OBJECTIVES: Patient Reported Outcomes (PROs) are increasingly used in clinical studies as well as requested by health authorities (EMA, FDA). This descriptive analysis aims to explore how PRO were used in the French affiliate of Roche. METHODS: 89 local studies performed between 2003 and 2013 were evaluated to determine how PROs were collected and evaluated. RESULTS: Forty (45%) out of 89 studies were clinical trials, 44 (49%) were non-interventional studies (NIS) and 5 were expanded access program studies. A total of 33 (37%) studies used at least one PRO; 13 (39%) were clinical trials and 20 (61%) were NIS. There were no PRO in the field of transplantation, 21% of the studies in oncology used PRO, 47% in virology, 60% in anemia and 72% in rheumatology. More than 3600 sites participated in the 33 studies using PRO: 17424 patients were included, 14237 patients (82%) answered at least one PRO and 12314 (86%) of the questionnaires received were considered to be analyzable for statistical analysis (non analyzable PROs were usually due to missing data). The median number of questionnaires per study was 2, (range 1-6) . While most studies in  virology and anemia used 1 or 2 questionnaires per study, rheumatology studies used at least 3 questionnaires per study. The median study duration was 12 months, (range 1-60). While 84% of PRO were collected at inclusion in the study, 73% of PROs were received after the 6 month visit, 65% at the 12 month visit and less than 50% after 12 months. Seven studies (21%) used PRO as a primary endpoint, and the percentage of patients who answered the questionnaire answered was higher (mean 91%) than when PRO was a secondary endpoint (mean 79%). CONCLUSIONS: Careful consideration should be given to number of questionnaires used per study as quality decreases as number of evaluations increases.