Choosing wisely: low-value services, utilization, and patient cost sharing.

THE JUNE 2012 ISSUE OF CONSUMER REPORTS includes a cover story entitled “5 Medical Tests You Don’t Need.” The story reflects a joint Choosing Wisely initiative by Consumer Reports and the American Board of Internal Medicine aimed at “encouraging physicians, patients and other health care stakeholders to think and talk about medical tests and procedures that may be unnecessary, and in some instances can cause harm.” The framing of this initiative as a way to improve quality and patient safety is important. For too long, efforts to reduce the use of low-value services have been decried by critics as rationing or as schemes to enhance insurance company profits. The rationing frame has often been motivated by political posturing or stakeholder financial interests and has helped perpetuate the consequences of unchecked health spending on individuals, families, and federal and state budgets. The Consumer Reports story reveals to the general public something many in the medical profession already know: While much health care spending does provide substantial individual and social value, some of it supports care of little or no value. Efforts to tie patient cost sharing to the benefit of the treatment in question and not just the cost through value-based insurance design (VBID) have recently proliferated within employee benefits circles. If co-payments are increased for low-value services and reduced for high-value services, standard economics predicts that patients will migrate from the former to the latter, making better use of health spending dollars. Several studies have found that patients who faced increases in medication co-payments decreased their use; of these, some also found that savings in pharmacy costs were offset by higher rates of emergency department utilization and hospitalization, so no money was saved overall— while rates of adverse events increased. These findings seemed to imply that reducing co-payments could have the reverse effect: increasing adherence and reducing emergency department utilization and hospitalization—better outcomes without higher costs. The logic behind this promoted efforts to reduce co-payments for high-value medications in high-risk populations. However, subsequent studies have found that increasing and decreasing co-payments do not have mirrorimage effects. Lowering co-payments does not improve utilization nearly as much—typically only 1 to 4 percentage points on baseline medication possession ratios (MPR) of 60% to 80%—an asymmetry that was not predictable from standard economic theory. This means that there would be 20 to 25 people whose adherence did not change for every completely nonadherent patient (MPR=0%) who became highly adherent (MPR 80%). A study in which patients who had acute myocardial infarction (AMI) were randomly assigned to standard co-payments or zero co-payments for statins, -blockers, and angiotensin-converting enzyme inhibitors found disturbingly low MPRs of 39% in the year following AMI in the control group with improvement to only 45% in the zero co-payment group, a difference that resulted in no significant reduction in the rate of total major vascular events or health care spending. There are several reasons for the asymmetry between the large effect of increasing co-payments and the small effect of lowering them. First, people tend to be loss averse, and as a result, co-payment increases are far more potent than co-payment decreases. Second, co-payment reductions every 30 or 90 days may be too infrequent to motivate daily medication adherence. Third, co-payment increases and decreases target different populations. Increases target adherent patients but decreases are meant to attract patients who are not taking medications. Those who do not take medication will not notice changes in prices they are not paying. These results imply that even though VBID may not be highly effective in increasing utilization of desired services, it could be effective in decreasing utilization of lowvalue services. Higher patient cost sharing would deter patient demand for certain types of low-value services: patients