This study investigates the relationships between health uncertainty and food consumption among low-income households in developing countries. The particular sample households are from Lima, Peru's bottom expenditure quartile. Several food and nonfood inputs are important factors affecting health status and health variance. In addition, there is strong evidence that expected health status and health risk affect the consumption of several food commodities;The model also provides a means of analyzing the effects of policy alternatives on food consumption, expected health status, and health risk. The results show that education programs and price subsidies for tubers and dairy products are the most efficient means of increasing food consumption and improving health among poor households in Lima. In addition, education programs and tuber subsidies dramatically lower health risk suggesting that policymakers give these two programs high priority.