OBJECTIVE: The spread of evidence-based care is an important challenge in healthcare. We evaluated spread of an evidence-based large-scale multisite collaborative care model for patients with depression and diabetes and/or cardiovascular disease (COMPASS). METHODS: Primary care patients with depression and comorbid diabetes or cardiovascular disease were recruited. Collaborative care teams used care management tracking systems and systematic case reviews to track and intensify treatment for patients not improving. Targeted outcomes were depression remission and response (assessed with the Patient Health Questionnaire-9) and control of diabetes (assessed by HbA1c) and blood pressure. Patients and clinicians were surveyed about satisfaction with care. RESULTS: Eighteen care systems and 172 clinics enrolled 3609 patients across the US. Of those with uncontrolled disease at enrollment, 40% achieved depression remission or response, 23% glucose control and 58% blood pressure control during a mean follow-up of 11 months. There were large variations in outcomes across medical groups. Patients and clinicians were satisfied with COMPASS care. CONCLUSIONS: COMPASS was successfully spread across diverse care systems and demonstrated improved outcomes for complex patients with previously uncontrolled chronic disease. Future large-scale implementation projects should create robust processes to identify and reduce expected variation in implementation to consistently provide improved care.