A Difference-In-Difference Analysis of Changes in Quality, Utilization and Cost Following the Colorado Multi-Payer Patient-Centered Medical Home Pilot
Published in: Journal of General Internal Medicine, 2015
Posted on RAND.org on November 20, 2015
BACKGROUND: Research on the effects of patient-centered medical homes on quality and cost of care is mixed, so further study is needed to understand how and in what contexts they are effective. OBJECTIVE: We aimed to evaluate effects of a multi-payer pilot promoting patient-centered medical home implementation in 15 small and medium-sized primary care groups in Colorado. DESIGN: We conducted difference-in-difference analyses, comparing changes in utilization, costs, and quality between patients attributed to pilot and non-pilot practices. PARTICIPANTS: Approximately 98,000 patients attributed to 15 pilot and 66 comparison practices 2 years before and 3 years after the pilot launch. MAIN MEASURES: Healthcare Effectiveness Data and Information Set (HEDIS) derived measures of diabetes care, cancer screening, utilization, and costs to payers. KEY RESULTS: At the end of two years, we found a statistically significant reduction in emergency department use by 1.4 visits per 1000 member months, or approximately 7.9 % (p = 0.02). At the end of three years, pilot practices sustained this difference with 1.6 fewer emergency department visits per 1000 member months, or a 9.3 % reduction from baseline (p = 0.01). Emergency department costs were lower in the pilot practices after two (13.9 % reduction, p < 0.001) and three years (11.8 % reduction, p = 0.001). After three years, compared to control practices, primary care visits in the pilot practices decreased significantly (1.5 % reduction, p = 0.02). The pilot was associated with increased cervical cancer screening after two (12.5 % increase, p < 0.001) and three years (9.0 % increase, p < 0.001), but lower rates of HbA1c testing in patients with diabetes (0.7 % reduction at three years, p = 0.03) and colon cancer screening (21.1 % and 18.1 % at two and three years, respectively, p < 0.001). For patients with two or more comorbidities, similar patterns of association were found, except that there was also a reduction in ambulatory care sensitive inpatient admissions (10.3 %; p = 0.05). CONCLUSION: Our findings suggest that a multi-payer, patient-centered medical home initiative that provides financial and technical support to participating practices can produce sustained reductions in utilization with mixed results on process measures of quality.