A pregnancy assistance system based on a blood glucose biosensor

Abstract

The present application relates to a kind of based on blood glucose biosensor pregnancy auxiliary system, belong to medical management technical field, especially, it is related to the extended application of biosensor.The system includes: a plurality of user terminals, for obtaining the data information related to user pregnancy, including the blood glucose data obtained based on blood glucose biosensor and the basic information entered;With the communication connection of user terminal server, wherein, the shared database of server construction can be divided into several subgroups based on the preset group division rule, so that each user can be based on its basic information and correspond to one subgroup in the server, the processed blood glucose data obtained by user in current gestation period can be compared with the processed blood glucose data obtained by other users in the same gestation period in corresponding subgroup, to analyze the current blood glucose management of the patient by determining the ranking way, wherein, server can determine the preset group division rule based on the category of basic information.

Description

Technical Field

[0001] This invention relates to the field of medical management technology, and in particular to a pregnancy support system based on a blood glucose biosensor. Background Technology

[0002] Gestational diabetes mellitus (GDM), one of the most common pregnancy complications worldwide, is characterized by increased insulin resistance in the second and third trimesters, leading to chronic hyperglycemia in the mother. This can result in obstetric complications such as macrosomia and preeclampsia, and significantly increase the long-term risk of metabolic syndrome in offspring. Traditional management relies on static indicators such as fasting plasma glucose (FPG) and glycated hemoglobin (HbA1c), but these methods have significant limitations: FPG only reflects instantaneous blood glucose levels and cannot capture the postprandial blood glucose fluctuation patterns unique to pregnancy; HbA1c is affected by the shortened lifespan of red blood cells, resulting in a systematic underestimation bias in the second and third trimesters. While the widespread adoption of continuous glucose monitoring (CGM) technology in recent years has increased data density, existing systems mostly use fixed thresholds (such as a target range of 3.9–7.0 mmol / L) for time-in-target (TIR) ​​calculations, failing to fully consider the dynamic changes in insulin sensitivity caused by gestational age, and lacking personalized reference standards for high-risk subgroups such as multiple pregnancies and obesity.

[0003] CN109692007A discloses a gestational blood glucose monitoring system based on blood glucose monitoring, including a real-time dynamic blood glucose monitoring device, a dynamic blood glucose monitoring workstation, and a blood glucose intelligent analysis server. This invention incorporates big data statistics and pre-sets threshold groups corresponding to pre-existing diabetes, diabetes complicated by pregnancy, or gestational diabetes to assess blood glucose control. Using received gestational user information and real-time blood glucose values, it generates a blood glucose diagnostic analysis report according to preset rules and sends this report to the dynamic blood glucose monitoring workstation, reminding relevant personnel to take timely action. This intelligent and accurate analysis of blood glucose control over a longer period provides excellent data support for subsequent blood glucose control.

[0004] Furthermore, since current research has not reached a consensus on the factors that cause gestational diabetes, it has been found that a large number of factors (especially hormone levels) can significantly affect insulin sensitivity, thereby causing insulin resistance.

[0005] In 2002, L. Barbour, J. Shao, L. Qiao, and others published an article titled "Human placental growth hormone causes severe insulin resistance in transgenic mice" in the American Journal of Obstetrics and Gynecology. This study, using a transgenic mouse model (5 mice overexpressing human placental growth hormone PGH vs. 6 control mice), revealed the role of PGH in insulin resistance. Experiments showed that the transgenic mice achieved PGH levels comparable to those of late pregnancy, weighed twice as much as the control group, had significantly increased bone mineral density, and experienced a slight decrease in body fat percentage. Transgenic mice showed a 4-fold higher fasting insulin level than the control group (1.57±0.22 vs 0.38±0.07 ng / mL, P<0.001), and a 7-fold higher insulin level 30 minutes after glucose stimulation (4.17±0.54 vs 0.62±0.10 ng / mL, P<0.0001). Insulin sensitivity was significantly reduced; transgenic mice showed only a slight decrease in blood glucose after insulin injection, while the control group showed a decrease of over 65% (P<0.001). The conclusion indicates that PGH, through fasting / postprandial hyperinsulinemia and insulin-responsive hypoglycemia, is highly likely a key mediator of insulin resistance during pregnancy.

[0006] Ryan, E., and Enns, L., published an article titled "Role of gestational hormones in the induction of insulin resistance" in *The Journal of Clinical Endocrinology and Metabolism* in 1998. Their research showed that placental prolactin (hPL) is closely associated with insulin resistance during pregnancy. hPL reduces glucose transport but does not affect insulin binding, which may lead to a post-binding defect in insulin action. Estradiol (E2) increases insulin receptor binding during pregnancy, but its effects may be offset by the influence of progesterone and cortisol, both of which reduce insulin binding and glucose transport.

[0007] In 2023, Hivert, M., White, F., Allard, C., and others published an article in Research Square entitled "Placental RNA sequencing implicates IGFBP1 in insulin sensitivity during pregnancy and in gestational diabetes." The study showed that IGFBP1 is positively correlated with insulin sensitivity. The research found that high expression of IGFBP1 in the placenta was associated with higher insulin sensitivity in mid-pregnancy, while low levels of IGFBP1 were associated with an increased risk of gestational diabetes. Circulating levels of IGFBP1 increased during pregnancy and decreased postpartum, suggesting that it may originate from the placenta and play a role in regulating insulin sensitivity during pregnancy.

[0008] In current clinical practice, blood glucose assessment and intervention decisions mainly face the following technical bottlenecks:

[0009] (1) Lack of dynamic adaptability of population reference standards: Existing clinical guidelines rely on static threshold assessments and cannot dynamically adjust the reference range according to the characteristics of pregnant women (such as age, metabolic indicators, and genetic background), resulting in assessment bias in heterogeneous populations.

[0010] (2) Insufficient multimodal data collaborative analysis: The fragmented storage of blood glucose monitoring data and non-time-series data such as pre-pregnancy metabolic indicators and pregnancy history hinders the modeling of the association between composite risk factors and real-time blood glucose fluctuations.

[0011] (3) Lack of longitudinal assessment system: Traditional methods mainly rely on single / single-day blood glucose testing, which lacks the ability to continuously track blood glucose changes throughout the entire pregnancy cycle and conduct cross-period comparative analysis.

[0012] (4) Fixed personalized early warning threshold: Early warning systems based on the population mean have difficulty identifying abnormal fluctuation patterns of individuals at specific gestational weeks and physiological states, resulting in false alarms / missed alarms.

[0013] The aforementioned shortcomings have led to prominent problems in the clinical translation of existing technical solutions, such as low compliance and limited early warning sensitivity. There is an urgent need to achieve technological breakthroughs in data integration, dynamic modeling, and personalized assessment.

