Variation in quality of care within health systems

Although many hospitals belong to health care systems, little is known about the quality of care provided by those systems, or whether characteristics of health care systems are related to the quality of care patients receive. Dimensions of the quality of care provided in 73 hospital systems were examined using hospital quality data publicly reported by the Centers for Medicare & Medicaid Services (CMS). The hospital systems consisted of six or more acute care hospitals and represented 1,510 hospitals. The study was designed to determine whether these dimensions of system quality could be reliably measured, to describe how systems varied with respect to quality of care, and to explore system characteristics potentially related to care quality.

Data were made available by CMS for 19 indicators of care quality for pneumonia, surgical infection prevention, acute myocardial infarction (AMI), and congestive heart failure.

At the system level, reliable measures (alphas > .70) were constructed for each of the four clinical areas, and these measures were combined into a single measure of quality (alpha = .85). Variability in system quality was substantial, ranging from 94% to 70% on the combined quality measure. On the clinical area measures, the smallest range was for AMI (99%-85%), whereas the largest was for surgical infection prevention (95%-54%). System ownership and system centralization were significant predictors of quality, accounting for 30% of variance in the combined quality measure. Geographic region, inclusion of teaching hospitals, and system size were unrelated to quality.

Systems vary greatly in terms of quality of care in each of the four clinical areas, with for-profit and more decentralized systems appreciably lower in quality of care. System-level quality measures and data could be used to compare processes within systems and to drive improvement efforts.