Introduction
Diagnosing mental health challenges in bereavement is controversial; however, regardless of one’s position on this matter, assessments of bereaved individuals continue to take place both in clinical and research contexts. We assert that as long as such assessments are occurring, it is critical for evaluations to account for contextual factors that are unique to bereavement. For instance, in the context of bereavement, certain behaviors may be related to cultural sanctions rather than reflections of clinical “symptoms.” A loss of interest in engaging in social activities may be because a griever has realistic concerns about how others will react to their loss or grief rather than anhedonia in the context of a frank clinical depression. Intricacies such as these complicate assessments in bereavement research and clinical work, and thus there is a need for more nuanced guidance when conducting evaluations in the context of bereavement.
Prior efforts to examine the prevalence of mental health challenges in bereaved populations have demonstrated that various syndromes (e.g., depression, anxiety, prolonged grief disorder, and posttraumatic stress disorder [PTSD]) can be empirically distinguished through factor analysis of symptom ratings (Boelen and van den Bout Reference Boelen and van den Bout2005; Spuij et al. Reference Spuij, Reitz and Prinzie2012). The network theory of psychopathology, which has received increasing attention in the context of bereavement, posits that these distinctions can be attributed to the fact that, within a symptom network, some symptoms are more causally related to one another than others (Borsboom and Cramer Reference Borsboom and Cramer2013; Robinaugh et al. Reference Robinaugh, Hoekstra and Toner2020). The presence of causal relationships also explains how symptoms ultimately present as syndromes. For example, the syndrome of depression may manifest in an individual because they are experiencing insomnia, which leads to fatigue, which leads to concentration problems, which leads to decision-making challenges, which then further exacerbates insomnia over time (Maccallum and Bryant Reference Maccallum and Bryant2020). Network theory also explains the overlap in symptoms across syndromes and syndrome comorbidity. A systematic review and meta-analysis of studies examining depression, anxiety, prolonged grief, and PTSD in bereaved adults found that among those with prolonged grief symptoms, 63% also had co-occurring depressive symptoms, 54% had co-occurring anxiety symptoms, and 49% had co-occurring posttraumatic stress symptoms (Komischke-Konnerup et al. Reference Komischke-Konnerup, Zachariae and Johannsen2021). It is difficult to determine whether these statistics reflect actual comorbidity or the complexities of diagnosing mental health challenges in bereavement.
Our own experience conducting diagnostic assessments through research with grieving individuals has punctuated the challenges of “diagnosing” in the context of bereavement. We have attempted to standardize decision rules related to various subtle but important considerations in symptom presentation. This paper summarizes our considerations of diagnosing depression in the context of bereavement, offering a guide to clinicians and researchers who are evaluating bereaved individuals. Specifically, it focuses on use of the Hamilton Depression Rating Scale (HAM-D), regarded as the gold-standard (Rohan et al. Reference Rohan, Rough and Evans2016) for assessing depression severity, based on our experience using this tool in a study of parents bereaved by cancer. While we are using terminology such as diagnosis and symptoms, this paper is not intended to be an affirmation that it is appropriate to diagnose bereaved individuals, and bereaved parents in particular, with depression; rather, it is meant to provide guidance about nuanced considerations when the HAM-D is being used in clinical and research settings in the context of bereavement.
The HAM-D and its administration
Published in 1960, the Hamilton Depression Rating Scale (HAM-D) has been considered an exemplar for assessing depression (Furukawa et al. Reference Furukawa, Akechi and Azuma2007; Moberg et al. Reference Moberg, Lazarus and Mesholam2001). Originally created with 17 items to score, the HAM-D is frequently used by clinicians and researchers to examine an individual’s depression symptoms. The scale has been shown to have high reliability and high discriminant, concurrent and construct validity (Potts et al. Reference Potts, Daniels and Burnam1990). Numerous versions of the HAM-D have been developed, varying in the “number, sequence and wording of items” (Carrozzino et al. Reference Carrozzino, Patierno and Fava Giovanni2020, 134). Adaptations range from including 6 to 36 items, in which some versions include symptoms of atypical depression or reverse neurovegetative symptoms (i.e., hypersomnia, increased appetite, and weight gain) (Carrozzino et al. Reference Carrozzino, Patierno and Fava Giovanni2020). Over the years, the HAM-D has been adapted for different populations and administered in different ways to increase validity and reliability (Carrozzino et al. Reference Carrozzino, Patierno and Fava Giovanni2020).