[0014] Furthermore, on the one hand, there are differences in understanding among those skilled in the art; on the other hand, the applicant studied a large number of documents and patents when making this invention, but due to space limitations, not all details and contents were listed in detail. However, this does not mean that the present invention does not possess the features of these prior art. On the contrary, the present invention already possesses all the features of the prior art, and the applicant reserves the right to add relevant prior art to the background art. Summary of the Invention

[0015] In view of the shortcomings of the prior art, the present invention provides a pregnancy assistance system based on a blood glucose biosensor to solve at least some of the above-mentioned technical problems.

[0016] This invention discloses a pregnancy support system based on a blood glucose biosensor, comprising: several user terminals for acquiring data related to the user's pregnancy, including blood glucose data acquired based on the blood glucose biosensor and entered basic information; and a server communicatively connected to the user terminals. The server constructs a shared database that can be divided into several subgroups based on preset grouping rules, so that each user can be assigned to a subgroup on the server based on their basic information. The processed blood glucose data acquired by the user during their current pregnancy can be compared with the processed blood glucose data acquired by other users in the corresponding subgroup during the same pregnancy to analyze the patient's current blood glucose management status by determining a ranking. The server can determine the preset grouping rules based on the categories of basic information.

[0017] This invention provides a pregnancy support system based on a blood glucose biosensor. It acquires pregnancy-related data through a user terminal and aggregates it into a shared database on a server, providing a scientific basis for personalized blood glucose management. Specifically, the system not only monitors the user's blood glucose levels in real time but also integrates basic health information (such as age and BMI) for comprehensive analysis. This integrated approach allows the system to generate personalized management recommendations based on each pregnant woman's specific situation. For example, for pregnant women with high-risk factors, the system can provide early warnings and targeted interventions. Furthermore, by comparing statistical data sets and reference statistical charts in the shared database, the system can assess the patient's blood glucose control from a more comprehensive perspective, thereby making more accurate medical decisions.

[0018] According to a preferred embodiment, the blood glucose biosensor can periodically collect raw blood glucose data from users based on continuous blood glucose monitoring technology. The blood glucose data can be uploaded to the server in the form of raw data and / or processed data. The processed data is preprocessed data obtained by the user terminal through preprocessing the raw data and / or analytical data obtained through complete analysis. The server can store all received blood glucose data in the form of analytical data.

[0019] This approach not only ensures data continuity and accuracy but also makes subsequent data analysis more efficient. In particular, all blood glucose data received by the server is converted into a standardized analytical data storage format. This standardized process simplifies data analysis and improves data utilization. The rich data generated by continuous monitoring helps reveal individual blood glucose fluctuation patterns, providing a reliable basis for developing personalized treatment plans. For example, by conducting detailed analysis of blood glucose fluctuation factors over a period of time, the system can better understand changes in a patient's insulin sensitivity, thereby adjusting treatment strategies and reducing the risk of GDM.

[0020] According to a preferred embodiment, the user terminal is equipped with a smart device for receiving raw data of the user's blood glucose values ​​collected by a blood glucose biosensor. The smart device can directly forward the raw data and / or process the raw data into processed data and then upload it to the server. The user can enter basic information through the input module of the smart device. The basic information entered is risk factors related to gestational diabetes.

[0021] The intelligent device configured in this invention can not only efficiently receive raw data collected by a blood glucose biosensor, but also flexibly process and upload it to a server according to actual needs. More importantly, the intelligent device allows users to conveniently enter basic information such as risk factors related to gestational diabetes mellitus (GDM). This basic information is crucial for a comprehensive understanding of the patient's background; it not only helps identify high-risk individuals but also guides personalized health management strategies. For example, by combining detailed family medical history and BMI data, the system can more accurately predict a patient's risk of developing GDM and provide corresponding preventative measures.

[0022] According to a preferred embodiment, the basic information entered by the user through the input module of the smart device includes time information, physical information and / or historical information, wherein the time information includes age and date of last menstrual period, the physical information includes pre-pregnancy waist-to-hip ratio and / or pre-pregnancy body mass index, and the historical information includes family history of diabetes and / or adverse pregnancy history.

[0023] This invention allows users to easily input basic data, including time information (such as age and last menstrual period date), physical information (such as pre-pregnancy waist-to-hip ratio and BMI), and historical information (such as family history of diabetes and adverse pregnancy history), through the input module of a smart device. This lays a solid foundation for a comprehensive assessment of the risk of gestational diabetes mellitus (GDM). In particular, considering that increasing age exacerbates insulin resistance, and that a higher BMI is often accompanied by obesity and its associated metabolic problems, these factors collectively increase the likelihood of GDM. At the same time, family history and previous adverse pregnancy experiences are equally important, reflecting the impact of genetic susceptibility and past health conditions on the current pregnancy. By recording and analyzing this information in detail, the system can not only identify high-risk individuals early but also provide crucial evidence for developing effective interventions. For example, for pregnant women with a high BMI, the system can recommend specific diet and exercise plans, thereby effectively reducing the maternal and infant health risks associated with GDM.

[0024] Preferably, the basic information of the present invention may also include hormone levels, especially hormone levels that are highly associated with insulin sensitivity or insulin resistance. These hormones may be, for example, placental growth hormone (PGH), human placental prolactin (hPL), estradiol (E2), and / or insulin-like growth factor binding protein 1 (IGFBP1).

[0025] According to a preferred embodiment, the user terminal can upload the user's basic information to the server when the user uses the device for the first time, so that the server can calculate the gestational age of the subsequently uploaded data based on the last menstrual period in the basic information. In this case, the server prioritizes data with complete gestational age data when building and / or updating the shared database.

[0026] This approach ensures that all data is interpreted within the correct gestational context, significantly improving the accuracy of data analysis. Especially when building and updating shared databases, prioritizing information with complete gestational data not only helps capture the full picture of blood glucose fluctuations throughout pregnancy but also improves the data integrity of the shared database. In the long term, this data management method helps accumulate rich clinical experience, promoting the development and improvement of knowledge systems in related fields. For example, by comparing blood glucose fluctuation patterns at different gestational stages, researchers can explore more deeply the key factors influencing the development and progression of GDM, opening new avenues for developing novel prevention methods.

[0027] According to a preferred embodiment, the server can group all data in the shared database based on a preset grouping rule. The preset grouping rule includes one or more combinations of several factors contained in the basic information. The type and / or number of factors in the grouping rule are determined based on the amount of data included in the shared database, and the grouping method can be adjusted based on a dynamic adjustment mechanism.

