care Plan parameters. Every clinician override may be recorded in the audit trail with the identity of the overriding clinician, the original system recommendation, the clinician’s modified value, the timestamp of the override, and optionally a reason for the override. The override data may be used by the machine-learning component to adapt to clinician preferences over time.

[0083] In some embodiments, the devices, systems, and methods may require clinician confirmation before accepting any recommendation generated by the system. In other embodiments, recommendations may be automatically accepted unless rejected by the clinician within a configurable time window. In all embodiments, the devices, systems, and methods may not take direct action on a patient, and all treatment actions may be performed by a clinician. The devices, systems, and methods may explicitly exclude automated insulin delivery, automated glucose (dextrose) delivery, autonomous medical decision-making, direct patient treatment actions, robotic infusion systems, and closed-loop insulin pumps.

[0084] In various embodiments, the devices, systems, and methods may allow healthcare professionals to override any system recommendation at any time during the treatment session. The override capability may ensure that the clinician retains full authority over all treatment decisions and that the system operates as a decision-support tool rather than an autonomous treatment device.Security and Data Management

[0085] In various embodiments, the devices, systems, and methods may include security and data management features. The platform may implement secure user authentication using role-based access control (“RBAC”), policy-based access control (“PBAC”), or combinations thereof, with optional additional security levels such as multifactor authentication, biometric authentication, or hardware token authentication. The platform may encrypt data during transmission (using TLS, SSL, or equivalent protocols) and / or during storage (using AES-256, RSA, or equivalent encryption algorithms). The platform may maintain compliance with healthcare data privacy regulations, including but not limited to the Health Insurance Portability and Accountability Act (“HIPAA”), the General Data Protection Regulation (“GDPR”), and applicable state and local privacy laws.

[0086] In various embodiments, the platform may maintain audit logging of all system access and activities, including login events, data access events, data modification events, and system configuration changes. The platform may undergo regular security assessments and updates to address emerging vulnerabilities. The data management features may include: structured and optionally segmented or sequestered storage; backup and disaster recovery capabilities; data export capabilities for research and analysis purposes; and anonymization tools for generating de-identified datasets suitable for research, regulatory reporting, or population-level outcomes analysis.

[0087] In various embodiments, the anonymization tools may remove or obscure patient identifiers while preserving clinically relevant data fields, thereby enabling researchers to analyze treatment patterns, outcomes, and population-level trends without compromising patient privacy. The data export capabilities may support standard data formats including CSV, HL7, FHIR, JSON, XML, or other formats suitable for interoperability with external research platforms, regulatory submission systems, or analytics tools.Reporting and Outcomes Analysis

[0088] In example embodiments, the devices, systems, and methods may include comprehensive reporting and outcomes analysis capabilities. The reporting module may provide a-priori delivery of patient treatment summaries to clinicians prior to scheduled treatment sessions, enabling clinicians to review relevant patient history, prior session data, and outstanding action items before beginning a session.

[0089] In additional example embodiments, the reporting module may provide longitudinal tracking of key metabolic parameters over time, including HbAlc trends, fasting glucose trends, fasting insulin trends, HOMA-IR trends, renal function trends, inflammatory marker trends, weight trends, and questionnaire score trends. The reporting module may generate visualizations such as time-series charts, trend lines, comparison tables, and summary dashboards to facilitate clinical interpretation of longitudinal data.

[0090] In various embodiments, the reporting module may further provide provider performance metrics that evaluate individual provider or practice group treatment patterns, recommendation adherence rates, and patient outcomes. The reporting module may provide population-level outcomes analysis that aggregates treatment data across multiple patients, clinics, or networks to identify best practices, outcome trends, and areas for improvement. The reporting module may also support custom report generation, enabling clinicians, administrators, or researchers to define custom queries, filters, and output formats for generating specialized reports tailored to specific clinical, operational, or research needs.

[0091] And in yet additional example embodiments, the reporting and outcomes analysis features may be accessible through the front-end interface and may be filterable by patient, provider, clinic, date range, treatment phase (induction, maintenance, tapering), insulin concentration category, or other parameters. In some embodiments, the reporting features may generate automated periodic reports (such as weekly, monthly, or quarterly summaries). In other embodiments, reports may be generated on demand by a clinician or administrator.Data Flows

[0092] In various embodiments, the devices, systems, and methods may process data through multiple flows as described herein.

[0093] Patient Intake Flow. A patient intake flow may receive patient metrics through manual entry, electronic medical record import, laboratory integration, or combinations thereof. The devices, systems, and methods may validate the completeness of the metrics, confirm that required fields are present, and verify that values fall within acceptable ranges. The devices, systems, and methods may store the metrics in a secure database within the data storage layer 130 and may trigger eligibility determination logic.

[0094] Eligibility Determination Flow. An eligibility determination flow may evaluate patient metrics against eligibility criteria, compute the EQVS, identify exclusion conditions, generate an eligibility status, and provide a clinician with a summary and override options. Eligibility criteria may include documented diabetes diagnosis, documented chronic kidney disease stage, HbAlc thresholds, fasting glucose, fasting insulin, HOMA-IR, renal function metrics, and exclusion conditions such as pregnancy, active cancer, or unstable psychiatric disorder. In some embodiments, the devices, systems, and methods may require inclusion of at least HbAlc, fasting glucose, and estimated glomerular filtration rate. In otherembodiments, the devices, systems, and methods may exclude one or more metrics and rely on clinician judgment.

[0095] PIR Care Plan Generation Flow. A PIR Care Plan generation flow may analyze patient metrics, determine an insulin sensitivity category, recommend an insulin concentration from Table 2, recommend a therapeutic glucose range, recommend a treatment frequency and duration, recommend medication monitoring and adjustment, recommend diet and exercise instructions, and generate a Care Plan for clinician approval. The Care Plan generation flow may generate Care Plans for induction phases, maintenance phases, or tapering phases.

[0096] Treatment Session Flow. A treatment session flow may proceed through the preinfusion, active infusion, and post-infusion phases described herein, collecting data, generating recommendations, scoring decisions, classifying the session, and generating a session summary.

[0097] Prescriptive Analytics Flow. A prescriptive analytics flow may receive glucose trend data, insulin response data, oral glucose (dextrose) response data, vital sign data, and laboratory values. The analytics flow may analyze patterns in the data and generate recommended adjustments to insulin bolus parameters, oral glucose (dextrose) dosing, session duration, treatment frequency, and medication regimens. The analytics flow may present recommendations to the clinician and record acceptance or rejection.

[0098] Machine-Learning Refinement Flow. A machine-learning refinement flow may receive session data and treatment decisions, score treatment decisions, classify sessions, incorporate “good” sessions into training data, exclude “bad” sessions, update model parameters, and generate improved recommendations for future sessions. The machinelearning refinement flow may adapt to clinician behavior patterns over time.Additional Embodiments

[0099] In various embodiments, the devices, systems, and methods may include automated reminders for glucose checks, vital sign measurements, or documentation requirements. The devices, systems, and methods may include time-based validation to ensure that required data is collected at appropriate intervals. The devices, systems, and methods may include trend visualization tools that display glucose, insulin, and vital signtrends over time. The devices, systems, and methods may include session-to-session comparison tools that allow clinicians to evaluate patient progress.

