Caring for depression in the dying is complex and challenging – survey of palliative physicians

Background

Depression is a distressing condition for people with advanced life-limiting illnesses. It can reduce the quality-of-life of those affected and others around them, exacerbate physical suffering and worsen psycho-existential distresses [1–4]. Not only does depression impact patient engagement with their nearest supporters, but depression can also negatively affect clinicians’ ability to deliver care [1, 5]. Despite its prevalence, there is evidence that depression has been under-assessed and, even when recognised, under-managed in the palliative care setting [6–9].

In the palliative care population, there is a sub-group of patients with very poor prognoses defined as an estimated life expectancy in the range of days to weeks. This sub-group is characterised by a high degree of frailty, often with significant symptom burden and rapidly declining functional status [10, 11]. The frailty, symptom burden (e.g. fatigue, confusion, and dysphagia), and limited time for interventions to take effect can make depression assessment, psychotherapies and administration of typical antidepressants (e.g. Selective Serotonin Reuptake Inhibitors [SSRIs] and Serotonin Noradrenaline Reuptake Inhibitors [SNRIs]) challenging for clinicians [12–14]. Subsequently, clinicians’ approaches to depression assessment and management for these people might differ from palliative patients with better prognoses.

While previous studies of clinicians’ approaches to depression assessment and management in the general palliative care population have been done in Australia and the United Kingdom, palliative physicians’ and psychiatrists’ approaches to depression care specifically for people with very poor prognoses have not been explored [8, 9, 15].

Methods

Results

Completed surveys were obtained from 110 individuals: 79 responses out of 522 members of ANZSPM (15.1%); and 31 out of 6655 RANZCP members (0.5%). Of the 110 responses, 72 respondents identified as having the primary specialty of palliative medicine and 32 with psychiatry. Due to the lack of response from the RANZCP members and hence the lack of representation of the Australasian psychiatry cohort, only results from those who identified themselves primarily as palliative physicians (n = 72) were reported (Table 1).

Participating clinicians were mainly specialist and fellows (73.6%); female (75.0%); aged 31–60-year-old (87.4%); primarily working in Australia (76.4%); graduated more than 10 years ago (88.9%); and working ≥20 clinical hours per week (90.2%). Most clinicians (n = 70; 97.2%) reported having encountered depression in people with very poor prognoses.

The majority (n = 42; 58.3%) of all palliative physicians reported that they screen for depression in general palliative care patients, while only 40.0% (n = 28 out of 70) of clinicians encountering patients with very poor prognoses reported screening for depression.

Among physicians who might screen for depression (answered “yes” or “depends”) in general palliative care patients, the primary screening method reported was clinical interview (n = 53; 93.0%), followed by asking the family/carers (n = 40; 70.2%), asking other health professionals involved in the care (n = 37; 64.9%), and the use of screening tools (n = 27; 47.4%). For the very poor prognoses group, while 68.6% (n = 48) of physicians reported no difference in the way of screening compared to the general palliative population, 18.6% (n = 13) reported a “difference”: taking a more reactive rather than proactive approach; being briefer in assessment; relying more on objective information sources; and emphasising less on somatic symptoms. Among those who reported to use screening tools (n = 27), the most commonly used tools was the ultra-short two-items questionnaire (n = 14; 51.8%) followed by a single-item questionnaire (n = 5; 18.5%) (e.g. asking “Are you depressed” and/or “Have you had little interest or pleasure in doing things”). Only one respondent reported to use Hospital Anxiety and Depression Scale.

For depression assessment, at least 80% of physicians would ascertain whether the depression episode is first or recurrent during assessment, regardless of whether the prognoses is very poor or not. All physicians who have encountered depressed patients with very poor prognoses have experienced uncertainty regarding the cause of depression. Most palliative physicians (n = 56; 80.0%) would treat the depressed mood despite the uncertain cause. The primary sources of assistance sought by palliative physicians in this context were from psychiatry (n = 33; 47.1%) and psychology (n = 29; 41.4%).