[0028] This invention introduces a shared database grouping mechanism based on preset population segmentation rules, considering a combination of factors, including but not limited to age, pre-pregnancy waist-to-hip ratio, BMI, family history of diabetes, and adverse pregnancy history. This grouping method fully considers the physiological differences between individuals and their impact on GDM risk, thus achieving more refined data classification. As the shared database grows, the system can dynamically adjust the grouping method according to actual conditions to maintain sufficient and representative sample sizes within each subgroup. This design not only improves the effectiveness of data analysis but also promotes the development of customized medical services for specific populations. For example, by comparing the glycemic management effects among different subgroups, researchers can explore the key factors influencing the occurrence and development of GDM more deeply, providing theoretical support for the development of new prevention methods.

[0029] When determining the preset group division rules, this invention can also be based on one hormone or a combination of multiple hormones. Preferably, the group division rules can be the similarities and differences in the influence trends of multiple hormones on insulin resistance or insulin sensitivity. Based on these group division rules, anisotropic and / or homotropic subgroups can be obtained. This technical solution achieves a three-dimensional synergistic effect in terms of hormone action mechanism and mathematical modeling by establishing dual classification rules for anisotropic and homotropic subgroups: First, the anisotropic subgroup adopts an antagonistic effect ratio model, the core of which is to transform the antagonistic effect between hormones into a nonlinear weight ratio. This design breaks through the limitation of traditional ratios that only reflect concentration ratios. By introducing the regression coefficient ratio (β1 / |β2|), the difference in biopotency is quantified into mathematical weights, enabling the model to capture the "threshold breakthrough effect" that occurs when the potency of pro-resistance hormones exceeds that of protective hormones—that is, when the ratio exceeds a critical value, a small change in concentration will trigger a step-like deterioration in insulin sensitivity. This nonlinear response characteristic allows the early intervention window to be 2-3 weeks earlier than the traditional linear model; Second, the product index of the homotropic subgroup is constructed by... The synergistic amplification factor mathematically simulates the positive feedback loop of hormone signaling pathways. When the product of the concentrations of two pro-resistance hormones reaches a critical value, their combined effect activates the oxidative stress pathway in placental tissue. This model achieves bioequivalence conversion of hormones of different dimensions through the standardized constant (1000) in the denominator, allowing the product exponent to be directly mapped to the molecular pathway activation threshold of mitochondrial dysfunction. Finally, the parallel application of the two subgroup division rules forms a complementary verification mechanism. When an individual exhibits antagonistic imbalance (R>2.5) in the heterotropic subgroup but does not reach the synergistic pathogenic threshold (PI<35) in the homotropic subgroup, the system can identify regional differences in hormone secretion caused by local ischemia in the placenta. This contradictory signal suggests that clinical assessment of placental blood perfusion is necessary rather than immediate initiation of drug treatment, thereby avoiding over-medical intervention. The root cause of the above-mentioned technical effects lies in transforming the biological action patterns of hormones (synergistic / antagonistic) into computational models (product / ratio) with different mathematical properties, and quantifying their efficacy differences through regression coefficients. This modeling approach enables metabolic load assessment to leap from judging a single hormone level to analyzing the dynamic balance of hormone networks, solving the technical defect of high false negative rates caused by neglecting the types of interactions between hormones in traditional methods.

[0030] According to a preferred embodiment, the data information stored in each subgroup of the server's shared database consists of analysis data from several users at different time points or periods that meet the same group division rules. The analysis data is a multi-dimensional comprehensive blood glucose evaluation index calculated based on blood glucose-related average blood glucose index, blood glucose fluctuation factor, postprandial peak gradient, and nighttime blood glucose load.

[0031] This multi-layered data structure not only encompasses the time-series characteristics of blood glucose levels but also reflects the quality and stability of blood glucose control. Through in-depth analysis of this complex data, the system can gain comprehensive insights into a patient's blood glucose management patterns, thereby enabling more accurate analysis of the user's blood glucose management. For example, by analyzing nighttime blood glucose load, doctors can identify the risk of nocturnal hypoglycemic events, adjust treatment plans in a timely manner, and avoid potential complications.

[0032] According to a preferred embodiment, the server can perform statistical analysis on all the analysis data of each subgroup stored in the shared database to obtain the statistical data set of each subgroup, and generate the corresponding multi-dimensional blood glucose comprehensive evaluation index change curve over time for each user in each subgroup, thereby forming a corresponding reference statistical chart for each subgroup. The server can perform spatiotemporal clustering analysis on the statistical data set and / or reference statistical chart of each subgroup to obtain reference interval boundary values ​​and / or continuous reference curves.

[0033] This invention not only provides individual patients with blood glucose control standards for populations with similar backgrounds as a reference, but also provides a powerful tool for large-scale epidemiological studies.

[0034] According to a preferred embodiment, when a user terminal uses the device for the first time, it uploads the user's basic information to the server, so that the server can determine the corresponding subgroup based on the user's basic information and a preset group division rule. The server can generate a hash code based on the user's basic information and compare it with the pre-stored subgroup feature code using Hamming distance to determine the target subgroup.

[0035] This automated workflow reduces the possibility of human error, ensuring that each user is correctly categorized into the most suitable subgroup. Once the target subgroup is identified, users can enjoy services specifically designed for them, making personalized health advice and customized treatment plans more feasible. For example, for pregnant women with high-risk factors, the system can provide more stringent blood glucose monitoring and management recommendations, effectively reducing the risk of GDM, and can learn from the effective blood glucose management experiences of other users in the corresponding subgroup, which is more tailored than generalized experiences.

[0036] According to a preferred embodiment, the user terminal and / or server can analyze the patient's current blood glucose management status by calculating the percentile ranking of the user's current multidimensional blood glucose comprehensive evaluation indicators. The user terminal and / or server can determine the warning level directly or indirectly with the help of other auxiliary functions based on the percentile ranking of one or more multidimensional blood glucose comprehensive evaluation indicators.

[0037] The early warning mechanism of this invention enables the system to quickly identify patients in a high-risk state, allowing for timely intervention. Percentile ranking not only reflects an individual's glycemic control level relative to the general population but also incorporates auxiliary functions to further refine the early warning criteria, improving the sensitivity and specificity of the early warning system. Attached Figure Description

[0038] Figure 1 This is a hardware connection diagram of the pregnancy assistance system provided by the present invention;

[0039] Figure 2 This is a schematic diagram of data transmission in the pregnancy assistance system provided by the present invention;

[0040] Figure 3 This is a schematic diagram illustrating how the server, as provided in this invention, groups a shared database.

[0041] Figure 4 This is a schematic diagram illustrating how the server provided by the present invention analyzes and statistically processes the data in each subgroup of a shared database.

[0042] Figure 5 This is an example of a multi-dimensional comprehensive blood glucose evaluation index changing over time (i.e., during pregnancy), provided by the present invention.

[0043] Figure 6 This is a schematic diagram of the hardware function allocation for analyzing a user's blood glucose management, provided by the present invention.

[0044] Figure 7 This is an enlarged analysis of the curve of the multidimensional blood glucose comprehensive evaluation index changing over time (i.e., during pregnancy), as provided by the present invention.

