UNDERSTANDING MULTIMORBIDITY AS A DRIVER OF HIGH-COST STATUS IN DIABETES: EVIDENCE FROM LONGITUDINAL CLAIMS DATA IN CHINA

Author(s)

Xing Chen, PhD1, Luying Zhang, PhD2, Wen Chen, PhD2.
1Shanghai Institute of Infectious Disease and Biosecurity, Fudan University, Shanghai, China, 2School of Public Health, Fudan University, Shanghai, China.
OBJECTIVES: Multimorbidity is highly prevalent among individuals with diabetes and is a key contributor to healthcare cost concentration. However, the relative contribution of specific conditions to high-cost (HC) status remains insufficiently understood. This study aimed to quantify the burden of multimorbidity and identify key cost-driving conditions among diabetic patients using machine learning.
METHODS: We conducted a retrospective cohort study using longitudinal health insurance claims and chronic disease management data from 79,910 diabetic patients in eastern China (2014-2019). Patients in the top 10% of annual healthcare expenditures in 2019 were classified as HC. Multimorbidity was defined based on 29 chronic conditions identified through literature review and utilization thresholds. Descriptive analyses compared HC and non-high-cost (NHC) patients. Six machine learning models were developed to predict HC status using 2018 multimorbidity profiles, with performance evaluated by AUC, accuracy, sensitivity, specificity, and F1 score. The best-performing model was interpreted using SHapley Additive exPlanations (SHAP).
RESULTS: A total of 7,991 (10.0%) patients were classified as HC. HC patients had a higher multimorbidity burden than NHC patients (mean number of conditions: 8.6 vs. 6.4). Mean annual expenditure was substantially higher among HC patients ($25,903 vs. $2,490). Multimorbidity-related care accounted for 34.2% of total expenditures and was over ten times higher in HC patients. Hypertension was the most prevalent condition among HC patients (91.3%), while cerebrovascular disease sequelae generated the highest per-patient costs ($8,869). The Extreme Gradient Boosting model achieved the best performance (AUC = 0.832). SHAP analysis identified chronic ischemic heart disease, cerebrovascular disease sequelae, and osteoporosis as the most influential predictors of HC status.
CONCLUSIONS: Multimorbidity is a central driver of healthcare cost concentration among diabetic patients. Cardiovascular, cerebrovascular, and skeletal conditions play a disproportionate role in shaping high-cost status.

Conference/Value in Health Info

2026-09, ISPOR Asia Pacific 2026, Bangkok, Thailand

Value in Health, Volume 55, Issue S1

Code

EE44

Topic

Economic Evaluation

Topic Subcategory

Cost/Cost of Illness/Resource Use Studies

Disease

SDC: Diabetes/Endocrine/Metabolic Disorders (including obesity)

Your browser is out-of-date

ISPOR recommends that you update your browser for more security, speed and the best experience on ispor.org. Update my browser now

×