For depression somatic symptom assessment, the majority (n = 37; 51.4%) of physicians reported including somatic symptoms in the general palliative care patients while excluding somatic symptoms in the sub-group with very poor prognoses (n = 29; 41.4%). Notably, in the setting of very poor prognoses, 30.0% (n = 21) of physicians reported “depends”: whether the somatic symptoms could be attributable to the nature of the terminal illnesses and associated interventions on an individual basis; and that somatic symptoms were still valuable to be considered in the “overall picture” of the patient.

For various treatment approaches for major depressive disorder in the setting of very poor prognoses (Table 2), most physicians reported using non-pharmacological approaches (n = 64; 91.4%), followed by the use of typical antidepressants (n = 63; 90%). When comparing the likelihood of using various depression interventions in the very poor prognoses sub-group as compared to the general palliative care cohort, the majority of physicians reported: no difference or less likely in using non-pharmacological interventions (both groups: n = 26; 37.1%), and less likely to use typical antidepressants (n = 36; 51.4%). For ECT in the setting of very poor prognosis, 72.9% (n = 51) and 14.3% (n = 10) of physicians reported to not use or less likely to use it respectively. There were bimodal distributions with the highest prevalence of “I don’t use” followed by “more likely to use” for treatment options of: atypical antipsychotics (n = 26; 37.1% - “I don’t use” and n = 20; 28.6% - “more likely”); benzodiazepines (n = 28; 40.0% - “I don’t use” and n = 24; 34.3% - “more likely”); and novel medication/experimental trials (n = 49; 70% - “I don’t use” and n = 12; 17.1% - “more likely”). Due to technical issues in the online survey platform, the psychostimulant item was initially not available for the first 28 participants, leading to the large proportion of non-response (n = 27; 38.6%) for this item. Despite this limitation, among the responders, the majorities answered “I don’t use” or “more likely to use” (both groups n = 18 out of 43; 41.9%).

For service linkage with psychiatry (Table 3), the majority of palliative physicians reported to request for psychiatry input in an interval of monthly or longer (n = 41; 56.9%) and being requested by psychiatry for palliative care input yearly or longer (n = 26; 36.1%). Two-thirds of the palliative physicians (n = 48) thought contract frequency with psychiatry should be more frequent.

Sixty-six respondents (91.7%) provided answers to the open-ended questions regarding key challenges or barriers to effective assessment and management of depression in palliative care patients with very poor prognoses. Respondents commented on the complexity of the clinical situation with interaction between physical, psychosocial, and spiritual dimensions. Reported key challenges and barriers are listed in Table 4, categorised under the domains of patient, clinician, health system, literature and society. On quantifying the various domain subcategories (Table 4), the three most frequently reported barriers were: the lack of therapeutic options that are rapidly effective (77.3%); the perceived frailty, burden and intolerance of depression assessment and management on the patient (71.2%); and the complexity in differentiating the symptoms of terminal illness from the somatic symptoms of depression (53.0%).

Discussion

This is the first study that captures palliative physicians’ practices and perceptions regarding depression care specifically in people with very poor prognoses of only days to weeks. As demonstrated by the survey, encountering depression in patients with very poor prognoses was common to palliative physician. However, despite the high prevalence of depression (up to 50%) in this population and the frequency of clinical encounters, only 40% of clinicians reported to screen for depression, with all clinicians reported to have experienced uncertainty when assessing the cause of depression [26]. This is reflected by the current study finding of the perceived challenging complexity of depression care in the very poor prognosis setting by clinicians. According to the literature, this complexity may be contributed to by the interplay of various domains of challenges reported in Table 4: 1) Patients’ frailty, co-existing symptom burden and associated end-of-life issues when time for intervention effects is poor [9, 14]; 2) Clinicians’ self-perceived limitations of psychiatry skills in the palliative care setting and incompetence in diagnostic differentiation [9, 27]; 3) Health system’s inadequacy of resources and access to required interventions in the local health services (e.g. mental health services) [8, 9]; 4) Heterogeneity of depression concept and the lack of evidence to guide practice in the literature for this context [26, 28]; and 5) Unsupportive societal attitudes that prevents the optimisation of depression care (e.g. stigma of mental illnesses, the “normalisation” or “acceptance” of depression at the end-of-life) [29, 30]. Each of these domains warrant future exploration for potential solutions to better optimise depression care in this setting.