[0045] List of reference numerals

[0046] 100: User terminal; 110: Blood glucose biosensor; 120: Smart device; 121: Input module; 200: Server; 400: Shared database. Detailed Implementation

[0047] The following is a detailed explanation with reference to the accompanying drawings.

[0048] like Figure 1As shown, this invention discloses a pregnancy support system based on a blood glucose biosensor 110, comprising: several user terminals 100 for acquiring data related to a user's pregnancy, and a server 200 communicatively connected to the user terminals 100. The server 200 can (with user authorization) collect data from each user through the user terminals 100, and aggregate this data to form a shared database 400. This allows any user terminal 100 to directly or indirectly obtain the blood glucose management status of its corresponding user when acquiring relevant data. In this invention, the user using the user terminal 100 typically refers to a pregnant woman. Further, as... Figure 2 As shown, data related to the user's pregnancy may include blood glucose data obtained using continuous glucose monitoring technology. The blood glucose data can be uploaded to the server 200 as raw data and / or processed data. The processed data may be preprocessed data obtained by simple preprocessing of the raw data and / or analytical data obtained by complex analysis.

[0049] Preferably, such as Figure 2 As shown, the user terminal 100 of the present invention may include a blood glucose biosensor 110 and a smart device 120 connected by communication, so that data information related to the user's pregnancy acquired by the blood glucose biosensor 110 can be sent to the smart device 120 for forwarding and / or data processing. Preferably, the blood glucose biosensor 110 may employ continuous blood glucose monitoring technology, for example, it may be configured as a continuous blood glucose monitor, which uses electrochemical sensing technology or optical detection technology to measure glucose concentration. The electrochemical sensor preferably uses a platinum-silver / silver chloride electrode system, with a sensing probe diameter ranging from 0.3 to 0.6 mm and a length controlled between 4 and 6 mm. In specific implementations, a disposable patch probe or a recalibrable probe may be selected. More preferably, the continuous blood glucose monitor of the present invention may use a disposable electrochemical probe with a diameter of 0.4 mm and a length of 5 mm, performing dynamic monitoring on a 14-day cycle, automatically recording blood glucose values ​​every 5 minutes and encrypting and packaging them, thereby achieving dynamic encrypted transmission.

[0050] Preferably, the smart device 120 can be configured as a smart mobile device to receive encrypted blood glucose data sent by the blood glucose biosensor 110. The smart mobile device can be a smartphone, a dedicated medical tablet, or a wearable device. The built-in processor may include a dual-core architecture and have an independent Secure Element in the local storage medium for storing biometric keys. Further, the built-in processor may integrate signal amplification circuitry and a noise filtering unit, with an analog-to-digital conversion accuracy selectable from 12-bit to 16-bit resolution. Preferably, a 14-bit high-precision ADC chip (such as the TIADS131M04) is used. More preferably, such as... Figure 2 As shown, users can input basic information through the input module 121 of their smart mobile devices. This basic information mainly includes risk factors related to gestational diabetes mellitus (GDM), such as time information, physical information, and / or historical information. Time information can include age and date of last menstrual period. The risk of GDM increases with maternal age, with a significant increase in risk for women over 30 years of age, especially those aged 35 and above, who are considered a high-risk group. Physical information can include pre-pregnancy waist-to-hip ratio and / or pre-pregnancy body mass index (BMI). Waist-to-hip ratio reflects the distribution of body fat; a higher ratio usually indicates more abdominal fat accumulation, which leads to increased insulin resistance and thus increases the risk of GDM. A higher BMI is usually associated with obesity, which causes insulin resistance and increases the likelihood of GDM. Furthermore, existing research has shown that multivariate logistic regression analysis indicates BMI is an independent risk factor for GDM. Furthermore, the physical information entered should ideally be measured before planning pregnancy to obtain the most accurate baseline data to assess the pregnant woman's pre-pregnancy health and potential risk factors. However, many women may not realize they are pregnant until weeks or even months later; in such cases, early pregnancy data can be used as an alternative. Historical information entered may include a family history of diabetes and / or a history of adverse pregnancy outcomes. Pregnant women with a family history of diabetes and / or an adverse pregnancy outcome have an increased risk of developing gestational diabetes. A family history of diabetes can primarily consider the woman's own medical history and that of first-degree relatives (or immediate family members other than her spouse), while an adverse pregnancy outcome can primarily consider a history of delivering macrosomic infants.

[0051] Preferably, the smart device 120 of the user terminal 100 can upload the user's blood glucose data and basic information to the server 200. The secure communication network architecture may include multi-level transmission channels. The main communication link uses wireless transmission conforming to the IEEE 802.11ax standard, and the backup link can use cellular network (4G / 5G) or Bluetooth Low Energy (BLE 5.2) connection. The network middleware deploys a medical-grade gateway device, which has a protocol conversion interface supporting the conversion of various medical data standards such as HL7 FHIR and DICOM. The data buffer storage capacity is configured as a 4GB to 16GB DDR4 module, preferably an 8GB dual-channel configuration.

[0052] Preferably, such as Figure 1 and Figure 3As shown, when server 200 receives data information uploaded by multiple user terminals 100 that is associated with the corresponding users, it can construct a shared database 400 based on this data information. When constructing the shared database 400, data information with complete pregnancy data is prioritized, that is, data information basically including the three stages of early pregnancy (before the end of the 13th week of pregnancy), mid-pregnancy (the 14th to the end of the 27th week of pregnancy), and late pregnancy (the 28th week of pregnancy and beyond). After delivery, pregnant women typically no longer use the assistance system of this invention. The data information with complete pregnancy data uploaded by the user of the assistance system during pregnancy can be used to construct or update the shared database 400. The user loses their user identity after delivery unless they become pregnant again to obtain a new user identity. Preferably, the user terminal 100 can upload the user's basic information to the server 200 upon the user's first use, so that the server 200 can calculate the gestational age of subsequently uploaded data based on the last menstrual period date in the basic information. "First-time use" refers to a confirmed pregnant woman entering her basic information into the user terminal 100 for the first time during the current gestational period, and the user terminal 100 will not re-upload the basic information during the current gestational period unless the user modifies the basic information. When constructing the shared database 400, the server 200 can also incorporate existing clinical or research data to enrich the data volume in the shared database 400. After the server 200 has constructed the shared database 400, data uploaded by the user terminal 100 can be used to update the shared database 400, and data with complete gestational age data is preferred for updating the shared database 400. Furthermore, if the server 200 discovers that the data uploaded by any user terminal 100 is missing data information for at least one stage, it can exclude it when building and updating the shared database 400, but can still provide corresponding services to the user using the user terminal 100. Further still, if the missing stage in the data uploaded by the user terminal 100 is a stage that the user has not yet reached in the current pregnancy cycle (for example, for a pregnant woman currently in the second trimester, the third trimester is a stage that has not yet been reached in the current pregnancy cycle, and therefore data information for that stage is missing), then the user's corresponding data information is temporarily stored until the missing stage's data information is complete, after which it can be used to build or update the shared database 400; otherwise, it is excluded. The exclusion scenarios mainly include: if, based on time calculations, the user has reached or passed the stage missing from their data information, but after a preset waiting period, the uploaded data information has not been received, then the user's corresponding data information can be excluded when building or updating the shared database 400, and simultaneously deleted from the temporary storage space.

