Objective: To assess the effect of possession of private health insurance on hospital use and outcomes in Western Australia.Method: Hospital morbidity records were extracted from the Western Australian (WA) Data Linkage System for all 22 major diagnostic categories (MDCs) for the period 1994–99, with follow-up to the end of 2000. Multivariate modelling techniques were used to estimate the effect of possession of private health insurance on hospital admission rates, average and total length of stay (LOS), cumulative incidence of admission at 30 days and one year, and case fatality at one year.Results: Possession of private health insurance had significant effects on hospital use and outcomes, even after adjustments for age, sex, aboriginality, socioeconomic status, location and comorbidity. Non-insured patients tended to have a higher overall hospital admission rate but a lower admission rate for surgical episodes, and they generally had a longer LOS although this difference was only greater than a day in three MDCs. Case fatality was higher in non-insured patients, but there was no systematic trend with regard to readmission rates.Conclusions: The study found that for all MDCs, other than those where treatment was required on an emergency basis, patients with private health insurance had improved access to surgical procedures. Either non-insured patients were disadvantaged in their access to surgery or the higher intervention rate in privately insured patients represented supplier-/consumer-induced demand which may not always have been to the patient's advantage or both may have occurred.