OBJECTIVES : Data from electronic medical records are frequently used for retrospective real-world studies of treatment patterns and outcomes. However, site-based retrospective chart review methodology has inherent limitations including extended timelines, greater funding requirements and high site burden. Thus, studies that utilize real-world direct-to-physician networks for data collection are warranted. The objective was to describe a drug utilization and effectiveness IRB-approved pilot study design and performance metrics, based on data from physician-reported medical charts.

METHODS : A US-based retrospective chart review pilot study was executed, using an existing on-line physician network. The protocol was centrally IRB-approved and eligible patients were identified from 2016-2019 who had ≥ 6 months of follow-up. A large panel of physicians, representative of the US market, had been previously screened on factors including years of practice, credentials, scientific publications and location. Stratified sampling, based on diseases treated and number of patients treated with the disease, was used to select physicians. Variables for the study included physician and patient characteristics, treatment patterns and response to treatment. Data were entered by treating physicians into an online platform, including automated, random and tailored edit checks and the subsequent issuing of any queries.

RESULTS : Forty-three of the 91 contacted physicians participated (47.3%), of which 24 (26.4%) provided data for patients included in the study. The medical chart data of 119/153 (77.8%) screened patients was included in the study. Data was collected (screened and abstracted) efficiently, within 2 months. Queries to the treating physicians on average were answered in 3-5 days. The final dataset had <1% of missing data.

CONCLUSIONS : This direct-to-physician, on-line approach for a retrospective chart review study was time and cost-efficient for collecting high quality real-world drug utilization and effectiveness data. This study demonstrates that direct-to-physician chart collection can be an effective and flexible alternative to traditional site-based chart review studies.