[0053] Preferably, the server 200 can update the shared database 400 periodically or irregularly. Specifically, the server 200 can update periodically at preset time intervals, or irregularly based on preset data volume or other factors. Figure 3 As shown, all data in the shared database 400 can be divided into several subgroups based on preset population segmentation rules. These preset rules may include one or more combinations of factors such as age, pre-pregnancy waist-to-hip ratio, pre-pregnancy body mass index (BMI), family history of diabetes, and adverse pregnancy history. The type and / or number of factors in the population segmentation rules can be determined based on the amount of data encompassed by the shared database 400 to ensure that the amount of data in each subgroup after segmentation meets statistical power requirements. Furthermore, when the shared database 400 is updated, newly added data information can be added to the corresponding subgroups according to the preset population segmentation rules to achieve data updates for the respective subgroups.

[0054] Preferably, when constructing group segmentation rules, the grouping criteria can be standardized by combining clinical guidelines and epidemiological research evidence. For example, for age grouping, the pregnancy risk stratification criteria recommended by the International Federation of Gynecology and Obstetrics (FIGO) can be used to divide pregnant women into four levels: 18–24 years old (stable reproductive function), 25–34 years old (optimal reproductive age), 35–39 years old (advanced maternal age), and ≥40 years old (very advanced maternal age). The age threshold of 35 years old for high-risk pregnancy has clear evidence-based medical support. For pre-pregnancy waist-to-hip ratio (WHR) grouping, for Asian populations, 0.80 is used as the upper limit of normal, with gradients of <0.75 (low waist-to-hip ratio), 0.75–0.79 (normal range), 0.80–0.84 (pre-central obesity), and ≥0.85 (pathological obesity). Other ethnic groups can adjust based on the standards of their respective ethnic groups. For pre-pregnancy body mass index (BMI) grouping, a modified Asian standard can be used: <18.5 kg / m². 2 (Low body weight), 18.5~22.9kg / m 2 (Standard group), 23.0~24.9kg / m 2 (Early stage of overweight), 25.0~29.9kg / m 2 (Grade I obesity), ≥30.0 ​​kg / m² 2 (Grade II obesity), with 23 kg / m² 2The cutoff criteria are stricter than the WHO criteria, making it more suitable for monitoring metabolic abnormalities during pregnancy. For grouping by family history of diabetes, a three-tiered kinship network model can be established, dividing individuals into those with no family history, those with a positive history in first-degree relatives (parents / siblings), and those with a positive history in second-degree relatives (grandparents / uncles). An epigenetic weighting coefficient is introduced, and the presence of ≥2 second-degree relatives with the disease is considered equivalent to the risk in first-degree relatives. For grouping by adverse pregnancy history, a cumulative risk scoring method can be used: no abnormal history (0 points), single abnormality (1 point: including ≥2 spontaneous abortions, gestational hypertension, or macrosomia delivery, etc.), multiple abnormalities (2 points: combined with fetal malformations / stillbirth, etc.), and recurrent pregnancy loss (≥3 points: three or more adverse outcomes). Furthermore, a dynamic adjustment mechanism can be established during population segmentation. When the sample size of a specific subgroup is lower than the statistical power requirement (e.g., n<30), clinical equivalence merging of adjacent levels is implemented. For example, merging 35–39 years old with ≥40 years old into an "advanced maternal age group," while simultaneously activating a data collection early warning system to selectively supplement data from weak subgroups. Furthermore, all division rules were verified for intergroup distribution differences using the Kolmogorov-Smirnov test to ensure that each subgroup has independent clinical intervention guidance value.

[0055] According to a preferred embodiment, the basic information uploaded by the user terminal 100 may also be a physical examination report (e.g., a prenatal checkup). The report may include some of the patient's hormone levels. Hormone levels highly correlated with insulin sensitivity or insulin resistance may be uploaded to the server 200, allowing the server 200 to use these hormone categories to pre-define population segmentation rules. For example, the uploaded hormone categories may include placental growth hormone (PGH), human placental prolactin (hPL), estradiol (E2), and / or insulin-like growth factor binding protein 1 (IGFBP1), etc.

[0056] Preferably, when determining the preset group division rules based on hormone categories, the server 200 can do so using one hormone or a combination of multiple hormones. Specifically, for relatively independent hormone categories (such as placental growth hormone), subgroups can be divided based on a single hormone level; for hormone categories with significant inter-correlation (such as human placental prolactin and estradiol, and human placental prolactin and insulin-like growth factor binding protein 1), subgroups can be divided based on a combination of multiple hormones. Furthermore, when dividing subgroups using a combination of multiple hormones (i.e., a complex hormone combination), the server 200 also needs to consider the correlation orientation between each hormone and insulin resistance, where correlation orientation refers to whether the association between the hormone and insulin resistance is positive or negative. If all hormones exhibit the same association direction, it is considered a homotropic subgroup. If at least one hormone exhibits a different association direction, it is considered a heterotropic subgroup. Compared to homotropic subgroups, heterotropic subgroups are more worthy of study. Among them, the subgroup formed by the combination of human placental prolactin and estradiol is a homotropic subgroup, while the subgroup formed by the combination of human placental prolactin and insulin-like growth factor binding protein 1 is a heterotropic subgroup.

[0057] Preferably, the grouping rules for a single hormone can be determined based on the dynamic threshold model of the corresponding hormone. For example, for placental growth hormone, a linear regression model can be used to establish the dynamic relationship between gestational age (GW) and hormone levels, i.e.:

[0058] Critical threshold = α × GW + β ± kσ

[0059] In the formula, α and β are the fitting coefficients of large-scale cohort studies; k is the standard deviation factor, set according to the clinical risk level; and σ is the standard deviation of placental growth hormone (PGH) levels.

[0060] For example, based on empirical values, when GW=14, α=0.183, β=2.41, and k=1.96, the cutoff value for early / mid pregnancy can be calculated as: the critical threshold value for GW=14 = 0.183×14 + 2.41 + 1.96×0.78 = 4.28 ng / mL. Therefore, the rule for classifying placental growth hormone into single hormone groups can be, for example: early pregnancy (<14 weeks) ≤ 4.2 ng / mL (normal proliferative phase), mid pregnancy (14–28 weeks) 4.3–7.8 ng / mL (metabolic adaptation phase), and late pregnancy (≥29 weeks) ≥ 7.9 ng / mL (insulin resistance triggering phase).

