Limitation of Real World Evidence Highlighted by Physician Survey
​[Posted on: Thursday, January 12, 2017]

Data collected from doctors, medical records, insurance claims databases and patient registries regarding use of medical products and disease conditions, called “Real World Evidence”, is often claimed to be reliable, robust and credible enough to support regulatory decisions. The 21st Century Cures Act requires FDA to create a process to use Real World Evidence (RWE) in support of market approval of products. However, FDA brass has raised concerns about the quality of such data collected in an uncontrolled manner with inherent biases and gaps. A recent physician survey by Medscape highlights many concerns about physician bias and preferences based on personal opinions when reporting data about medical practice. Physicians make prescription decisions based on the insurance coverage available for a given patient, budgetary considerations of their organization, and relationships with the manufacturers. Significant numbers of physicians alter the description of the disease to meet insurance coverage requirements, and do not disclose errors. The patient pools for a given practice is also influenced by the physician’s preference regarding the type of patients they feel comfortable treating. Similar findings were made by European regulators for a similar program to evaluate the quality of RWE in support of market approvals. While in theory RWE present a very interesting and appealing way to generate data on medical products, without a robust prospective plan, such data has major limitations. The RWE data is unquestionably very useful for meeting post-marketing commitments and identifying safety signals, but for use in market approval or label revisions to add new uses via the 505(b)(2) pathway, prospective design of RWE collection process and robust tools to evaluate the gaps and biases in the available information are critical to get regulatory blessing. Use of RWE is both easy and tricky. Easy because it is readily available and with electronic health records becoming universal over the few years, it is the low hanging fruit for the industry. But it is tricky due to the several limitations in the real life collection of data. The silver lining is that with FDA being pushed to define the process under with RWE can be used; we should soon see guidance from the regulators about the expected elements and potential applications.