[0100] In various embodiments, the devices, systems, and methods may integrate with wearable devices, infusion pumps, continuous glucose monitors, laboratory information systems, or electronic medical record systems. The devices, systems, and methods may import data automatically or may require manual confirmation. In some embodiments, the devices, systems, and methods may exclude one or more integration modules and rely solely on manual entry. In other embodiments, the devices, systems, and methods may require inclusion of at least one integration module.

[0101] In other various embodiments, the devices, systems, and methods may be configured to exclude machine learning and rely solely on rule-based logic. In other embodiments, the devices, systems, and methods may exclude real-time recommendations and provide only static guidance. In still other embodiments, the devices, systems, and methods may exclude oral glucose (dextrose) recommendations, insulin concentration recommendations, treatment scoring, session classification, electronic medical record integration, or laboratory integration.

[0102] In more embodiments, the devices, systems, and methods may explicitly exclude automated insulin delivery, automated glucose delivery, autonomous medical decisionmaking, direct patient treatment actions, robotic infusion systems, or closed-loop insulin pumps. These exclusions ensure that the devices, systems, and methods may operate as a decision-support tool and that all treatment actions are performed by a clinician.

[0103] In still more example embodiments, the devices, systems, and methods may require manual confirmation for every recommendation generated during a treatment session. In other embodiments, the devices, systems, and methods may automatically accept recommendations unless the clinician rejects them within a configurable time window. In some embodiments, the devices, systems, and methods may generate recommendations only at fixed intervals, such as every four minutes, every six minutes, or every eight minutes. In other embodiments, the devices, systems, and methods may generate recommendations only at the start of a session or only in response to specific clinical triggers. In still other embodiments, the devices, systems, and methods may generate recommendations only for insulin dosing, only for glucose (dextrose) dosing, only for session duration, or only fortreatment frequency. These fallback embodiments provide explicit support for narrower claim amendments in any jurisdiction.

[0104] In various example embodiments, the devices, systems, and methods may include a platform or system for administering PIR comprising: a secure web-based platform for planning and documenting PIR; digital flowsheets for capturing treatment session data; prescriptive analytics for recommending insulin and dextrose dosing; treatment scoring and evaluation capabilities; and machine learning for continuous improvement of recommendations.

[0105] In various embodiments, the devices, systems, and methods may include a device, method, or system for optimizing PIR through: collection and analysis of patient metabolic parameters; generation of personalized treatment protocols; at least near-real-time monitoring and documentation of treatment sessions; adaptive recommendations based on patient response; and evaluation of treatment decisions and outcomes.

[0106] In various embodiments, the devices, systems, and methods may include a device, method, or system using machine learning for improving PIR by: scoring individual treatment decisions; evaluating complete treatment sessions; incorporating successful treatment patterns into recommendation algorithms (optionally generated by artificial intelligence, by a computer system’s enabling logic reduced to hardware and / or software, and / or by manual selection); and continuously refining recommendations based on clinical outcomes.

[0107] In various embodiments, the software platform may: serve as a central web-based repository for PIR infusion data through digital flowsheets; help medical providers capture all necessary information at correct times; administer validated medical questionnaires; integrate with laboratories to receive patient lab results; have prescriptive analytics capabilities to recommend insulin dose titration, dextrose dosing titration, infusion length, and infusion frequency; allow healthcare professionals to override recommendations; perform dosing calculations for initial dose and intra-infusion modifications; score decision points in infusion by comparing medical professional decisions to patient outcomes; in one example embodiment, grade infusions as “good” or “bad” or differently with greater granularity; and use “good” infusions to refine the learning algorithm and disregard “bad” infusions.

[0108] Turning now to logic, which may reside locally to device(s) and / or system(s) for the disclosed methods, systems, and device, in one example embodiment may include a portal, such as a secure, web-based portal. Such a portal and / or other repository of accessible and enabling logic, may be used by physician and / or clinician for PIR treatment planning, data collection, and / or guidance, whether before, during, and / or after administering the PIR treatment session(s), wherein such logic may be employed as well.

[0109] In this disclosure, discussed is what is called the “Well Cell Portal application,” but naturally this is just one example embodiment and may vary minorly or widely in other embodiments processing and / or accessing the same or equivalent resources and logic.

[0110] PIR Care Plans: The portal may capture specific physiological metrics of the patient, analyze said metrics, and / or provide the physician with guidance on the creation of a PIR Care Plan, including the one or more of the following example recommendations: the target therapeutic range for blood glucose level during PIR treatment, dosing recommendations for IV insulin (Humulin R U100, BLA 018780 or Novolin R U100, BLA 019938), and / or oral glucose (dextrose).

[0111] As relevant patient metrics are collected during baseline eligibility consultation, routine lab work collections (e.g., at least every 12 weeks), and at each PIR treatment session, the patient’s PIR Care Plan may be adjusted as the treating physician deems necessary, with guidance from the portal.

[0112] PIR Treatment Session Administration: The Well Cell Portal software may also used during PIR treatment sessions to collect values obtained during the session (e.g., blood glucose levels and vitals) and / or guide decisions made by and / or to be made by the clinician during the session (e.g., oral glucose administration, IV insulin infusion pump settings of bolus volume and / or frequency interval, etc.). The treating physician may have the choice to simply ignore the Well Cell Portal recommendations based on expert medical judgment or otherwise.

[0113] The physician decisions on quantity and time duration of IV insulin administration and / or oral glucose intake (and each of the titration adjustments to the patient’s dosing) over the course of the PIR treatment session may be input into the Well Cell Portal, i.e., device or system. The insulin information may be collected manually / imported from an infusion pump (e.g., KI 92860) along with blood glucose readings that are entered / imported into the WellCell Portal but obtained from continuous glucose monitor (CGM) or glucometer. The Well Cell Portal software may then analyze these metrics and provide ongoing dosing option recommendations to the physician.PIR Treatment Workflow with Well Cell Portal

[0114] Potential patients must have an in-clinic consultation with the treating physician to determine if he / she qualifies for the Physiologic Insulin Resensitization (PIR) treatment paradigm. Importantly, the treating physician must have previously received training from Well Cell Global on the guidelines of the PIR treatment paradigm.

[0115] Eligible patients will have received a documented diagnosis of diabetes mellitus at least 6 months prior to screening, as well as a documented diagnosis of CKD stage 3b, 4, or 5 based on eGFR.

[0116] Any patients with the following conditions are not considered eligible for PIR treatment: pregnant, trying to conceive, or breast feeding; provider determined medically unstable; active cancer; or unstable psychiatric disorder; active dialysis; immobile; or eating disorder.

