Facing escalating health care expenditures, the governments of countries with national health insurance programs are trying to control or even to reduce health care utilization. Little research has examined the effects of decreased health care utilization on health outcomes. Applying a natural experiment design to the Taiwan population between 2000 and 2004, which includes the 2003 SARS epidemic when an average 20% decline in health care utilization occurred, this study examines the association between a decline in health care utilization and health outcomes measured by cause-specific mortality rates. We analyse the monthly mortality rates caused by infectious diseases, cancer, diabetes mellitus, nervous system diseases, cerebrovascular diseases, heart and other vascular diseases, respiratory system diseases, digestive system diseases, genitourinary system diseases and accidents. Models control for age, sex, month and year effects. Results show the heterogeneous effect of reduced health care utilization on health outcomes. Patients with diabetes mellitus or cerebrovascular diseases are vulnerable to short-term reductions in health care; compared with the non-SARS period, mortality caused by diabetes mellitus and cerebrovascular diseases significantly increased during the SARS epidemic by 8.4% and 6.2%, respectively. No significant change in mortality rates caused by the other diseases or accidents is found. This study suggests that governments of countries where health care utilization and spending are similar to or inferior to those in Taiwan should carefully evaluate the impact of policies that attempt to reduce health care utilization. Furthermore, when an area encounters an epidemic, governments should be aware of the negative consequences of voluntary restraints on access to health care that accompany decreases in utilization.