This paper uses the Nepal Living Standards Survey, a nationally representative sample of households from 1996, to investigate the determinants of household out-of-pocket health expenditures. The analysis uses a multi-equation joint estimation to control for endogeneity of sickness and provider choice. The results of this analysis indicate several interesting findings. First, common unobserved factors were found to be statistically significant determinants of illness, choice of provider, and health expenditures, and may cause bias to parameter estimates if not controlled. Second, the income elasticity is estimated to be 1.10, with income having both a direct effect on health expenditure, and an indirect effect through likelihood of illness and the type of provider that is chosen. Third, housing and sanitary conditions were found to have a substantial effect on illness, and as a result, out-of-pocket health care expenditures. Fourth, despite the fact that urban, ill individuals who seek care are more likely to utilize care in more expensive settings, average health care expenditure among the urban sample was found to be substantially lower than among the rural sample, partly due to a lower likelihood of reporting illnesses and injuries and of using any type of health care provider. Copyright © 2002 John Wiley & Sons, Ltd.