[0061] Preferably, for the co-correlated subgroups in the compound hormone combination, the division rule can be determined according to the product index (PI). Taking the co-correlated subgroup of hPL+E2 as an example, its product index (PI) can be calculated by the following formula:

[0062]

[0063] Here, w1 and w2 are the weighting coefficients of hPL and E2, respectively, which can be assigned weights using principal component analysis and / or the Delphi method. Empirically, w1 and w2 can be set to values ​​of 0.68 and 0.71, respectively.

[0064] Furthermore, the threshold division based on the product exponent (PI) is as follows:

[0065] Low metabolic load group: PI < μ-1.5σ;

[0066] Compensatory equilibrium group: μ-1.5σ≤PI<μ+1.5σ;

[0067] Synergistic pathogenicity group: PI ≥ μ + 1.5σ,

[0068] In the formula, μ and σ are the mean (μ = 26.5) and standard deviation (σ = 6.3) of the PI of the healthy control group.

[0069] For example, for the co-associated subgroup (hPL+E2) in the compound hormone combination, the following group division rules can be adopted: product index <17 (low metabolic load group); 17≤product index<36 (compensatory balance group); product index≥36 (co-pathogenic variable group);

[0070] Preferably, for the anisotropic subgroup in the compound hormone combination, the division rule can be determined based on the antagonistic effect ratio (R). Taking the anisotropic subgroup of hPL+IGFBP1 as an example, its antagonistic effect ratio (R) can be calculated by the following formula:

[0071]

[0072] In the formula, β1 and β2 are the regression coefficients of hPL and IGFBP1 on HOMA-IR, respectively, and can be weighted using principal component analysis and / or the Delphi method. Empirically, β1 and β2 can be taken as 0.78 and -0.53, respectively.

[0073] Furthermore, the threshold division based on the antagonistic effect ratio (R) is as follows:

[0074]

[0075] In the formula, R0 is the baseline equilibrium value of the antagonistic effect ratio; IQR(R) is the interquartile range of the healthy population; MAD(R) is the median absolute deviation; and k corresponds to the 95% confidence interval.

[0076] For example, for the anisotropic subgroup (hPL+IGFBP1) in the compound hormone combination, R0 can be 1.123, IQR(R) can be 1.2, and MAD(R) can be 0.85. Therefore, C can be calculated. lower = -1.229 (since it is a negative value, the physiological lower limit of 0.8 can be taken), C upper =2.79≈2.8, and the following group division rules can be adopted: when the antagonistic effect ratio is ≥2.8, the hPL insulin resistance effect is dominant; when 0.8≤antagonistic effect ratio<2.8, the dual hormone antagonistic equilibrium is reached; when the antagonistic effect ratio is <0.8, the IGFBP1 protective effect is dominant.

[0077] Preferably, such as Figure 3 As shown, the shared database 400 stores analysis data from several users at different time points or periods that meet the same group division rules in each subgroup. This analysis data can be obtained by user terminal 100 and / or server 200 after processing the raw data. The raw data consists of blood glucose values ​​automatically recorded by blood glucose biosensor 110 at preset intervals within a preset sampling period. Furthermore, by comprehensively analyzing the blood glucose values ​​recorded over a period of time (e.g., 24 hours), analysis data that can be stored in the shared database 400 can be obtained. The hardware for processing raw data can be selected based on the computing power configuration of the user terminal 100. For example, for user terminals 100 with insufficient computing power, raw data can be directly uploaded to the server 200 for processing; for user terminals 100 with average computing power, raw data can undergo simple preprocessing on the user terminal 100 to form preprocessed data, so that the server 200 can perform subsequent processing on the received preprocessed data; for user terminals 100 with sufficient computing power, raw data can undergo complex analysis processing on the user terminal 100 to obtain analyzed data, so that the server 200 can directly receive the final analyzed data. The hardware for processing raw data can also be selected and adjusted based on various factors such as user needs and / or the load of the server 200.

[0078] Preferably, such as Figure 3 As shown, the data analysis can be a comprehensive evaluation index calculated based on one or more parameters related to blood glucose. These parameters may include the average glycemic index, glycemic variability factor, postprandial peak gradient, and nighttime glycemic load.

[0079] Preferably, the comprehensive evaluation index can be configured as a multidimensional comprehensive glycemic index (CGCI), whose mathematical expression is as follows:

[0080] CGCI=α·AGI+β·GVF+γ·PPG+δ·NGL,

[0081] In the formula, AGI is the average glycemic index; GVF is the glycemic variability factor; PPG is the postprandial peak gradient; NGL is the nocturnal glycemic load; and α, β, γ, and δ are the weighting factors of each item.

[0082] Furthermore, the average glycemic index (AGI) is calculated using a sigmoid function (logistic function) to represent the average blood glucose level (G) over a period of time. avg Mapping to the (0,1) interval reflects the degree of deviation of the overall blood glucose level from the ideal range. The specific calculation formula is as follows:

[0083]

[0084] In the formula, G avg λ is the mean blood glucose level (mmol / L); λ is the sensitivity adjustment factor, which can be set to 0.3 to adjust the steepness of the S-curve and control the sensitivity to changes in blood glucose; θ is the clinical anchor value, which can be referenced from the ADA guidelines to set the ideal fasting blood glucose level at 5.5.

[0085] Furthermore, the glucose variability factor (GVF) is measured by the coefficient of variation (the ratio of the standard deviation of blood glucose σ to the mean blood glucose G). avg The ratio of τ to τ quantifies the amplitude of blood glucose fluctuations, and logarithmic transformation is performed on high-fluctuation data (>τ) for correction. The specific calculation formula is as follows:

[0086]

[0087] In the formula, σ is the standard deviation of blood glucose (mmol / L); G avg τ represents the mean blood glucose level (mmol / L); τ is the cardiovascular risk inflection point threshold, which can be taken as 36% based on the DECODE study.

[0088] Furthermore, the postprandial peak gradient (PPG) is used to measure the rate of postprandial blood glucose rise and metabolic recovery capacity. It can be calculated by the ratio of the "peak-baseline difference" to the "meal interval," with the specific calculation formula as follows:

[0089]

[0090] In the formula, G peak,t The peak blood glucose level 2 hours after a meal (mmol / L); G pre,t Baseline value before meal (mmol / L); ΔT t The interval between two meals is 1 hour; n is the standard number of meals, which is usually 3.

