Elderly diabetes death risk early warning method and system

Abstract

The embodiment of the invention discloses a senile diabetes death risk early warning method and system. The method comprises the following steps: identifying an annual change percentage sequence and an annual average change percentage sequence of the diabetes death rate of the senile of each age group in a prediction region through a Joinpoint regression model; on the basis of estimating the relative risk degree of the death rate of the elderly diabetes mellitus under the age, the period and the queue effect by using a Bayesian age-period-queue model, constructing an absolute death rate prediction sequence of each age group in combination with the death rate of the elderly diabetes mellitus based on the current year; and based on the absolute death rate prediction sequence, the annual change percentage sequence and the annual average change percentage sequence, generating a structured early warning index set of different age groups, and according to the structured early warning index set, generating a differentiated early warning instruction for the community health service center in the prediction area. According to the invention, through multi-dimensional modeling, dynamic prediction and structured early warning, accurate early warning of the senile diabetes death risk is realized.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of medical early warning, and relates to but is not limited to an old-age diabetes death risk early warning method and system. BACKGROUND

[0002] With the continuous growth of the population aged 65 and above, the trend of population aging is becoming increasingly prominent. Diabetes, as a common chronic disease, has a particularly high incidence rate in the elderly population, which has posed significant pressure on the public health system, medical resource allocation, and social economy. A number of studies have shown that diabetes-related mortality is on the rise globally. Since the elderly are more susceptible to diabetes complications, their disease burden and mortality risk are further increased, so it is urgent to strengthen the systematic management and effective intervention of old-age diabetes to improve patient outcomes and reduce social burden. SUMMARY

[0003] Based on the above problems, the embodiments of the present application provide an old-age diabetes death risk early warning method and system, which aims to realize the precise early warning of old-age diabetes death risk by fusing multi-dimensional modeling, dynamic prediction and structured early warning.

[0004] The technical scheme of the embodiments of the present application is implemented as follows: In a first aspect, embodiments of this application provide a method for early warning of mortality risk in elderly people with diabetes. The method includes: using a Joinpoint regression model to analyze the age-specific mortality rates of elderly people with diabetes in different age groups up to the current year in the prediction region, obtaining the annual percentage change sequence and the annual average percentage change sequence of mortality rates for each age group; using a Bayesian age-period-cohort model to model the number of elderly people with diabetes deaths and the corresponding number of people exposed during the year up to the current year in the prediction region, estimating the age effect for each age group, the period effect for each calendar year, and the relative risk under the cohort effect for each birth cohort; and analyzing the age effect for each calendar year, the relative risk under the cohort effect for each birth cohort, and the relative risk under the cohort effect for each age group in the prediction region up to the current year. The mortality rate of elderly people with diabetes was analyzed to obtain the baseline absolute mortality rate for each age group. Based on the baseline absolute mortality rate, the relative risk under the age effect, the relative risk under the period effect of each calendar year, and the relative risk under the cohort effect of each birth cohort, the absolute mortality rate of elderly people with diabetes in each age group in the prediction area for multiple future years was predicted, resulting in the absolute mortality rate prediction sequence for each age group. The absolute mortality rate prediction sequence, the annual percentage change sequence and the annual average percentage change sequence of elderly people with diabetes mortality rate for each age group were analyzed to obtain the structured early warning indicator set for each age group. Based on the structured early warning indicator set for each age group, differentiated early warning instructions were triggered and generated for community health service centers in the prediction area.

[0005] In some embodiments, the process of obtaining the age-specific mortality rate of elderly people with diabetes in different age groups in the prediction region up to the current year includes: substituting the obtained number of deaths of elderly people with diabetes in different age groups in each year of the prediction region and the average population of the prediction region in different age groups in each year of the prediction region into the age-specific mortality rate calculation formula to obtain the age-specific mortality rate of elderly people with diabetes in different age groups in each year; wherein, the age-specific mortality rate calculation formula is: ; in, For the year age group Age-specific mortality rates; For the year age group The number of deaths from diabetes in the elderly; For the year age group The average population during the same period; This is the population proportion base.

[0006] In some embodiments, the Joinpoint regression model is used to analyze the age-specific mortality rates of elderly diabetes in different age groups in each year of the prediction area ending in the current year, to obtain a sequence of annual percentage changes and a sequence of average annual percentage changes of the mortality rates of elderly diabetes in each age group, including: for each age group, using the Joinpoint regression model to perform piecewise linear fitting on the age-specific mortality rates of elderly diabetes in that age group in each year of the prediction area ending in the current year, to obtain a plurality of turning points corresponding to the mortality rates of elderly diabetes in that age group; based on the plurality of turning points, outputting the sequence of annual percentage changes and the sequence of average annual percentage changes of the mortality rates of elderly diabetes in that age group.

[0007] In some embodiments, the Bayesian age-period-cohort model is used to model the number of death cases of elderly diabetes in different age groups and the corresponding annual exposed population in each year of the prediction area ending in the current year, to estimate the relative risk under the age effect of each age group, the period effect of each calendar year, and the cohort effect of each birth cohort, including: using the Bayesian age-period-cohort model to analyze the number of death cases of elderly diabetes in different age groups and the corresponding annual exposed population in each year of the prediction area ending in the current year, to obtain the age effect coefficient of each age group, the period effect coefficient of each calendar year, and the cohort effect coefficient of each birth cohort in the prediction area; performing exponential transformation on the age effect coefficient of each age group, the period effect coefficient of each calendar year, and the cohort effect coefficient of each birth cohort in the prediction area, respectively, to obtain the relative risk under the age effect of each age group, the relative risk under the period effect of each calendar year, and the relative risk under the cohort effect of each birth cohort.

