INTRODUCTION
Diagnosing depression in the palliative care context has been associated with challenges that include difficulties in distinguishing depression from sadness and advanced disease, the stigma of psychiatric diagnoses, and clinician-related factors such as the belief that depression is an expected part of dying, insufficient skills and therapeutic nihilism (Block, Reference Block2000; Pessin et al., Reference Pessin, Olden and Jacobson2005; Noorani & Montagnini, Reference Noorani and Montagnini2007). A perhaps less emphasized challenge relates to the conceptual ambiguity of depression, which has been complicated by the numerous definitions and classifications that have been applied to depression over time, in reflection of the prevailing psychiatric and social paradigms (Kendell, Reference Kendell1976; Farmer & McGuffin, Reference Farmer and McGuffin1989; Boyce & Hadzi-Pavlovic, Reference Boyce, Hadzi-Pavlovic, Parker and Hadzi-Pavlovic1996; McPherson & Armstrong, Reference McPherson and Armstrong2006; Misbach & Stam, Reference Misbach and Stam2006; Paykel, Reference Paykel2008; Jansson, Reference Jansson2011). Although the concept of major depression, arising from the advent of the third edition of the Diagnostic and Statistical Manual of Mental Disorders (DSM–III) in 1980 (American Psychiatric Association, 1980) is widely used in clinical practice and research (McPherson & Armstrong, Reference McPherson and Armstrong2006), its conceptual foundation has been criticized and validity challenged (Kendler & Gardner, Reference Kendler and Gardner1998; van Praag, Reference van Praag2000; Parker, Reference Parker2005; Galatzer-Levy & Galatzer-Levy, Reference Galatzer-Levy and Galatzer-Levy2007; Horwitz & Wakefield, Reference Horwitz and Wakefield2007; Mulder, Reference Mulder2008; Jacob, Reference Jacob2009).
Research has identified that medical practitioners experience difficulties in reconciling the biomedical and psychosocial models of understanding depression (Thomas-MacLean & Stoppard, Reference Thomas-MacLean and Stoppard2004; Schumann et al., Reference Schumann, Schneider and Kantert2012). Similar difficulties have been identified in the palliative care setting, with previous research showing two distinct ways in which palliative medicine specialists conceptualize depression, as pathology and as contextual phenomena (Ng et al., Reference Ng, Crawford and Chur-Hansen2014a ). The tension between these concepts is clinically important, as they may signify different assessment and treatment approaches (Ng et al., Reference Ng, Crawford and Chur-Hansen2014b ).
In the absence of somatic diagnostic markers, assessment by psychiatrists has often been referred to as the “gold standard” for diagnosing depression (Golden et al., Reference Golden, McCartney and Haggerty1991; Power et al., Reference Power, Kelly and Gilsenan1993; Stiefel et al., Reference Stiefel, Die Trill and Berney2001; Jefford et al., Reference Jefford, Mileshkin and Richards2004), although psychiatrists only assess a minority of patients who may be depressed. Understanding how psychiatrists conceptualize depression may be informative and of relevance when compared with other practitioners in palliative medicine. As there are no published empirical studies on this subject, the present research aimed to investigate this employing a qualitative approach. The research question was “What are psychiatrists' conceptualizations of depression in the palliative care setting?”
METHODS
Participants and Recruitment
Our study targeted consultant psychiatrists who practiced in Australia and had clinical involvement in the palliative care setting. Recruitment took place through the Royal Australian and New Zealand College of Psychiatrists (RANZCP), which claims approximately 85% of practicing psychiatrists in Australia as its members (RANZCP, 2013). A recruitment notice was placed in an RANZCP electronic bulletin that was distributed to its members. An email was also sent via RANZCP to the members of its Section of Consultation–Liaison Psychiatry, which was anticipated to capture members who had the most clinical involvement with palliative care services.
From those who responded to the notice and email, participants were selected using purposive sampling that took into consideration gender, geographical location (state), and duration of qualification as a psychiatrist. Recruitment continued until data saturation was reached. This occurred after nine psychiatrists had been interviewed. Their demographic characteristics are shown in Table 1.
