(C-220) Global Feasibility Assessment of Real-World Databases for Long-Term Safety of Vitiligo Treatments

Background: Recent discoveries on the pathophysiology of vitiligo led to new approved treatments for the condition globally. These new topical and systemic therapies, along with various procedures, underscore the need for real-world evidence to address drug safety knowledge gaps.

Objectives: To assess the relevance and reliability of United States (US) and European vitiligo data sources that could answer questions on the long-term safety of approved products.

Methods: Ten secondary data sources from Austria, France, Germany, Italy, Spain, United Kingdom (UK), and the US were systematically investigated. Data sources were required to capture vitiligo diagnosis to be eligible for the study. A detailed feasibility questionnaire was developed to capture data source characteristics, operational aspects, patient management, sample size, and variables using the Structured Process to Identify Fit-For-Purpose Data (SPIFD) framework and other qualitative frameworks for regulatory studies. Variables were assessed for their availability and expected completeness using a two-step approach with 3-level rankings (good: >70%, moderate: 40-70%, and low: < 40%). Feasibility of rules-based Natural Language Processing methodology to extract clinical characteristics in unstructured patient notes was explored in 1 US data source. Data suitability was assessed according to the total 70-point score of SPIFD 5-point scale and priority 1-3 scale for study requirements.

Results: Ninety percent of data sources (n=9) met the inclusion criteria for this study. Most patient demographics had good completeness in all sources, while ethnicity was only available with good completeness in the US and UK. Vitiligo clinical characteristics, such as diagnosis codes and their dates, had good completeness in 6 sources, moderate completeness in 1 source, and unknown completeness in 2 sources. Location of vitiligo and phenotype had moderate to good completeness in 4 sources. All of the assessed data sources had good completeness of treatment variables except for 3 European sources. Six promising sources (scored ≥63 out of 70) were identified for pharmacoepidemiologic-based vitiligo research from the US (n=3, 1 EMR-linked-claims, 1 claims, and 1 EHR data sources) and Europe (n=3, 1 German claims, 1 French data system-linked claims, and 1 UK EMR-linked-claims data sources).

Conclusions: The recommended data sources are promising candidates for causal and descriptive evaluations for US and European regulatory drug safety studies in vitiligo patients. This study highlights the importance of using qualitative frameworks for data feasibility assessments.