[0008] In some embodiments, the relative risk under the period effect of each calendar year, the relative risk under the cohort effect of each birth cohort, and the mortality rate of elderly diabetes in each age group in the prediction area in the current year are analyzed to obtain the baseline value of the absolute mortality rate of elderly diabetes in each age group, including: obtaining the mortality rate of elderly diabetes in each age group in the prediction area in the current year; substituting the mortality rate of elderly diabetes in each age group in the prediction area in the current year, the relative risk under the period effect of each calendar year, and the relative risk under the cohort effect of each birth cohort into a baseline value back-calculation formula to calculate the baseline value of the absolute mortality rate of elderly diabetes in each age group; wherein the baseline value back-calculation formula is: ; wherein, is the baseline value of the absolute mortality rate of elderly diabetes in age group ; is the mortality rate of elderly diabetes in age group in the prediction area in the current year ; for the period effect of the current year for the period effect of the current year for the period effect of the current year for the cohort effect of the birth cohort for the cohort effect of the birth cohort for the cohort effect of the birth cohort

[0009] In some embodiments, the absolute mortality baseline value of senile diabetes of each age group, the relative risk for the age effect of each age group, the relative risk for the period effect of each calendar year, and the relative risk for the cohort effect of each birth cohort are used to predict the absolute mortality of senile diabetes of each age group in a plurality of future years of the prediction region, to obtain an absolute mortality prediction sequence of each age group, including: sorting the relative risk for the period effect of each calendar year to obtain a historical relative risk sequence for the period effect, and sorting the relative risk for the cohort effect of each birth cohort to form a historical relative risk sequence for the cohort effect; using a time trend extrapolation strategy to analyze the historical relative risk sequence for the period effect and the historical relative risk sequence for the cohort effect, respectively, to obtain a period effect prediction value of the prediction region in a plurality of future years and a cohort effect prediction value of the birth cohort that will enter the elderly period in the future; and substituting the absolute mortality baseline value of senile diabetes of each age group, the relative risk for the age effect of each age group, the period effect prediction value of the prediction region in a plurality of future years, and the cohort effect prediction value of the birth cohort that will enter the elderly period in the future into a mortality prediction formula to calculate the absolute mortality prediction sequence of each age group of the prediction region in a plurality of future years.

[0010] In some embodiments, the mortality prediction formula is: ; wherein, is the absolute mortality baseline value of senile diabetes of the age group; is the absolute mortality of senile diabetes of the age group in the prediction year; is the absolute mortality of senile diabetes of the age group in the prediction year; is the relative risk for the age effect of the age group; is the relative risk for the period effect of the future year; is the relative risk for the period effect of the future year; is the relative risk for the cohort effect of the birth cohort that will enter the elderly period in the future. is the relative risk for the cohort effect of the birth cohort that will enter the elderly period in the future.

[0011] ​​​​​​In some embodiments, the absolute mortality prediction sequence of each age group, the annual change percentage sequence of the old diabetes mortality of each age group, and the annual average change percentage sequence are analyzed to obtain a structured early warning indicator set of each age group, including: for each age group, extracting a death prediction value of a future year from the absolute mortality prediction sequence of the age group, and analyzing the annual change percentage sequence and the annual average change percentage sequence of the old diabetes mortality of the age group to obtain a short-term trend slope and a long-term trend direction of the age group; comparing the death prediction value of the future year, the short-term trend slope, and the long-term trend direction with a preset death risk threshold, a trend slope threshold, and a trend direction threshold corresponding to the age group respectively to obtain a structured early warning indicator set of the age group including a risk level grade, a trend urgency grade, and a trend stability grade.

[0012] In a second aspect, the embodiments of the present application provide an old diabetes death risk early warning system, the system comprising: a first analysis module configured to analyze, by using a Joinpoint regression model, old diabetes age-specific mortality of different age groups in each year of a prediction area up to a current year to obtain an annual change percentage sequence and an annual average change percentage sequence of the old diabetes mortality of each age group; a modeling module configured to model, by using a Bayesian age-period-cohort model, old diabetes death cases of different age groups in each year of the prediction area up to the current year and corresponding annual exposed population to estimate relative risks under age effects of each age group, period effects of each calendar year, and cohort effects of each birth cohort; a second analysis module configured to analyze the relative risks under the period effects of each calendar year and the cohort effects of each birth cohort and the old diabetes mortality of each age group in the current year of the prediction area to obtain baseline values of the absolute mortality of the old diabetes of each age group; a prediction module configured to predict, based on the baseline values of the absolute mortality of the old diabetes of each age group, the relative risks under the age effects of each age group, the relative risks under the period effects of each calendar year, and the relative risks under the cohort effects of each birth cohort, the absolute mortality of the old diabetes of each age group in a plurality of future years of the prediction area to obtain an absolute mortality prediction sequence of each age group; a third analysis module configured to analyze the absolute mortality prediction sequence of each age group, the annual change percentage sequence of the old diabetes mortality of each age group, and the annual average change percentage sequence to obtain a structured early warning indicator set of each age group; and a generation module configured to trigger and generate a differentiated early warning instruction of a community health service center in the prediction area according to the structured early warning indicator set of each age group.

[0013] The technical scheme provided by the embodiments of the present application has at least the following beneficial effects: In the method provided by the embodiment of the present application, during the execution of the method, first, the change inflection points and rates of the mortality of the elderly with diabetes in different historical stages in each age group in the prediction area are identified by a Joinpoint regression model, the long-term trend is converted into a quantifiable and comparable annual change percentage sequence, so that the evolution characteristics of the mortality in different age groups are accurately captured; second, the Bayesian age-period-cohort model is used to decompose the mortality change into independent effects of three dimensions of age, period and cohort, and estimate the relative risk, which not only reveals the differential effects of aging, era background and generation experience on the disease risk, but also provides a structured risk parameter system for prediction, and based on the mortality of the elderly with diabetes in the current year, the relative risk under the estimated period and cohort effects, an absolute mortality prediction sequence of each age group is constructed; then, the absolute mortality prediction sequence is deeply fused with the change percentage sequence and the annual average change percentage sequence to form a structured early warning index set of each age group covering the continuity of the trend, variability and risk level, the integrated system takes into account the long-term regularity and short-term fluctuation, and enhances the systematicness and explanatory power of the early warning signal; finally, based on the early warning index set of each age group, a differentiated early warning instruction for the community health service center is automatically generated. In this way, through the fusion of multi-dimensional modeling, dynamic prediction and structured early warning, the fine assessment, prospective prediction and precise intervention of the mortality risk of the elderly with diabetes are realized, and a full-chain scientific basis from data analysis to decision support for the prevention and control of diabetes in the prediction area is provided.

