In 2008, Blue Cross Blue Shield of North Carolina (BCBS-NC) initiated a value-based insurance design (VBID) program that reduced coopayments for hypertension, hyperlipidemia, diabetes, and congestive heart failure medications. Over 700,000 employer-sponsored insurance-based policyholders were subject to the program. (A refresher in VBID is here.) In the February 2014 issue of Health Affairs, Matthew Maciejewski and colleagues studied BCBS-NC's program. Maciejewski et al. framed their examination as an investigation of the "business case" for VBID.
The business case for population-based VBID is particularly important because group health plans and issuers of group and individual health insurance have been covering certain preventive services without patient cost sharing since 2012, as a result of VBID provisions in the Affordable Care Act. Services that are exempt from cost sharing include those with A and B recommendations from the US Preventive Services Task Force (such as screening for high blood pressure and colorectal cancer) and women’s and children’s health services.
It would be especially valuable and impressive if a change in medication utilization could more than pay for itself through a reduction in non-drug utilization. However, it would be unwise to judge VBID by this criterion alone. If medications at least partly substitute for or delay utilization with potentially higher chances of complications and life disruption (e.g., surgery), that may be a substitution worth making even if it doesn't save money to the health system.
Indeed, the evidence that VBID is cost savings is mixed. Maciejewski et al.'s analysis comes down on the side of no cost savings from VBID. They compared the change in expenditures from one year before to one and two years after the BCBS-NC program between those with one of the relevant conditions and subject to the program and a matched sample of patients from employers not subject to it. The VBID "treatment" group experienced two copay reductions for selected drug classes relevant to the conditions listed above: elimination of copay for generic medications and a reduction in copay for brand-name medications. The comparison group experienced only the brand-name copay reduction.
Two years after the population-based implementation we studied, total expenditures were not reduced. Thus, to inform VBID implementations going forward, there is an urgent need to establish the conditions under which it would be cost-neutral. The business case for value-based insurance design may be more compelling over the long term or in higher-risk subgroups, such as older subgroups or patients for whom cost is an important barrier to improved medication adherence.
Though their point estimate was a reduction in total spending for VBID-subject patients, it was not statistically significant. One potential interpretation is that the additional cost of medications were offset about dollar-for-dollar by reduced non-medication spending. Consequently, while a (cost savings) business case for VBID from these results is weak, a case could be made that the program helped shift care to a modality that was less disruptive and detrimental to patients.
There are, of course, some limitations, which the authors acknowledge. These include that it's possible cost savings occur after two years of program experience; It's also possible there is something idiosyncratic about the North Carolina population or conditions examined, threatening the generality of the findings; This was not a randomized design, so bias from omitted variables is possible.
The analysis also only considered costs to the insurer and patients, not to the employer or society. If, for example, use of medications as substitutes for other care reduces absenteeism, that would be a benefit to the employer. If it increased quality of life, that would be a benefit to patients. (Data for such analyses were not available to the researchers.)
A cost-savings business case for VBID is hard to make, but that alone does not imply that VBID is not a more broadly worthwhile feature of health insurance.
For more about recent research on and context of VBID, see Michael Millenson's post at Forbes.
* One year before the program the first and second year after it was implemented.