In outcome studies, healthcare costs and health status are two focuses of primary interest. These two variables interact with each other: change in health status results in healthcare utilization and healthcare utilization has impact on health status. This interaction necessities a statistical simultaneous modeling. Ignoring such endogeneity could lead to inappropriate conclusions. Often outcomes data from clinical trials consist of panel data in which outcomes are repeatedly measured over time. The nature of panel data complicates statistical modeling to control for this endogeneity bias. OBJECTIVE: To present a statistical model and a estimation procedure that are most suitable to tackling the endogeneity between healthcare costs and health status by using panel data from clinical trials. METHOD: A simultaneous error-component model is established as: <1> <2> where The dependent variable and are the health expenditure and health status of individual i (i=1,2, . . ., N) at time t (t=1, . . ., T); is the vector of the common exogenous variables in both the cost and health equations; is the vector of explanatory variables included in the cost equation but not the health equation; is the vector of explanatory variables included in the health equation but not the cost equation. ?’s are the unknown parameters; and and are the unobserved disturbances which consist of , the component from an individual unit, , the component from a time period, , the independently identically distributed random disturbance. Given the structure of the disturbances in the model, we propose the error-component three-stage least-squares method to estimate this model. IMPLICATION: This paper provides an example of modeling costs and health status simultaneously using panel data from clinical trials.