PHS91 - Do Medical Homes Affect Healthcare Utilization Rates? A Longitudinal Evaluation on Medical Home Implementation in Italy

Abstract

This study compares healthcare utilization rates between patient populations served by general practitioners (GPs) affiliated with newly established medical homes (MHs) and those not affiliated with MHs. This longitudinal study of administrative healthcare data from the Local Health Authority of Parma, Emilia-Romagna, Italy included residents ≥14 years residing in Parma ≥1 year prior to entry in the study. During the study period (1/1/2010-12/31/2016), 16 MHs were established at various times, with the first one opening in December 2011. The exposure of interest for any patient at any time was assignment to a GP affiliated with an MH. Utilization measures evaluated were Emergency Department (ED) visits and hospital admissions, including ambulatory care sensitive conditions (ACSCs). The association between MH healthcare and each outcome was evaluated through a series of Cox proportional hazards models with a time-dependent MH status. To adjust for the strong influence on healthcare utilization rates of patients treated by a GP with a disposition to join an MH, the models were adjusted for GP type (i.e., GP who joined an MH during the study period vs. GP who had not). 412,394 patients contributed exposure time to non-MH healthcare and 124,801 contributed to a MH healthcare (median follow-up time 4.5 years and 3.9 years, respectively). Most patients (65.4%) were followed for the entire 6-year period. Compared to patients not exposed to MH healthcare, those exposed had a lower risk (HR; 95% CI) of ED visits (0.96; 0.95-0.97) and ordinary hospital admissions (0.98; 0.96-0.99). No difference was observed among admissions for ACSCs. These findings suggest MH healthcare is associated with improvements in healthcare utilization rates. Additional healthcare utilization outcomes should be evaluated and further longitudinal analyses, including adjustments for other potential confounders, should be conducted as more MHs are implemented and additional years’ data become available.