Background: Widespread restructuring of health delivery systems is underway in the United States to reduce costs and improve the quality of health care.
Objective: To describe studies evaluating the impact of system-level interventions (incentives and delivery structures) on the value of US health care, defined as the balance between quality and cost.
Research design: We identified articles in PubMed (2003 to July 2014) using keywords identified through an iterative process, with reference and author tracking. We searched tables of contents of relevant journals from August 2014 through 11 August 2015 to update our sample.
Subjects: We included prospective or retrospective studies of system-level changes, with a control, reporting both quality and either cost or utilization of resources.
Measures: Data about study design, study quality, and outcomes was extracted by one reviewer and checked by a second.
Results: Thirty reports of 28 interventions were included. Interventions included patient-centered medical home implementations (n=12), pay-for-performance programs (n=10), and mixed interventions (n=6); no other intervention types were identified. Most reports (n=19) described both cost and utilization outcomes. Quality, cost, and utilization outcomes varied widely; many improvements were small and process outcomes predominated. Improved value (improved quality with stable or lower cost/utilization or stable quality with lower cost/utilization) was seen in 23 reports; 1 showed decreased value, and 6 showed unchanged, unclear, or mixed results.Study limitations included variability among specific endpoints reported, inconsistent methodologies, and lack of full adjustment in some observational trials. Lack of standardized MeSH terms was also a challenge in the search.
Conclusions: On balance, the literature suggests that health system reforms can improve value. However, this finding is tempered by the varying outcomes evaluated across studies with little documented improvement in outcome quality measures. Standardized measures of value would facilitate assessment of the impact of interventions across studies and better estimates of the broad impact of system change.