OnCo
ideasIdea

Power trials to detect a benefit patients would value, not the smallest detectable one

A trial can be designed to detect a tiny improvement that is statistically real but too small to matter. Protocols should state up front what size of benefit would be worth having, and be built to detect that.

Protocols pre-declare a minimal clinically important benefit using the ESMO-MCBS or ASCO Value Framework thresholds (e.g., a hazard ratio and absolute gain in median OS or PFS), justify the sample size against it, and report whether the observed effect met it. Regulators and HTA bodies use the pre-declared threshold in review, discouraging trials sized to detect trivial differences.

Hypothesis
Requiring a pre-declared meaningful benefit will reduce the number of approvals with marginal absolute gains and will shift sample sizes and endpoints toward those that capture patient-relevant differences.
Rationale
Analyses of approved cancer drugs show many with small absolute survival gains; sample-size inflation makes trivial effects significant. Pre-declaration is the standard in non-inferiority trials and could be symmetric.
What would test it
Audit recent approvals against ESMO-MCBS grades; pilot pre-declaration in cooperative-group protocols and track whether the observed effects meet it.
Maturity
speculative
Who has to act
regulator
Cost to try
Small (under $1M)
Years to first evidence
2
Bottlenecks it attacks
  • Trial design, endpoints and cost · A phase 3 trial takes years and hundreds of millions of dollars, and often answers a question that has already moved on.
  • Prices and value · New cancer drugs routinely cost over $150,000 a year, often for months of benefit. Systems cannot afford them and patients go bankrupt.

Key papers

1top

Connected

9top

Pages like this

not linked directly; found by shared links