[0117] The Well Cell Portal may take and / or require one or more of the following metrics as input, whether at the baseline eligibility consultation, every 12 weeks or other minimum set time period, or more frequently as the treating physician deems necessary: (a) HbAlc; (b) Fasting blood glucose; (c) Fasting insulin level; (d) HOMA-IR (Homeostatic Assessment of Insulin Resistance score); (e) Creatinine; (f) Blood-urea-nitrogen (BUN); (g) Glomerular filtration rate (eGFR); (h) Urine albumin creatinine ratio (UACR); (i) C-Peptide level; (j) Hs CRP; (k) DNA Methylation; (1) Blood Pressure: Systolic / Diastolic sitting both arms manual / automatic cuff; (m) Uipid Profile: Total Cholesterol, HDU-C, LDL-C, Triglycerides; (n) Weight in lbs.; (o) Well Cell Provider Questionnaire (clinician facing); (p) Validated Questionnaires: RAND-SF-36 (patient facing), Goldberg Depression (patient facing), Michigan Neuropathy Screening (if applicable, Part 1 patient facing and Part 2 clinician facing foot examination), VF-14 QCU (if applicable, patient facing), and / or PDQ-39 (if applicable, Part 1 patient facing and Part 2 clinician facing).

[0118] Some of these metrics may be imported from an Electronic Medical Record (EMR) rather than being manually inputted; however, all questionnaires are completed directly within the Well Cell Portal software.PIR Care Plan Creation

[0119] If the patient is determined by the physician to be eligible for PIR treatment, a PIR Care Plan may be created showing how often treatments will occur and the protocol regarding insulin delivery, such as discrete, continuous, or otherwise. The PIR Care Plan may also include any medication to be monitored / adjusted / discontinued (if applicable), recommended diet and / or exercise instructions, and goals for transitioning the patient out to more time between PIR infusion treatments.

[0120] An initial target therapeutic blood glucose range and the insulin concentration to be administered in the first PIR treatment is determined. The insulin concentration administered during the first PIR treatment may be selected based on: (1) the patient’s degree of insulin resistance, (2) their HbAlc, and / or (3) their compliance with treatments to date (e.g., medications, diet, exercise, etc.). Table 3 below may guide physicians in selecting the insulin concentration to be used. For non-compliant patients, the concentration may be decreased based on the physician’s discretion. The blood glucose target level at the end of the infusion, referred to as the “Session Landing,” may be determined by the physician in advance and is typically focused at bringing the patient to a euglycemic level, typically within the last 30 minutes of the infusion.leaf’s Kit InAitts ITSHBAisPIR Tteattnm nemcn A:ss Ranee: Mtetoely Resistant GRREX PIR Sessscsx Ssei Ranee:Senath* &$...74 BLS PIR Tteattam Session Tarset Range:ISMSC- Saasifeve 57- $4 strrs PIR Treatrnm Semen Tsrsst fame 5.4 AS P5L Trtsatn-ent Sesssx: Raw: < > ? SLi PIR Tteate-ettt Sessren. Tarset Range:Table 3 - Well Cell Portal Recommendations for PIR Therapeutic Blood Glucose Level and Respective PIR Insulin Concentration Reference Names

[0121] Table 3’s last three rows are used as the patient's HbAlc and insulin resistance improve over the course of PIR treatment.

[0122] The following Table 4 defines the volumes of insulin and saline 0.9% that may be required to achieve each of the recommended PIR insulin concentrations.2.0 IS.OBJL GREEN l.OsaL UOisl Wife 07 ask UAtsL WHITE 0.22 SBLCW & V ;X;L USUsaL LUKA 0.0?>2>xst lORUSnsLTable 4. Volumes of Insulin and Saline required to achieve PIR Insulin Reference Concentrations

[0123] Moving on, PIR Care Plans may be broken down into two distinct phases. First the “Induction Phase” may be at or about 4 weeks and focused on upregulating glucose metabolism, re-establishing insulin sensitivity, and / or reducing medications. This phase may start with PIR treatment sessions being administered twice per week, on consecutive days, for 4 weeks (these are termed “Upregulation” appointments and may be referred to as UR1-UR8). Of course, other example induction phase routines may be used.

[0124] Second is the “Maintenance Phase.” PIR treatment sessions post the induction phase may initially occur weekly (and may be termed “T7” appointments) and focused on maintaining upregulated glucose metabolism, insulin sensitivity, and / or diabetes medication reductions. If improvement is sustained, the physician may choose to extend time between PIR treatment sessions (e.g., biweekly or “T14 appointments” rather than weekly “T7”) and / or reduce the length of each individual PIR treatment session (e.g., 2-hour total infusion durations rather than 3 hours or otherwise) to identify the minimum PIR treatment needed to maintain the improved patient outcome.PIR Treatment Infusion SessionsPre-Treatment Preparation Instructions for PIR Treatment Sessions

[0125] Patients may be instructed to fast, e.g.. 4 hours, before their PIR treatment session begins. Patient pre-treatment medication instructions may be highly individualized; however, patients with morning appointments, for instance, may be asked to take half their normal dose of diabetic medications the night before and take no diabetic medications the morning of PIR treatment sessions. For afternoon appointments, patients may be asked to take half their normal dose of diabetic medications in the morning leading up to treatment. Otherwise, there are generally no changes made to any other prescribed medications. Patients with a subcutaneous insulin pump may be asked to turn off the pump, e.g., 2 hours, before the treatment or as directed by the treating physician.Starting a PIR Treatment Session in the Well Cell Portal

[0126] A webpage, such as the homepage, of the Well Cell Portal application may show each physician appointment(s) for a day or otherwise in anticipation of those patients arriving at their clinic. When a patient presents for an appointment at the clinic, that patient’s appointment button may be clicked in the Well Cell Portal and the application may ask for a confirmation of the appointment type, e.g., UR1-8, T7, etc. Upon confirming, a blank infusion flowsheet record may appear to help begin documenting the appointment, such as depicted in FIG. 1.Pre-Infusion Data Entry and Infusion Plan Determination

[0127] The infusion flowsheet may show basic patient information, e.g., height, weight (to be entered prior to start of PIR treatment session), indication type (T1DM, T2DM, CKD, etc.), and any allergies. From there, clinical staff may collect information about the patient’s most recent meal (if applicable) or indicates the patient has been fasting. See FIG. 2.