Palliative physicians reported to less likely screen for depression and have ambivalence in depression assessment methods (e.g. approach to somatic symptoms of depression) in the very poor prognosis setting compared to the better prognosis setting. Diagnosing depression in the setting of very poor prognosis can be challenging as the symptoms of terminal illnesses (e.g. fatigue and weight loss) can confound the somatic symptoms of depression [17]. Importantly, this study shows that while clinicians may perceive somatic symptoms of depression to be less useful in depression diagnosis, somatic symptoms are still important to be considered during the overall depression assessment as they can affect the appropriateness of intervention choices. It may be desirable for clinicians to be trained with the various approaches to somatic symptoms such as Endicott Criteria to enable better diagnostic differentiation and depression assessment [31]. While they reported to generally intervene less in this setting (compared to patients with better prognoses), it is worth noting the bimodal distributions of clinicians not-using and more-likely-to-use certain non-typical pharmacological interventions (e.g. psychostimulants, atypical antipsychotics, benzodiazepines and novel medications such as ketamine) that have more augmentation and rapid-onset potentials than typical antidepressants [18–20, 32]. This may reflect clinicians’ attitudes where clinicians who were trained and aware of how to leverage the potential benefits of these non-typical treatments while minimising intolerance were more likely to embrace their use. Whereas, clinicians who lacked training or resources for these treatments did not tend to use them. Comparable to the study findings in the United Kingdom primary care and palliative settings, inadequately equipped clinicians may have a nihilistic attitude and ambivalence towards depression screening and assessment [8, 29, 30]. The low reported usage of ECT was likely related to clinicians perceiving the intervention to be over-burdensome for people with very poor prognoses [33]. Subsequently, palliative physicians and their multidisciplinary team members should be trained with the necessary skills to screen, assess, and administer first-line rapidly effective depression interventions in low-burden manners [34]. This may be facilitated by better linkage and integration of the psychiatry services into the palliative care services [17, 35].

Similar to the United States palliative physician cohort, near 70% of current survey’s respondents expressed desires for better collaboration with the psychiatry services [36]. At the clinical and health service levels, some strategies to improve palliative care and psychiatry collaboration might include: integrative multidisciplinary team [15, 36–38]; joint development of a tiered-referral model tailored to local health service needs [39]; and integrated clinician training via workshops and experiential training [35, 37, 38]. For research, palliative care and psychiatry researchers must collaborate to address barriers to the currently limited evidence base. On top of the barriers to depression care processes identified in this survey, other challenges include the effects of depression and terminal illnesses on participants’ ability to consent and engage with research activities, and the ethical concerns of trial participants receiving potentially ineffective therapies [40, 41]. There is a need for integrated palliative care and psychiatry research that explores appropriate depression screening and assessment strategies and potentially rapidly effective interventions using feasible and inclusive trial designs in the very poor prognosis setting (e.g. n-of-1, Bayesian response-adaptive-randomisation, or well-designed prospective case-controlled studies) [26, 35, 42]. Developing consensus approaches between palliative care and psychiatry via Delphi and updating the existing guideline based on the currently limited evidence to guide depression care specifically for people with very poor prognoses need to be considered [16, 17]. Overall, better collaboration between palliative care and psychiatry is urgently required, optimising timely access to needed interventions, complementing the shortfalls of both disciplines, and ultimately improving care to affected patients [35, 37, 43].

Conclusions

Palliative physicians perceived depression care in people with very poor prognoses to be complex and challenging. The lack of screening, heterogeneity in the depression assessment, and the generally reduced likelihood of intervening for depression in the very poor prognosis setting compared to that of better prognosis highlighted the need for better collaboration between palliative medicine and psychiatry in health service delivery, clinician training, and research.

Supplementary Information

Authors’ contributions

Under the supervision of DC, MA, and BD, WL designed the survey and collected the data. WL analysed the data with statistical support of SC and qualitative analysis support of MD. WL drafted the manuscript and DC, MA, BD, MD and SC provided critical revisions throughout. The author(s) read and approved the final manuscript.

Funding

The survey was funded by the Translational Cancer Research Network Clinical PhD Scholarship Top-up award, supported by the Cancer Institute New South Wales, Australia.

Availability of data and materials

The datasets used and analysed during the current study are available from the corresponding author on reasonable request.

Declarations