Table 1. Participant characteristics
aSome participants worked in multiple settings and/or areas in psychiatry.
bSome participants worked in both palliative care and psycho-oncology services, and reported commonalities in the nature of the work involved in these two services.
Data Collection
The first author conducted all interviews between February and April of 2011. She was introduced to participants as a psychiatrist with an interest in palliative care psychiatry who was conducting the research as part of a doctoral degree. She did not have a personal or working relationship with any of the participants, although she had previously met the local participants within the professional circle. Local participants were interviewed in person in a private setting at their workplace, while interstate participants were interviewed over the phone. All participants consented to their interviews being audiotaped.
Interviews were semistructured and assisted by an interview guide. This comprised participant details (training background, details of psychiatric practice and involvement in palliative care), followed by in-depth exploration of content areas that included concept of depression, its causality, pathogenetic mechanisms, timing and onset of symptoms, course, and treatment. Interviews were open ended, and participants were given the freedom to develop their responses. Consistent with the practice of qualitative interviewing (Braun & Clarke, Reference Braun and Clarke2013), the contents of each interview were utilized to guide and modify subsequent interviews. Participants were assumed to hold different concepts of depression, and the term “depression” was used without specification other than the clinical usage of the term.
An audit trail was kept to record interim analysis, and constant comparison was performed to determine when data saturation was reached, that is, when no new themes for the research question were identified (Morse, Reference Morse1995; Bowen, Reference Bowen2008). The interviews ranged from 47 to 76 minutes in duration, with a mean of 63 minutes.
Data Analysis
All interviews were transcribed verbatim by the first author and deidentified. Three participants agreed to review their transcripts for participant validation, and they made no modifications as to content.
The first author performed thematic analysis as described by Braun and Clarke (Reference Braun and Clarke2006), which involved the stages of data familiarization, generation of initial codes, search for themes, review of themes, definition and labeling of themes, and finally, report writing. Coding was performed deductively in reference to interview content areas and inductively to emergent contents relevant to the research question. The criteria for quality outlined by Braun and Clarke (Reference Braun and Clarke2006) were employed as a reference for rigor. The final themes were verified against the raw data by two coauthors (ACH and GC) to enhance the trustworthiness of our analysis. NVivo software (v. 9; QSR International, 2011) was utilized to facilitate data organization.
Ethics
This study received institutional approval from the University of Adelaide Human Research Ethics Committee (no. H-086-2010).
RESULTS
Three overarching themes were identified: (1) depression means different things, (2) depression is conceptualized using different models, and (3) depression is the same concept within and outside of the palliative care setting.
Depression Means Different Things
As a clinical term, participants saw depression as representing different things and as unhelpful since it could neither provide an understanding of the problem nor guide intervention. Depression as a heterogeneous concept and as being variously conceptualized by clinicians formed the two subthemes, which are illustrated in Table 2.
Table 2. Illustrative data extracts for the theme: Depression means different things
Depression Is a Heterogeneous Concept
Participants considered depression to be a heterogeneous concept, but individual participant views were varied in terms of conceptual breadth. Those who conceptualized depression in broad terms of being an illness and emotional state and/or symptom warned against adopting a narrow view of equating depression with “Major Depressive Disorder” or an illness requiring antidepressants, because of the risk of thinking that depression falling outside of such criteria was insignificant or not requiring assistance. In comparison, those with narrow concepts of depression as an illness considered it to be overdiagnosed and often represented a mislabeling of distress and normal sadness, thus exposing patients to inappropriate treatment with antidepressants.
