Blinding data collectors in investigations of experimental pain: let’s get real

Clinical trials often state investigators were blinded without describing or testing the blinding procedures. Here we examined the blinding of data collectors to participants’ assigned condition within a crossover investigation of sex differences in exercise-induced hypoalgesia and conditioned pain modulation. Participants (N = 33; 52% women; 23 years old (SD = 5.01)) were assigned to a random order of three conditions: (1) control - quiet rest and trapezius isotonic saline injections; (2) algesic injections - quiet rest and trapezius hypertonic saline injections; and (3) exercise - quadriceps exercise and trapezius isotonic injections. Data collectors recorded participants’ pressure and heat pain thresholds and pain ratings before and after a condition was administered by different personnel, but they did not observe the experimental condition. Thus, each session required multiple research personnel. At the end of each session, data collectors recorded their beliefs regarding the injection and exercise condition, their confidence level in their beliefs (0-100% scale), and the reason/s for their beliefs. Seven participants (21.2% of sample; 57.1% women) had vasovagal reactions to the injections, which caused missing data. Of the remaining participant sessions, investigators correctly identified isotonic and hypertonic injections 90% and 73% of the time, respectively, and quiet rest and exercise 94% and 69% of the time, respectively. Accordingly, substantial agreement was detected between the data collectors’ beliefs and the actually administered condition for both the injections (Kappa = 0.65, p < .001) and exercise (Kappa = 0.67, p < .001). The average confidence of the data collectors in their beliefs was 60.17% (SD = 34.54) for injections and 62.19% (SD = 34.95) for exercise. Participants’ pain responses were the primary reason for data collectors’ beliefs, but the multitude of reasons reported supports the challenges of successful blinding. Researchers should consider the impact and probability of success of blinding procedures.