Background: Methodological differences between evaluations of medical home adoption might complicate readers' ability to draw conclusions across studies.
Objectives: To study whether associations between medical home adoption and patient care are affected by methodological choices.
Design, setting, and subjects: Among 71 practices participating in the Pennsylvania Chronic Care Initiative (a medical home pilot), we estimated cross-sectional and longitudinal associations between 4 definitions of "medical home adoption" [National Committee on Quality Assurance (NCQA) recognition in year 3, Medical Home Index scores at baseline and 3, and within-practice changes in Medical Home Index scores between baseline and year 3] and utilization and quality.
Measurements: Six utilization and 6 quality measures.
Results: In cross-sectional analyses at year 3, NCQA recognition was associated with higher rates of nephropathy monitoring (7.23 percentage points; confidence interval, 0.45-14.02), breast cancer screening (7.48; 2.11-12.86), and colorectal cancer screening (8.43; 2.44-14.42). In longitudinal analyses, NCQA recognition was associated with increases in hospitalization rates (2.75 per 1000 patient-months; 0.52-4.98). In baseline cross-sectional analyses, higher Medical Home Index scores were associated with fewer ambulatory care-sensitive hospitalizations (-0.61 per 1000 patient per month; -1.11 to -0.11), all-cause emergency department visits (-6.80; -12.28 to -1.32), and ambulatory care-sensitive emergency department visits (-5.60; 10.32 to -0.88). There were no statistically significant associations between any other measure of medical home adoption and quality or utilization.
Conclusions: The findings of medical home evaluations are sensitive to methodological choices. Meta-analyses, narrative reviews, and other syntheses of medical home studies should consider subdividing their findings by analytic approach.