[0014] It should be understood that the foregoing general description and the following detailed description are only exemplary and explanatory, but not limiting the technical solutions provided by the embodiments of the present application. BRIEF DESCRIPTION OF DRAWINGS

[0015] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the drawings needed in the embodiment description will be briefly introduced. Obviously, the drawings in the following description are only some embodiments of the present application, and other drawings can be obtained by those skilled in the art without creative labor on the basis of these drawings. Figure 1 A flowchart of a method for early warning of mortality risk of the elderly with diabetes provided by the embodiment of the present application; Figure 2 A composition structure diagram of a system for early warning of mortality risk of the elderly with diabetes provided by the embodiment of the present application. DETAILED DESCRIPTION

[0016] In order to make the objects, technical solutions and advantages of the embodiments of the present application clearer, the following will clearly and completely describe the technical solutions in the embodiments of the present application with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are some but not all of the embodiments of the present application. The following embodiments are used to describe the present application, but not to limit the scope of the present application. Based on the embodiments in the present application, all other embodiments obtained by a person of ordinary skill in the art without creative work fall within the scope of the present application.

[0017] In the following description, "some embodiments" are related to a subset of all possible embodiments, but it can be understood that "some embodiments" can be the same subset or different subsets of all possible embodiments, and can be combined with each other without conflict.

[0018] It should be noted that the terms "first", "second", "third" involved in the embodiments of the present application are only to distinguish similar objects, and do not represent the specific order of the objects. It can be understood that "first", "second", "third" can be interchanged with specific order or sequence as allowed, so that the embodiments of the present application described herein can be implemented in an order other than that illustrated or described herein.

[0019] Those skilled in the art can understand that, unless otherwise defined, all terms (including technical terms and scientific terms) used herein have the same meaning as generally understood by those skilled in the art to which the embodiments of the present application belong. It should also be understood that terms such as those defined in a general dictionary should be understood as having a meaning consistent with the meaning in the context of the prior art, and should not be interpreted in an idealized or overly formal sense unless specifically defined as such herein.

[0020] Embodiment one: Referring to Figure 1 Fig. 1 is a flowchart of a method for early warning of the risk of death from diabetes in the elderly according to an embodiment of the present application, and the following will be described in conjunction with Figure 1 Step 101: Using the Joinpoint regression model, the age-specific mortality rate of diabetes in the elderly in different age groups in the prediction area in the past years up to the current year is analyzed to obtain the sequence of annual change percentage and the sequence of annual average change percentage of the mortality rate of diabetes in the elderly in each age group.

[0021] ​In some embodiments, the prediction area can be a certain region, a certain city, or a certain country, and the present application does not make any limitation thereto; correspondingly, the years up to the current year, taking the current year as 2025 for example, the years up to the current year can be: 2020, 2021, 2022, 2023, 2024, and 2025, i.e. a plurality of years before 2025, and the plurality of years can be determined according to actual needs, and the present application does not make any limitation thereto.

[0022] In some embodiments, the obtaining process of the age-specific mortality rate of elderly diabetes in different age groups in the prediction area in the years up to the current year includes: The number of deaths of elderly diabetes in different age groups in the prediction area in each year and the average population in the same period in different age groups in the prediction area are substituted into the age-specific mortality rate calculation formula to calculate the age-specific mortality rate of elderly diabetes in different age groups in each year.

[0023] The age-specific mortality rate calculation formula is: Formula (1); Wherein, is the age-specific mortality rate in the age group of in the year is the number of deaths of elderly diabetes in the age group of in the year is the average population in the same period in the age group of in the year is the average population in the same period in the age group of in the year is the population proportion base.

[0024] In some embodiments, the population aged ≥65 years in the prediction area is divided into four age groups (age groups) every five years, i.e. 65-69 years old, 70-74 years old, 75-79 years old, and 80-84 years old.

[0025] For example, the number of deaths of elderly diabetes in the age group of 65-69 years old in the prediction area in 2021 is 300, and the average population in the same period in the age group of 65-69 years old in the prediction area in 2021 is 20000, is the population proportion base, such as 10 5 Correspondingly, by substituting these information into the above formula (1), the age-specific mortality rate of elderly diabetes in the age group of 65-69 years old in the prediction area in 2021 is 1500.

[0026] ​It should be noted that the joinpoint regression model is a statistical method for analyzing the change of trend in time series data, which divides the time series into several linear segments by identifying the joinpoints in the data, each segment has a different slope, so that the trend change in different time periods can be described more accurately. Correspondingly, the annual percent change (APC) refers to the percentage of trend change in a certain period of time. In the joinpoint regression model, each linear segment has an APC, which represents the growth or decline rate of the trend in that segment; the average annual percent change (AAPC) refers to the average percentage of trend change during the entire study period, and the AAPC is a weighted average of all APCs, which can more comprehensively reflect the overall trend.

[0027] In some embodiments, the above step 101 can be implemented in the following steps 1011 and 1012: Step 1011, for each age group, using the joinpoint regression model, the age-specific mortality rate of elderly diabetes in the prediction area in the past years up to the current year is segmented linearly fitted to obtain multiple turning points corresponding to the mortality rate of elderly diabetes in the age group.

[0028] Step 1012, based on the multiple turning points, output the annual percent change sequence and the average annual percent change sequence of the mortality rate of elderly diabetes in the age group.