[0128] A physician may designate a target therapeutic blood sugar range depending, at least partially, on the patient’s existing PIR Care Plan and prior infusions. If the patient has not had prior infusions, the Well Cell Portal may provide guidelines for the target therapeutic blood sugar range, the “landing” or blood glucose value intended at the end of the PIR treatment session, and / or infusion bolus volume (sometimes referred to as “V” setting in an infusion pump menu). Please note that the Well Cell Portal may recommended infusion interval (sometimes referred to as “Q” setting) for patients initially as Q6, i.e., infusion bolus every 6 minutes). However, the physician or clinician may adjust this value based on prior infusions and clinical judgment.Table 5 - Well Cell Portal Recommended Infusion Pump Settings Based on Patient’s Initial Blood Glucose

[0129] For rows for Resistant and Moderately Resistant categories, also see Table 3. The Q setting for all patients initially may be Q6 (e.g., infusion bolus every 6 minutes) or otherwise for all or some patients.

[0130] The physician may choose from one of the Well Cell Portal recommended target ranges or can specify a unique maximum and minimum for that patient. As an example, patients who state they become diaphoretic even at high blood sugar levels (e.g., they state they become diaphoretic at 240 and their HbAlc is 14) may require a custom range.

[0131] The physician may chose the insulin type and / or concentration for a patient and duration of the PIR infusion session. These data may be determined prior to beginning an infusion, such as during the initial patient intake in a visit.

[0132] While these fields are being completed, the patient’s vitals may be taken by a person working with the physician and entered into the application (e.g., temperature, blood pressure, pulse rate, respiratory rate). The patient’s subcutaneous insulin pump may be turned off and the timing / dosing of the patient’s last diabetes medications, including insulin, may be recorded in the Well Cell Portal, such as in the session comments section. Some or all data may be used by the physician to decide whether it is appropriate to perform a PIR infusion.

[0133] FIG. 3 is an example screenshot of \input of the physician’ s decisions for blood glucose target / landing, insulin type / concentration, vitals, IV line details, and infusion pump used or to be used.

[0134] If for any reason, the physician or their QHP are uncomfortable with the patient’s vitals, no treatment should be administered, and appropriate medical attention provided to the patient. If vitals are appropriate for care, then an IV may be administered.Placing IV and Administering Initial Oral Dextrose

[0135] Once a physician has decided to perform treatment on the patient during their appointment, an IV may be placed and information about the needle, site, and insulin pump may be entered. At this point, all further information in the flowsheet may be automatically signed with the initials of the physician entering the data.

[0136] Once the IV line is placed, initial blood glucose may be measured using, for example, either the patient’s existing CGM or glucometer. It is recommend that an infusion on a patient does not commence unless a patient’ s blood glucose level exceeds 120 mg / dL. Associated software may prevent initiation of a PIR infusion if a blood glucose value is under 120, unless overridden in some embodiments, and / or may generate an additional line for asubsequent blood glucose reading to be inputted after glucose is administered. FIG. 4 displays an infusion flowsheet showing an out-of-range blood glucose level.

[0137] There may be a physician override to this halt as just mentioned. A physician may manually confirm the override (multiple confirmations required) which allows for proceeding with the infusion. This action is attributed to the physician who initiated the override and may be stored in associated logs in the portal.

[0138] Glucose (dextrose) may be administered orally to fuel glucose metabolism uptake and to keep blood glucose level in a prescribed target therapeutic range as trained and determined by the treating physician. The physician may determine the patient’s target therapeutic blood glucose range by factoring in several individual items on a patient-by-patient basis (avoidance of relative hypoglycemia or hyperglycemia, gastroparesis, medication impacts, patient-reported comfort, etc.).

[0139] Based on the patient’s initial blood glucose level and the decision of insulin concentration and targeted therapeutic blood glucose range, associated, enabling logic, such as in the Well Cell Portal, may recommend the following oral intake of dextrose prior to the start of the insulin infusion as shown below in Table 6.s to- totsssss stesteg Igssss} wgsted Wste tow: tssgssto Sseod tosse:teas Wites ms? to toitof toirw, siter 5Ssss. to sttaW, sites- tsteg-ws- teg-teg te<5g-teg itoteg m to w? sto® sttotoK tarn sstosst iteg-teg Ateteftete-r sssi sm i tote”;Atotat (taw sste stsst te&to sTable 6. Well Cell Portal recommendations for initial dextrose dosing based on patient's initial blood glucose levelInitiating the PIR Insulin Infusion

[0140] Once the provider has decided to start the infusion, the infusion pump may be turned on and the physician may select a chosen insulin dose volume and frequency intervalfrom the device library menu. The physician may have 5 different bolus volume settings (indicated in column “V”) from which to choose, ranging, for example, from 0.20-0.60 mL and 7 different frequency settings, for example, (indicated in column “Q”) to choose from ranging, for example, from 3 minutes apart to 9 minutes apart. Of course, variations in V, Q, and time ranges are possible with different example embodiments. As noted above, the Well Cell Portal may recommend bolus volume “V" setting based on the patient’s blood glucose level at the start of the session. And even if the Well Cell Portal recommends a default infusion interval (referred to as “Q” setting) for all patients initially as Q6 (i.e., infusion bolus every 6 minutes) or otherwise, the clinician may adjust this value based on prior infusions and / or clinical judgment. See FIG. 7 of infusion flowsheet showing data capture of physician decisions for pump settings and dextrose administration.

[0141] After the pump settings are selected, the provider may start the infusion on the pump, logging the infusion information into the Well Cell Portal, and beginning the treatment.Patient Monitoring and Titration by Oral Dextrose and Infusion Pump Settings

[0142] Vitals may be taken hourly at a minimum and input, optionally required, into the Well Cell Portal for the PIR treatment session. Each individual “hour” on the infusion flowsheet may have a maximum allowable time of 60 minutes. This time may be broken down any way a clinician or provider or patient chooses (e.g., 4 sets of 15 minutes, 3 sets of 20, etc.). If an attempt to exceed the hour threshold is made, the system may prevent that number from being added and may issue an audio and / or visual warning.

[0143] Throughout the first PIR treatment session, blood sugar levels may be monitored using a CGM or glucometer, such as at approximately every 14 minutes then 16 minutes, and every 25 and then 35 minutes for subsequent PIR treatment sessions, assuming stability. A physician may navigate between past and present infusion sessions at any time.

[0144] Selecting the “Previous” button at the top may take the user to the previous infusion session recorded in the system that they may use for reference in guiding the current infusion session. They may traverse back and forth between saved sessions, and the “Current” button may return them to the active infusion session.

[0145] If the patient’s blood sugar level is trending higher or lower than the physician-defined target therapeutic range, the Well Cell Portal may alert the physician. The physician may clear the alarm and maintain course or administer oral glucose, and / or change infusion pump settings in order to achieve the targeted therapeutic blood glucose range. See FIG. 5 of infusion flowsheet showing first hour data capture for blood glucose level (B / S), infusion pump settings (Q and V, for interval duration and bolus volume settings, respectively), Glucose (oral dextrose taken), and Pump (identification number of infusion pump used).

