Investigating socioeconomic deprivation and antibiotic prescribing among older medicare patients using an instrumental variable approach

Abstract

Background: Socioeconomic deprivation has been associated with antibiotic overprescription in the US; however, prior studies could not quantify a causal relationship due to endogeneity. This study examines how socioeconomic deprivation is related to the rate of antibiotic days supplied to older Medicare Part D beneficiaries, utilizing an Instrumental Variable (IV) approach.

Methods: Data from the Medicare Part D and the Social Deprivation Index (SDI) repositories were analyzed. To address potential endogeneity and omitted variable bias in the relationship between SDI and antibiotic prescribing, we used the maximum Earned Income Tax Credit as an IV. Bivariate Moran’s I assessed the spatial correlation between SDI and antibiotic prescribing across geographic regions. The IV analysis then examined the relationship between predicted SDI and antibiotic days supplied (ln). Linear regression models estimated associations between SDI and its components, and antibiotic days supplied, adjusting for prescriber, beneficiary, and geographic factors.

Results: Among 161,164, there was no significant spatial dependence between SDI and antibiotic days supplied (P = 0.0656). In the IV model, a one-unit increase in SDI was associated with a 0.582 (SE = 0.164, P < 0.0005) increase in antibiotic days supplied (ln). Higher unemployment and single-parent family rates were linked to increased antibiotic days supplied, while crowded housing was associated with a reduction.

Conclusion: This study identified that socioeconomic deprivation may influence antibiotic days supplied to Medicare Part D beneficiaries. Findings highlight the need for targeted public health interventions to address the socioeconomic factors contributing to excess antibiotic use.