Is care really shared? A systematic review of collaborative care (shared care) interventions for adult cancer patients with depression

A diagnosis of cancer impacts individuals’ psychological and physical wellbeing. Prevalence estimates of major depression (16%), minor depression and dysthymia (22%) in cancer patients are higher than in the general population. [1, 2] In palliative settings, the prevalence of depression approaches 49%. [3] Inadequate treatment of depression results in poorer adherence to anti-cancer treatments, decreased tolerance of cancer treatment side-effects, higher use of health care resources including increased hospital re-admissions, may adversely impact on interpersonal relationships and reduced overall survival. [3‐9] Co-existing depression therefore poses a significant burden for patients, families and the health system. Cognitive behavioural therapy (CBT) is effective in treating depression [10] with a meta-analysis (n = 198 studies, 22, 238 patients) reporting medium to large effect sizes sustained 6–12 months post intervention. A recent meta-analysis also confirmed the effectiveness of antidepressant medication in the treatment of depression [11] both in combination with psychological therapy and as primary treatment. [12] To guide evidence-based treatment, there has been a renewed effort to promote routine distress screening as a first step to improve detection and hence treatment for depression in cancer care. [13] Such strategies have included the development of a clinical pathway for identification and management of depression in adults with cancer, an international first. [14]

Despite increased evidence about effective treatments, many patients still do not seek treatment for their depression. For instance, a meta-analysis of 53 studies (n = 12,052) found in a research context less than 60% of distressed cancer patients engage in psychological treatment. [15] Uptake in routine care is even lower. [16] At a system level access to treatment is constrained by a shortfall in the psycho-oncology workforce resulting in long waiting lists and geographical disparities in access. [17‐19] Practical constraints such as transport, inconvenience and cost have also been suggested. [20, 21] Normalisation of distress and attribution of somatic depression symptoms to cancer by clinicians also means depression goes untreated and patients’ attitudes to mental illness are also likely to play a role. [22] Patient factors such as negative attitudes to mental health and stigma also reduce patient willingness to access treatment. [23]. To address perceived barriers, models of care that encompass systematic identification of depression in cancer patients, reduce dependence on specialist input and enable timely access to evidence-based treatment have been proposed.

The collaborative care model, based on the principles of chronic disease management, has been successfully implemented to treat medical conditions including depression. [24, 25] Core components of collaborative care for depression are: i) a multi-professional approach to patient care, ii) a structured management plan tailored to depression symptom severity, iii) scheduled patient follow-ups and iv) enhanced inter-professional communication. [26] The model endorses a multi-disciplinary approach involving joint participation of primary and specialist care physicians in planned delivery of care over and above routine discharge and referral. [23] A key aspect of effective collaborative care is case management [27] in which a member of the clinical team works closely with the main treating physician and monitors patient progress including adherence to psychological and pharmacological treatments, initiating treatment changes as necessary. [28] A meta-analysis of collaborative care interventions (n = 37 studies, 12,355 patients with depression receiving primary care) found that interventions that included a mental health professional as a care provider within the clinical team reported the largest effect sizes. [29]

The evidence base for collaborative care for depression among patients with cancer is rapidly developing. A recent meta-analysis of studies (n = 8) purporting to be collaborative care interventions concluded that the interventions were significantly more effective than usual care (standardized mean difference = − 0.49, p = 0.003), with remission rates for depression higher in the intervention groups at 12 months. While promising, the collaborative care interventions identified in that meta-analysis varied in content, intensity, and number of components, thus making it difficult to determine the relative contribution of each component to overall effectiveness. [30] Prior to wider implementation of the model, greater understanding of the determinants essential to model success for the treatment of depression in oncology is required. The aim of this systematic review was to determine the fidelity of the depression collaborative care models trialled in oncology to the recommended collaborative care criteria. Specifically, the review sought to identify reported intervention components, delivery models and role responsibilities.