[0146] As shown in Table 7, methods, systems, and device have further enabling logic that recommends the following with respect to oral dextrose titration and infusion pump settings based on the inputted blood glucose (and assuming vitals remain stable):Ores essrejfssirere's ®<re<3fei'.OSi res "<fe !S f'OjSOSO jSXtSSSB ?re T3;g*t>-:iS oiood sfere-s:Ctesre s&wsg g&sSOite'W:Pa^sreSCstffifss isre-iivrejM RR^fgM!Cress &ue (ta W wisHs&sf WKJ ®rcio«s4Sg-2s W'^ Cgss;o miWss 8^»W i 52r-iSS<re.^X W-W W"W S8g~15S W*s? icsSiX«KS KW MM X s®»^ ® V i iswsm soS■sorere*® teife sofosisn cscre® dssSssre MM ssifssfcss X ssSssg fef x isrcressscfis are so Ks> sssgm iresjs* reJvSsw sssre®?Table 7. Well Cell Portal recommendations for titration of oral dextrose dosing based on patient's blood glucose level during PIR infusion session *See source document Table 6 for complete titration dosing table including all blood glucose ranges.

[0147] During an infusion session, a pump may need to be paused (e.g., if blood sugar level drops below 90 mg / dL). When that occurs, selecting the “Pump Pause” button at the interval check may force a line to show that the pump has been paused. The time a pump is paused is still considered part of the appointment, and the flowsheet may consider that time to be a portion of an hour (e.g., if the pump is paused for 20 minutes, then the infusion hour mayallow only 40 minutes of active pump time before requiring a new hour and vitals to be inputted). See FIG. 6, which is a snapshot or screenshot of Well Cell Portal infusion flowsheet showing paused infusion pump.Physician Commentary during PIR Infusion Session

[0148] To capture events, activities, and observations during a PIR infusion session, several comment fields are present in the infusion flowsheet record. The primary comment field may be located in the bottom right corner of the record on the screen and floats down the page. At any point during the infusion, any user may enter the comments window, read previous comments, and add additional comments at their discretion. Added comments may be timestamped and their authors logged. Comments (and other fields of the infusion session) may be edited / deleted later; however, all information may be saved and stored in the databases for auditing purposes and completeness of data. See FIG. 7, which is a snapshot or screenshot of Well Cell Portal infusion flowsheet showing pop-up session comments box.

[0149] In the event that a medication must be administered by the staff during an infusion, there is a comment bar near the bottom of the page that may allow a user to enter the time, the medication used, and / or any relevant comments. This information may be saved as part of the flowsheet record as well. If more than one medication is needed, each medication may receive an individual line to maintain separation. See FIG. 7, which is a snapshot or screenshot of Well Cell Portal infusion flowsheet showing session comments, session medications given, and comments for next session.

[0150] The disclosed methods, systems, devices may also include another section for commentary regarding future infusions. If a physician chooses to make modifications to a patient’s infusion or care plan as a result of the data from this infusion (e.g., a change in concentration, a change in duration, or an additional questionnaire needing completion, etc.), then they may put this information in the “Comments for next session” bar. These comments may be translated over to the next infusion under the “Last Session Notes” bar at the top of the infusion flowsheet for reference.Concluding a PIR Infusion Session

[0151] During the course of the treatment session, the physician may review the patient's previous PIR flowsheets to help determine how the patient reacted to glucose and pumpsettings, and to help sharpen the plan for going forward. As the patient becomes more sensitive to insulin and glucose, adjustments may be made by the provider to the dosing and timing to optimize the amount of glucose and insulin the patient is administered to achieve their desired clinical outcomes. Typically, patients are able to increase the amount of glucose uptake their cells are able to process in a shorter infusion time period, all determined and managed by the physician. Data inputted into the Well Cell Portal will reflect changes in insulin delivered, infusion time, and / or serum glucose results.

[0152] In approximately the last 30 minutes of the infusion or otherwise, the treating physician may begin to reduce the glucose and potentially increase the IV insulin (“V” setting) to return the patient to their predefined “landing” point. This is typically a euglycemic level; however, each patient’s landing is individualized and predetermined by the trained treating physician.

[0153] Final comments may be collected on any patient behavioral responses during the course of the treatment session, including any medications delivered accordingly (e.g., antiemetics, IV dextrose, etc.) during the course of the treatment session and inputted into the Well Cell Portal.

[0154] As the infusion is reaching its end, the provider has administered the amount of insulin and glucose they are satisfied with and the patient is comfortable back to the blood sugar target therapeutic range the provider defined in the care plan; the pump is then turned off.

[0155] The IV may be disconnected, removed, and the IV site wrapped with gauze and coban wrap. The patient is encouraged to stand up and walk around. The clinician may take about 10-15 minutes to speak with the patient and review the treatment with them and how they felt about the process. Comments may be taken about how the patient feels physically and are referred to in guiding future PIR treatment sessions. The clinician may sign off on the session record and a document or record is generated by the Well Cell Portal software. See FIG. 9. Well Cell Portal, which shows an end-of infusion flowsheet.

[0156] The patient may receive individualized instructions regarding food / drink and / or medications to follow the infusion based on the physician's determination. Before the patient leaves, the next PIR appointment may be scheduled in accordance with the patient’s PIR CarePlan, and may differ as the patient moves through the induction phase to the maintenance phase of their care.Software Development Description

[0157] Modules of the Well Cell Portal are listed below and discussed as they cooperatively work on computer systems to enable the disclosed methods, systems, and devices disclosed herein.

[0158] Data Collection module may be comprised of two sub-modules: (i) Manual collection (user interface) of all data described in the patient assessment section above. This module may provide the user with as many dropdown or system-selected options as possible. The system also has anticipated ranges for auto-validation of the entered data. These options minimize the possibility of data entry errors and streamline data collection; and (ii) Interface to the Electronic Medical Records, insulin pumps, and to the glucose monitors / glucometer to capture insulin and glucose (including amount, dose, date, and time).

[0159] Reports module may allow for operational and usage reports.

[0160] Analysis / Recommendation module analyzes all collected data (manually and via interfaces) and presents the physician with recommendations for continued treatment.

[0161] Administrative module allows the user / administrator to define the system parameters, configure interfaces to insulin pumps and glucose monitors, set various defaults, lookup tables, users, and security.Back-End Description of The Well Cell Portal Application

[0162] All source code is developed on developer local machines using IDEs such as VSCode or other options. The programming language utilized is TypeScript. The code is version controlled using private GitHub repositories. The deployment logic for the application is created using Terraform and GitHub Actions, which orchestrate the deployment of the application to AWS, Azure, and / or other cloud-based system. Deployments occur in response to code merges / pushes to different git branches (development, production, etc.).

[0163] The backend API of the platform is based on Nest.IS / Node.js and containerized with the images sent to AWS ECR (Elastic Container Registry) and run on AWS ECS (Elastic Container Service). The ECS containers sit behind an AWS Elastic Load Balancer,which handles the SSL termination and load balancing of requests to the backend. A Postgres RDBMS is run via AWS RDS for all data persistence. The frontend of the application is based on the React. js library and the compiled JavaScript bundle is hosted on AWS S3 with a CloudFront CDN service for optimized distribution.