[0029] In some embodiments, the annual percent change sequence and the average annual percent change sequence of the mortality rate of elderly diabetes in the age group can be calculated according to the age group. Here, the joinpoint regression model describes the change trend of mortality rate in a period of time by identifying the turning points and fitting the mortality rate of elderly diabetes with straight lines between the turning points, and calculates the APC sequence and the AAPC sequence of the mortality rate of elderly diabetes. Here, for each turning point, the corresponding slope can be calculated, which represents the change rate of the mortality rate of elderly diabetes before and after the turning point. By analyzing the joinpoints and the slopes of the trend lines, the change of the mortality rate of elderly diabetes can be evaluated.

[0030] Here, the joinpoint regression model is used to fit the mortality rate of elderly diabetes, where an APC>0 in the APC sequence indicates that the mortality rate of elderly diabetes is increasing, and an APC<0 indicates that the mortality rate of elderly diabetes is decreasing; APC=AAPC indicates that the mortality rate of elderly diabetes has no turning point, and the trend of diabetes mortality rate is monotonically decreasing or increasing.

[0031] Step 102, using the Bayesian age-period-cohort model, the number of elderly diabetes death cases in different age groups and the corresponding number of exposed population in the prediction area in each year up to the current year are modeled to estimate the relative risk under the age effect of each age group, the period effect of each calendar year and the cohort effect of each birth cohort.

[0032] In some embodiments, the Bayesian age-period-cohort model (BAPC) is a statistical method for analyzing and predicting the trends of disease burden, incidence or mortality over time, which helps to understand the effects of different factors on health outcomes by separating the independent effects of age, period and birth cohort.

[0033] In some embodiments, the above step 102 can be implemented by the following step 1021 and step 1022: Step 1021, using the Bayesian age-period-cohort model, the number of elderly diabetes death cases in different age groups and the corresponding number of exposed population in the prediction area in each year up to the current year are analyzed to obtain the age effect coefficient of each age group, the period effect coefficient of each calendar year and the cohort effect coefficient of each birth cohort in the prediction area.

[0034] In some embodiments, the Bayesian age-period-cohort model is used to model based on the historical data of the prediction area up to the current year, including the number of elderly diabetes death cases in different age groups and the corresponding number of exposed population in each year. The Bayesian age-period-cohort model aims to estimate and decompose three effect coefficients; among them, the age effect reflects the change of death risk with age, the period effect reflects the risk change experienced by all age groups in a specific calendar year, and the cohort effect reflects the lifelong risk change faced by a specific birth cohort due to its unique growth and life experience.

[0035] Step 1022, the age effect coefficient of each age group, the period effect coefficient of each calendar year and the cohort effect coefficient of each birth cohort in the prediction area are respectively subjected to exponential transformation to obtain the relative risk under the age effect of each age group, the relative risk under the period effect of each calendar year and the relative risk under the cohort effect of each birth cohort.

[0036] Here, the Bayesian age-period-cohort model can be fitted by using the intrinsic estimator (IE) to estimate the effect coefficients of age, period and cohort, and to perform exponential transformation on the effect coefficients to calculate the relative risk (RR) for evaluating the death risk and its change rule, which is expressed as follows: Formula (2); wherein, represents the diabetes mortality rate of the th age group (age period), the th period, and the th birth cohort, is the intercept representing the mortality reference level under the age, period, and cohort baseline, represents the age effect coefficient, represents the period effect coefficient, represents the cohort effect coefficient, is the error term.

[0037] It should be noted that in the past research of Bayesian age-period-cohort model, the period effect is often calculated by 5-year average rate, but this may have some problems, that is, the 5-year average mortality rate will average the incidence risk in a period to each age group and cohort, and this averaging effect may mask the real differences between age groups and cohorts; there will be time overlap between adjacent periods, which will cause confusion of period effect; it may not capture the short-term trend. Rapid increase or decrease in mortality rate may be averaged, so as to accurately reflect the real period effect. Therefore, the single-year incidence rate is used to calculate the period effect in the present application to improve the accuracy and reliability of the analysis.

[0038] It should be noted that the age effect is the most important source of variation, which reflects the biological and social development process of individuals with aging, and represents the development and change in the whole life process. Since the mortality risk of the disease increases with the aging of the human body, the trend of the change of the disease mortality rate presents a certain rule over time. Therefore, the age in the present application refers to the number of years from birth to death due to diabetes. Correspondingly, the period effect refers to the mortality risk of the disease in different periods, which includes a series of complex historical events and environmental factors, public health interventions and medical technology breakthroughs, which can produce period effect, thereby reducing the mortality rate of the disease in the period. The period studied in the present application refers to the year of death due to diabetes. The cohort effect refers to the mortality risk of the disease in different birth cohorts, and the birth cohort experiences different historical and social conditions at different stages of its life process, so it has different exposure of social economic, behavioral and environmental risk factors. Therefore, the cohort effect is an important part of the age-period-cohort model analysis, and the birth cohort studied in the present application refers to the year of birth of the death due to diabetes.

[0039] Step 103: Analyze the period effect of each calendar year, the relative risk under the cohort effect of each birth cohort, and the mortality rate of elderly people with diabetes in the current year in the predicted region to obtain the baseline value of the absolute mortality rate of elderly people with diabetes in each age group.

[0040] In some embodiments, step 103 above can be implemented by the following steps 1031 and 1032: Step 1031: Obtain the elderly diabetes mortality rate for each age group in the current year for the predicted region.

[0041] Step 1032: Substitute the mortality rate of elderly people with diabetes in the current year, the relative risk under the period effect of each calendar year, and the relative risk under the cohort effect of each birth cohort into the baseline value back-calculation formula to obtain the baseline value of absolute mortality rate of elderly people with diabetes in each age group.

