Abstract
BACKGROUND: This proof-of-concept retrospective case study investigated whether patient-reported outcomes (PRO) instruments, designed to capture symptomatic adverse event data, could identity a known exposure-response (ER) relationship for safety characterized in an original FDA analysis of an approved anti-cancer agent. PRO instruments have been designed to uniquely quantify the tolerability aspects of exposure-associated symptomatic adverse events. We explored whether standard ER analyses of clinician-reported safety data for symptomatic adverse events could be complemented by ER analysis using PRO data that capture and quantify the tolerability aspects of these same symptomatic adverse events. METHODS: Exposure-associated adverse event data for diarrhea were analyzed in parallel in 120 patients enrolled in a clinical trial using physician reported Common Terminology Criteria for Adverse Events (CTCAE) and patient-reported symptomatic adverse event data captured by the National Cancer Institute's (NCI) PRO Common Terminology Criteria for Adverse Events (PRO-CTCAE) instrument. Comparative ER analyses of diarrhea were conducted using the same dataset. Results from the CTCAE and PRO-CTCAE ER analyses were assessed for consistency with the ER relationship for diarrhea established in the original NDA using a 750-patient dataset. The analysis was limited to the 120-patient subset with parallel CTCAE and PRO-CTCAE assessments. RESULTS: Within the same 120-patient dataset, ER analysis using dense, longitudinal PRO-CTCAE-derived data was sensitive to identify the known ER relationship for diarrhea, whereas the standard CTCAE based ER analysis was not. CONCLUSIONS: ER analysis using PRO assessed symptomatic adverse event data may be a sensitive tool to complement traditional ER analysis. Improved identification of relationships for safety, by including quantification of the tolerability aspect of symptomatic adverse events using PRO instruments, may be useful to improve the sensitivity of exposure response analysis to support early clinical trial dosage optimization strategies, where decision making occurs within limited small patient datasets.