OBJECTIVES: The objectives were to obtain inpatient admission cost distributions of cervical cancer by stage of the disease and to identify significant predictors of cost. METHODS: 1,559 inpatient admission records for the period 1994-1997 were obtained from MedStat s private insurance databases. Costs were measured by total payments. Probability densities of cost were calculated nonparametrically for each stage. Regression analyses of cost per stage were conducted using both nonparametric and ordinary least squares (OLS) procedures. RESULTS: The 95% quantile of cost ranged from $12,748 for LSIL (low-grade squamous intraepithelial lesions) to $44,505 for Stage IV cervical cancer. The OLS results show that Health Maintenance Organizations (HMO) and Point of Service (POS) plan types have significantly lower costs than the comprehensive coverage plan (p <0.05) for all five stages of cervical cancer. The largest difference was seen in Stage IV with savings of $10,964 for HMO and $11,544 for POS both compared to the comprehensive coverage plan. Nonparametric regression analysis confirms this finding of statistically significant lower costs for HMO and POS for all stages except HMO in Stage 2/3 and POS in Stage IV. CONCLUSIONS: It is not clear if and to what extent the cost advantages of HMO and POS relative to the comprehensive coverage plan compromise the quality of care. The information from the study may be useful for predicting future costs of treatment for cervical cancer. This may also help policy makers in the economic evaluation of cervical cancer screening programs by providing benchmark values of treatment cost savings made possible because of earlier detection of cervical cancer.