Once the best measures have been selected, there are multiple challenges to implementing measures in routine practice, particularly considering the wide range of populations and the multiple levels of the health system at which measurement needs to be useful. In this session, Patrick Remington, associate dean for public health at the University of Wisconsin School of Medicine and Public Health, spoke about how to measure the health status of the U.S. population on a county-by-county basis. Stefan Gildemeister, director of the Health Economics Program and state health economist in the Minnesota Department of Health, addressed the measurement issue from a state health reform perspective. Carole Romm, who led, in a consulting capacity, the accountability and quality efforts for the Oregon Health Authority’s Medicaid program, described a measurement framework for coordinated care in another state effort.
To begin his presentation, Patrick Remington described the simple logic model that informs his work on measuring health status. This logic model assumes that programs and policies impact health factors—what used to be called health determinants—and that changes in health factors, over the long run, impact health outcomes. Health outcomes are measured using measures of morbidity and mortality, while health factors consist of health behaviors, clinical care, social and economic factors, and the physical environment. Remington then described the metrics that he and his colleague David Kindig chose to measure each of these inputs and the weighting that each metric is given in the final calculation (see Figure 6-1). He noted that one of the keys to developing this model was deciding on a set of measures and acknowledging that the final list is not perfect but could be sufficient. He also remarked that an important factor in selecting metrics is that the data are readily accessible and preferably available at no cost from every one of the more than 3,000 counties in the nation. Remington commented that together, these metrics provide a good assessment of the three-part aim.