Background: Randomized controlled trials are of limited value in studying the comparative effectiveness of the complex care changes that are needed to improve medical care quality, cost, and patient experience. They are of even less value in studying the implementation and spread process. DIAMOND Study exemplifies a rigorous observational design and unobtrusive subject recruitment for studying the impact and implementation of the DIAMOND Initiative. Methods: The Study was designed to answer the questions of the Initiative and in partnership with it through a jointly developed staggered implementation, multiple baseline design. Patients filling a new antidepressant prescription from a participating clinic were recruited by phone from data provided by all state health plans and surveyed at baseline and six months. Clinic leaders were surveyed about their change priority and capability, the extent of depression system implementation at baseline and 1 and 2 years post-implementation, and both implementation and operating costs. Results: Of 24,000 patient names submitted, 60% could be contacted in time, 66% of contacted agreed to be screened, 26% of screened were eligible, and 99.7% of eligibles consented to participate and have their health care utilization data audited. There were few differences between these groups. Of 553 surveys of physician and administrative leaders of each of 85 clinics, 550 completed surveys (99.4%) were obtained, showing variable priority and few depression systems pre-implementation. Medical group average costs of implementing DIAMOND were $45,000 and operational costs = $160/patient/month. Conclusions: Scientifically strong observational studies of natural experiments are possible if created and conducted in partnership.