Depression is everybody’s business, with the World Health Organization estimating that 350 million people around the world are directly affected by it (WHO, 2015). It is a major cause of distress and disability, and for many is a recurring condition.

People who are depressed can also be helped. One approach is ‘collaborative care’. In this context, this refers to a closely specified way of managing long-term ill-health, developed over a period of years in the USA. It involves a team of health professionals working together in a structured way, with one assuming the role of care manager. This person proactively works with patients, keeps in close touch with primary care colleagues, and receives supervision from a specialist.

In a Cochrane Review of collaborative care for depression and anxiety (Archer et al, 2012), completed by members of the same research team as the new study highlighted in this blog, collaborative care was found to be associated with significant improvements in outcomes compared with usual care. A justification offered for this new study is that much of the research in this area has been conducted in the USA, meaning that evidence is needed to establish whether collaborative care for common mental health problems also works in other parts of the world, including the UK.

Methods

This is a large cluster RCT which investigated the clinical and cost effectiveness of collaborative care provided in primary care settings for people with moderate to severe depression. Cluster trials involve the randomisation of whole groups, rather than individuals, and here 51 primary care practices located in three parts of England were randomly allocated to one of the study’s two arms:

1. Collaborative care
2. Care and treatment as usual

Five hundred and eighty one adults who met ICD-10 criteria for being currently depressed took part, of whom 276 were allocated to the collaborative care arm. This is a number exceeding the target determined through the study’s power calculation.

In studies of this type which examine complex interventions involving lots of moving parts, considerable care is needed to specify what happens in the experimental arm and to make sure that critical ingredients are consistently applied. Collaborative care involved:

• Care management, encompassing 6-12 contacts between care managers and patients over a period of no more than 14 weeks. The norm for all but the first contact was via telephone. Contacts were designed to be structured, including:
  • Formal assessments of mood using the Hospital Anxiety and Depression Scale (HADS);
  • Help for patients to manage any prescribed antidepressant medication;
  • Direct support for patients with behavioural activation, a brief psychosocial intervention (Ekers et al, 2008) which aims to engage people in activities bringing improvements in mood.
• Care managers keeping in close contact with GPs, using a structured protocol.
• Care managers, all of whom had previously been prepared to work as paraprofessional mental health workers in primary care, receiving structured supervision from mental health specialists representing a number of different professional groups.

In contrast, ‘usual care’ for the 305 patients allocated to this arm was not pre-specified, but involved GPs continuing with whatever comprised their standard practice. This could involve (for example) prescribing antidepressants, and referring patients on for specialised assessment and ongoing care and treatment.

In complex intervention studies, of which this is an example, it is not possible to blind participants to the type of intervention they are receiving, unlike (for example) in trials of new medicines. It is possible, however, to take steps to ensure that researchers recruiting participants, and then gathering outcome data, do not know which arm participants will be (and then have been) allocated to. This was done in this study, along with care being taken to conceal from researchers during primary care practice recruitment the sequence through which practices were allocated to the different arms.

All clinical outcomes were assessed using self-report tools, the primary outcome selected being severity of depression at four months’ follow-up measured using the widely-used PHQ-9. Secondary outcomes included depression at 12 months, along with anxiety, overall and health-related quality of life and participant satisfaction variously measured at four and/or 12 month time points. Depression, anxiety and quality of life were also measured at 36 months.

Two other important components completed the study. First, costs were calculated using health economic techniques, involving estimates of the costs of participants’ use of health and social care services, the costs of providing collaborative care and participants’ quality-adjusted life years (QALYs). Second, a process evaluation was completed to assess factors which may have moderated participant outcomes, to investigate the mechanisms through which collaborative care may have exerted an effect, and to explore its implementation. This involved a number of elements, including calculations using participant self-report and demographic, socioeconomic and service use data, scrutiny of audio-recorded collaborative care sessions, and analyses of interviews with care managers, GPs and supervisors.