[0042] The formula for calculating the baseline value is as follows: Formula (3); in, For age groups The baseline absolute mortality rate for elderly people with diabetes; For predicting the region in the current year younger age group The mortality rate of elderly people with diabetes; For the current year The period effect The relative risk level below; For birth cohort queue effect The relative risk level below.

[0043] This approach first uses a Bayesian age-period-cohort model to decompose the age, period, and cohort effects of relative risk from historical data. Then, by using a formula derived from the baseline value, the additional relative risk impact of the current specific year (period effect) and the specific birth population (cohort effect) is subtracted from the currently observed actual age-specific mortality rate. The resulting absolute mortality baseline not only eliminates the interference from specific period events (such as sudden changes in medical policy and public health emergencies) and specific cohort experiences (such as childhood nutritional status and historical exposure), but also more purely reflects the inherent disease mortality risk level of different age groups, determined by biological aging and long-term chronic disease courses. This absolute mortality baseline can serve as a more stable reference for fair comparisons across periods and cohorts, and as a crucial baseline input parameter in future mortality risk prediction models, unaffected by short-term fluctuations.

[0044] Step 104, predicting the absolute mortality of senile diabetes of each age group in the prediction area in future years based on the absolute mortality baseline value of senile diabetes of each age group, the relative risk of age effect of each age group, the relative risk of period effect of each calendar year, and the relative risk of cohort effect of each birth cohort, to obtain an absolute mortality prediction sequence of each age group.

[0045] In some embodiments, the above step 104 can be implemented by the following steps 1041 to 1043: Step 1041, sorting the relative risk of period effect of each calendar year by year to obtain a relative risk sequence of historical period effect, and sorting the relative risk of cohort effect of each birth cohort by birth year to form a relative risk sequence of historical cohort effect.

[0046] Step 1042, using a time trend extrapolation strategy to analyze the relative risk sequence of historical period effect and the relative risk sequence of historical cohort effect respectively, to obtain the predicted period effect value of the prediction area in future years and the cohort effect prediction value of the birth cohort that will enter the old age in the future.

[0047] Step 1043, substituting the absolute mortality baseline value of senile diabetes of each age group, the relative risk of age effect of each age group, the predicted period effect value of the prediction area in future years, and the cohort effect prediction value of the birth cohort that will enter the old age in the future into the mortality prediction formula to calculate the absolute mortality prediction sequence of each age group in the prediction area in future years.

[0048] In some embodiments, first, the estimated period and cohort relative risk are arranged in chronological order to form a relative risk sequence of historical period effect and a relative risk sequence of historical cohort effect respectively. Then, a time trend extrapolation strategy (such as a time series model or curve fitting) is used to analyze the relative risk sequence of historical period effect and the relative risk sequence of historical cohort effect respectively to predict the period effect value of future years and the cohort effect value of the new birth cohort that will enter the old age in the future. Finally, these prediction values, the absolute mortality baseline value of each age group, and the relative risk of age effect of each age group are substituted into the mortality prediction formula to calculate the absolute mortality prediction sequence of senile diabetes of each age group in the prediction area in future years.

[0049] This enables dynamic and structured predictions of future disease burden. By extrapolating trends from historical periods and cohort effects, it simulates the evolution of future social environments, advancements in medical technology (period effects), and life course risks of specific generations (cohort effects). Simultaneously, by combining inherent age risk patterns and established absolute mortality baselines, it can systematically generate age- and year-specific future mortality prediction sequences.

[0050] The mortality prediction formula is as follows: Formula (4); in, For age groups The baseline absolute mortality rate for elderly people with diabetes; For predicting the year younger age group The absolute mortality rate of diabetes in the elderly; For age groups age effect The relative risk level below; For future years The period effect The relative risk level below; Birth cohorts for future old age queue effect The relative risk level below.

[0051] In some embodiments, the predicted absolute mortality rate for a specific age group is obtained by summing the baseline absolute mortality rate for each age group, the relative risk of the age effect for that age group, the predicted period effect for the corresponding future year, and the predicted cohort effect for the corresponding future birth cohort, and then taking the exponent. This transforms the complex prediction process into a clear and repeatable formula. This log-linear formula not only ensures the theoretical interpretability of the prediction results but also guarantees the efficiency and stability of the calculation. Thus, it becomes clear how baseline risk, inherent age risk, the impact of generational changes, and intergenerational characteristics of the population interact to determine the final predicted value, greatly enhancing the auditability, verifiability, and feasibility of flexible simulations under multiple scenario assumptions.

[0052] Step 105: Analyze the absolute mortality prediction sequence, the annual percentage change sequence and the annual average percentage change sequence of elderly diabetes mortality in each age group to obtain a set of structured early warning indicators for each age group.

[0053] Here, the corresponding analysis can be performed for each age group, thereby obtaining a structured early warning indicator set for each age group; wherein the structured early warning indicator set includes but is not limited to: risk level (such as: high, medium, low), trend urgency (such as: rapid deterioration, slow change, improvement), and trend stability (such as: trend stable, trend fluctuation, trend unknown) three dimensions of grade indicators.

[0054] In some embodiments, the above step 105 can be implemented by the following step 1051 and step 1052: Step 1051, for each age group, extract the death prediction value of future years from the absolute mortality prediction sequence of the age group, and analyze the annual change percentage sequence and the annual average change percentage sequence of the age group of the elderly diabetes mortality, to obtain the recent trend slope and long-term trend direction of the age group.

[0055] Step 1052, compare the death prediction value of future years, the recent trend slope and the long-term trend direction with the corresponding preset death risk threshold, trend slope threshold and trend direction threshold of the age group, respectively, to obtain the structured early warning indicator set of the age group including risk level grade, trend urgency grade and trend stability grade.

