Characteristics of Hospital Emergency Room Visits for Mental and Substance Use Disorders.

OBJECTIVE: This study compared characteristics of visits to emergency rooms (ERs) for mental and substance use disorders and for physical health conditions to establish a baseline against which to measure changes after full implementation of the Affordable Care Act (ACA) and parity legislation. METHODS: The retrospective, cross-sectional analysis fit a logistic regression model to pooled data comprising 193,526 observations from National Hospital Ambulatory Medical Care Survey (NHAMCS) data from 2005 to 2011. RESULTS: ER visits for mental or substance use disorders increased from 27.9 per 1,000 ER visits in 2005 to 35.1 in 2011. Homeless persons and nursing home residents had the highest rates of such visits-173.7 and 95.2 per 1,000 ER visits, respectively, in 2011. CONCLUSIONS: Understanding differences in profiles of ER visits on the basis of the reason for the visit can inform the design of more cost-effective policies to guide ER intake, after further implementation of the ACA and parity legislation.

Determinants of U.S. Prescription Drug Utilization using County Level Data.

Prescription drugs are the third largest component of U.S. healthcare expenditures. The 2006 Medicare Part D and the 2010 Affordable Care Act are catalysts for further growths in utilization becuase of insurance expansion effects. This research investigating the determinants of prescription drug utilization is timely, methodologically novel, and policy relevant. Differences in population health status, access to care, socioeconomics, demographics, and variations in per capita number of scripts filled at retail pharmacies across the U.S.A. justify fitting separate econometric models to county data of the states partitioned into low, medium, and high prescription drug users. Given the skewed distribution of per capita number of filled prescriptions (response variable), we fit the variance stabilizing Box-Cox power transformation regression models to 2011 county level data for investigating the correlates of prescription drug utilization separately for low, medium, and high utilization states. Maximum likelihood regression parameter estimates, including the optimal Box-Cox λ power transformations, differ across high (λ = 0.214), medium (λ = 0.942), and low (λ = 0.302) prescription drug utilization models. The estimated income elasticities of -0.634, 0.031, and -0.532 in high, medium, and low utilization models suggest that the economic behavior of prescriptions is not invariant across different utilization levels.