High-deductible health plans are here to stay and for the employed population are often the only insurance choice they have. The major question regarding these plans is not whether they work to reduce health spending, they clearly seem to do that, it is whether they negatively affect quality by deterring receipt of appropriate as well as inappropriate care. New research at the National Bureau of Economic Research examines patients’ responses to high-deductible plan incentives in regard to medication use. (NBER Paper) The researchers compared drug use by persons covered by two large companies’ high-deductible plans, one of which subjected medications to the deductible, and one which did not, with drug use for members in 19 other large firms’ non-high-deductible plans. The study was focussed on drugs for chronic diseases, which account for a high percent of all drug use. Both of the high-deductible plan employers in the study shifted all employees to that form of coverage in 2005. The high-deductible plans in both cases were accompanied by health-savings accounts, which ameliorate the effect of the increased cost-sharing. The authors looked at use of generic drugs, the timing of drug purchases and reduced medication adherence. The study primarily covered the first and second year after the switch, which may be slightly misleading, because these plan designs are more complex and it can take covered persons some time to understand their coverage better and figure out how to optimize their benefits. Other studies have shown shifts in behavior from year-to-year after implementation of a high-deductible plan.
The primary drugs looked at were those for cholesterol control, hypertension and diabetes. Presumably just not obtaining or filling prescriptions for these drugs would be an inappropriate response by the members in the high-deductible plans, because that could lead to worse health outcomes in the longer term. The researchers expected to see larger behavior changes in the company where drugs were subject to the deductible, but even in the other high-deductible plan company, they believed there could be a response because of the potential for fewer physician visits or general greater sensitivity to cost. (Note that measures used for adherence, days supply purchased or used, has problems because it does not identify pill-splitting, which is very common. In addition, pharmacies don’t always send claims in when the cost of a prescription is below the maximum cost-share, which is often the case with generics, because the pharmacy isn’t getting paid anything by the plan in those cases.) At baseline in 2004, many patients appeared to be non-adherent, in other words they did not seem to be filling enough prescriptions to meet their drug needs. After implementation of the high-deductible plan, patients in those plans appeared to alter the timing of their drug purchases, filling more prescriptions during times of lower cost-sharing, i.e. after the deductible was met. Patients in the comparator traditional plans did not show these changes. In addition, the members in the high-deductible plan which included drugs under the deductible had an overall reduction in drug use. These patients also had higher rates of generic drug use. While the authors suggest that the results likely suggest poorer health outcomes, they investigate no data to confirm this. That is the crucial question, for if health status or health outcomes are no worse over an extended period, you have to wonder if these patients really needed the drugs in the first place.