[0056] In some embodiments, for each age group, first, extract the death prediction value of future T years (such as T = 1, 3, 5) from the absolute mortality prediction sequence of the age group; and based on the annual change percentage sequence of the elderly diabetes mortality of the age group in the last N years (for example N = 5), calculate its recent trend slope by linear regression; and according to the positive and negative of the annual average change percentage sequence of the elderly diabetes mortality of the age group and the statistical significance, determine its long-term trend direction (up, down or stable). Then, compare the death prediction value of future T years, the recent trend slope, and the long-term trend direction with the preset death risk threshold, the trend slope threshold, and the trend direction stability threshold for the age group, respectively. Finally, according to the comparison result, determine the grade of the age group in the risk level (for example: high, medium, low), the trend urgency (for example: rapid deterioration, slow change, improvement), and the trend stability (for example: trend stable, trend fluctuation, trend unknown) three dimensions, that is, form the structured early warning indicator set of the age group composed of the three dimension grades.

[0057] In this way, based on the death prediction value of each age group, the recent trend change and the long-term trend direction, combined with the historical benchmark, a multi-dimensional level evaluation is performed to generate a set of structured early warning indicators. It can realize comprehensive and quantitative analysis of the death risk of the elderly with diabetes from the static level to the dynamic trend, that is, it can convert complex mortality data into intuitive risk, urgency and stability levels, and provide direct and operable decision basis for hierarchical and precise public health intervention.

[0058] Step 106, according to the structured early warning indicator set of each age group, triggering and generating the differentiated early warning instruction of the community health service center in the predicted area.

[0059] In some embodiments, according to the structured early warning indicator set of each age group, based on the preset early warning rule, a differentiated early warning instruction is generated and directed to each community health service center in the predicted area; wherein the differentiated early warning instruction matches the age-specific risk characteristics of the population in the predicted area. That is, the differentiated early warning instruction generated and issued to each community health service center in the predicted area matches the risk level and intervention focus.

[0060] The elderly diabetes death risk early warning method provided by the embodiments of the present application, in the execution process of the method, first, through the Joinpoint regression model, the change inflection point and rate of the death rate of the elderly with diabetes in each age group in the predicted area in different historical stages are identified, the long-term trend is converted into a quantifiable and comparable annual change percentage sequence, thereby accurately capturing the evolution characteristics of the death rate of different age groups; secondly, by using the Bayesian age-period-cohort model, the change of the death rate is decomposed into the independent effects of three dimensions of age, period and cohort, and the relative risk is estimated, which not only reveals the differential influence of aging, era background and generation experience on disease risk, but also provides a structured risk parameter system for prediction, and based on the current year's death rate of the elderly with diabetes, combined with the estimated relative risk under the period and cohort effect, an absolute death rate prediction sequence of each age group is constructed; then, the absolute death rate prediction sequence is deeply integrated with the change percentage sequence and the annual average change percentage sequence to form a structured early warning indicator set of each age group covering the continuity of the trend, variability and risk level, which takes into account the long-term regularity and short-term fluctuation, and enhances the systematicness and explanatory power of the early warning signal; finally, based on the early warning indicator set of each age group, the differentiated early warning instruction for the community health service center is automatically generated. In this way, through the integration of multi-dimensional modeling, dynamic prediction and structured early warning, the fine evaluation, forward prediction and precise intervention of the death risk of the elderly with diabetes are realized, and a full-chain scientific basis from data analysis to decision support is provided for diabetes prevention and control in the predicted area.

[0061] Based on the foregoing embodiments, the embodiments of the present application further provide an old-age diabetes death risk early warning system, referring to Figure 2 As shown in the figure, the old-age diabetes death risk early warning system 200 comprises: A first analysis module 201 is configured to analyze the age-specific death rates of old-age diabetes in different age groups in the prediction area in previous years up to the current year by using a Joinpoint regression model, to obtain a sequence of annual percentage changes and a sequence of annual average percentage changes of the death rates of old-age diabetes in different age groups.

[0062] A modeling module 202 is configured to model the number of death cases of old-age diabetes in different age groups in the prediction area in previous years up to the current year by using a Bayesian age-period-cohort model, to estimate the relative risk under the age effect of each age group, the period effect of each calendar year, and the cohort effect of each birth cohort.

[0063] A second analysis module 203 is configured to analyze the relative risk under the period effect of each calendar year, the cohort effect of each birth cohort, and the death rate of old-age diabetes in each age group in the prediction area in the current year, to obtain the baseline value of the absolute death rate of old-age diabetes in each age group.

[0064] A prediction module 204 is configured to predict the absolute death rate of old-age diabetes in each age group in the prediction area in future years based on the baseline value of the absolute death rate of old-age diabetes in each age group, the relative risk under the age effect of each age group, the relative risk under the period effect of each calendar year, and the relative risk under the cohort effect of each birth cohort, to obtain a sequence of predicted absolute death rates in each age group.

[0065] A third analysis module 205 is configured to analyze the sequence of predicted absolute death rates in each age group, the sequence of annual percentage changes and the sequence of annual average percentage changes of the death rates of old-age diabetes in each age group, to obtain a set of structured early warning indicators in each age group.

[0066] A generation module 206 is configured to trigger and generate differentiated early warning instructions for community health service centers in the prediction area according to the set of structured early warning indicators in each age group.

[0067] It should be noted that the description of the above system embodiments is similar to that of the above method embodiments, and has similar beneficial effects as the method embodiments. For technical details not disclosed in the system embodiments of the present application, please refer to the description of the method embodiments of the present application for understanding.

[0068] It should be noted that, in the embodiments of the present application, if the above-mentioned old diabetic death risk early warning method is implemented in the form of a software function module and sold or used as an independent product, it can also be stored in a computer-readable storage medium. Based on this understanding, the technical solutions of the embodiments of the present application can be embodied in the form of a software product, and the computer software product is stored in a storage medium, and includes a number of instructions for causing an electronic device to execute all or part of the methods described in the embodiments of the present application. The aforementioned storage medium includes: a U disk, a mobile hard disk, a read-only memory (ROM), a magnetic disk or an optical disk, and various media that can store program codes. Thus, the embodiments of the present application are not limited to any specific combination of hardware and software.

