（图2）对于从医院活着出院的时间，HR为0.91，与通常剂量蛋白组相比，高剂量蛋白组HR<1（危害）的概率为92%，在信息量弱、怀疑性早的模型中。
Background:The EFFORT Protein trial assessed the effect of high vs usual dosing of protein in adult ICU patients with organ failure. This study provides a probabilistic interpretation and evaluates heterogeneity in treatment effects (HTE).
Methods:We analysed 60-day all-cause mortality and time to discharge alive from hospital using Bayesian models with weakly informative priors. HTE on mortality was assessed according to disease severity (Sequential Organ Failure Assessment [SOFA] score), acute kidney injury, and serum creatinine values at baseline.
Results:The absolute difference in mortality was 2.5% points (95% credible interval -6.9 to 12.4), with a 72% posterior probability of harm associated with high protein treatment. For time to discharge alive from hospital, the hazard ratio was 0.91 (95% credible interval 0.80 to 1.04) with a 92% probability of harm for the high-dose protein group compared with the usual-dose protein group. There were 97% and 95% probabilities of positive interactions between the high protein intervention and serum creatinine and SOFA score at randomisation, respectively. Specifically, there was a potentially relatively higher mortality of high protein doses with higher baseline serum creatinine or SOFA scores.
Conclusions:We found moderate to high probabilities of harm with high protein doses compared with usual protein in ICU patients for the primary and secondary outcomes. We found suggestions of heterogeneity in treatment effects with worse outcomes in participants randomised to high protein doses with renal dysfunction or acute kidney injury and greater illness severity at baseline.
