Table 2.
Bayesian meta-analysis of endotracheal intubation and mortality outcomes
| Priors* | Empiric priors | Hypothetical priors | |||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|
| Mean (95% CrI) | Pr(RR <1) | Pr(RR <0.95) | Pr(RR >1) | Pr(RR >1.05) | Mean (95%CrI) | Pr(RR <1) | Pr(RR <0.95) | Pr(RR >1) | Pr(RR >1.05) | ||
| Intubation | |||||||||||
| Non-informative/neutral | 0.83 (0.70 to 0.97) | 0.99 | 0.96 | 0.01 | 0.00 | 0.84 (0.72 to 0.97) | 0.99 | 0.95 | 0.01 | 0.00 | |
| Enthusiastic | 0.76 (0.68 to 0.85) | 1.00 | 1.00 | 0.00 | 0.00 | 0.82 (0.70 to 0.94) | 1.00 | 0.98 | 0.00 | 0.00 | |
| Skeptical: minimal | 0.82 (0.74 to 0.98) | 1.00 | 1.00 | 0.00 | 0.00 | 0.83 (0.71 to 0.96) | 0.99 | 0.97 | 0.01 | 0.00 | |
| Skeptical: moderate | 0.89 (0.81 to 0.99) | 0.96 | 0.90 | 0.01 | 0.00 | 0.84 (0.73 to 0.98) | 0.99 | 0.95 | 0.01 | 0.00 | |
| Mortality | |||||||||||
| Non-informative/neutral | 0.90 (0.73 to 1.13) | 0.83 | 0.68 | 0.17 | 0.08 | 0.91 (0.75 to 1.11) | 0.83 | 0.68 | 0.17 | 0.08 | |
| Enthusiastic | 0.83 (0.69 to 0.99) | 0.98 | 0.93 | 0.02 | 0.01 | 0.87 (0.71 to 1.05) | 0.93 | 0.82 | 0.07 | 0.03 | |
| Skeptical: minimal | 0.91 (0.79 to 1.04) | 0.87 | 0.73 | 0.09 | 0.02 | 0.92 (0.78 to 1.11) | 0.81 | 0.62 | 0.19 | 0.08 | |
| Skeptical: moderate | 0.96 (0.84 to 1.09) | 0.68 | 0.47 | 0.25 | 0.09 | 0.96 (0.85 to 1.08) | 0.74 | 0.44 | 0.26 | 0.08 | |
Crl=credible interval; Pr(RR)=posterior probability of relative risk.
Informative mean effect priors for the intubation outcome were: Enthusiastic 0.71 (95% CrI 0.61 to 0.82), skeptical: minimal 0.83 (0.72 to 0.95), skeptical: moderate 0.94 (0.82 to 1.07). Informative mean effect priors for the mortality outcome were: Enthusiastic 0.64 (95% CrI 0.44 to 0.92), skeptical: minimal 0.92 (0.76 to 1.10), skeptical: moderate 1.0 (0.84 to 1.20). See supplemental eTable 1 for expanded rationale and justification for mean effect priors. Informative between study variance priors were selected as previously described25 and include intubation (log normal distribution 3.93, 1.912) and mortality (log normal distribution −4.17, 1.552).