Does Medication Cost-Sharing Predict Emergency Room Use for Asthma?

Research Objective:  Growing evidence suggests that reduced cost-sharing for secondary prevention therapies can be cost-effective. And yet, 1 in 5 Americans is now insured with a high-deductible policy, based on the theory that moral hazard causes excess healthcare utilization when cost-sharing is low. The purpose of this study was to determine the relationship between relative level of cost-sharing for daily preventive asthma medications and risk of an asthma-related visit to the emergency department (ED).

Study Design:  This study consisted of secondary analysis of Medical Expenditure Panel Survey (MEPS) data. The primary outcome was the occurrence of an ED visit due to asthma during the year of data collection, according to the household questionnaire. The predictor was the proportion of total payments for preventive asthma medications paid by the participant or family, according to the pharmacy questionnaire. Quintiles were examined due to non-normality of the cost-sharing data. The sample was then risk-adjusted for race/ethnicity, because being Black and/or Hispanic was associated with both cost-sharing level and ED use. Statistical analysis was performed with the complex samples logistic regression procedure in SPSS v21.

Population Studied:  MEPS employs stratified cluster sampling and is representative of the civilian, non-institutionalized U.S. population. Data from 2010 and 2011 were pooled, because asthma-related ED visits were assumed to be relatively rare. Participants over age 18 and currently taking daily preventive asthma medication at the time of interview were included.

Principal Findings:  Contrary to the hypothesis, increased level of cost-sharing for preventive asthma medications was associated with decreased odds of asthma-related ED visit in binary analysis. Relative to the first cost-sharing quintile, the second, third, and fourth quintiles had 90.2% (p=.031), 77.5% (p=.042), and 85.3% (p=.032) decreased odds of asthma-related ED visit. In multivariate analysis, the inclusion of race/ethnicity removed the association between cost-sharing quintile and odds of ED visit for asthma (p≥.098 for all quintiles). Prescription cost data were missing for 33% of all eligible cases; 41.3% of Black and/or Hispanic cases had missing prescription data compared to 25.3% of others. Cases with missing prescription data were also significantly likelier to have an asthma-related ED visit.

Conclusions:  Among adult MEPS participants with reported preventive asthma medication use in 2010-2011, there was no difference in likelihood of an asthma-related ED visit across cost-sharing quintiles once the effect of race/ethnicity was taken into account. However, Blacks and Hispanics were notably likelier to have missing prescription cost data, and cases with missing prescription data were at greater risk of asthma-related ED visit. Therefore, the results may not include persons at risk of ED use who could not afford their prescriptions at all.

Implications for Policy or Practice:  This study is important because it fails to disconfirm earlier studies which found that lower cost-sharing for chronic disease medications may reduce ED visits. It also highlights the importance of minimizing missing data in large, national datasets, especially when the missing data are distributed unevenly across relevant risk factors. Future analyses should further examine the relationship between prescription cost-sharing and likelihood of purchasing needed medications, and other sources of cost-sharing data should be explored.