Prospectively measuring placebo and nocebo effects: development of the Stanford Expectancy of Treatment Scale (SETS)

Patient expectations have shown to impact positive and negative response to treatments. In clinical trials, positive and negative expectancies may drive placebo and nocebo responses, which can blur the difference between real and sham treatments and reduce statistical power to detect true treatment effects. We sought to develop and validate an instrument that measures patients' expected response to new clinical interventions. Our goal was to design an instrument that was: 1) quick to administer, 2) easily understandable by patients, 3) able to detect positive and negative treatment expectancies, and 4) applicable to a broad range of intervention modalities. The Stanford Expectations of Treatment Scale (SETS) was developed from a pool of 22 questions. Initial factor analysis and internal reliability analysis were conducted by recruiting 200 patients who were preparing to undergo inpatient surgery. In the scale development phase, two factors emerged: positive expectations and negative expectations. Three items were chosen for each subscale that most highly correlated with the respective factor. Internal reliability (Cronbach's alpha) was 0.86 for the positive expectation subscale, and 0.81 for the negative expectation subscale. Both subscales demonstrated good internal reliability. We assessed the initial predictive validity of the scale on a new group of 84 surgery and chronic pain patients. These individuals completed the 6-item SETS, and two weeks following their treatment, they provided reports of their improvement and experience of adverse events. Pre-treatment positive expectancies were significantly correlated with post-treatment beneficial response (r = 0.30, p = 0.007). Pre-treatment negative expectancies were significantly correlated with post-treatment severity of adverse effects (r = 0.45, p = 0.00001). We show that the SETS is an easy-to-use, effective measure of patient expectations that can prospectively measure placebo and nocebo responses in clinical interventions.