Back to Case Studies
BayesianRare Disease

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.