Abbreviated forms have also been created (Luckenbaugh et al. Reference Luckenbaugh, Ameli and Brutsche2015). Bech et al. (Reference Bech, Gram and Dein1975) created the six-item subscale of the Hamilton Depression Rating Scale (HAM-D6), which evaluates the six core symptoms of depression: depressed mood, work and activities, general somatic symptoms, guilt, psychic anxiety, and psychomotor retardation. The HAM-D6 proved to have significant advantages compared to the longer scales; it is faster to administer, and it excludes symptoms which may be more easily attributed to antidepressant side effects (i.e., concentration problems, nausea, and sexual dysfunction) (Timmerby et al. Reference Timmerby, Andersen and Søndergaard2017). Most importantly, the HAM-D6 has been shown to be superior to the HAM-D17 “in terms of scalability (each item contains unique information regarding syndrome severity), transferability (scalability is constant over time and irrespective of sex, age, and depressive subtypes), and responsiveness (sensitivity to change in severity during treatment)” (Timmerby et al. Reference Timmerby, Andersen and Søndergaard2017, 141). Used by clinicians and researchers for over 40 years, the HAM-D6 “displays clinical validity (the total score is strongly associated with the global perception of severity evaluated by clinical experts)” (Timmerby et al. Reference Timmerby, Andersen and Søndergaard2017, 147).
There have been various approaches to administering the HAM-D. The original HAM-D only provided assessors with general, limited instructions to rate each of its items. Hamilton claimed that the assessment depended “entirely on the skill of the interviewer in eliciting the necessary information” (Hamilton Reference Hamilton1960, 56), meaning the scores were solely based on assessor judgment and expertise (Carrozzino et al. Reference Carrozzino, Patierno and Fava Giovanni2020). This reliance on the individual assessor’s skills and judgment, along with the ambiguity of the item descriptions, led to poor item reliability (Moberg et al. Reference Moberg, Lazarus and Mesholam2001). To address these limitations and standardize administration, structured interview guides were created. One of the first was the Structured Interview Guide for the Hamilton Depression Rating Scale (SIGH-D) (Williams Reference Williams1988), which detailed specific interview questions to facilitate score determination. Evaluation of each item begins with a single prompt question that is intended to elicit enough information about the severity and frequency of a given symptom to provide a rating. Follow-up questions are provided and can be generated by the rater until sufficient details have been obtained (Williams Reference Williams1988).
While the SIGH-D demonstrated increased item-reliability (Williams Reference Williams1988), it does not provide specific guidelines to assess the dimensions of symptom frequency and intensity. For example, for the “depressed mood” item, the description for a score of 1 reads: “Indicated only on questioning (occasional, mild depression)” (Williams Reference Williams1988), which may lead to additional questions (i.e., What if the patient has occasional severe depression where they are unable to get out of bed? What if the patient spontaneously reported that they are depressed, but only occasionally and mildly depressed?). The Depression Rating Scale Standardization Team (DRSST) (2003) thus developed a more structured interview guide, the GRID-HAMD, which utilizes a grid format to independently assess the dimensions of frequency and intensity and clarify item content (Williams et al. Reference Williams, Kobak and Bech2008). In contrast to the SIGH-D, for the GRID-HAMD, an individual can receive a score of 1 for depressed mood in three instances: if they have mild depression (described as “feelings of sadness, discouragement, low self-esteem, pessimism”) (DRSST 2003, 7 of 27) that occurs occasionally (“infrequent; less than 3 days; up to 30% of the week”) (DRSST 2003, 7 of 27), if they have mild depression that occurs much of the time (“often; 3–5 days; 31%–75% of the week”) (DRSST 2003, 7 of 27), or if they have moderate depression (described as “clear nonverbal signs of sadness (such as tearful-ness), feelings of hopelessness, helplessness, or worthlessness about some aspects of life DRSST 2003, 7 of 27) that occurs occasionally. Additionally, many of the score descriptions in the SIGH-D are outdated and only apply to hospitalized patients, despite the fact that the HAM-D is now more frequently used in outpatient settings. To increase utility and relevance among different populations, the GRID-HAMD removed inpatient-specific references (Williams et al. Reference Williams, Kobak and Bech2008). For example, a score of 4 in the “work and activities” item was changed from “in hospital, no activities except ward chores” to “unable to work; needs help performing self-care activities; unable to function without assistance” (Williams et al. Reference Williams, Kobak and Bech2008, 121).
