Using primary and routinely collected data to determine prevalence and patterns of multimorbidity in rural China: a representative cross-sectional study of 6474 Chinese adults.

Background: In China, rising chronic diseases has coincided with the increasing burden of multimorbidity, particularly for vulnerable populations. Limited primary data are available to understand the prevalence and patterns of multimorbidity, especially in resource-limited rural areas. This study aims to conduct robust evaluations of the prevalence and patterns of multimorbidity among rural adults in China, and to compare the differences in prevalence and patterns when using primary data alone versus in combination with routinely collected data. Methods: This cross-sectional study was conducted in three provinces in China, with two counties per province and 40 villages per county, resulted in a total of 240 villages. Participants were randomly selected and stratified by sex and age in each village. Multimorbidity, defined as the coexistence of two or more diseases in same individual, was assessed through data collection involving primary data (face-to-face questionnaire, physical examination and fasting blood sample collection) and routinely collected data (health insurance claims, hospital electronic records and infectious disease surveillance system). Multimorbidity prevalence and patterns were compared based on 1) primary data alone and 2) primary data complemented by routinely collected data. Findings: A total of 6474 individuals participated in this study (50.9% women, mean age 57.1). Combining routinely collected data with the primary data increased the prevalence of all single disease conditions. Multimorbidity prevalence rose from 35.7% with primary data alone to 44.4% with the addition of routinely collected data. The top three dyad multimorbidity patterns (hypertension with heart disease, stroke, or chronic digestive diseases) remained consistent between the two ascertainment methods, while triad pattern rankings had a substantial shift. According to blood pressure measurements, over 40% of participants had elevated blood pressure and may have undiagnosed hypertension. Over 20% may have undiagnosed mental health disorders base on the questionnaires, and nearly 10% with undiagnosed chronic kidney disease as indicated by blood testing. Interpretation: The utilisation of primary data combined with routinely collected data provided a robust estimation of multimorbidity burden in three rural regions in China. Yet, the prevalence may still have been underestimated due to inaccuracies in self-reported data and underdiagnosis of diseases. Future research should incorporate routinely collected data for more robust epidemiological evidence of multimorbidity. Funding: Harbin Medical University Leading Talent Grant (31021220002) and National Natural Science Foundation of China (72074065 and 72474063).