Fig. 6: The estimated aHR and 95% CI on eICU and MIMIC, comparing pooled analysis, our FL-TTE and meta-analysis with fixed effects and random effects. | npj Digital Medicine

Fig. 6: The estimated aHR and 95% CI on eICU and MIMIC, comparing pooled analysis, our FL-TTE and meta-analysis with fixed effects and random effects.

From: Federated target trial emulation using distributed observational data for treatment effect estimation

Fig. 6

The third column on the right side is the p-value and significance level of the Z-test on whether the estimated aHR is significantly different with 1.0 (reference value indicating the treatment does not alter the risk compared to no treatment). The fourth and fifth columns denote the test statistic and p-value of the Z-test on whether the estimated aHR is significantly different from the results of pooled analysis. Our FL-TTE had less-biased estimates than the meta-analysis in three types of outcomes (28-day mortality, Time to ICU discharge, and Time to cessation of mechanical ventilation). *p < 0.05; **p < 0.01; ***p < 0.001; not significant (“ns”) with p ≥ 0.05.

Back to article page