Skip to main content

Table 2 Combined operating characteristics of the design including a frequentist efficacy evaluation and the continuous Bayesian safety decision rule

From: Accelerating clinical development of HIV vaccine strategies: methodological challenges and considerations in constructing an optimised multi-arm phase I/II trial design

Simulation scenario

Proportion of outcomes in 10,000 simulations of one trial arm (%)

Safety hypothesis

Immunogenicity hypothesis

Vaccinations stopped (unsafe)

Vaccinations not stopped (safe), inefficacious at final

Vaccinations not stopped (safe), efficacious at final

Overall probability of erroneous conclusion

Unsafe (0.70)

Inefficacious (0.50)

95.5

4.2

0.3

4.5

Unsafe (0.70)

Efficacious (0.80)

95.6

0.3

4.1

4.7

Safe (0.95)

Inefficacious (0.50)

4.8

91.4

3.8

8.6

Safe (0.95)

Efficacious (0.80)

5.3

6.8

87.9

12.1

  1. Simulation methods: Simulation of independent binomial distributions for safety and efficacy, respectively. Simulation of participant outcomes for one trial arm with continuous Bayesian monitoring of the safety endpoint, repeated in 10,000 trial simulations per simulation scenario.
  2. Bayesian continuous monitoring of the safety endpoint. Decision rule for safety outcome: Vaccine considered unsafe at the first analysis where the posterior probability that the vaccine is below the target level (Psafe <0.95) exceeds 95%. Enthusiastic prior: beta (6,0.3).
  3. Fixed-sample frequentist analysis of immunogenicity endpoint with 23 participants: conclusion that vaccine strategy is efficacious at final analysis if lower bound of one-sided exact 95% confidence interval of observed proportion >0.5. Proportion of trials with erroneous outcomes shown in bold. The overall risk of erroneous conclusion corresponds to the sum of erroneous trial outcomes.