Real-World Data In Medical Devices: What Makes Healthcare Data Good Enough For Regulatory Use

Real-world data from healthcare sources holds immense potential for medical device manufacturers, but its value hinges on several crucial factors. Senior Manager at Intuitive Surgical, Kiran Chitambare, emphasizes that data quality, regulatory compliance, and clinical research standards are paramount. These include ICH-GCP, 21 CFR Part 11, CDISC, and CDASH.

The increasing volume of data from electronic health records, device registries, insurance claims, medical records, digital health technologies, remote monitoring systems, and patient inputs forms real-world data or RWD. This data can help researchers understand device performance across diverse patient populations, clinical settings, and extended timeframes.

However, more data does not automatically translate to better evidence. Regulatory bodies like the FDA are concerned with the reliability, relevance, and scientific sufficiency of the data to answer specific regulatory questions. Therefore, the "good enough" standard for regulatory use is not just about database size. It's about the credibility of the information in contributing to safety, performance, or effectiveness evidence while maintaining data integrity and traceability.

The question of regulatory relevance often comes first. Even a vast dataset may be invaluable only if it represents the intended patient population, captures clinically meaningful outcomes, establishes device exposure, and extends follow-up periods for related complications. Thus, researchers need to prioritize relevance over available data size.

Accuracy in real-world data involves more than just correcting data-entry errors. It requires establishing a reliable link between patients, procedures, devices, and clinical outcomes. This can be complex when data are sourced from multiple systems. A patient record combined with device registry, claims system, or remote monitoring platform data must be accurately linked.

The same applies to medical devices that may undergo model, software, component changes, or manufacturing lot variations. Identifying these specifics is crucial for determining whether a safety signal applies broadly or to a specific model, version, or production lot. Missing data in real-world datasets is another concern. The issue isn't just whether information is missing; it's what is missing and whether it could materially alter the conclusion.

Regulators may assess whether key endpoints are adequately captured, whether sufficient follow-up exists, whether adverse events are documented, and whether important variables are consistently available across the patient population. For example, if a study on device procedure outcomes shows significant data loss before the relevant follow-up period, the remaining patients might have excellent outcomes. However, understanding what happened to the lost patients is critical for accurate conclusions.