(368) Comparisons of pharmacy-dispensed medications in the 2022 Medical Expenditure Panel Survey and the Healthcare Integrated Research Database

Background: : Drug shortages pose serious public health risks that can be difficult to predict. The U.S. Agency for Healthcare Research and Quality’s (AHRQ) Medical Expenditure Panel Survey (MEPS) provides national data on pharmacy-filled medications each year. However, the ability of the MEPS to inform public health responses to drug shortages is limited due to timeliness and small sample sizes. Pharmaceutical claims databases may provide timely, reliable data on pharmaceutical use, but their representativeness is uncertain.

Objective: : To compare the utility of data on prescription drug dispensing events from two sources to inform drug shortage occurrence and impacts.

Methods: : We used data from the Medical Expenditure Panel Survey (MEPS) and the Healthcare Integrated Research Database (HIRD®), a large, commercial health insurance claims database to identify drugs dispensed in 2022. We categorized National Drug Codes (NDCs) associated with each event into 14 Multum Lexicon® therapeutic categories. For each data source and therapeutic category, we calculated the number of fills, the number of unique NDCs, and the average number of fills per NDC within each therapeutic category. We used standardized mean differences (SMDs) to compare the share of fills and NDCs within each therapeutic category across the two data sources.

Results: : MEPS recorded 191,514 fills, a value roughly 14 times higher than the 163.5 million HIRD-recorded fills. MEPS recorded 8 percent more unique NDCs than did the HIRD (7,661 vs. 7,038, respectively). The shares of total fills across data sources differed most for anti-infectives and metabolic agents (SMDs = 0.23 and 0.11, respectively). Counts of NDCs within therapeutic categories were more similar (SMDs <= 0.02). The number of fills per NDC in each therapeutic category ranged between 8 and 48 in the MEPS and 6,000 and 34,801 in the HIRD.

Conclusions: : Results suggest data on pharmacy-dispensed medications from the HIRD are highly similar to those reported in the MEPS, despite the widely varying measurement methodologies. The larger number of NDCs reported in the MEPS suggests it more fully represents national variation in prescription drug benefit coverage. At the same time, the substantially larger number of fills per NDC in the HIRD demonstrates that claims databases can facilitate reliable analysis of shortages affecting narrow drug classes. Future research should explore how claims data can complement nationally representative data to promote timely detection and response to drug shortages.

Sponsorship:: Carelon Research, RAND Corporation