Geographic variation, physician characteristics, and diabetes care disparities in a metropolitan area, 2003-2004.

Quality improvement begins with quality measurement. Quality improvement continues by examining variation in patient standards of care. Diabetes is a complex chronic disease requiring aggressive care to improve and prolong life. Identifying and determining what factors explain care variation-including physician office location, physician characteristics, practice factors, and the translation of knowledge over time-is an area of important research. Determining physician practice characteristics' impact on these diabetes measures is an important step in understanding diabetes care in greater Kansas City. Aggregated diabetes HEDIS data from numerous private insurance plans from greater Kansas City, and associated practice factors from these same providers serves as a large and representative source of information to evaluate the impact of these factors on diabetes care. Using both multivariate and logistic methods, we find that the variation in care is largely explained by physician office location, but there is statistical explanatory significance for physician age on A1c testing rates. Also, nephropathy screening rates are positively related to whether a physician is a member of a group or a solo practice. The location of a practice has significant effects for diabetes care because physician office location approximates to some extent patient characteristics. This is not to say that physician practice factors are unimportant, rather that diabetes care is complex, and requires a fruitful interaction between physician and patient. Quantifying the interaction between an informed, activated patient and a prepared, proactive physician is difficult to observe and measure, and evaluating testable hypotheses about this interaction is correspondingly difficult.