Challenges with using the guide in the bereavement context
In a randomized controlled trial investigating the efficacy of grief counseling interventions for bereaved parents, we chose to use the HAM-D6, as opposed to the HAM-D17, in blinded assessments of depression symptoms to minimize participant burden. We decided that the GRID-HAMD structured guide was the best fit to assist in rating the six symptoms because it offers clarification on the frequency and intensity of symptoms. As we conducted these interviews, we recognized that reports of changes and behaviors that in a clinically depressed person might be diagnostic may reflect other factors in the bereaved, such as a change in priorities, social expectations surrounding grief, or avoidance of grief activators. Our team met regularly to discuss the nuances of rating items in the context of grief, and specifically following the loss of a child to cancer. Through comprehensive discussions surrounding conducting and scoring these qualitative interviews, we created guidelines and probes to consider (Table 1) when interviewing bereaved individuals and grieving parents in particular. We present these considerations for each item of the HAM-D6 below, illustrating clinical distinctions through case examples from our research with parents bereaved by cancer.
Symptom assessment guidelines
The following section will expand on existing guidelines for each of the HAM-D6 items, including depressed mood, work and activities, general somatic symptoms, guilt, psychic anxiety, and psychomotor retardation. In addition to pointing out the nuances between depression and grief, this guide will offer probes and suggestions to future assessors on how to modify the diagnostic interview specifically for parents bereaved by cancer (Table 1) (*Note that gender-neutral pseudonyms and pronouns are used throughout the paper to protect participants’ confidentiality.).
Table 1. Suggested probes to evaluate symptoms of depression in parents bereaved by cancer
Depressed mood
This item assesses “feelings of sadness, hopelessness, helplessness, and worthlessness” (DRSST 2003, 7 of 27).
Assessment of depressed mood in the context of bereavement can be immensely challenging because grief and sadness commonly overlap. While grievers often endorse deep sadness, they simultaneously may experience moments of positive emotions tied to memories of the deceased. Sadness in the context of grief tends to come in waves and is often associated with reminders of the loss (American Psychiatric Association 2013). Depression, on the other hand, tends to consist of persistent and pervasive low mood or misery, and depressed individuals may have difficulty experiencing positive emotions (American Psychiatric Association 2013).
When investigating the “depressed mood” item in the HAM-D6 with bereaved individuals, assessors should pay particular attention to and, if necessary, probe the griever about whether the sadness is specifically related to missing the deceased and whether the experience of sadness comes and goes in waves. To illustrate the distinction, in our study, Fallon* described a sense of constant gloom, expressing that for the past four years, “…even in moments of joy, there’s still…this cloud of sadness that is on everything…; it’s an underlying part of every day. If we’re out to dinner or visiting with people or sitting in church, there’s always…the uneasiness of something not being right.” In contrast, Marley* described their sadness as being “ephemeral,” experiencing it when thinking about the loss but knowing that “those feelings will pass…the good and the bad come and go.”
The Diagnostic and Statistical Manual of Mental Disorders, Fifth Edition (DSM-5) also differentiates between the thought patterns of the two groups; it explains that grieving individuals may be preoccupied with memories or thoughts of the deceased, while depressed individuals may be overly self-loathing, have feelings of worthlessness, and be more pessimistic (American Psychiatric Association 2013). Feelings of worthlessness and lower self-esteem are characteristic of depression and not grief (American Psychiatric Association 2013). In order to parse out this distinction, in this item, it is important to be especially attuned to indications of worthlessness or helplessness. These constructs, compared to hopelessness alone, may be more useful indicators of depression. Grievers, especially those earlier on in their bereavement, commonly report feeling hopeless about a fulfilling future.
Typically, when administering the HAM-D6, it is important to pay attention to non-verbal cues such as tearfulness, slumped posture, frowning, and infrequent eye contact. In the context of non-bereaved individuals, spontaneous tearfulness is often an indicator of depressed mood. Yet, for grievers, tearfulness is commonly an indicator of yearning and sadness about the loss and not necessarily a reflection of low mood. Thus, in the context of bereavement, tears and frowning can be noted, but the context should be considered (i.e., was the person speaking about their loss or the deceased individual when crying?). That said, because tracking the context can be challenging, we have found that further probing about worthlessness and helplessness are more informative than non-verbal cues in making a score determination for the depressed mood item.
