Patient Responses to Incentives in Consumer-Directed Health Plans: Evidence from Pharmaceuticals

Background:  Consumer-directed health plans (CDHPs), characterized by high deductibles and health savings accounts, are meant to encourage more efficient use of health care services, as beneficiaries face the full cost of services before reaching the deductible.  Earlier research has shown that health care costs for individuals in such plans fall in the first years of enrollment. It remains unclear whether cost reductions represent indiscriminate reductions in health care utilization versus more strategic decision-making.

Objective: We investigate the effects of CDHP enrollment on utilization of pharmaceuticals for chronic diseases.  Specifically, we estimate the extent to which CDHP enrollees make “smarter” decisions, including switching to lower-priced drugs and shifting the timing of purchase to periods of lower cost sharing, versus more shortsighted decisions such as reducing medication adherence.

In addition, we examine how CDHP responses differ across drug classes with varying availability of low cost alternatives and between firms with CDHPs where pharmaceuticals are subject to the deductible versus exempt from the deductible.

Data:  We use 100% pharmaceutical claims and enrollment data from 2004-2006 for 21 large firms, including: (1) a large firm shifting all employees to CDHP in 2005 in which pharmaceuticals are subject to the deductible; (2) a large firm shifting all employees to a CDHP in 2005 in which pharmaceuticals are exempt from the deductible; and (3) 19 large firms not offering high deductible health plans or CDHPs between 2004 and 2006.  Our sample includes employees using drugs treating high cholesterol, hypertension, or type 2 diabetes in 2004.

Methods:   We focus on firms shifting all employees to CDHPs to avoid selection bias from CDHP take-up. We examine pharmaceuticals because they are used repeatedly, allowing us to trace changes in utilization before CDHP adoption and over the course of the plan-year. We use a differences-in-differences strategy, estimating differential changes in pharmaceutical utilization for firms adopting CDHPs in 2005 relative to firms only offering traditional health plans. We perform inference using a permutation test approach, comparing estimated treatment effects for the CDHP firms to estimated placebo effects that assign treatment to each of the 19 comparison firms.  

Findings:  Overall, we find evidence that employees in the CDHP with pharmaceuticals subject to the deductible reduce adherence, use more low cost drugs, and shift the timing of drug purchases to periods with lower cost sharing. The low-cost drug response is only present for classes where there are equally effective low-cost alternatives available.  These results imply that enrollees possess some awareness of benefit design and availability of low cost drugs, but also suggest that enrollees either discount or are not cognizant of the adverse consequences of poor adherence.

Employees in the CDHP with pharmaceuticals exempt from the deductible also exhibit differential reductions in adherence and shifts in the timing of purchase relative to firms with traditional plans; however, the magnitudes are smaller and often statistically insignificant. The presence of a utilization response may suggest a lack of understanding of plan design or reflect increased cost sharing for physician visits where drugs are prescribed.