Objectives: Previous research has identified that clinician-reported outcomes (ClinRO) may be affected by knowledge of treatment assignment. We sought to evaluate whether physician reporting of adverse events (AE) may have been affected by knowledge of treatment assignment.

Methods: We used data from the control arms of two randomized registration commercial trials in multiple myeloma with identical active control arms and similar populations. The trials were conducted within one year of each other, but one trial was open-label and one was double-blind. Both trials used the CTCAE for adverse event reporting. We compared the incidence of reporting of twelve relevant AEs known to be associated with the control drugs through six months of treatment across trials. These AEs included directly observable events (e.g. rash) and less observable events (e.g. insomnia). To evaluate the impact of blinding, we estimated the average treatment effect on the treated (ATT), using propensity scores (PS) with inverse probability weighting (IPW). The ATT was calculated with logistic regression models, with the double-blind trial as the reference group (OR>1 indicates higher incidence in the open-label trial).

Results: The unadjusted incidence of any of the twelve AEs in the double-blind trial was 76.4%, vs 82.3% in the open-label trial. The OR for the ATT was 1.39 (95% CI: 0.92 to 2.08). For rash, the OR was 1.50 (95% CI: 0.96 to 2.34) and for back pain it was 1.82 (95% CI: 1.13 to 2.94). However, incidence for some AEs including diarrhea (OR 0.77, 95% CI: 0.52 to 1.14) and nausea (OR 0.82, 95% CI: 0.52 to 1.30) was higher in the double-blind trial.

Conclusion: Overall, we did not find consistent evidence of higher clinician reporting of AEs if treatment assignment was known. Future work assessing this association in more datasets across more disease populations is warranted.