THE COST OF TOO MUCH TREATMENT: CAUSAL EVIDENCE ON POLYPHARMACY AND HEALTHCARE EXPENDITURES AMONG PATIENTS WITH DIABETES
Author(s)
Nan Peng, PhD1, ruolan wei, BS2, Ermo Chen, Ph.D.3, Jinbo Zhang, BS4, Dongning Yao, Ph.D.2, Jing Wu, PhD5, Gordon Liu, PhD3, Beini Lyu, Ph.D.3.
1phd, School of Pharmaceutical Science and Technology,Tianjin University, Tianjin, China, 2School of Pharmacy, Nanjing Medical University, Nanjing, China, 3Institute for Global Health and Development, Peking University, Beijing, China, 4School of International Pharmaceutical Business, China Pharmaceutical University, Nanjing, China, 5Chinese Pharmaceutical Association, Tianjin, China.
1phd, School of Pharmaceutical Science and Technology,Tianjin University, Tianjin, China, 2School of Pharmacy, Nanjing Medical University, Nanjing, China, 3Institute for Global Health and Development, Peking University, Beijing, China, 4School of International Pharmaceutical Business, China Pharmaceutical University, Nanjing, China, 5Chinese Pharmaceutical Association, Tianjin, China.
OBJECTIVES: Healthcare systems face growing pressure to improve efficiency as resources remain limited and the economic burden of chronic diseases continues to rise. Diabetes requires long-term pharmacological management for glycemic control and comorbidity care, making polypharmacy common in routine practice. While multiple medications may reflect appropriate treatment intensification, excessive medication burden may increase treatment complexity, medication-related risks, and potentially avoidable healthcare expenditures. This study aimed to estimate the causal effect of the number of medications used on healthcare expenditures among patients with diabetes.
METHODS: We used population-based longitudinal data from 73,766 adults with diabetes in eastern China between 2015 and 2021. The primary exposure was the number of medications used, defined as a measure of medication burden and regimen complexity. To address endogeneity between disease severity and medication count, we applied an instrumental variable strategy using practice-level prescribing preference as quasi-exogenous variation in medication use. We estimated the causal effect of medication count on total healthcare expenditures using IV-based generalized linear models and two-part models. Secondary analyses examined outpatient, inpatient, and medication expenditures.
RESULTS: We identified a nonlinear relationship between medication count and healthcare expenditures. At lower levels of medication count, additional medications were associated with reduced total expenditures, suggesting that appropriate treatment intensification may prevent costly healthcare use. However, once medication count exceeded a critical threshold, further increases were associated with sharply rising expenditures, mainly driven by higher inpatient and medication-related costs. Findings were consistent across alternative model specifications and sensitivity analyses.
CONCLUSIONS: This study provides causal evidence of a threshold effect between medication burden and healthcare expenditures among patients with diabetes. While moderate treatment intensification may reduce costs, excessive medication burden may increase avoidable healthcare spending. These findings support risk-based medication review, prescribing optimization, and system-level incentives to promote efficient medication use.
METHODS: We used population-based longitudinal data from 73,766 adults with diabetes in eastern China between 2015 and 2021. The primary exposure was the number of medications used, defined as a measure of medication burden and regimen complexity. To address endogeneity between disease severity and medication count, we applied an instrumental variable strategy using practice-level prescribing preference as quasi-exogenous variation in medication use. We estimated the causal effect of medication count on total healthcare expenditures using IV-based generalized linear models and two-part models. Secondary analyses examined outpatient, inpatient, and medication expenditures.
RESULTS: We identified a nonlinear relationship between medication count and healthcare expenditures. At lower levels of medication count, additional medications were associated with reduced total expenditures, suggesting that appropriate treatment intensification may prevent costly healthcare use. However, once medication count exceeded a critical threshold, further increases were associated with sharply rising expenditures, mainly driven by higher inpatient and medication-related costs. Findings were consistent across alternative model specifications and sensitivity analyses.
CONCLUSIONS: This study provides causal evidence of a threshold effect between medication burden and healthcare expenditures among patients with diabetes. While moderate treatment intensification may reduce costs, excessive medication burden may increase avoidable healthcare spending. These findings support risk-based medication review, prescribing optimization, and system-level incentives to promote efficient medication use.
Conference/Value in Health Info
2026-11, ISPOR Europe 2026, Vienna, Austria
Value in Health, Volume 29, Issue 12S
Code
HSD30
Topic
Health Service Delivery & Process of Care, Real World Data & Information Systems
Disease
Diabetes/Endocrine/Metabolic Disorders (including obesity)