The Determinants of Care Quality in a Health Maintenance Organization:

In this paper we assess the determinants of care quality in a health maintenance organization. Our research relies on a novel source of information on care quality, the “errors” detected by an artificial intelligence program that scans a patient’s medical records for deviations from best practice. We examine the determinants of these “errors” for a population of 50,000 commercial insurance patients over the course of a year. We intend to use this data (together with other information about physicians, their incentive contracts and their patient populations) to answer three questions. First, do physicians with "expensive" practice styles generate more or fewer errors than other physicians? Second, do the cost-cutting incentives written into physician contracts have any effect on the propensity to generate errors. The third question concerns the role of more conventional quality metrics employed in the industry. At the time of the study, the HMO collected some of the commonly used quality metrics (largely HEDIS measures of preventative care) and used these to offer financial rewards to physicians for sustaining high levels of care quality. We ask whether there is a relationship between the dimensions of care quality rewarded by the HMO and the more comprehensive quality measures generated by the computer program.