Collaborative chronic care models (CCMs) are effective in serious mental illnesses, which has been shown in extensive randomized clinical trials. Much of their effectiveness comes from the emphasis on flexibility: They are implemented according to local needs, capabilities, and priorities. Collaborative chronic care models also provide support: for redesigned work roles that promote “anticipatory” continuous care, for self-management, and for clinical decision making at a local level.
In 2013, the VA Office of Mental Health and Suicide Prevention (OMHSP) began an initiative to enhance care coordination in general mental health clinics with mixed-diagnosis populations. It established interdisciplinary teams in each VA medical center throughout the US. Although providing centrally developed guidance, VAOMHSP gave facilities “broad latitude” to develop their team processes. In 2015, VAOMHSP adopted the CCM.
But most of the data on how well CCMs work for mental health conditions come from depression treatment in primary care—and the effects seem to be inconsistent. So researchers from Veterans Affairs Boston Healthcare System and others partnered with VAOMHSP to find out whether the CCM model would be effective in a general clinical setting.
They recruited 9 VA facilities for a 2-year study conducted in 3 waves. The implementation strategy was based on the premise that “health care is a complex adaptive system rather than a highly controlled machine.” That is, it would work best if local solutions for local challenges could be developed in accordance with evidence-based guidance. The multifaceted approach included an external facilitator who provided guidance and quality improvement expertise and an on-site internal facilitator to direct the implementation.
Outcomes aside, patients deserve the conversation
In the study, 5,596 veterans treated by outpatient general mental health teams were included in hospitalization analyses. A randomly selected sample of 1,050 (including 210 women) was identified for health status interviews.
The researchers found a “robust” and sustained reduction in mental health hospitalization. However, the effects on self-reported health outcomes were “limited,” the researchers say. The mental component score (the primary intervention outcome) did not change statistically significantly with implementation support in adjusted or unadjusted models, nor did other interview measures. The researchers say they saw no difference in the way veterans were treated between higher and lower implementing teams.
In post hoc analyses, though, patients with more complex problems, defined as receiving treatment for ≥ 3 mental health diagnoses in the previous year, did show statistically significant improvements in the facilitation year (by a magnitude of 0.31 SD). By contrast, those with ≤ 2 diagnoses declined nonsignificantly during the same time. The researchers note that other studies have found that CCM-based teams in patient-centered medical homes have also shown more benefit among higher morbidity patients.
Overall, the model was shown to be effectively implemented with “practical, scalable support” for clinicians. Another benefit was that teams performed better, the researchers found. They assessed team function at baseline and during the second 6 months on measures, including communication, cohesion, role clarity, and team primacy (prioritizing team over individual goals). The subscales showed high ratings for cohesion and communication at baseline, which did not change with implementation support. However, role clarity and team primacy improved significantly. The researchers conclude that under typical practice conditions CCMs can help the clinicians help the sickest patients.