Objective: Signal detection and, ultimately, regulatory approval depend on high-quality, valid and reliable data. The subjective rating scales utilized in CNS clinical trials may be vulnerable to spurious ratings and intentional or unintentional manipulation of ratings by investigators at screening or baseline visits. The objective of this study was to evaluate the feasibility of utilizing a patient reported outcome as a quality assurance measure for evaluation of the quality of a clinician rated primary efficacy measure in a CNS clinical trial. Methods: A proprietary ratings surveillance system was utilized in a multi-center, double blind, randomized, placebo-controlled clinical trial in which the Hamilton Anxiety Rating Scale (HARS) was the primary efficacy measure. The patient rated Beck Anxiety Inventory (BAI) was added to the baseline visit for quality assurance purposes. Based on published guidelines of the expected relationship between HARS and BAI scores, a computer program flagged aberrant ratings and three flags with the same rater triggered a teaching intervention. The ratings surveillance system was intended both to detect aberrant rating patterns and to deter intentional inflation of ratings in order to qualify subjects. Results: The clinical trial is ongoing. 91 pairs of HARS and BAI ratings have been examined from the randomization visit. 61/91 (67%) pairs were flagged for discordance, in most cases (79%) due to disproportionately high HARS scores compared to the BAI. In 8 cases, the BAI was under 10 with the HARS 22 or greater. In 11 cases, there were at least 3 flags for the same rater and the pattern of discordance was considered to be of sufficient clinical significance to warrant a teaching intervention. Conclusion: Use of patient reported outcomes as a measure of quality of clinician reported outcomes appears to be a feasible tactic in a site-based ratings surveillance quality assurance system.