(207) Understanding the Utility of Historical Cohort Studies Via Retrospective Chart Review in Inflammatory Bowel Disease: A Review of Global Studies

Background: There is a growing number of therapies available in inflammatory bowel disease (IBD), with diverse treatment preferences in clinical practice. This presents a need to collect real-world evidence (RWE) to understand the natural history and treatment outcomes. Retrospective chart review studies can collect RWE when existing data sources are not viable. However, it is important to understand study-specific considerations.

Objectives: To describe recent chart review studies in IBD indications, and inform on study design and data quality considerations, and share synthesized data metrics and learnings.

Methods: A review of 9 recent global historical cohort chart review studies in IBD populations was undertaken to summarize objectives and outcomes and describe outcome assessment methods, data missingness and data quality.

Results: From the 9 studies, there were 3015 patients from 12 countries (in North America, Europe, and Asia-Pacific). There was 1 natural history study and 8 incident user studies (index date being treatment initiation) describing treatment outcomes (effectiveness, safety, treatment patterns and health resource utilization [HRU]). Chart reviews were chosen because secondary data sources were not available, or missing key outcomes (e.g., disease severity, adverse events [AEs], reasons for treatment modifications, concomitant medications, endoscopy data). Patient follow-up (min-max) was < 0.5 – 8 years, and disease assessment frequency varied among patients within studies. Data on demographics (e.g., age, sex), medical history (e.g., diagnosis, disease location and severity) and treatment patterns (e.g., treatment discontinuation/modifications) had low missingness ( < 10% for most variables). There was no standardized method of assessing effectiveness (e.g., clinical remission) across sites. Thus, multiple methods had to be collected (up to 5 methods/outcome) and built into algorithms, so outcomes could be compared among methods. Effectiveness outcomes typically had low missingness (≤20%), except when data from procedures were needed for assessments (e.g., endoscopic outcomes), where fewer patients had data available (15-60%). For safety outcomes, AE type/severity were typically able to be captured from charts, even though Common Terminology Criteria for AEs was not used by sites. For AEs and HRU outcomes, due to limitations in chart review methodology, it was unclear if any events were not reported in charts.

Conclusions: Retrospective chart review studies in IBD populations generated RWE that led to many disseminations to inform disease management and to understand treatment outcomes. It is key to understand methodologic and operational considerations when deciding if retrospective chart review design is appropriate.