This article examines associations of socio-demographic and health-care indicators, and the statistic 'mortality amenable to health care' (amenable mortality) across the US states. There is over two-fold variation in amenable mortality, strongly associated with the percentages of state populations that are poor or black. Controlling for poverty and race with bi- and multi-variate analyses, several indicators of health system performance, such as hospital readmission rates and preventive care for diabetics, are significantly associated with amenable mortality. A significant crude association of 'uninsurance' and amenable mortality rates is no longer statistically significant when poverty and race are controlled. Overall, there appear to be opportunities for states to focus on specific modifiable health system performance indicators. Comparative rates of amenable mortality should be useful for estimating potential gains in population health from delivering more timely and effective care and for tracking the health outcomes of efforts to improve health system performance.