[0069] It should be understood that the "one embodiment" or "an embodiment" mentioned throughout the specification means that the specific features, structures or characteristics related to the embodiment are included in at least one embodiment of the present application. Therefore, "in one embodiment" or "in an embodiment" appearing throughout the specification does not necessarily refer to the same embodiment. In addition, these specific features, structures or characteristics can be combined in one or more embodiments in any suitable manner. It should be understood that in various embodiments of the present application, the size of the sequence number of each process does not mean the order of execution, and the execution order of each process should be determined according to its function and inherent logic, and should not constitute any limitation on the implementation process of the embodiments of the present application. The sequence number of the above-mentioned embodiments of the present application is only for description, not representing the advantages and disadvantages of the embodiments.

[0070] It should be noted that in this paper, the term "include", "contain" or any other variant thereof is intended to cover non-exclusive inclusion, so that the process, method, article or device including a series of elements not only includes those elements, but also includes other elements not explicitly listed, or includes elements inherent to such process, method, article or device. Without more limitations, the element defined by the statement "including a" does not exclude the presence of other identical elements in the process, method, article or device including the element.

[0071] In several embodiments provided in the present application, it should be understood that the disclosed devices and methods can be implemented in other manners. The embodiments described above are merely exemplary, and the unit division is merely logical function division, and there can be other division manners in actual implementation. For example, a plurality of units or components can be combined or integrated into another system, or some features can be ignored or not executed. In addition, the displayed or discussed coupling, direct coupling or communication connection between the components can be indirect coupling or communication connection through some interfaces, and can be electrical, mechanical or other forms.

[0072] The units described as separate components can or can not be physically separate, and the components shown as units can or can not be physical units; they can be located in one place, or distributed on a plurality of network units; and part or all of the units can be selected according to actual needs to achieve the purpose of the embodiments of the present application.

[0073] In addition, each functional unit in each embodiment of the present application can be integrated into one processing unit, or each unit can be a separate unit, or two or more units can be integrated into one unit; the integrated unit can be realized in the form of hardware, or in the form of hardware plus software functional unit.

[0074] Alternatively, the integrated unit of the present application, if realized in the form of a software function module and sold or used as an independent product, can also be stored in a computer readable storage medium. Based on this understanding, the technical solutions of the embodiments of the present application can be embodied in the form of a software product, which is stored in a storage medium, and includes a plurality of instructions for causing an apparatus to perform all or part of the methods described in the embodiments of the present application. The foregoing storage medium includes: mobile storage devices, ROM, magnetic disks or optical disks, and various other media that can store program codes.

[0075] The methods disclosed in the several method embodiments of the present application can be combined arbitrarily without conflict, to obtain new method embodiments.

[0076] The features disclosed in the several method or system embodiments of the present application can be combined arbitrarily without conflict, to obtain new method embodiments or system embodiments.

[0077] The above merely provides the implementation of the present application, but the protection scope of the present application is not limited thereto, any person skilled in the art can easily think of the change or replacement within the technical range disclosed by the present application, which should be covered in the protection scope of the present application. Therefore, the protection scope of the present application should be subject to the protection scope of the claims.

Claims

1. A method for early warning of mortality risk in elderly people with diabetes, characterized in that, The method includes: The Joinpoint regression model was used to analyze the age-specific mortality rates of elderly people with diabetes in different age groups in the prediction area up to the current year, and the annual percentage change series and annual average percentage change series of elderly people with diabetes mortality rates in each age group were obtained. Using a Bayesian age-period-cohort model, we modeled the number of elderly diabetes deaths and the corresponding number of people exposed in the middle of the year for different age groups in the prediction region up to the current year, and estimated the age effect for each age group, the period effect for each calendar year, and the relative risk under the cohort effect for each birth cohort. The relative risk under the period effect of each calendar year, the cohort effect of each birth cohort, and the mortality rate of elderly people with diabetes in the current year in the predicted region were analyzed to obtain the baseline value of the absolute mortality rate of elderly people with diabetes in each age group. Based on the baseline absolute mortality rate of elderly people with diabetes in each age group, the relative risk under the age effect of each age group, the relative risk under the period effect of each calendar year, and the relative risk under the cohort effect of each birth cohort, the absolute mortality rate of elderly people with diabetes in each age group in the prediction area in the next few years is predicted, and the absolute mortality rate prediction sequence of each age group is obtained. The absolute mortality prediction series, the annual percentage change series and the annual average percentage change series of elderly diabetes mortality in each age group were analyzed to obtain a set of structured early warning indicators for each age group. Based on the structured early warning indicator set for each age group, differentiated early warning instructions are triggered and generated for community health service centers within the prediction area.

2. The method according to claim 1, characterized in that, The process of obtaining age-specific mortality rates for diabetes in different age groups over the predicted region up to the current year includes: The number of elderly people with diabetes in different age groups in each year of the predicted region and the average population of the same period in each age group of the predicted region are obtained and substituted into the age-specific mortality rate calculation formula to obtain the age-specific mortality rate of elderly people with diabetes in different age groups in each year; wherein, the age-specific mortality rate calculation formula is: ； in, For the year age group Age-specific mortality rates; For the year age group The number of deaths from diabetes in the elderly; For the year age group The average population during the same period; This is the population proportion base.

3. The method according to claim 1 or 2, characterized in that, Using the Joinpoint regression model, we analyzed the age-specific mortality rates of elderly people with diabetes in the prediction region up to the current year, obtaining the annual percentage change series and annual average percentage change series of mortality rates for each age group, including: For each age group, the Joinpoint regression model is used to perform piecewise linear fitting on the age-specific mortality rate of elderly people with diabetes in the prediction area up to the current year, and obtain multiple inflection points corresponding to the mortality rate of elderly people with diabetes in that age group. Based on multiple inflection points, the annual percentage change and annual average percentage change sequence of elderly diabetes mortality rate in this age group are output.

