A Method to Estimate Prevalence of Diabetes for Adults Based on Insurance Data

Abstract

Background: Insurance data could be a good source to get prevalence of diabetes. However, because of the low treatment rate of diabetes, the prevalent cases from the insurance data are just part of all prevalent cases in general population. Therefore, it is important to explore a method to fill in the gap between the insurance data and the general population. Objective: This study aims to offer a method to estimate the prevalence of diabetes among population based on insurance data, and reported the prevalence of Weifang city, China. Methods: By borrowing the methods on population deaths estimation based on the hospital deaths, we constructed a possibility model by using the data from the 2018 Social Factors of Chronic Disease (SFCD) of Weifang to assess the probability (Pi) of diabetes cases having treatment in hospital, which means when there is one patient in the insurance, there will be 1/(Pi) patients in the population. Then based on the model, we used the diabetes patient in 2020 insurance database to estimate the prevalent cases in population. Results: The possibility model is Logit P= −3.4484+0.0520∗A+0.2520∗G−0.7011∗I. The female or the people having urban employees insurance were more possible to have treatment. The model estimated that the prevalence of diabetes in the whole population in Weifang City was 7.29% for adults. Conclusions: This study proposed a method to estimate diabetes prevalence among population based on the insurance data. The prevalence of diabetes in the whole population in Weifang City was 7.29% for adults.