The demand for health care is estimated using household survey data from poor neighborhoods in New York City. A theoretical model is developed such that people pay for alternative types of medical care with money and time. Implications of the model are examined by performing Tobit regressions. The time needed to consume alternative types of care are explanatory variables as are earned and nonearned income and selected socio-demographic variables. Time plays a significant role in determining demand in these populations. This has important policy implications in light of the continued spread of private health insurance and the possible enactment of National Health Insurance. The effects of changes in clinic locations, waiting room policies, and substituting income maintenance for direct provision of care are examined for increasing access of the poor to health care.