Parametric estimation of medical care costs under conditions of censoring

This paper is concerned with a set of parametric estimators that attempt to provide consistent estimates of average medical care costs under conditions of censoring. The main finding is that incorporation of the inverse of the probability of an individual not being censored in the estimating equations is instrumental in deriving unbiased cost estimates. The success of the approach is dependent on the amount of available information on the cost history process. The value of this information increases as the degree of censoring increases.