4. The method according to claim 1, characterized in that, Using a Bayesian age-period-cohort model, the number of elderly diabetes deaths and the corresponding mid-year exposure population in the prediction region across different age groups up to the current year are modeled. The age effect, the period effect of each calendar year, and the relative risk under the cohort effect for each birth cohort are estimated, including: Using the Bayesian age-period-cohort model, we analyzed the number of elderly diabetes deaths and the corresponding number of people exposed during the year in different age groups in the prediction region up to the current year, and obtained the age effect coefficient, period effect coefficient and cohort effect coefficient of each birth cohort in the prediction region. The age effect coefficients for each age group, the period effect coefficients for each calendar year, and the cohort effect coefficients for each birth cohort within the prediction area are subjected to exponential transformation to obtain the relative risk under the age effect for each age group, the relative risk under the period effect for each calendar year, and the relative risk under the cohort effect for each birth cohort.

5. The method according to claim 1, characterized in that, The period effect of each calendar year, the relative risk under the cohort effect of each birth cohort, and the mortality rate of elderly people with diabetes in the current year in the predicted region were analyzed to obtain the baseline values ​​of the absolute mortality rate of elderly people with diabetes in each age group, including: Obtain the elderly diabetes mortality rate for each age group in the current year for the predicted region; The baseline mortality rate for diabetes in the elderly in the current year is calculated by substituting the relative risk under the period effect of each calendar year and the relative risk under the cohort effect of each birth cohort into the baseline value back-calculation formula. The baseline value back-calculation formula is as follows: ； in, For age groups The baseline absolute mortality rate for elderly people with diabetes; For predicting the region in the current year younger age group The mortality rate of elderly people with diabetes; For the current year The period effect The relative risk level below; For birth cohort queue effect The relative risk level below.

6. The method according to claim 1, characterized in that, Based on the baseline absolute mortality rate of diabetes in the elderly for each age group, the relative risk under the age effect for each age group, the relative risk under the period effect for each calendar year, and the relative risk under the cohort effect for each birth cohort, the absolute mortality rate of diabetes in the elderly for each age group in the prediction region is predicted for multiple years in the future, resulting in the absolute mortality rate prediction sequence for each age group, including: The relative risk under the period effect corresponding to each calendar year is sorted by year to obtain the relative risk sequence under the historical period effect. The relative risk under the cohort effect corresponding to each birth cohort is sorted by birth year to form the relative risk sequence under the historical cohort effect. Using a time trend extrapolation strategy, we analyzed the relative risk series under the historical period effect and the relative risk series under the historical cohort effect, respectively, to obtain the predicted values ​​of the period effect in the prediction area in the next few years and the predicted values ​​of the cohort effect of the birth cohort that will enter old age in the future. The absolute mortality rate baseline for diabetes in each age group, the relative risk under the age effect for each age group, the period effect prediction value of the prediction area in multiple future years, and the cohort effect prediction value of the birth cohort that will enter old age in the future are substituted into the mortality prediction formula to calculate the absolute mortality rate prediction sequence for each age group in the prediction area in multiple future years.

7. The method according to claim 6, characterized in that, The mortality prediction formula is: ； in, For age groups The baseline absolute mortality rate for elderly people with diabetes; For predicting the year younger age group The absolute mortality rate of diabetes in the elderly; For age groups age effect The relative risk level below; For future years The period effect The relative risk level below; Birth cohorts for future old age queue effect The relative risk level below.

8. The method according to claim 1, characterized in that, Analysis of the absolute mortality prediction series, the annual percentage change series and the annual average percentage change series of elderly diabetes mortality rates for each age group yielded a structured early warning indicator set for each age group, including: For each age group, the predicted mortality value for future years is extracted from the absolute mortality prediction sequence for that age group. The annual percentage change sequence and the annual average percentage change sequence of elderly diabetes mortality for that age group are analyzed to obtain the short-term trend slope and long-term trend direction for that age group. The predicted mortality values, recent trend slope, and long-term trend direction for future years are compared with the preset mortality risk threshold, trend slope threshold, and trend direction threshold corresponding to the age group, respectively, to obtain a structured early warning indicator set for the age group, including risk level, trend urgency level, and trend stability level.

9. A risk warning system for mortality from diabetes in the elderly, characterized in that, The system includes: The first analysis module is used to analyze the age-specific mortality rate of elderly people with diabetes in different age groups in the prediction area up to the current year using the Joinpoint regression model, and to obtain the annual percentage change series and annual average percentage change series of elderly people with diabetes mortality rate in each age group. The modeling module is used to model the number of elderly diabetes deaths and the corresponding number of people exposed in different age groups in the prediction region up to the current year using a Bayesian age-period-cohort model, and to estimate the age effect, period effect and relative risk under the cohort effect of each age group, each calendar year and each birth cohort. The second analysis module is used to analyze the period effect of each calendar year, the relative risk under the cohort effect of each birth cohort, and the mortality rate of elderly people with diabetes in the current year in the predicted region, so as to obtain the baseline value of the absolute mortality rate of elderly people with diabetes in each age group. The prediction module is used to predict the absolute mortality rate of elderly people with diabetes in the prediction area in the next few years based on the baseline absolute mortality rate of elderly people with diabetes in each age group, the relative risk under the age effect of each age group, the relative risk under the period effect of each calendar year, and the relative risk under the cohort effect of each birth cohort, so as to obtain the absolute mortality rate prediction sequence of each age group. The third analysis module is used to analyze the absolute mortality prediction sequence, the annual percentage change sequence and the annual average percentage change sequence of elderly diabetes mortality in each age group, and to obtain a set of structured early warning indicators for each age group. The generation module is used to trigger and generate differentiated early warning instructions for community health service centers within the prediction area based on a structured early warning indicator set for each age group.

Images