Consumer-directed health plans reduce utilization of essential drugs

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

By Peter J. Huckfeldt, Amelia Haviland Ateev, Mehrotra Zachary Wagner, and Neeraj Sood
National Bureau of Economic Research, NBER Working Paper No. 20927, February 2015


Prior studies suggest that consumer-directed health plans (CDHPs) - characterized by high deductibles and health care accounts - reduce health costs, but there is concern that enrollees indiscriminately reduce use of low-value services (e.g., unnecessary emergency department use) and high-value services (e.g., preventive care). We investigate how CDHP enrollees change use of pharmaceuticals for chronic diseases. We compare two large firms where nearly all employees were switched to CDHPs to firms with conventional health insurance plans. In the first firm’s CDHP, pharmaceuticals were subject to the deductible, while in the second firm pharmaceuticals were exempt. Employees in the first firm shifted the timing of drug purchases to periods with lower cost sharing and were more likely to use lower-cost drugs, but the largest effect of the CDHP was to reduce utilization. Employees in the second firm also reduced utilization, but did not shift the timing or use of low cost drugs

From the Introduction

Overall, we find evidence that employees in the CDHP with pharmaceuticals subject to the deductible use more low cost drugs and shift the timing of drug purchases to periods with lower cost sharing. The majority of the response, however, is to reduce overall drug utilization. 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 decreased utilization relative to firms with conventional coverage. The presence of a utilization response may suggest a lack of understanding of plan design or reflect secondary effects where increased cost sharing leads to fewer physician visits where drugs are prescribed. The persistence of the response into the second year of the plan (presumably a sufficient time period for enrollees to understand plan design) may suggest this response is due to cost sharing for physician visits.

From the Conclusion

We investigate the effects of CDHPs on use of pharmaceuticals that treat and prevent chronic diseases including high cholesterol, hypertension, and type 2 diabetes. We compare changes in utilization among employees in two firms that shift all of their employees to CDHPs relative to employees in firms that offered no CDHP; in one firm (CDHP firm 1) pharmaceuticals are subject to the deductible and in the other pharmaceuticals are exempt from the deductible (CDHP firm 2). We find that enrollees in CDHP firm 1 change utilization of pharmaceuticals along each hypothesized margin: enrollees shift the timing of purchase to periods with lower cost sharing, increase the use of lower cost drugs (but only when such drugs were equally or more effective), and reduce overall pharmaceutical utilization. The majority of the response, however, is focused on the reduction in overall utilization. In addition, most of these effects persist into the second year of the CDHP. In CDHP firm 2, we observe insignificant changes in utilization for statins, but statically significant or marginally significant reductions in utilization of antihypertensive drugs and diabetes drugs. We observe no effect, however, on the percentage of days supply purchased or used that were low cost. The utilization effects for CDHP firm 2 are larger and statistically significant for all drug categories in the second year of the CDHP. We interpret the persistence of these effects, combined with the absence of a change in use of low cost drugs, as suggestive evidence of a “gatekeeper” effect, where higher cost sharing for physician visits result in reductions in pharmaceutical use. Moreover, some of the reduction in utilization in CDHP firm 1 could also be driven by a reduction in access to physician visits, a phenomenon observed in the RAND Health Insurance Experiment (Leibowitz, Manning, and Newhouse 1985).



By Don McCanne, MD

One of the most important changes taking place in the financing of health care is the widespread adoption of high-deductible health plans, not only with the plans offered in the ACA insurance exchanges, but also with employer-sponsored health plans. When high-deductible plans are combined with a personal health account (health reimbursement arrangement or health savings account) then they are referred to as consumer-directed health plans (CDHPs). This study further informs us on the perverse consequences of this model of health care financing.

Three different paragraphs, each describing this study, were selected for today’s quote to help the reader get through the fog of this fairly complex, technical study. Basically, patient utilization rates for drugs in three chronic conditions (hypercholesterolemia, hypertension, and Type 2 diabetes) were measured under three different scenarios: a large employer with a CDHP in which pharmaceuticals applied to the deductible, requiring full payment until the deductible was reached, a large employer with a CDHP in which pharmaceuticals were exempt from the deductibles, requiring only copayments, and a third group of 19 other firms that offered only traditional coverage and no CDHP.

The primary finding was that patients in CDHP plans decreased utilization of pharmaceuticals for these disorders compared to patients in traditional health plans. This was true whether or not drug purchases applied to the deductible. It is crucial to note that having a personal health account from which funds could be drawn to pay for these drugs did not prevent patients from decreasing utilization of these important drugs.

It is easy to understand why patients declined to purchase drugs for which they had to pay the full cost, even though those drugs were important in their medical management. We have innumerable studies that show that patients will decline care that they should have when they are faced with significant out-of-pocket expenses, demonstrating why high-deductible plans are a such a perverse method of controlling health care spending.

But why would patients with high-deductible plans reduce their use of beneficial drugs if they were exempt from the deductible, requiring them to pay only the copayments? One reason may be that the intrusions and complexities of private insurer cost-shifting arrangements, such as through CDHPs, may be training patients to decline care except in the most dire of circumstances. Another reason may be that the physician visit required to obtain the prescription must be paid in full until the deductible is met, causing patients to drop off of their therapeutic program. Again, the availability of an employer-funded health reimbursement arrangement did not prevent patients from declining to fill their prescriptions (although an empty personal health account is not very helpful).

The bottom line is that too many patients are not receiving drugs and other care that they should have merely because of the increasing transition to CDHPs, with their high-deductible barriers to care. We should abandon CHDPs and shift to first dollar coverage in a program that controls spending through patient-friendly efficiencies - a single payer national health program.