Standard incremental cost-effectiveness ratios (ICERs) calculated as the ratio of the difference in mean costs and mean effects at the group level may be biased in nonrandomized studies due to unbalanced characteristics of the two groups under study.