Table 3.
Multiple Imputationa Results Using Rubin's Methodb |
SII-NVX-CoV2373 (N = 340) | NVX-CoV2373 (N = 110) |
---|---|---|
Estimated Mean | ||
GMEUc | 139937.78 | 153200.54 |
95% Confidence Interval | (116839.75, 167602.05) | (124050.99, 189199.68) |
Estimated Ratio | ||
GMEU Ratioc | 0.91 | |
95% Confidence Interval | (0.79, 1.06) |
Multiple Imputation model with classification variables vaccine, age group, sex, comorbidity status and continuous covariates log baseline titer used to impute 50 values for each missing value.
Rubin's method in PROC MIANALYZE used to pool estimates and SE across the 50 multiply imputed datasets.
Pooled ANCOVA results, LS Mean and it's 95% CI values by treatment were used to generate the GMEU and 95% CI and the differences in LS Means and corresponding 95% CI limits were used to obtain GMEU Ratio and 95% CI using back transforming to the original scale.