[0091] Furthermore, the nocturnal glycemic load (NGL) comprehensively assesses the level of glycemic exposure (AUC) and the risk of hypoglycemia (penalty) during the night (e.g., 00:00–06:00), where the specific calculation formula is as follows:

[0092]

[0093] In the formula, the AUC term is the area under the curve, reflecting the cumulative damage to tissues caused by chronic hyperglycemia; the penalty term is triggered when the lowest nocturnal blood glucose Gmin is less than η (which can be taken as 3.9 mmol / L), simulating the acute harm of hypoglycemia; κ is the penalty coefficient, which makes a single hypoglycemic event equivalent to an increase in AUC of 1 mmol / L, and can be taken as 5.

[0094] Preferably, the weighting coefficients α, β, γ, and δ can be set in the order α > β > γ > δ. This is because average blood glucose (index) is a strong predictor of glycated hemoglobin (HbA1c) and is directly related to microvascular complications, while the others, as independent risk factors, contribute progressively less to increasing risk. Furthermore, in special circumstances, such as nocturnal metabolic disorders, specific weighting coefficients can be assigned individually as a warning.

[0095] Preferably, such as Figure 4 As shown, server 200 can perform statistical analysis on all the data from each subgroup stored in shared database 400 to obtain statistical data sets for each subgroup, and can generate corresponding multi-dimensional blood glucose comprehensive evaluation index change curves over time (i.e., during pregnancy) for each user in each subgroup (e.g., ...). Figure 5 As shown in the figure, a corresponding reference statistical chart is formed for each subgroup. Further, the server 200 can perform spatiotemporal clustering analysis on the statistical data set and / or reference statistical chart for each subgroup. For each subgroup, a sliding window segmentation is performed according to gestational week, with a window width of 4 weeks and a sliding step size of 1 week, ensuring that adjacent windows overlap by 3 weeks to smooth physiological fluctuations. Within each gestational week window, the probability distribution function of the CGCI value is calculated using the kernel density estimation method, and several percentiles (e.g., the 5th, 10th, 25th, 50th, 75th, 90th, and 95th percentiles) are extracted as reference interval boundary values. For non-integer gestational week data points, a gestational week-percentile mapping function is constructed based on a cubic spline interpolation algorithm, such that any gestational week t∈[1,40] can be mapped by the function f(t)=a0+a1t+a2t. 2 +a3t 3 +Σβ i (tk i ) + 3 Generate a continuous reference curve, where k iFor nodes with known gestational ages, coefficients are determined through least squares fitting. Reference interval data is stored in a distributed database in a binary tree structure, with each node containing the gestational age start point, subgroup code, and percentile value matrix, enabling fast retrieval with O(log n) time complexity.

[0096] Preferably, for users who are pregnant, the user terminal 100 can upload the user's basic information to the server 200 upon the user's first use, so that the server 200 can determine the corresponding subgroup based on the user's basic information and preset grouping rules. Further, the server 200 can generate a 128-bit hash code based on the user's basic information and compare it with the pre-stored subgroup feature code using Hamming distance to determine the target subgroup. Since the preset grouping rules are generally based on relatively stable basic information, they usually do not change the basic information except for errors in data entry. Therefore, once a user is assigned a subgroup, they are generally not regrouped unless the preset grouping rules are adjusted based on a dynamic adjustment mechanism.

[0097] Preferably, such as Figure 6 As shown, for a user who is pregnant, the user terminal 100 can upload the user's blood glucose data to the server 200, and can also download the statistical data set and / or reference statistical chart of the corresponding subgroup from the server 200, so that the user terminal 100 and / or the server 200 can obtain the user's blood glucose management status through analysis. The hardware for analyzing the user's blood glucose management status can be selected and adjusted based on various factors such as the computing power configuration of the user terminal 100, user needs and / or the load of the server 200.

[0098] Preferably, when analyzing the blood glucose management of a user during pregnancy, the user terminal 100 and / or server 200 can call the gestational age-percentile mapping function f(t) to calculate the percentile values ​​for each gestational age (e.g., {p5, p...)). 10 ,…,p 95 The algorithm uses linear interpolation to accurately calculate the percentile rank of the user's current CGCI value (i.e., c0). For example, if c0 ≤ p5, then the percentile rank PR = 5%; if p5 ≤ p5, then the percentile rank PR = 5%. <c0≤p 10 Then, using the linear interpolation formula PR=5+5*(CGCI-p5) / (p 10- p5), and so on until c0>p 95 PR = 100%. Preferably, user terminal 100 and / or server 200 may be configured with several percentile thresholds to issue an alert when the percentile ranking of a user's CGCI value exceeds or exceeds the percentile threshold multiple times consecutively. Different percentile thresholds may correspond to different alert levels. Further, such as... Figure 6As shown, user terminal 100 and / or server 200 may also be configured with auxiliary functions to further improve the accuracy of the warning.

[0099] Preferably, the user terminal 100 and / or server 200 can generate a fitting curve based on the user's historical data sequence using the Loess method (locally weighted scatter smoothing). The weight function is set to w i =(1-|tt) i | 3 ) 3 The window spans 8 weeks to capture medium- to long-term trends in glycemic control, where the user's historical data sequence is the user's historical CGCI sequence {c 1, c2,…,c0} and their corresponding gestational weeks {t1,t2,…,t0}. Preferably, user terminal 100 and / or server 200 can calculate the current CGCI value (i.e., c0) relative to the subgroup median (i.e., p). 50 The standard deviation multiple Z = (c) 0- p 50 ) / σ, where σ is the standard deviation of the current gestational age window for the subgroup.

[0100] For example, the auxiliary system of the present invention can be set with two levels of dynamic early warning thresholds. The first level of early warning is triggered when the user's CGCI value percentile ranking exceeds the 90th percentile of the target subgroup for three consecutive times, or when the Z value is >2.0 for two consecutive weeks. The second level of early warning is activated when any of the following conditions are met:

[0101] a) The percentile ranking of CGCI value exceeds the 95th percentile of the target subgroup, accompanied by NGL penalty.

[0102] b) Slope of the Loess curve And the current PR is ≥85%;

[0103] c) In a single measurement, the logarithmic transformation value of GVF, log(GVF), exceeds the subgroup mean by 3σ.

[0104] In one example, the curve showing the change of a user's multidimensional blood glucose comprehensive evaluation index over time (i.e., during pregnancy) is as follows: Figure 5 As shown, when analyzing blood glucose management using user terminal 100 and / or server 200, it was found that in week 34 (i.e. Figure 5 An anomaly exists at point A in the diagram. Figure 5 Further magnified analysis of point A in the middle can yield the following results. Figure 7 , Figure 7The graph shows the changes in the user's multidimensional comprehensive blood glucose assessment indicators over time at week 34. A comparison with the CGCI value at the 90th percentile of the (first) target subgroup clearly shows that the user's CGCI values ​​for three consecutive weeks (days 3-5 of week 34) have exceeded the 90th percentile of the (first) target subgroup, thus meeting the triggering conditions for a Level 1 warning. Upon triggering the warning, server 200 can filter out individuals with similar backgrounds from the current target subgroup and send the intervention plan selected by this group as a suggestion to user terminal 100. This allows the user to restore their blood glucose to normal levels after implementing effective intervention measures based on the suggestion (e.g., ...). Figure 5 (The curve changes after point A shown).

