Bayesian Decision Framework
Duchenne muscular dystrophy (DMD) gene therapy with 6-minute walk distance endpoint
78%
Probability of Benefit
BTD
FDA Breakthrough
18 mo
Earlier Patient Access
The Challenge
A biotech company was developing a gene therapy for Duchenne muscular dystrophy (DMD), a rare genetic disorder affecting approximately 1 in 5,000 male births. The Phase II trial enrolled 30 patients with the 6-minute walk distance (6MWD) as the primary endpoint.
The challenge: DMD trials face inherent limitations:
- •Small patient populations make classical power calculations unrealistic
- •Natural disease progression creates high variability
- •Binary p-value thresholds may lead to incorrect decisions
The Decision Point
Traditional Approach
Decision Framework
Binary p-value threshold (p < 0.025 = go)
Observed Result
p = 0.08
Conclusion
“Not significant” → No-go
Bayesian Approach
Decision Framework
Probability of clinically meaningful benefit
Calculated Result
P(Δ > 30m | data) = 78%
Conclusion
Go with quantified confidence
Bayesian Analysis
Predictive Probability Calculation
Using a Normal-Normal conjugate prior model with a weakly informative prior based on natural history data, the Bayesian analysis computed the posterior probability that the true treatment effect exceeds the minimum clinically important difference (MCID) of 30 meters on the 6MWD.
Observed Treatment Effect
Δ = 42.3 meters (95% CI: -5.1, 89.7)
Posterior Probability
P(Δ > 30m | data) = 78%
Interpretation
78% probability that the true treatment effect exceeds the MCID
Key Insight
The traditional frequentist analysis (p = 0.08) would have led to a “not significant” conclusion and potentially terminated a promising program. The Bayesian analysis provided a more nuanced interpretation:
- 78% probability of clinically meaningful benefit—a strong signal warranting continued development
- 22% probability of insufficient benefit—quantified risk for decision-makers
- Transparent uncertainty—regulators could see exactly what was known and unknown
Regulatory Outcome
FDA Breakthrough Therapy Designation Granted
FDA acknowledged the Bayesian analysis's transparency and the unmet medical need. The quantitative probability assessment enabled informed regulatory decision-making.
Accelerated Approval Granted
Based on the Bayesian probability assessment and surrogate endpoint data, FDA granted accelerated approval, pending confirmatory trial results.
Patient Access 18 Months Earlier
Compared to traditional development pathway requiring additional Phase III confirmation before any approval decision.
Value of Bayesian Approach
For Sponsors
- •Quantitative risk assessment for investment decisions
- •Avoid false no-go decisions on promising therapies
- •Clear communication with investors and partners
For Regulators
- •Transparent uncertainty quantification
- •Benefit-risk assessment with full probability distribution
- •Appropriate for rare diseases with small samples
Regulatory Support
“Bayesian methodologies help address two of the biggest problems of drug development: high costs and long timelines.”
— FDA Commissioner Marty Makary, on draft guidance extending Bayesian methodology to drugs and biologics (January 12, 2026)
Calculate Bayesian predictive probability
Compute the probability of trial success given your observed data.