Work and activities
This item assesses “loss of interest or pleasure and impairment in functioning at work inside and outside of the home, leisure activities, and family and social relationships” (DRSST 2003, 13 of 27).
This item can be one of the most challenging items to score when evaluating bereaved individuals. With loss of interest or pleasure and evaluating changes in behaviors at work and in other important activities, it is crucial to explore the reason(s) why these changes have occurred. For example, it is quite common for altered behaviors and interests to reflect changes in priorities and perspectives among bereaved individuals. The assessor should ask whether there has been a shift in priorities regarding what is important to them and whether there are new things that engage them. For instance, bereaved individuals may be less motivated to attend large social events that do not serve to deepen relationships and prefer to spend time with individuals to whom they feel close. The assessor should also evaluate whether changes in behaviors reflect avoidance of reminders of their loss or activation of their grief rather than more pervasive anhedonia.
Similarly, grievers may report changes in social relationships and activities for a variety of reasons that are not a reflection of anhedonia or impaired social functioning. For instance, Sam* reported a decrease in social interactions, but explained this was because they are “much more selective about who they spend their time with.” Bereaved individuals may also report legitimate concerns about how others will react to expressions of their grief or challenges with speaking about the deceased. While at times bereavement-related social challenges can lead to anxiety that ultimately impairs social functioning, this should not automatically be considered a reflection of depression.
Another nuance that should be considered is the loss of interest or pleasure in activities that remind the griever of the deceased. These activities activate grief or no longer result in pleasure because the positive affect was derived purely from the shared experience (e.g., “I used to really enjoy playing catch with my child but now I have no interest in playing catch because they are no longer here to throw a ball with me”). Such “grief-related anhedonia” would not reflect a clinical depression but instead is directly caused by the loss. For example, Cameron* described how challenging it was to bake a cheesecake, which was their “specialty,” following the loss of their daughter. They shared, “I tried to make a cheesecake several times…but it was difficult for me to really do it.” After months of trying, they were finally able to make a cheesecake “successfully” for their other daughter’s birthday. Even so, they explained, “I don’t find all that joy in making those things anymore. People don’t ask for them as much either.” When asked if making a cheesecake was difficult because it reminded them of their daughter who passed away, Cameron answered “yes” and described their daughter’s influence on their love for baking: “She loved food… She would help me in the kitchen…and she would always say ‘tell me how to do this; show me how to do that’ and I would show her.”
To distinguish whether the loss of interest in certain activities stems from “grief-related anhedonia” or general anhedonia, it is important to explore their level of interest and engagement in activities that are not linked to the deceased. This often is challenging for grievers whose relationship with the deceased was central to their sense of identity. For instance, Jordan* remarked, “I used to enjoy painting and drawing. I don’t know what I like anymore. I’m trying to figure that out.” When asked if they have tried to take up painting again, Jordan answered, “I try but it’s something I did with [my son]. Every time I try to do it on my own it feels…wrong. Like I don’t get the same joy in it that I used to.” The assessor then probed more to see if they engaged in any activities without their deceased child, and Jordan expressed, “I’ve been a [parent] for so long, I can’t remember the last time I did anything on my own… Being a [parent], you put so much emphasis on your kids and you’re not important anymore, they are. Then when you lose one and you have another one who’s a teen, who wants nothing to do with you, I don’t know who I am anymore. I’m stuck in this feeling of ‘what now?’” Though it took Jordan some time to identify sources of interest outside of activities they did with their child, they eventually mentioned that they enjoy watching television with their spouse and listening to music.
