Integration of mental health/substance abuse and primary care.

Objectives: To describe models of integrated care used in the United States, assess how integration of mental health services into primary care settings or primary health care into specialty outpatient settings impacts patient outcomes and describe barriers to sustainable programs, use of health information technology (IT), and reimbursement structures of integrated care programs within the United States.

Data sources: MEDLINE, CINAHL, Cochrane databases, and PsychINFO databases, the internet, and expert consultants for relevant trials and other literature that does not traditionally appear in peer reviewed journals.

Review methods: Randomized controlled trials and high quality quasi-experimental design studies were reviewed for integrated care model design components. For trials of mental health services in primary care settings, levels of integration codes were constructed and assigned for provider integration, integrated processes of care, and their interaction. Forest plots of patient symptom severity, treatment response, and remission were constructed to examine associations between level of integration and outcomes.

Results: Integrated care programs have been tested for depression, anxiety, at-risk alcohol, and ADHD in primary care settings and for alcohol disorders and persons with severe mental illness in specialty care settings. Although most interventions in either setting are effective, there is no discernible effect of integration level, processes of care, or combination, on patient outcomes for mental health services in primary care settings. Organizational and financial barriers persist to successfully implement sustainable integrated care programs. Health IT remains a mostly undocumented but promising tool. No reimbursement system has been subjected to experiment; no evidence exists as to which reimbursement system may most effectively support integrated care. Case studies will add to our understanding of their implementation and sustainability.

Conclusions: In general, integrated care achieved positive outcomes. However, it is not possible to distinguish the effects of increased attention to mental health problems from the effects of specific strategies, evidenced by the lack of correlation between measures of integration or a systematic approach to care processes and the various outcomes. Efforts to implement integrated care will have to address financial barriers. There is a reasonably strong body of evidence to encourage integrated care, at least for depression. Encouragement can include removing obstacles, creating incentives, or mandating integrated care. Encouragement will likely differ between fee-for-service care and managed care. However, without evidence for a clearly superior model, there is legitimate reason to worry about premature orthodoxy.