OBJECTIVES: This paper applies the inverse probability weighted (IPW) least-squares method to estimate the effects of treatment on total medical cost, subject to censoring, in a panel-data setting. Two analyses were performed to examine how patient-and treatment related variables explain total medical costs for older persons newly diagnosed with lung cancer. METHODS: IPW pooled ordinary-least squares (POLS) and IPW random effects (RE) models are used. Because total medical cost might not be independent of survival time under administrative censoring, unweighted POLS and RE can not be used with censored data, to assess the effects of certain explanatory variables. Even under the violation of this independency IPW estimation gives consistent asymptotic normal coefficients with easily computable standard errors. A traditional and robust form of the Hausman test can be used to compare weighted and unweighted least squares estimators. RESULTS: The methods are applied to a sample of 201 Medicare beneficiaries diagnosed with lung cancer between 1994 and 1997. Regional stage decreased total cost of care almost 68% according to IPW POLS and 41% according to IPW RE compared to in situ or local stage cancer. A person who received radiation only decreased the total medical cost relative to mean cost for surgery plus adjuvant therapies. The estimates with respect to IPW POLS and IPW RE are 72% and 49%. The Hausman Tests, in comparison between POLS and IPW POLS, RE and IPW RE models suggest that there exist no bias due to censoring. CONCLUSION: Measurement of treatment cost is especially important in the evaluation of medical intervention, in the analysis of clinical trials and in social experiment. Currently, statistical methods that are applicable to administrative data-which is often censored- are under developed. We offered a model which solves possible selection bias due to censoring.