[0164] End users access the Well Cell Portal application by retrieving the front-end JavaScript application to their browser, which then real-time renders the UI as a single-page application (SPA) and makes API requests to the backend for the application data. All communication between the browser frontend and the backend API is performed using TLS 1.3 with 256-bit AES encryption.Current Pathway for Data CapturePIR Infusion Flowsheet

[0165] After a flowsheet is completed and signed by the provider (and the flowsheet is locked), a JSON containing the information from the infusion, including patient information and appointment information, is sent to an HTTP API endpoint presented from AWS API Gateway. The API POST request is validated using a JWT authorization method with the issuer as AWS Cognito.

[0166] Once the handshake is completed, the API Gateway routes the information contained within the request to an AWS Lambda function. From this point onward, the data does not have exposure to the Internet and is contained entirely within the AWS cloud infrastructure, except for one singular case that will be described below.

[0167] This AWS Lambda function strips the JSON containing the flowsheet information from the HTTP request and writes the information to two separate files with the name of the patient and the date, and adds the file to an S3 bucket for storage. This S3 bucket collects all incoming flowsheets and organizes them by clinic and then by date. Only the handler Lambda function can upload to this bucket, and this bucket is not public. The two generated files are identically named, except one is a TXT file and the other is a JSON file. This is done because each file has a separate Lambda trigger. These actions are all logged in AWS CloudWatch under the appropriate log groups. This step is primarily for error handling in the database, in which data may not be added in directly. This ensures that no data is lost and can always be remapped into the database later if necessary.

[0168] The TXT file creation triggers an additional Lambda function that separates the incoming files into buckets organized by clinic, with date subdirectories in the bucket. This helps keep the data accessible to individual clinics without any possibility of clinics seeing PHI of outside patients.

[0169] These buckets are only accessible to the Lambda function and are not public facing. These actions are logged in AWS CloudWatch.

[0170] The JSON file creation triggers a Lambda function that engages in the mapping of the flowsheet information into the database. The function collects the information from the JSON file and uses Python's Psycopg2 module to log into an RDS instance running on a PostgreSQL engine that houses the databases for each active clinic. Each individual clinic has its own database to keep data separate, using the bridge model in which one RDS instance hosts multiple clinics' databases. Clinics will be able to engage in more advanced reporting using only their clinical data at a later time. This helps keep that data siloed for them. The data is mapped to the database and each query that adds data is logged using a Python logger that is captured via AWS CloudWatch logs.

[0171] At the time where the functionality is being developed, the appropriate pathways to access data will be generated for an Internet-based request to pull information from the buckets and databases. This model helps ensure that in the event of a single-clinic data breach, the other databases and buckets are unaffected to minimize data leakage.

[0172] The RDS instance only has the ability to be accessed and written to via the Lambda function as described above. The security group that covers the RDS instance has an incoming rule that allows the security group containing only the Lambda function to have access, inside the VPC where they both reside. There are no other active points of entry, which allows the data to reside safely in the RDS databases.Temporary Point of Entry

[0173] There is a provision for a temporary direct entry to read and access the databases for a user outside of the Lambda function pathway described previously. A bastion host in AWS EC2 is also included in the security group with direct access to the RDS database. This bastion host has the capability to access the databases directly using administrative credentials. This is used periodically to cross-check relations or field lengths, etc. The bastionhost is connected to the Internet with a public-facing IPv4 address but is only accessible on port 22 (SSH) using a .PEM access key that is securely stored. This access method is temporary only, and when not being actively used, all incoming traffic is blocked from this IP address and the host is also disconnected from the RDS instance.Questionnaires

[0174] The flow of questionnaires is very similar to the flowsheet workflow above, except the questionnaires are presented to a distinct API endpoint and use separate Lambda functions to process, store, and map the questionnaires to the databases. Logs are also kept separate for invocations of these Lambda functions. The S3 buckets and RDS databases used are the same.Data Aggregation

[0175] The RDS instances will have another set of Lambda functions that will regularly push de-identified data to a larger data warehouse housed in AWS. This larger database can then be used for broad-spectrum analyses, clinical retrospective research, and other applications. This database must be accessible via the Internet but will be exclusively interrogable from the Well Cell Portal, after the user has authenticated themselves for access under a role that only permits access to the de-identified data.Analysis / Recommendation Module

[0176] During a PIR infusion session, certain data from the PIR infusion flowsheet — such as target blood glucose, current blood glucose, previously inputted Q and V settings, etc. — may be used as inputs to the recommendation engine. These inputs are passed during the infusion through an API endpoint to the AWS API Gateway. From there, they are collected by a Lambda function and fed into the recommendation engine. The source code for the engine will be securely stored in AWS. The engine will accept these inputs and return a suggested Q, V, and dose of dextrose for the physician to view prior to the next blood glucose check. The physician can either accept those suggestions or reject them via pressing a button that accompanies the window with the recommendations.More Example Embodiments

[0177] In various example embodiments of devices, a device may comprise one or more processors, memory storing executable instructions, and optionally a display devicecomprising audio output and / or visual output, the executable instructions when executed by the one or more processors causing the device to provide a digital flowsheet configured to receive, in real time or near real time, data comprising at least one of blood glucose value(s), insulin administration parameter(s), and glucose amount(s); validate completeness of the data received, and optionally validate physiological plausibility, including cross-variable coherence of the data received; and generate prescriptive analytics tailored to rhythmic bolus physiologic insulin resensitization, wherein the prescriptive analytics comprise analytics for insulin bolus volume, bolus frequency, and glucose dosing. For “glucose amount(s),” which may be just dextrose, that may be oral, IV, enteral, parenteral, sublingual or any other administration route, whether in various device embodiments or various example method and / or system embodiments. In various embodiments, the device may further comprise a rules engine configured to apply deterministic logic to generate the prescriptive analytics; a machine -learning engine configured to refine treatment recommendations based on prior treatment-session data; an alert module configured to detect out-of-range blood glucose values and generate alerts; a clinician-override module configured to accept and record clinician-entered modifications to the prescriptive analytics; a laboratory-integration module configured to receive and process laboratory values; a questionnaire-administration module configured to present validated medical questionnaires; a treatment-scoring engine configured to evaluate clinician decisions; and components configured to validate completeness and physiological plausibility, capture infusion pump settings including interval duration (Q) and / or bolus volume (V), capture session comments and medications administered, generate recommendations for treatment-session duration or frequency, insulin concentration selection, and achieving a target landing blood glucose, capture vital signs, compute or receive an eligibility qualification verification score, and generate a care plan comprising a target therapeutic blood glucose range, an initial insulin dosing protocol, and an oral dextrose administration protocol.