[0105] Figure 7 The diagram illustrates the 90th percentile curves for two different target subgroups formed based on two different population segmentation rules. The first target subgroup could be, for example, the anisotropic subgroup (hPL+IGFBP1) within a combination hormone regimen, and the second target subgroup could be, for example, the homotropic subgroup (hPL+E2) within the same combination hormone regimen. Because the two target subgroups use different population segmentation rules, even target subgroups with some shared factors (such as hPL) may have significantly different 90th percentile curves. This results in different comparison standards for the analytical data when users are included in different target subgroups, leading to different analytical results. For example, Figure 7 The curve showing the change of the user's multidimensional comprehensive blood glucose assessment index over time at week 34 triggers a Level 1 alert when compared with the CGCI value of the 90th percentile of the first target subgroup, but does not trigger an alert when compared with the CGCI value of the 90th percentile of the second target subgroup. Therefore, when analyzing the user's blood glucose management, user terminal 100 and / or server 200 can retrieve data from one or more target subgroups matched with the user, thereby achieving multidimensional analysis and comparison to avoid missing alarm opportunities.

[0106] Preferably, when the user terminal 100's smart device 120 has a display, the warning information can be displayed visually on the display of the smart device 120, such as a mobile phone screen.

[0107] It should be noted that the specific embodiments described above are exemplary. Those skilled in the art can devise various solutions inspired by the disclosure of this invention, and these solutions all fall within the scope of this invention and its protection. Those skilled in the art should understand that this specification and its accompanying drawings are illustrative and do not constitute a limitation on the claims. The scope of protection of this invention is defined by the claims and their equivalents. This specification contains multiple inventive concepts; phrases such as "preferred" or "according to a preferred embodiment" indicate that the corresponding paragraph discloses an independent concept. The applicant reserves the right to file divisional applications based on each inventive concept. Throughout the text, the feature introduced by "preferred" is only an optional mode and should not be construed as mandatory. Therefore, the applicant reserves the right to abandon or delete relevant preferred features at any time.

Claims

1. A pregnancy assistance system based on a blood glucose biosensor, characterized in that, It includes: Several user terminals (100) are used to acquire data related to the user's pregnancy, including blood glucose data acquired based on a blood glucose biosensor (110) and basic information entered. A server (200) is connected to the user terminal (100) for communication, wherein, The shared database (400) related to gestational blood glucose management constructed by the server (200) can be divided into several subgroups based on preset grouping rules, so that each user can be assigned to a subgroup in the server (200) based on their basic information. The processed blood glucose data obtained by the user assigned to the corresponding subgroup during the current pregnancy can be compared with the processed blood glucose data obtained by other users in the corresponding subgroup during the same pregnancy, so as to analyze the user's current blood glucose management status by determining the ranking. The server (200) can determine the preset grouping rules based on the category of basic information. The grouping rules are the similarities and differences in the influence trends of various hormones on insulin resistance or insulin sensitivity. Based on the grouping rules, heterotropic and / or homotropic subgroups are obtained. The blood glucose biosensor (110) can periodically collect raw data of the user's blood glucose values ​​based on continuous blood glucose monitoring technology. The blood glucose data can be uploaded to the server (200) in the form of raw data and / or processed data. The processed data is the preprocessed data obtained by the user terminal (100) through preprocessing the raw data and / or the analytical data obtained through complete analysis. The server (200) can store all received blood glucose data in the form of analytical data. The data information stored in each subgroup of the shared database (400) of the server (200) is the analytical data of several users at different time nodes or periods that meet the same group division rules. The analytical data is a multi-dimensional blood glucose comprehensive evaluation index calculated based on the average blood glucose index, blood glucose fluctuation factor, postprandial peak gradient, and nighttime blood glucose load related to blood glucose.

2. The system according to claim 1, characterized in that, The user terminal (100) is equipped with a smart device (120) for receiving raw data of the user's blood glucose values ​​collected by the blood glucose biosensor (110). The smart device (120) can directly forward the raw data and / or process the raw data into processed data and upload it to the server (200). The user can enter basic information through the input module (121) of the smart device (120). The basic information entered is the risk factors related to gestational diabetes.

3. The system according to claim 2, characterized in that, The basic information entered by the user through the input module (121) of the smart device (120) includes time information, physical information and / or historical information. The time information includes age and last menstrual period, the physical information includes waist-to-hip ratio and / or body mass index before pregnancy, and the historical information includes family history of diabetes and / or adverse pregnancy history.

4. The system according to claim 3, characterized in that, The user terminal (100) can upload the user's basic information to the server (200) when the user uses it for the first time, so that the server (200) can calculate the gestation period of the data information uploaded later based on the last menstrual period in the basic information. When the server (200) builds and / or updates the shared database (400), it will give priority to data information with complete gestation data.

5. The system according to claim 4, characterized in that, The server (200) can group all the data in the shared database (400) based on a preset grouping rule. The preset grouping rule includes one or more combinations of several factors contained in the basic information. The type and / or number of factors in the grouping rule are determined according to the amount of data included in the shared database (400), and the grouping method can be adjusted based on a dynamic adjustment mechanism.

6. The system according to claim 5, characterized in that, The server (200) can perform statistics on all the analysis data in each subgroup stored in the shared database (400) to obtain the statistical data set of each subgroup, and generate the corresponding multi-dimensional blood glucose comprehensive evaluation index change curve over time for each user in each subgroup, thereby forming a corresponding reference statistical chart for each subgroup. The server (200) can perform spatiotemporal clustering analysis on the statistical data set and / or reference statistical chart of each subgroup to obtain the reference interval boundary value and / or continuous reference curve.

7. The system according to claim 1, characterized in that, When the user terminal (100) uses the user for the first time, it uploads the user's basic information to the server (200) so that the server (200) can determine the corresponding subgroup based on the user's basic information and the preset group division rules. The server (200) can generate a hash code based on the user's basic information and determine the target subgroup by comparing it with the pre-stored subgroup feature code using Hamming distance.

8. The system according to claim 6, characterized in that, The user terminal (100) and / or server (200) can analyze the user's current blood glucose management status by calculating the percentile ranking of the user's current multi-dimensional blood glucose comprehensive evaluation index. The user terminal (100) and / or server (200) can determine the warning level directly or indirectly with the help of other auxiliary functions based on the percentile ranking of one or more multi-dimensional blood glucose comprehensive evaluation indicators.

Images