Estimation of medical costs by copula models with dynamic change of health status

► A semiparametric model is developed for more accurate estimation of medical costs. ► A dynamic change of health status via Markov transition is taken into account. ► A copula is used to model the relation between a sojourn and its associated cost. ► A multivariate local likelihood method is employed to estimate model parameters. ► The asymptotic properties of the estimators are established.