Results

The headline finding is that people with depression who then went on to receive collaborative care had lower estimated mean depression scores than that of people receiving usual care once adjustments for baseline depression scores had been made:

• At 4 months, scores were 1.33 PHQ-9 points lower (n=230; 95% CI 0.35 to 2.31, p=0.009)
• At 12 months, scores were 1.36 PHQ-9 points lower (n=275; 95% CI 0.07 to 2.64, p=0.04)

Here are some other, selected, findings from those also reported:

• At 4 (but not 12) months, quality of mental health but not physical health was significantly better for those in the collaborative care group, with no differences being found between groups for anxiety
• A key process evaluation finding is that outcomes for people in receipt of collaborative care were predicted by how much behavioural activation they undertook
• In terms of longer-term follow-up, at the 36 month point, differences between groups could no longer be detected
• Collaborative care was calculated to have cost £272.50 per person receiving it, with people in this group being significantly more satisfied with the care they had received than those receiving usual care
• Collaborative care offered gains in quality-adjusted life years, and in the health economics analysis was judged as affordable.

Conclusions

Here is the study team’s own conclusion, directly extracted from their report:

Collaborative care improves depression up to 12 months after initiation of the intervention, is preferred by patients over usual care, offers health gains at a relatively low cost, is cost-effective compared with usual care and is mediated by patient activation. Future work should test enhanced intervention content not collaborative care per se.

Strengths and limitations

This was a big, complex, study running over a period of years which set out to definitively improve the evidence base for the care and treatment of people with depression in primary care. It succeeded in its aim, producing important new knowledge of an effective, acceptable and affordable approach to organising services and providing interventions. Future NICE guidelines, we can safely assume, will make good use of the findings reported here.

Future lines of research are also opened up, with a view to refining the content of interventions (of which behavioural activation is one) which might be most helpfully incorporated within the organisational framework which collaborative care offers.

Promoting the widespread uptake of collaborative care over time presents a rather different challenge, likely to occupy the energies of managers, practitioners and researchers concerned with the workings of health systems, knowledge mobilisation and service improvement.

The trial was registered and the protocol published in advance, promoting transparency. The cluster design minimised the chances of the principles and practices of collaborative care seeping out to influence usual care. Throughout, care was taken to reduce biases:

• As already noted, researchers with the task of recruiting practices and participants were not told which of the two arms each would be allocated to
• During the collection of outcome data, researchers were also formally unaware of which arm each participant had been in
• Care managers’ fidelity to the collaborative care model was monitored and variances noted, and the views and experiences of practitioners relating to the implementation of collaborative care were sought in a process evaluation
• Public and patient involvement (PPI) over the life of the study was secured in a number of complementary ways, including through the addition of a PPI advisor as an investigator and trial management group member. Part of this person’s role was the review of all documents destined for use by service users.

In terms of limitations, as is the case in all studies of this type, participants will have known whether or not they were receiving collaborative care. This may have influenced self-reported data.

Summary

Collaborative care combining a multidisciplinary approach, structured care management and supervision from specialists is an effective, acceptable and affordable approach to the care of people with depression.

Conflicts of interest

BH reports knowing members of the CADET research team through shared professional activities, such as participating in the work of Mental Health Nurse Academics UK.

Links

Primary paper

Richards DA, Bower P, Chew-Graham C, Gask L, Lovell K, Cape J, et al. (2016) Clinical effectiveness and cost-effectiveness of collaborative care for depression in UK primary care (CADET): a cluster randomised controlled trial. Health Technol Assess 2016;20(14).

Other references

Archer J, Bower P, Gilbody S, Lovell K, Richards D, Gask L, Dickens C, Coventry P. (2012) Collaborative care for depression and anxiety problems. Cochrane Database of Systematic Reviews 2012, Issue 10. Art. No.: CD006525. DOI: 10.1002/14651858.CD006525.pub2.

Ekers, D., Richards, D. and Gilbody, S. (2008) A meta-analysis of randomized trials of behavioural treatment of depression. Psychological Medicine, 38, 5, 611-623. [PubMed abstract]

Richards DA, Hughes-Morley A, Hayes RA, Araya R, Barkham M, Bland JM, Bower P, Cape J, Chew-Graham CA, Gask L, Gilbody S, Green C, Kessler D, Lewis G, Lovell K, Manning C, Pilling S. (2009) Collaborative Depression Trial (CADET): multi-centre randomised controlled trial of collaborative care for depression – study protocol. BMC Health Services Research, 9, 188.

World Health Organization (2015) Depression, http://www.who.int/mediacentre/factsheets/fs369/en/

Photo credits

• daniel zimmer CC BY 2.0
• Highways Agency CC BY 2.0
• Anna and Michal CC BY 2.0
• Richard Clifford CC BY 2.0
• Chad Kainz CC BY 2.0