OnCo
ideasIdea

A public rulebook for when an external or synthetic control arm is acceptable

Sometimes a trial cannot randomise, so the new drug is compared with past patients' records. Clear published rules on when that is allowed, and how it must be done, would replace case-by-case guesswork.

Regulators publish a binding checklist for externally controlled trials: pre-registration of the comparator cohort and analysis before unblinding; target-trial emulation framing; minimum data quality (endpoint ascertainment, line of therapy, index date); quantitative bias analysis and tipping-point reporting; and the settings where it is acceptable (rare disease, effect size large relative to plausible confounding, randomisation infeasible). Submissions meeting the rulebook receive predictable review.

Hypothesis
A published rulebook will increase the share of externally controlled oncology submissions that are accepted without a confirmatory-trial demand, while post-approval confirmatory results will not disagree with the external-control estimate more often than they do today for single-arm accelerated approvals.
Rationale
External controls are already used informally and inconsistently. Cases where the historical comparison misled (later overturned by randomised data) share identifiable features (immortal-time bias, mismatched lines of therapy, endpoint drift) that a rulebook can exclude.
What would test it
Retrospectively apply the draft rulebook to past single-arm approvals with later randomised confirmation; check whether rulebook-compliant cases had smaller estimate discrepancies. Then pilot prospectively for two years.
Maturity
early clinical
Who has to act
regulator
Cost to try
Small (under $1M)
Years to first evidence
2
Bottlenecks it attacks

Connected

6top

Pages like this

not linked directly; found by shared links