Impact of "Defensive Medicine" on the Costs of Diabetes and Associated Conditions

Abstract

Background: Geographic variation in health care spending is typically attributed to differences in patient health status and provider practice patterns. While medicolegal considerations (i.e., "defensive medicine") anecdotally impact health care spending, this phenomenon is difficult to measure. The purpose of this study was to explore the association between the medicolegal environment and Medicare costs for diabetes and associated conditions of interest to vascular surgeons. Specifically, we hypothesized that an adverse medicolegal environment is associated with higher per capita Medicare costs for diabetic patients.

Methods: Medicare data including the most recent (2018) Medicare Geographic Variation Public Use Files and Chronic Conditions Data Files were linked to National Practitioner Data Bank files from the preceding 5 years (2013-2017), in addition to the US census data and American Medical Association workforce statistics. The state-level medicolegal environment was characterized by K-means clustering across a panel of metrics related to malpractice payment magnitude and prevalence. Per capita Medicare spending for diabetes was compared across 5 distinct medicolegal environments. Costs were standardized and risk-adjusted to account for known geographic variation in health care costs and patient population. Analysis of variance was applied to unadjusted data, followed by multivariate regression modeling. Readmission rates, per capita imaging studies, per capita tests, per capita procedures, and lower extremity amputation rates were compared between the least litigious quintile from the K-means clustering and the 2 most litigious quintiles.

Results: The median unadjusted Medicare per capita expenditure on diabetic patients was $15,963 ($14,885-$17,673), ranging from $13,762 (Iowa) to $21,865 (D.C.). A 1.6-fold variation persisted after payment standardization. Cluster analysis based on malpractice-related variables yields 5 distinct medicolegal environments, based on litigation frequency and malpractice payment amounts. Per capita spending on diabetes varied, ranging from $15,799 in states with low payments and infrequent litigation to $18,838 in states with the most adverse medicolegal environment (P < 0.05). After cost standardization and risk adjustment with multiple linear regression, malpractice claim prevalence (per 100 physicians) remained an independent predictor of states with the highest diabetes mellitus spending (P = 0.022). Moreover, diabetic patients in states with adverse medicolegal environments had more procedures, imaging studies, and readmissions (P < 0.05 for all) but did not have significant differences in amputation rates compared to less litigious states.

Conclusions: An adverse medicolegal environment is independently associated with higher health care costs but does not result in improved outcome (i.e. amputation rate) for diabetic Medicare beneficiaries. Across states, a 1% increase in lawsuits/100 physicians was associated with a >10% increase in risk-adjusted standardized per capita costs. These findings demonstrate the potential contribution of "defensive medicine" to variation in health care utilization and spending in a population of interest to vascular surgeons.