Collectively, participants defined depression variously as an emotional state, a symptom, and an illness. As an emotional state, it was thought to be ubiquitous in the palliative care setting and to reflect a normal response to adverse life events. It was also considered to be an inevitable stage of grief that must be traversed to reach acceptance. As a symptom, depression signaled a potential cause of suffering and a need for assessment and intervention, akin to other symptoms addressed in palliative care. Depression as a symptom was divorced from the notions of normality or pathology. Participants emphasized the need to distinguish between depression as a symptom and as an illness. Confusing the two was thought to account for inflated prevalence estimates of depressive illness in research, which did not reflect their observations in clinical practice. Depressive illness was considered to be a syndrome with anhedonia as its cardinal feature, to be a biological illness, pathological, and not an emotional reaction or part of the dying process. Participants objected to depressive illness being interpreted as a spiritual issue or as a normal part of dying, both of which would deny patients the opportunity for intervention. They explicitly distinguished depressive illness from other forms of depression.
Depression Is Conceptualized by Clinicians in Various Ways
Participants believed that clinicians differed in their conceptualizations of depression, but also noted overlap between disciplines as well as variability within them. Conceptual differences were primarily attributed to differences in the clinicians' training backgrounds—for example, between palliative medicine specialists with general practice and those with physician training backgrounds. Participants articulated the potential for shaping clinicians' conceptualization through education, either through explicit training or more implicitly through collaborative clinical work with psychiatrists. Some participants observed an approximation of concepts between themselves and their palliative medicine colleagues after long periods of working together. Despite noting conceptual differences among clinicians, disagreement over diagnosis or management was not encountered, with participants indicating that their views seemed readily accepted by their palliative medicine colleagues.
Depression Is Conceptualized Using Different Models
Participants described three conceptual models that respectively viewed depression as a spectrum, a dichotomy, and a mixture of spectrum and subtypes (Table 3).
Table 3. Illustrative data extracts for the theme: Depression is conceptualized using different models
Spectral Model of Depression
In the spectral model, depression was seen as a dimensional rather than a categorical construct, spanning from normal emotional reaction to depressive illness that included major depression and its melancholic, catatonic, and psychotic forms. In this model, depression had considerable overlap with other psychological states such as distress, grief, and demoralization. The spectrum therefore encompassed both normal and pathological depressive states, and the ill-defined boundaries between these made differentiation of clinical depression difficult in the less severe portion of the spectrum. For participants who described this model, this ambiguity gave rise to conflicting concerns about medicalizing and stigmatizing patients for understandable emotional responses on the one hand, and denying them appropriate treatment for depression on the other. The DSM diagnosis of “Adjustment Disorder” was a disputed construct because it called for a judgment to delineate unacceptable from acceptable depressive responses to previously unchartered adversities, when such division was considered arbitrary within this spectral perspective. Similarly, some participants saw the concept of major depression as categorical, rigid, and incompatible with the dimensional nature of depression.
Dichotomous Model of Depression
In the dichotomous model, depression was broadly divided into reactive and endogenous types, which were considered to be separate processes with different manifest qualities. Reactive depression was regarded as an understandable response to a situation, likened to an “extreme type of sadness.” However, it was emphasized that its understandability did not imply a lesser severity. Participants analogously compared reactive depression with pain, which, despite being an expected response to tissue pathology, caused suffering and could be ameliorated through appropriate treatment. Endogenous depression was also referred to as melancholic depression, and was viewed as a biological illness. Very ill patients were seen to be more vulnerable to developing this type of depression, but differentiating this from the effects of advanced disease could be challenging.
Mobility was a feature within both the spectral and dichotomous models. This refers to the view held by some participants that one type of depression could transform into another—for example, reactive depression turning into melancholic depression. Some also expressed the corollary that timely intervention for reactive depression could prevent a melancholic progression.
Mixed Model of Depression
The mixed model viewed depression in both spectral and typological terms. Melancholic and psychotic depressions were seen to be distinct biological depressive illnesses, which sat separately from a continuum of depressive presentations that spanned from “normal reactions” to non-melancholic depression. Non-melancholic depression was considered to be a heterogeneous disorder underpinned by a variety of interplaying biopsychosocial etiological factors. Participants applying this model referred to and endorsed the hierarchical model, as proposed by Parker (Reference Parker2000), which distinguished psychotic, melancholic, and non-melancholic depressions as three types of depression, with the latter being a spectral construct.