Changes in work circumstances, including unemployment, avoidance of work, decreased productivity, and presenteeism, should also be more deeply explored. Some individuals who served as the deceased’s caregivers during illness may have had discontinued work to fulfill this role prior to their loss. Others may be impacted by factors related to the work environment, such as bereavement leave policies, variable grief literacy, and concerns about working in an unsupportive environment. For instance, Blake,* who stopped working the day their child was diagnosed, expressed, “The thought of going to the office right now and not being able to disappear if something triggered me would be frightening.” Would the decision not to return to work in this case reflect an impairment in functioning or a rational understanding of the environment to which they would have been returning? Instead of returning to work, Blake devoted their time to creating a foundation in honor of their deceased child. They were able to concentrate and productively advance the foundation. When they spoke about the foundation, they became animated and smiled sharing, “What’s nice about the foundation [is that] I can go at my own speed and do it however I want to do it. I just feel more comfortable being the captain of my own ship right now.” Because the change in work circumstances for Blake reflected concerns about the work environment and a shift in priorities, this item would not be scored as present. A more pervasive lack of interest or reduction in capacity should be present to endorse this item.
Somatic symptoms, general
This item assesses “tiredness, loss of energy, fatigue, and muscular aches and pains” (DRSST 2003, 19 of 27).
Assessment of somatic symptoms in bereaved individuals should involve questioning whether symptoms occur in the context of surges of yearning or reminders of the loss, or instead, are more non-specific and pervasive. For example, when asked if they experienced any aches or pains in the past one to two weeks, Sam reported feeling like “there’s an elephant on [their] chest” and being unable to take a deep breath specifically when reminded of their loss: “My heart will sink for a minute when I look at my phone to say, ‘I wonder if she texted me,’ and then remember that she can’t.” In addition to understanding whether grief or loss reminders occurred around the time of the somatic symptoms, it is valuable to evaluate the duration of the symptoms. In addition to being grief-related, the chest pressure Sam described was transient.
More pervasive and enduring physical symptoms may reflect depression. Fallon reported, “I have pressure on my chest constantly… Physically, my body has, for the past four years now, completely changed to where my bones hurt like I have the flu…I can’t lay in bed comfortably. It feels like something is wrong constantly through my whole body.” Fallon also reported intense exhaustion, saying “I have not slept in four years since I lost my [child].” Grief can indeed be physically exhausting, so reports of fatigue are not unique indicators of depression. Additionally, with any depression assessment, it is important to differentiate depression-related physical symptoms from medically related physical symptoms. For example, if an individual were to report severe exhaustion related to their recent diagnosis of Lyme disease, we would not consider this a depressive symptom. Overwhelming loss of energy or extreme aches not clearly linked to a known injury or medical explanation should be considered when scoring this item. However, as a rule-of-thumb, if the feelings are transient and directly related to grief activation, it may not be appropriate to consider them depressive somatic symptoms.
Guilt
In this item, guilt is defined as “the sense of having done something bad or wrong and is accompanied by feelings of regret or shame” (DRSST 2003, 8 of 27). By GRID-HAMD guidelines, assessors should only rate guilt if it is “excessive or unrealistic” (DRSST 2003, 8 of 27).
The HAM-D was initially designed for clinically depressed individuals experiencing some level of impaired functioning. With this consideration in mind, the GRID-HAMD guidelines state that “realistic self-reproach is not rated (e.g., feeling bad to some degree about falling behind in work or not attending to children when this is really a problem)” (DRSST 2003, 8 of 27). In other words, if an individual is actually unable to take care of their children due to their depressive symptoms and feels guilty about this, such “realistic” guilt would not be rated. It would follow that this guideline can apply in the context of bereavement; if an individual could not take care of their children because of debilitating grief symptoms, associated feelings of guilt would be “realistic.”
However, there are other nuances in scoring the guilt item that should be considered in the context of bereavement. Grievers commonly experience guilt as they reflect on whether behaviors they engaged in could have prevented their loss, such as medical treatment decisions. It can be difficult to assess whether guilt over past actions is realistic, but assessors may consider whether they feel pulled to offer the griever a more compassionate perspective.
Assessors should also consider the frequency and duration of feelings of guilt in scoring this item; transient feelings of guilt would be less likely to increase the individual’s score. For example, Blake explained, “As a [parent], I feel like you’re just supposed to protect your children and I feel guilt or sadness that we didn’t pick up [on their illness]. That’s gonna be with me forever. But guilt doesn’t drive my day, it’s just one of the many feelings and thoughts I have over [my child’s] whole incredible illness and diagnosis; it’s ‘what could we have done to prevent it?’ I always think of what we could have done.” While Blake endorsed frequent guilt that could be considered unrealistic, it would be rated mild because it does not “drive their day.” A higher rating would be appropriate for Jamie’s* guilt: “It’s a guilt that I don’t think is ever going to go away…I feel like I should’ve saved her.” When they elaborated, they discussed feeling like they were being punished: “You think back on things you did that you might not be as proud of or that you wish you’d done differently and you think ‘oh, that must be it; I must be being punished for something…’ and feeling sad is a way of making me continue to pay for whatever it is.” Additionally, when asked about severity and frequency, Jamie explained that the guilt is “always there,” and that it causes impairment, saying, “those feelings will talk me right out of doing something.”
