Canadian regulators laid out guidance that they hope will help sponsors use real-world evidence to speed life-saving drugs to those who need them most.

Health Canada is eager to bring real-word evidence into drug reviews. But it points out that data can be compromised by bias or poor study design. For now, it recommends using either prospective or retrospective data.

Prospective data collection requires similar rigor to a clinical trial, with “transparency and clearly defined procedures” essential, the agency says. Retrospective data collection needs those elements but also some of its own.

Researchers should understand the limits of looking back on previously collected data and try to mitigate the risks of poor quality. “All data should be traceable to the source, and source data should be available whenever possible,” the document states. “Curated data will be preferred in all instances.”

Data elements should be well-defined to “identify and assess exposures, outcomes, endpoints and covariates of interest, and validation of these study aspects should be included,” the agency says.