(SLC-040) Application of observational research methods to real-world studies for rare disease treatments: a scoping review

Background: Rare diseases, affecting fewer than 200,000 people annually but encompassing approximately 7,000 conditions globally, present challenges in generating treatment efficacy and safety evidence through conventional RCTs due to small populations and heterogeneity. Real-world evidence (RWE) from observational studies has been proposed to address these challenges yet concerns regarding randomization and masking persist.

Objectives: This scoping review identifies observational research methods used in the past five years to evaluate rare disease treatments and explores their application in generating effectiveness and safety evidence.

Methods: We searched Ovid MEDLINE and EMBASE for studies published between 2018 and 2023. Reference lists of included articles and relevant literature were also screened. Observational studies assessing the effectiveness or safety of rare disease treatments were included. Independent reviewers conducted screening and data extraction.

Results: A total of 83 studies were included, nearly half of which were conducted in Asia (n = 41, 49%), followed by Europe (n = 23, 28%) and North America (n = 10, 12%). Rare oncology conditions (n = 41) were the most frequently studied. 36% of the studies (n = 30) were multicenter by design, while 64% (n = 53) were single center. Sample size limitations were notable, with 71% (n = 59) of studies involving fewer than 100 patients. Majority of studies included data sources from medical records and or databases (n=68), as well as registries (n = 9), with the Surveillance, Epidemiology, and End Results (SEER) database commonly utilized for rare oncology conditions. Single-arm studies comprised nearly half (n = 41), often supplementing the absence of a control arm with external controls from historical data or clinical trials, as well as self-controlled designs. Among the multi-arm, comparative studies (n=42), propensity score matching, and cox proportional hazards models were the most used methods to address small sample sizes and confounding. Design adaptations, including multicenter collaborations, database networks, linked registries, and extended retrospective timeframes, mitigated sample size limitations; however, issues such as missing data and selection bias persisted.

Conclusions: Observational studies play a crucial role in rare disease research, offering pragmatic alternatives to RCTs. Standardizing methodologies is essential to enhance the consistency and reliability of RWE in informing healthcare decisions and policies for rare disease patients.