[0178] In various example embodiments of methods, a computer-implemented method may comprise receiving, in real time or near real time by a computing system, data comprising at least one of blood glucose value(s), insulin administration parameter(s), and glucose amount(s); validating completeness of the data received, and optionally validating physiological plausibility, including cross-variable coherence of the data received; and generating prescriptive analytics tailored to rhythmic bolus physiologic insulin resensitization, the prescriptive analytics comprising analytics for insulin bolus volume, bolusfrequency, and glucose dosing. In various embodiments, the method may further comprise generating an alert when a blood glucose value is out-of-range; applying a rules engine to generate deterministic recommendations; applying a machine-learning engine to refine treatment recommendations based on prior treatment-session data; receiving laboratory values via a laboratory-integration module; receiving questionnaire responses via a questionnaire-administration module; receiving vital signs including temperature, blood pressure, heart rate, and respiratory rate; recording clinician-entered overrides of the prescriptive analytics; scoring clinician decisions using a treatment-scoring engine; storing, subsequent to the validating, the data received, the data validated, and / or the prescriptive analytics; tracking one or more metabolic parameters; validating physiological plausibility including verifying human-possible physiological ranges, biologically realistic rates of change, and associations between insulin bolus events and glucose measurements; capturing infusion pump settings including interval duration (Q) and / or bolus volume (V); capturing session comments, medications administered, and comments for a next session; generating recommendations for treatment-session duration or frequency, insulin concentration selection, and a target landing blood glucose; computing or receiving an eligibility qualification verification score; and generating a care plan comprising a target therapeutic blood glucose range, an initial insulin dosing protocol, and an oral dextrose administration protocol.

[0179] In various example embodiments of systems, a computing system may comprise one or more processors, memory storing executable instractions, and non-transitory storage, the executable instructions when executed by the one or more processors causing the computing system to receive, in real time or near real time, data comprising at least one of blood glucose value(s), insulin administration parameter(s), and glucose amount(s); validate completeness of the data received, and optionally validate physiological plausibility, including cross-variable coherence of the data received; and generate prescriptive analytics tailored to rhythmic bolus physiologic insulin resensitization, the prescriptive analytics comprising analytics for insulin bolus volume, bolus frequency, and glucose dosing. In various embodiments, the computing system may further be configured to generate an alert when a blood glucose value is out-of-range; apply a rales engine to generate deterministic recommendations; apply a machine-learning engine to refine treatment recommendations; receive laboratory values via a laboratory-integration module; receive questionnaire responses via a questionnaire-administration module; receive vital signs; recordclinician-entered overrides; score clinician decisions; store the data received, the data validated, and / or the prescriptive analytics; track metabolic parameters; validate completeness and physiological plausibility including enforcing required PIR-treatment parameters, verifying time-dependent entries, verifying human-possible physiological ranges, verifying biologically realistic rates of change, and verifying associations between insulin bolus events and glucose measurements; provide a digital flowsheet that captures infusion pump settings including interval duration (Q) and / or bolus volume (V); provide a digital flowsheet that captures session comments, medications administered, and comments for a next session; generate recommendations for treatment-session duration or frequency, insulin concentration selection, and a target landing blood glucose; compute or receive an eligibility qualification verification score; and generate a care plan comprising a target therapeutic blood glucose range, an initial insulin dosing protocol, and an oral dextrose administration protocol.

Claims

CLAIMSWhat is claimed is:

1. A device comprising :one or more processors, memory storing executable instructions, optionally a display device comprising audio output and / or visual output, wherein the executable instructions when executed by the one or more processors causing the device to:provide a digital flowsheet configured to receive, in real time or near real time, data comprising at least one of blood glucose value(s), insulin administration parameter(s), and glucose amount(s);validate completeness, and optionally physiological plausibility, including crossvariable coherence of the data received; andgenerate prescriptive analytics tailored to rhythmic bolus physiologic insulin resensitization, wherein the prescriptive analytics comprise analytics for insulin bolus volume, bolus frequency, and glucose dosing.

2. The device of claim 1, further comprising a rules engine configured to apply deterministic logic to generate the prescriptive analytics.

3. The device of claim 1, further comprising a machine-learning engine configured to refine treatment recommendations based on prior treatment-session data.

4. The device of claim 1 , further comprising an alert module configured to detect out-of-range blood glucose values and generate alerts.

5. The device of claim 1, further comprising a clinician-override module configured to accept and record clinician-entered modifications to the prescriptive analytics.

6. The device of claim 1, further comprising a laboratory-integration module configured to receive and process laboratory values associated with a patient.

7. The device of claim 1, further comprising a questionnaire-administration module configured to present validated medical questionnaires to a patient.

8. The device of claim 1 , further comprising a treatment-scoring engine configured to evaluate clinician decisions at individual decision points.

9. The device of claim 1, further comprising to validate the completeness and the physiological plausibility.

10. The device of claim 9, wherein to validate the completeness comprises to verify that timedependent entries occur within predefined intervals.

11. The device of claim 9, wherein to validate the physiological plausibility comprises to verify that the data received falls within human-possible physiological ranges.

12. The device of claim 9, wherein to validate the physiological plausibility comprises to verify that the blood glucose values exhibit biologically realistic rates of change.

13. The device of claim 1, wherein to validate the completeness comprises to enforce entry of required physiologic-insulin-resensitizati on-treatment parameters prior to generation of the prescriptive analytics.

14. The device of claim 1, wherein to validate the completeness comprises to verify that the insulin bolus events are associated with corresponding glucose measurements.

15. The device of claim 1 , wherein the digital flowsheet is configured to capture infusion pump settings including interval duration (Q) and / or bolus volume (V).

16. The device of claim 1, wherein the digital flowsheet is configured to capture session comments, medications administered, and comments for a next session.

17. The device of claim 1, wherein the prescriptive analytics further comprise recommendations for treatment-session duration or frequency and / or insulin concentration selection.

18. The device of claim 1, wherein the prescriptive analytics further comprise recommendations for achieving a target landing blood glucose.

19. The device of claim 1, wherein the digital flowsheet is configured to capture vital signs comprising at least two or more of temperature, blood pressure, heart rate, and respiratory rate.

20. The device of claim 1, wherein the digital flowsheet is configured to receive, in real time or near real time, data comprising the blood glucose value(s), the insulin administration parameter(s), and the glucose amount(s).

21. The device of claim 1, further comprising an analytics engine configured to compute or receive an eligibility qualification verification score based on patient laboratory values, questionnaire responses, and / or treatment-session data.

22. The device of claim 1, wherein the device is further configured to generate a care plan comprising a target blood glucose range, an initial insulin dosing protocol, and an oral dextrose administration protocol.

23. A computer-implemented method comprising:receiving, in real time or near real time by a computing system, data comprising at least one of blood glucose value(s), insulin administration parameter(s), and glucose amount(s);validating completeness of the data received, and optionally validating physiological plausibility, including cross-variable coherence of the data received;generating prescriptive analytics tailored to rhythmic bolus physiologic insulin resensitization, the prescriptive analytics comprising analytics for insulin bolus volume, bolus frequency, and glucose dosing.