Depression is the Same Concept Within and Outside of the Palliative Care Setting
Participants considered depression to be the same condition when occurring in the palliative care setting as in other settings, explainable using the same etiological framework and responsive to the same treatments (Table 4). They also believed the prevalence of depression to be similar inside and outside of the palliative care setting, with the reported higher prevalence in palliative care attributed to a methodological artifact arising from measuring depressive symptomatology instead of disorders. The artificiality of distinguishing depression occurring in the palliative care setting was emphasized by a participant who referred to this as a “mock concept.”
Table 4. Illustrative data extracts for the theme: Depression is the same concept within and outside of the palliative care setting
Nevertheless, participants described two peripheral differences about depression in the palliative care setting, involving context and ease of diagnosis (Table 4). Contextual differences referred to the influence of advanced illness and dying on the presentation of depression. Separation, loss, grief, death and dying, and existential concerns were cited as dominant issues for depressed patients in this setting, and together with issues of family dynamics formed a major focus in intervention. Although participants applied the same biopsychosocial etiological framework to depression in this setting, specific contributory factors were considered to more commonly relate to advanced illness and its psychological and social consequences, such as uncontrolled symptoms, brain disease, organ failure, treatments, disability, loss of autonomy, existential crisis, and sense of being abandoned by family and treating clinicians. Diagnosing depression in this setting was considered harder due to the difficulty of distinguishing between symptoms of depression from those of life-limiting diseases and dying, and from psychological adaptation, distress, and demoralization. Judgment was required to interpret symptoms in the illness and dying context, and in some situations to determine the relevance, if any, of making a diagnosis.
DISCUSSION
In this study, psychiatrists explicitly identified depression as a heterogeneous concept that requires specification in order to be meaningful. Individual interpretations of the term varied from the relatively narrow construal of an illness, to broader ones that also encompassed ideas of emotional state and symptom. These interpretations differed in their philosophical bases: restricting the concept of depression to a depressive illness makes a demarcation based on pathology, whereas broader concepts demarcate depression based on clinical significance (i.e., potential benefit from clinical intervention). Herein lies a source of tension, with proponents of the former criticizing the mislabeling of sadness and distress as depression, and proponents of the latter criticizing the dismissal of depression that did not meet illness criteria. Such disputes therefore reflect differences in the conceptual breadth of the term “depression,” rather than differences in conceptualization of depressive illness. In fact, depressive illness was consistently viewed as a biological type of depression, distinguished from depressive reactions and considered to exist as different types.
The relationship of depressive illness to other depressive presentations was more ambiguous and was conceptualized by participants using three models. In the spectral model, depressive presentations were considered along a continuum, with no clear demarcation between depressive emotional reactions and depressive illness. This model resembles other dimensional approaches to conceptualizing affective disorders (Akiskal & Pinto, Reference Akiskal and Pinto1999; Angst et al., Reference Angst, Sellaro and Merikangas2000; Ayuso-Mateos et al., Reference Ayuso-Mateos, Nuevo and Verdes2010). In contrast, the dichotomous and mixed models considered depressive illness to be distinct categories and conceptually demarcated from depressive reactions. The dichotomizing of endogenous/psychotic depressions from reactive/neurotic depressions has been criticized, among other reasons, for its weak boundary of distinction based on life stressors, and has lost currency with terminological and paradigmatic shifts (Paykel, Reference Paykel2008). It is therefore notable that participants continued to find the dichotomous model useful despite its supersedence in recent decades by more contemporary classifications. The continuing relevance of the dichotomous model is also supported by a recent study of psychiatrists, whose explanations to patients and treatment recommendations differed according to the two types of depression (Mizushima et al., Reference Mizushima, Sakurai and Mizuno2013). Despite using DSM terminologies, participants considered these to be inadequate and conceptualized depressive illness as phenotypic subtypes, such as the melancholic, non-melancholic, and psychotic subtypes of the hierarchical model (Parker, Reference Parker2000), rather than as major depression.