Collaborative care interventions have the potential to improve management of depression for people with cancer. However, before wider implementation, there are several important questions that need to be addressed, including clearer definition of the elements of collaborative care (including roles, communication processes, governance), which collaborative care components are essential to treatment outcomes and who is best placed to deliver care. We sought to explore these questions by systematically describing the components of depression collaborative care interventions. Eight collaborative care RCTs were identified in this review. Five interventions were developed specifically for patients with cancer and three interventions were adapted from a more general collaborative care intervention for management of depression in older patients.

All studies identified in our review highlighted the benefit of systematic, intensive depression treatment and ongoing symptom monitoring, and reported decreased depression symptoms in comparison to usual care. However, in contrast to collaborative care interventions more generally, the review found the interventions were primarily delivered in the hospital setting rather than in primary care. While the hospital-based care model makes sense for those receiving ongoing cancer treatment, this model does little to address the preference of patients for care closer to their home and issues of access, particularly for those patients not in active treatment. [43] Similarly, the continued dependence on hospital specialists for the management of patients does not ameliorate the scarcity of psycho-oncology resources available in many cancer services.

In the literature, collaborative care is used to describe a range of interventions of varying intensity from simple interventions to encourage compliance with medication (typically delivered via telephone) to complex multi-component interventions that incorporate psychological and pharmacological treatments and intensive monitoring and follow up. [29] The studies identified in this review were consistent with this heterogenous depiction of collaborative care, ranging from symptom monitoring and patient navigation, anti-depressant adherence models to interventions involving proactive follow-up by care co-ordinators working closely with psychiatrists integrating pharmacological medical and psychological treatments as part of individualized treatment plans. As a result, treatment dose within the interventions identified varied in level of intensity as well as whether patients were able to access psychological and pharmacological treatment. Importantly, a number of collaborative care models were reliant on anti-depressant medication as the primary treatment. This is of concern given current gold standard depression treatment recognises the importance of combined cognitive behavioural therapy and anti-depressant medication, [44] and the fact that patients may be reluctant to take antidepressant medication. [45] Despite this, all studies documented clear pathways to patient management and follow up protocols to ensure patients had access to appropriate levels of care.

Despite the focus of hospital-based care, the single study in our review that provided care in the primary care setting, [38] demonstrated the acceptability among people diagnosed with cancer to attend primary care appointments for depression management. Other studies such as the Sharpe study [40] purported to provide the option for patients to choose depression care in either the cancer centre or primary care, however in truth this was limited to anti-depressant prescribing rather than overall psychological treatment and care coordination, and the authors did not evaluate differences in outcome based on whether GPs or oncologists were responsible for anti-depressant management. Lack of GP engagement across studies is surprising given the increasing evidence GPs can play an important role in collaborative care models in terms of continuity of care and reduced patient burden due to excessive travel. From a health service perspective, inclusion of GPs will in part address some of the psycho-oncology workforce shortages that mean that many patients do not receive depression treatment in a timely manner. Under this model of care, less complex patients’ pharmacological treatment can be managed in the community, with psychiatry treating only those patients with persistent symptoms.

A key feature of collaborative care models is a team-based approach to delivery of care. Previous reviews have highlighted the importance of a multi-disciplinary approach to care, with inclusion of case managers with mental health training reported to increase the efficacy of the model. [29] Our review found in the context of cancer, care co-ordination was primarily undertaken by nurses trained to deliver the intervention. However, given their limited specialist mental health training and consistent with the findings of previous work conducted by our Group that nurses are inadequately equipped to deliver depression care, [46] there was a high level of psychiatry case management incorporated into all studies, with psychiatrists providing both staff supervision and treatment decision-making. The time commitment required for ongoing psychiatry oversight limits long term sustainability of the proposed model in routine care as it is predicated on sufficient psychiatry staff to provide the high level of supervision. Modifications of the model to include other (non-medical) mental health staff such as clinical psychologists, such as demonstrated in the study by Steel [34] in the delivery of psychological treatment and supervision of care co-ordinators in a stepped care model, with psychiatric review reserved for those patients in need of more intensive psychiatric care, is one way to reduce is one way to reduce the need for routine psychiatry oversight.