Anxiety, psychic
This item assesses “apprehension, fear, panic, and worry, as well as irritability” (DRSST 2003, 16 of 27).
It is very common and understandable for bereaved individuals to worry about the future, including the well-being of other important people in their lives as well as how they will cope with their grief and the physical absence of the deceased moving forward. Assessors should inquire about how distressing or impairing the anxiety is. For instance, Jordan’s son died from brain cancer, and Jordan reported that whenever their daughter complains of getting a headache, they feel a pit in their stomach, “overthink” the situation, and fear the worst. Jordan’s reaction is understandable, and thus the assessor can focus on investigating the frequency and severity of the anxiety to make a score determination. Indicators of severity may include how strongly the anxiety impacts behavior. For example, Blake shared their worry that something else bad would happen in their family and thought about booking separate flights from their spouse in case the plane crashed. In the GRID-HAMD scoring guidelines, severe anxiety is characterized as “fearing the worst” (DRSST 2003, 16 of 27). If Blake were to take steps to book two separate flights, this impairment would raise their symptom intensity to “severe.” Assessors should thus distinguish between “I booked us separate flights” and “I thought about booking us separate flights” when evaluating symptom severity.
Psychomotor retardation
This item assesses “retardation in movement and speech observed during interview” (DRSST 2003, 14 of 27).
Throughout the interview, assessors should observe if the individual has frequent lengthy pauses, a significantly reduced rate of speech, or slow movements. Yet when assessing psychomotor retardation in the context of bereavement, one should consider the effort required for emotion regulation, which may slow down speech and increase fatigue.
Conclusion
In order for bereavement research and clinical care to advance, there is a need to operationalize different clinical presentations among grievers. Diagnosticians should be careful not to pathologize behaviors that may be linked to typical grief and should recognize that “symptoms” operate on a continuum. On the other hand, it is also important to recognize when bereaved individuals are experiencing debilitating depression that could benefit from treatment. While treating the depression will not take away their grief, it may reduce unnecessary suffering and, in fact, allow them to more meaningfully process and connect with their grief. In this paper, we described nuanced factors for assessors to consider when administering the HAM-D6 to bereaved individuals, and, in particular, grieving parents. Our sharing of these considerations is not intended to promote diagnosis of depression in bereavement but rather to highlight the unique contextual factors that distinguish symptoms of depression from common experiences of grievers when applying an assessment tool such as the HAM-D6. While such validated measures can be constraining, they can also have clinical utility (First Reference First2010); that is, they increase standardization in research, help clinicians communicate with each other, advance the field more generally to understand the varying struggles bereaved individuals experience, and systemically facilitate access to services via managed care. Our efforts to disentangle subtle distinctions in emotional reactions and behaviors reflect just how complex responses to bereavement can be. We, therefore, humbly recognize that there is much more to learn through clinical encounters as well as qualitative and quantitative research initiatives. We hope this paper propels additional efforts to understand the intricacies of grief to prevent pathologizing and misdiagnosis and to increase grief literacy.
Acknowledgments
We are grateful to Drs William Breitbart, Talia Zaider, Lamia Barakat, and Justin Baker as well as Danielle Rosenkilde, Nithya Ramaswamy, Nidhi Mali, Cameka Woods, Ashley Anil, Shannon Hammer, Shoshana Mehler, and all of the dedicated interventionists for their contributions to the study that inspired this paper. We are also grateful to the bereaved parents who have given so much to this study and inspire us to continue learning how to best understand and support grief.
Funding
This research was supported by the National Institute of Nursing Research grant R01NR019637 (Lichtenthal) and National Cancer Institute grant P30CA008748 (Vickers). The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health.
Competing interests
The authors declare none. The authors note that Dr. Wendy Lichtenthal is on the editorial board of Palliative & Supportive Care.