24. The computer-implemented method of claim 23, further comprising generating an alert when a blood glucose value is out-of-range.

25. The computer-implemented method of claim 23, further comprising applying a rules engine to generate deterministic recommendations.

26. The computer-implemented method of claim 23, further comprising applying a machinelearning engine to refine treatment recommendations based on prior treatment-session data.

27. The computer-implemented method of claim 23, further comprising receiving laboratory values via a laboratory-integration module.

28. The computer-implemented method of claim 23, further comprising receiving questionnaire responses via a questionnaire-administration module.

29. The computer-implemented method of claim 23, further comprising receiving vital signs including temperature, blood pressure, heart rate, and respiratory rate.

30. The computer-implemented method of claim 23, further comprising recording clinician- entered override(s) of one or more of the prescriptive analytics.

31. The computer-implemented method of claim 23, further comprising scoring clinician decisions using a treatment-scoring engine.

32. The computer-implemented method of claim 23, further comprising storing, subsequent to the validating, the data received, the data validated, and / or the prescriptive analytics.

33. The computer-implemented method of claim 32, further comprising tracking one or more metabolic parameters in the data subsequent to receipt and validation.

34. The computer-implemented method of claim 23, wherein the validating further comprises validating physiological plausibility.

35. The computer-implemented method of claim 34, wherein the validating of the completeness comprises enforcing entry of required PIR-treatment parameters prior to the generating of the prescriptive analytics.

36. The computer-implemented method of claim 34, wherein the validating of the completeness comprises verifying that time-dependent entries occur within predefined intervals.

37. The computer-implemented method of claim 34, wherein the validating of the physiological plausibility comprises verifying that data received falls within humanpossible physiological ranges.

38. The computer-implemented method of claim 34, wherein the validating of the physiological plausibility comprises verifying that the blood glucose value(s) exhibit biologically realistic rates of change.

39. The computer-implemented method of claim 34, wherein the validating of the physiological plausibility comprises verifying that each insulin bolus event is associated with a corresponding glucose measurement.

40. The computer-implemented method of claim 23, wherein the digital flowsheet is configured to capture infusion pump settings including interval duration (Q) and / or bolus volume (V).

41. The computer-implemented method of claim 23, wherein the digital flowsheet is configured to capture session comments, medications administered, and comments for a next session.

42. The computer-implemented method of claim 23, wherein the generating of the prescriptive analytics further comprises recommending treatment-session duration and / or frequency.

43. The computer-implemented method of claim 23, wherein the generating of the prescriptive analytics further comprises recommending an insulin concentration associated with a reference name for a specific insulin concentration or range.

44. The computer-implemented method of claim 23, wherein the generating of the prescriptive analytics further comprises recommending a target landing blood glucose.

45. The computer-implemented method of claim 23, further comprising computing or receiving an eligibility qualification verification score based on laboratory values, questionnaire responses, and / or treatment-session data.

46. The computer-implemented method of claim 23. further comprising generating a care plan comprising a target therapeutic blood glucose range, an initial insulin dosing protocol, and an oral dextrose administration protocol.

47. The computer-implemented method of claim 23, wherein the receiving, in real time or near real time by the computing system, of the data comprises the blood glucose value(s), the insulin administration parameter(s), and the glucose amount(s).

48. A computing system comprising one or more processors, memory storing executable instructions, and non-transitory storage, the executable instructions when executed by the one or more processors causing the computing system to:receive, in real time or near real time, data comprising at least one of blood glucose value(s), insulin administration parameter(s), and glucose amount(s);validate completeness of the data received, and optionally validate physiological plausibility, including cross-variable coherence of the data received; andgenerate prescriptive analytics tailored to rhythmic bolus physiologic insulin resensitization, the prescriptive analytics comprising analytics insulin bolus volume, bolus frequency, and glucose dosing.

49. The computing system of claim 48, further configured to generate an alert when a blood glucose value is out-of-range.

50. The computing system of claim 48, further comprising a rules engine configured to generate deterministic recommendations.

51. The computing system of claim 48, further comprising a machine-learning engine configured to refine treatment recommendations based on prior treatment-session data.

52. The computing system of claim 48, further comprising a laboratory integration module configured to receive laboratory values.

53. The computing system of claim 48, further comprising a questionnaire administration module configured to receive questionnaire responses.

54. The computing system of claim 48, further configured to receive vital signs including temperature, blood pressure, heart rate, and respiratory rate.

55. The computing system of claim 48, further configured to record clinician-entered overrides of one or more of the prescriptive analytics.

56. The computing system of claim 48, further comprising a scoring engine configured to score clinician decisions.

57. The computing system of claim 48, further comprising storage for the data received, the data validated, and / or the prescriptive analytics.

58. The computing system of claim 57, further configured to track one or more metabolic parameters in the data stored.

59. The computing system of claim 48, wherein the executable instructions when executed by the one or more processors causes the computing system to validate the physiological plausibility.

60. The computing system of claim 59, wherein to validate the completeness comprises to enforce entry of required physiologic-insulin-resensitization parameters prior to generation of the prescriptive analytics.

61. The computing system of claim 59, wherein to validate the completeness comprises to verify that time-dependent entries occur within predefined intervals.

62. The computing system of claim 59, wherein to validate the physiological plausibility comprises to verify that data received falls within human-possible physiological ranges.

63. The computing system of claim 59, wherein to validate the physiological plausibility comprises to verify that the blood glucose value(s) exhibit biologically realistic rates of change.

64. The computing system of claim 59, wherein to validate the physiological plausibility comprises to verify that each insulin bolus event is associated with a corresponding glucose measurement.

65. The computing system of claim 48, wherein the computing system is configured to provide a digital flowsheet that captures infusion pump settings including interval duration (Q) and / or bolus volume (V).

66. The computing system of claim 48, wherein the computing system is configured to provide a digital flowsheet that captures session comments, medications administered, and comments for a next session.

67. The computing system of claim 48, wherein to generate the prescriptive analytics further comprises to recommend treatment-session duration and / or frequency.

68. The computing system of claim 48, wherein to generate the prescriptive analytics further comprises to recommend an insulin concentration associated with a reference name for a specific insulin concentration or range.

69. The computing system of claim 48, wherein to generate the prescriptive analytics further comprises to recommend a target landing blood glucose.

70. The computing system of claim 48, further configured to compute or receive an eligibility qualification verification score based on laboratory values, questionnaire responses, and / or treatment-session data.

71. The computing system of claim 48, further configured to generate a care plan comprising a target therapeutic blood glucose range, an initial insulin dosing protocol, and an oral dextrose administration protocol.

72. The computer-implemented method of claim 48, wherein to receive, in real time or near real time by the computing system, of the data comprises the blood glucose value(s), the insulin administration parameter(s), and the glucose amount(s).