Additionally, over-reliance on specialist psychiatrist involvement for treatment decision-making further underscores the perception among oncology clinicians that the mental wellbeing of patients is not the responsibility of the cancer team and is further reinforced when oncologists are informed of their patient’s depression diagnosis but not engaged in any delivery of depression care. Similarly, although primary care physicians provided pharmacological treatment to their patients as part of the intervention, we were unable to identify any studies that incorporated specific training for GPs about prescribing and follow up. Upskilling of non-mental health providers in evidence-based management of depression is essential for a truly collaborative model to be sustainable. However, in the studies identified in this review, communication pathways between hospital-based care providers and GPs failed to engage GPs in shared decision-making as much of the communication was via written reports or telephone calls initiated by the care co-ordinator (nurse). This resulted in passive engagement with treatment decision-making whereby GPs implemented recommendations for individual patients rather than establishing ongoing communication where treatment decisions were jointly shared between care providers, taking into account not only the patient’s cancer and current depression episode, but each patient’s wider medical and social history available from their primary care provider. Attention to clinical governance and management of shared care is necessary to achieving positive patient outcomes. The funding models and organisation of primary care and specialist care vary internationally and influence the governance of such models. This context necessitates closer examination of the processes and ingredients of findings regarding collaborative care when evaluating transferability to diverse settings. Evidence from “shared care” for primary mental disorders indicates that substantial cross-organisational commitment is required, with consistently delivered interventions, and with close attention to staff selection, training and supervision. [47] Implementation and longterm sustainability of shared care models is contingent on funding models that support clinical pathways that incorporate primary and tertiary care, the introduction of information technology to facilitate information sharing, and collaborative interdisciplinary practice models.

The results of this review need to be considered in light of a number of limitations. Firstly, we were unable to conduct the planned meta-analysis to compare between intervention components due to the small number of studies identified. The review however does extend the recent meta-analysis conducted by Li and colleagues [30] that demonstrated overall efficacy of the model, as we were able to qualitatively explore the components of each intervention and compare the roles of health professionals across interventions. Our analysis of intervention components was however limited by the intervention descriptions detailed in study publications; the authors may have used a more detailed intervention protocol when conducting their study and explicit references to previously published standardised protocols were included in assessment of interventions. Similarly, we were unable to determine the quality of training provided as part of the interventions or assess the quality of care provided by each member of the care team. Generalisability of review results are also limited to interventions published in English. All studies identified also excluded patients with co-morbid psychiatric and substance use disorders, limiting generalisability of the study results given the high prevalence of co-occurring mental health conditions in the general population. Despite these limitations, this review highlights that, collaborative care interventions for depression in oncology do vary according to key criteria such as level of inter-professional engagement and communication, inclusion of psychological and pharmacological treatment options and documented follow up.

In the context of research, collaborative care interventions for the management of depression in cancer have demonstrated efficacy over usual care. However, the sustainability of the level of hospital staff engagement raises questions about the model’s utility in routine care. Current models of care reflect hospital-based multi-disciplinary models of care. Greater engagement with oncology and primary care is required to reduce the over-reliance on specialist psychiatry services. Models that utilise existing community-based services, including clinical psychologists in partnership with GPs have the potential to enhance depression care. Given the aim of collaborative models is to provide more care in the community, future models need to consider the training needs of primary care providers and there needs to be greater emphasis placed on inter-professional communication and shared decision-making.

Future research is required to explore patient and provider acceptability as well as pragmatic non-inferiority trial designs that incorporate process evaluations to identify implementation strategies and practice change required to move these models of care from research into clinical practice.