Five hundred data sets are generated under each of the nine scenarios and the results are displayed in Figure 3 and Table 1, which, respectively, summarize the root mean square error (RMSE) of the estimated regression parameter and the coverage and widths of the 95% uncertainty intervals.
The uncertainty metrics including percentage of coverage and average width were used to evaluate the precision of the modeling approaches.
LMEM showed the best performance with regard to coverage and median width.
When randomization of twins was to opposite treatment arms, ordinary logistic regression and GEE performed similarly with respect to coverage and median width.
using simulations in a large empirical diagnostic study, the six approaches were compared in terms of coverage and the width of the 95% confidence intervals.
When there was no within-birth correlation, ANOVA and LMEM performed similarly with regard to coverage and median width, regardless of the randomization method, sample size, proportion of twins, and effect size.
When within-birth correlation was present and randomization of twins was to the same arm, GEE tended to be a compromise between ANOVA and LMEM with regard to coverage and median width.
For continuous outcomes, using LMEM is recommended in comparison to either ANOVA or GEE, taking into consideration the following operating characteristics: median bias, mean squared error, coverage, and median width.
For continuous outcomes, while the coverage never fell below 0.93, and the type I error rate never exceeded 0.07 for any method, overall linear mixed-effects models performed well with respect to median bias, mean squared error, coverage, and median width.
