Validation Issues Could Dull Synthetic Data Benefits
Is it real, or is it synthetic? When it comes to drug development data in the future, the answer could be the latter. Datasets that replicate the statistical patterns of real patient populations but carry no risk of privacy disclosures could enable the use of synthetic control arms. However, the biopharma industry and regulators will need assurance that these datasets accurately reflect the population from which they were derived.
You may also be interested in...
Agency is going outside its ‘comfort zone’ to leverage different streams of existing real-world data sources to quickly assess the impact of potential treatment approaches, Amy Abernethy says; pandemic experience could lead to a more open embrace of RWE beyond the current public health emergency.
A randomized COVID-19 study spanning the whole of the UK could help to bolster the UK’s reputation as a location for future clinical trials. It is also being seen partly as the result of the UK regulator's "decisive action" in fast-tracking trials for potential COVID-19 therapies.
It may be easier to use observational study data to replicate randomized controlled trial results in certain therapeutic areas or where drugs were studied against each other instead of placebo, the US FDA’s David Martin says; the agency-funded DUPLICATE real-world evidence demonstration project has completed replication work on the first five trials, all of which are diabetes drug cardiovascular outcomes studies.