Care Management Implementation and Patient Safety

Purpose: To (1) examine the association between the intensity of care management (CM) implementation in hospitals and hospital performance on 4 selected indicators of patient safety among hospitalized Medicare recipients and (2) assess the extent to which these relationships are moderated by hospital organizational and environmental context. Design and Methods: Primary data from the 1997 mailed survey of hospital quality improvement practices; secondary data from the Medicare Inpatient Database, American Hospital Association's Annual Survey of Hospitals, Area Resource File; and 2 proprietary data sets compiled by Solucient Inc. were used in the study. A cross-sectional analysis of 1784 community hospitals was performed to assess the relationship between CM implementation intensity and 4 hospital-level patient safety indicators. Two-stage instrumental variables estimation, in which predicted values (instruments) of 4 CM implementation intensity variables plus control (exogenous) variables were used to estimate hospital-level patient safety indicators. Results: Results of main effects provided inconsistent support to our hypotheses that greater implementation of CM would result in improved patient safety. For example, more extensive use of statistical and process management tools was positively associated with 3 of the 4 patient safety indicators. Similarly, hospitals reporting a greater number of clinical guidelines in use exhibited significantly worse values on 3 of the 4 patient safety indicators. However, results also indicate that forces external and internal to the hospital condition the association of particular CM activities and patient safety indicators: specifically data use, statistical and process management tool use, and organizational emphasis on continuous quality improvement. For example, we observed that hospitals' relative focus on improving systems and processes of care lead to better patient safety results as the financial position of the hospital improved. Similarly, hospitals with more extensive use of statistical and process management tools in high managed care penetration markets were more likely to experience fewer adverse patient safety incidents than hospitals using the same practices in markets with lower managed care penetration. Implications: It is unlikely that CM per se will improve patient safety for hospitalized patients without a commensurate fit with the financial, strategic, and market imperatives faced by the hospital. In designing approaches to CM, managers and physicians need to consider both the internal and external conditions necessary to support effective CM implementation.