Compared with palliative medicine specialists (Ng et al., Reference Ng, Crawford and Chur-Hansen2014a ), psychiatrists in this study more explicitly articulated depression as a heterogeneous concept, distinguished depressive illness from other depressive presentations, and subtyped depressive illness. The psychiatrists also saw depression to be the same concept within and outside of the palliative care setting, which contrasts with the distinction made by palliative medicine specialists between de novo and recurrent/persisting depressions (Ng et al., Reference Ng, Crawford and Chur-Hansen2014a ). However, both groups considered the boundaries of depression to be unclear and diagnosis to be difficult, and shared concerns about the medicalization of human emotions and underrecognition of depressive disorders.
This study finds that contemporary psychiatrists do not subscribe to a unitary model of depression, but understand it as a heterogeneous concept comprised of fairly uniform views of depressive illness and other less clearly defined depressive presentations. They also utilize phenotypic subtyping of depressive illness. Given the apparent clinical influence of psychiatric opinions (Golden et al., Reference Golden, McCartney and Haggerty1991; Power et al., Reference Power, Kelly and Gilsenan1993; Stiefel et al., Reference Stiefel, Die Trill and Berney2001; Jefford et al., Reference Jefford, Mileshkin and Richards2004), these findings suggest that depressive illness could be more explicitly distinguished in clinical parlance and conceptualization from other depressive syndromes or symptoms, rather than coalescing these under the single term of “depression.” Furthermore, the unitary and amorphous concept of major depression could be refined with consideration of phenotypic subtypes, in order to promote a more specific approach to assessment and management. These clinical practice recommendations are in line with those advocated by other authors in the psychiatric literature (Parker, Reference Parker2000; Parker, Reference Parker2005; Jacob, Reference Jacob2009; Restifo, Reference Restifo2012; Reference Restifo2013). In research, depression as a singular notion could similarly be replaced by specification of depressed mood as a symptom and various subtypes of depressive illness, in order to produce more meaningful data. Distinguishing between depressive symptoms and illness may also circumvent the ambiguity arising from the usage of broad and narrow concepts of depression, and emphasize the clinical significance of depression both at the level of a symptom requiring intervention and at the level of an illness. In health professional education, greater emphasis could be placed on assessment for depressive illness based on phenomenological and contextual appraisal, rather than on diagnostic criteria or severity scales, although the latter remain useful screening measures. Whether depression is different in the palliative care setting, in causality or form, is not known, but it may be a point of conceptual difference between palliative medicine specialists (Ng et al., Reference Ng, Crawford and Chur-Hansen2014a ) and psychiatrists. Having more refined concepts of depression in both clinical practice and research may assist in elucidating this matter.
It was recognized throughout the conduct of this study that the first author's profession as a psychiatrist, specialized interest in palliative care psychiatry, and the nature of her own concepts of depression could potentially influence the data collection and analysis processes. Care was taken to avoid conceptual assumptions during interviews, and all interview transcripts were read by the coauthors during the data collection phase, and no undue interviewer influence was noted on participants' responses. In analysis, the final themes were verified by the coauthors with nonpsychiatric backgrounds (palliative medicine [GBC] and psychology [ACH]) as a measure of trustworthiness. As there may be local differences in training and practice, the findings of this study may not be transferable to countries where these are very different from the Australian context. However, the findings of our study should have widespread relevance given the clinical importance of depression in palliative care, the challenges of its diagnosis and management in this setting, and its conceptual ambiguity in clinical practice and research. By recruiting psychiatrists with an intimate understanding of the palliative care context, this study describes the concepts of those who are providing clinical guidance in this area and may hopefully serve to stimulate further debate and research on the concepts of depression within the palliative care community.
ACKNOWLEDGMENTS
We wish to thank all the participants for their generous input of time and intellectual contributions. We also thank the Royal Australian and New Zealand College of Psychiatrists for its assistance in the recruitment process, and Dr. Rajan Nagesh for his comments on the manuscript.
COMPETING INTERESTS
The authors state that there are no competing financial interests to declare.
FUNDING
This research received no specific grant from any funding agency in the public, commercial, or not-for-profit sectors.
