Résumé : BACKGROUND:: The medical home (MH) model has prompted increasing attention given its potential to improve quality of care while reducing health expenditures. OBJECTIVES:: We compare overall and specific health care expenditures in Belgium, from the third-party payer perspective (compulsory social insurance), between patients treated at individual practices (IP) and at MHs. We compare the sociodemographic profile of MH and IP users. RESEARCH DESIGN:: This is a retrospective study using public insurance claims data. Generalized linear models estimate the impact on health expenditures of being treated at a MH versus IP, controlling for individual, and area-based sociodemographic characteristics. The choice of primary care setting is modeled using logistic regressions. SUBJECTS:: A random sample of 43,678 persons followed during the year 2004. MEASURES:: Third-party payer expenditures for primary care, secondary care consultations, pharmaceuticals, laboratory tests, acute and long-term inpatient care. RESULTS:: Overall third-party payer expenditures do not differ significantly between MH and IP users (&OV0556;+27). Third-party payer primary care expenditures are higher for MH than for IP users (&OV0556;+129), but this difference is offset by lower expenditures for secondary care consultations (&OV0556;-11), drugs (&OV0556;-40), laboratory tests (&OV0556;-5) and acute and long-term inpatient care (&OV0556;-53). MHs attract younger and more underprivileged populations. CONCLUSIONS:: MHs induce a shift in expenditures from secondary care, drugs, and laboratory tests to primary care, while treating a less economically favored population. Combined with positive results regarding quality, MH structures are a promising way to tackle the challenges of primary care. Copyright © 2013 by Lippincott Williams & Wilkins.