Are preferences for cyclooxygenase-2 inhibitors influenced by the certainty effect?

Abstract

To test whether the widespread use of cyclooxygenase-2 (COX-2) inhibitors may be mediated in part by the certainty effect, i.e., by a perception that COX-2 inhibitors eliminate the risk of serious gastrointestinal (GI) events in contrast to merely reduce their risk. Patients' preferences for conventional nonsteroidal antiinflammatory drugs (NSAID) and COX-2 inhibitors for treatment of arthritis were predicted across a range of absolute risks of GI events using an Adaptive Conjoint Analysis survey. Preferences for COX-2 inhibitors were much stronger when the risk of serious GI events was eliminated in contrast to reduced, even though the absolute risk reduction (ARR) was the same. Few patients (22%) preferred COX-2 inhibitors when the risk associated with NSAID and COX-2 inhibitors was 4% and 2%, respectively (ARR = 2%). In contrast, the majority (90%) preferred COX-2 inhibitors when the risk associated with NSAID and COX-2 inhibitors was 2% and 0%, respectively (ARR = 2%). We obtained similar findings regardless of the ARR. The willingness shown by older adults to pay for COX-2 inhibitors may reflect a misperception of the risk of toxicity associated with these medications.