Enhanced management of depression can improve outcomes for patients when implemented through quality improvement resources increasingly available to community practices. Depressed patients in our intervention practices reported significantly milder symptoms of depression and had higher response and remission rates at follow up than patients in the usual care practices. The effect sizes are comparable to those of other trials on depression in primary care.22
We identified elements that possibly influence clinical outcomes. The intervention patients had more telephone contacts and more visits. Intervention patients reported that the clinicians were more likely to assess suicide risk, offer educational materials, and assist with self management goals. Counselling and adherence to antidepressants were similar between the groups, suggesting that specific but modest support for patients result in better outcomes for depression and higher patient ratings for quality of care.
Although clinical outcomes were enhanced, these effects were modest. At least four features of our study may help to explain this. Firstly, rather than identifying potentially ambivalent patients through screening, all patients were identified during routine care by the clinicians, had accepted the diagnosis of depression, and had agreed to be managed with drugs or by counselling. Secondly, the usual care clinicians performed well on process and outcome measures compared with those in other trials,22,23
creating a high standard of comparison. Thirdly, there was potential for attenuation of the intervention tasks compared with their direct implementation by researchers using a strict protocol.24
Finally, our intervention was modest compared with other recent large trials. Unlike our trial, IMPACT (improving mood-promoting access to collaborative treatment), PIC (partners in care), and PROSPECT (prevention of suicide in primary care elderly: collaborative trial) offered increased access to mental health services at no or reduced cost.20,23,25
Given these design features and dependence of the intervention on training manuals and existing quality improvement resources rather than research protocols and transient resources, the superior outcomes compared with usual care are noteworthy. We acknowledge that remission rates at six months indicate that most patients had at least partial symptoms. These patients continued to receive support and may have improved further, as found in longer term studies.20,23,26
Evaluation interviews in our study were completed by six months after patient enrolment in keeping with the commitment to provide the model to usual care practices in that time.
We suggest that by recruiting 60 practices in diverse locations, our trial extends the generalisability of earlier findings on efficacy of telephone support and primary care specialty cooperation beyond special settings. Many of these earlier studies took place in health maintenance organisations, academic sites, or veterans administration practices.27-31
Our trial adds to this knowledge by showing that modest resources delivered through established programmes can approach outcomes achieved by more research intense, resource rich interventions. These findings take on added importance in light of the negative result from the Hampshire depression project in which the intervention was based on clinician education but did not include telephone support or psychiatry for patients.32
In addition, some US government bodies have called for more intensive monitoring of patients prescribed antidepressants.33
Our model provides this, including assessment of suicidal thoughts.
The generalisability of our findings to primary care are constrained because the practices had access to established quality improvement programmes and care management staff, resources which are becoming more widely, but not yet universally, available. Attention to quality improvement is growing.8,9,34
Primary care trusts in the United Kingdom may be able to assume this function, and in some countries professional societies and regional health authorities could play a part.
Our study shows the feasibility of doing rigorous research in community practice settings relying on established resources to deliver the intervention rather than on transient research teams. Using this approach would enable future research in primary care to explore sustainability and dissemination of interventions. In addition, materials created to support these interventions could be more readily modified for application outside the research setting.
What is already known on this topic
Trials have shown improved outcomes for depression in primary care patients using models for the management of chronic illness
These models involve more systematic follow up and monitoring, patient telephone support, and cooperation between primary care and psychiatry
Implementation of these models has usually depended on research teams and has not been sustainable when research support ends
What this study adds
Evidence based models of depression management can be implemented in primary care with support from existing quality improvement resources
These models can improve outcomes for depression
Our model for depression produced significantly better outcomes and more favourable patient responses on quality of care than usual care. The model requires only modest changes from practices and creates a framework to study long term sustainability and dissemination of evidence based care.