OBJECTIVES: Randomized clinical trials (RCT) have a set of inclusion and exclusion criteria enabling to target the patient population that will be analyzed. New therapies in oncology trend to personalized medicine. The objective was to assess the extent to which French medico-administrative database (PMSI) can be used to assess transferability of patient characteristics from RCT in oncology to real life setting.

METHODS: Patient characteristics in non-small-cell lung cancer (NSCLC), melanoma and renal cell carcinoma (RCC) were taken from published RCTs of nivolumab. Whilst some are common across RCTs (age, sex, performance status), several are tumor specific (histology, tumor mutations, risk factors, prior therapies, metastases, surgery). Using PMSI, a retrospective cohort study was conducted with patients having received nivolumab in 2017 for NSCLC, melanoma, and RCC. PMSI was queried to retrieve patient characteristics from RCTs for each tumor location.

RESULTS: In 2017, 15,787 patients were treated with nivolumab in France. PMSI allowed to capture 3 out of 10 (3/10) published baseline RCT characteristics in NSCLC, 3/9 in melanoma and 5/11 in RCC. Age and sex were directly retrievable. Metastatic sites, prior intravenous therapies (if above Diagnosis-Related Group) and surgeries can also be retraced. However, performance status, histology, tumor mutations, risk factors, and oral therapies are not recorded in PMSI, which constitutes a major limitation to assess transferability of patient characteristics.

CONCLUSIONS: PMSI is an exhaustive database enabling the assessment of real-life use of oncology drugs in French hospital setting. Indeed, PMSI allows to analyze drug consumption per patient as well as some patient characteristics that can be directly retrieved or recalculated. However, there is limited tumor-specific information available. Thus, in order to assess transferability of RCT patient characteristics to real-life setting, additional sources are needed.