OBJECTIVES:   In a simulated dataset, evaluate incremental cost-effectiveness ratios (ICERs) adjusted for covariates and missing data using three different regression models.  The regression parameter of interest is the incremental net monetary benefit (INMB).  Models are ANCOVA, mixed effects (ME), and joint mixed effects and log time-to-dropout (joint ME), a selection model. METHODS:   Traditional cost-effectiveness analysis (CEA) uses the incremental cost-effectiveness ratio (ICER), a measure with statistical issues and limitations with missing data.  Regression analysis can estimate incremental net monetary benefit (INMB) and avoid these statistical issues while adjusting for covariates as well as missing data.  The cost-effectiveness acceptability curve (CEAC) generated from a family of these regressions can identify an ICER adjusted for the factors included in the INMB regressions (the ICER is the point on the CEAC where the probability of being cost-effective is 50%).  Data were simulated to include missing at random (MAR) and missing not at random (MNAR).  Simulated treatment effect provided a “true” INMB for model evaluations that included bias (absolute difference from “true”), precision (ratio of variances), and CEACs with willingness-to-pay (λ) values from $0 to $100K. RESULTS:   The ANCOVA and ME models produced the least biased estimates.  At λ = $50K, bias was $1.3K, $1.4K, and 2.3K, and precision was 1.27, 0.90, and 1.24 for ME, ANCOVA, and joint ME, respectively.  The joint ME model performed best when missingness was high. CONCLUSIONS:   Once the CEACs had been generated, deriving ICERs adjusted for covariates and missing data from those CEACs based upon INMB regressions proved easy and feasible.  The models used in this simulation analysis performed differently under alternative missingness conditions and were sensitive to nonresponse mechanisms.  All estimates were poor when missingness was high; suggesting prevention of missing data should be a goal of research.