Abstract
Objectives
Older adults with psychiatric illnesses often have medical comorbidities that require symptom management and impact prognosis. Geriatric psychiatrists are uniquely positioned to meet the palliative care needs of such patients. This study aims to characterize palliative care needs of geriatric psychiatry patients and utilization of primary palliative care skills and subspecialty referral among geriatric psychiatrists.
Methods
National, cross-sectional survey study of geriatrics psychiatrists in the United States.
Results
Respondents (n = 397) reported high palliative care needs among their patients (46-73% of patients). Respondents reported using all domains of palliative care in their clinical practice with varied comfort. In multivariate modeling, only frequency of skill use predicted comfort with skills. Respondents identified that a third of patients would benefit from referral to specialty palliative care.
Conclusions
Geriatric psychiatrists identify high palliative care needs in their patients. They meet these needs by utilizing primary palliative care skills and when available referral to subspecialty palliative care.
Introduction
The population of older adults in the United States (age >65) is expected to nearly double between 2020 and 2060 increasing from 54 million to nearly 95 million.1,2 This increasing population of older adults in the United States is also expected to enjoy greater longevity, with projected increases in life expectancy at age 60 through 2060. 3 However, in parallel with these demographic shifts, the burdens of physical and mental health conditions associated with aging will increase.
Over half of all adults ages 65 and older in the United States have multiple chronic medical conditions such as cancer, heart disease, or renal disease. 4 Multimorbidity—the presence of 2 or more chronic health conditions—has increased rapidly in the population. 5 Chronic illnesses and multimorbidity are associated with poor function, symptom burden, disability, and poor quality of life among affected older adults.6,7 Furthermore, the prevalence of neurocognitive disorders such as Alzheimer’s disease and related dementias is expected to increase significantly in the coming decades. 8
The medical and psychiatric needs of older adults are closely intertwined. The majority of older adults accessing psychiatric care have medical comorbidities.9-12 Medical and psychiatric comorbidity are closely linked and bidirectionally impactful among older adults. For instance, psychiatric disorders may predispose older adults to greater physical morbidity burden and vice versa. 13
Palliative care is recognized as the standard of care for older adults with serious medical illness including those with multiple chronic conditions and a limited prognosis.14,15 It aims to improve patient quality of life through symptom management, psychosocial support, and collaborative medical communication and decision-making. 15 While palliative care includes end-of-life care, it also spans the trajectory of serious illness and is delivered alongside curative or disease-modifying therapies. As with psychiatry, the need for palliative care services is greater than the palliative care specialist workforce. Given the inadequate workforce of palliative care specialists, there has been significant and longstanding interest among palliative care educators, researchers, and advocates to train non-palliative care specialists to deliver palliative care. In contrast to specialist care, palliative care delivered by non-palliative care clinicians is referred to as “primary palliative care.” 16
Many older adults receiving psychiatric care have unmet palliative care needs. Recent data from geriatric psychiatry inpatient and outpatient settings suggest a high physical symptom burden among patients. 17 Among older adults with neurocognitive disorders, a core domain of geriatric psychiatry with profound public health importance, palliative care is considered a key component of high-quality care. 18 Furthermore, older adults with psychiatric illness experience disparities in end-of-life and palliative care including differential access to care, less advance care planning, and fragmented care. 19 Though some descriptive examples exist, there are not operationalized nor disseminated models for integrating palliative care into psychiatric settings.20,21
Geriatric psychiatrists are well positioned to provide primary palliative care to older adults with medical and psychiatric comorbidity and to steward these patients to specialty palliative care when necessary. 22 Geriatric psychiatrists have well developed communication skills; a focus on quality-of-life; close attunement to psychological, social, and existential concerns; and often maintain longitudinal and close relationships with patients and caregivers. These aspects of geriatric psychiatry align well with core components of palliative care including advance care planning and caregiver support.23,24 Furthermore, many geriatric psychiatrists receive some training in palliative care as a component of their fellowship training. Approximately 70% of geriatric psychiatry fellowship programs provide some didactic teaching on palliative care topics. 25 However, there have been virtually no operationalized primary palliative care training interventions geared towards or adapted to psychiatrists specifically. Rather, the majority of primary palliative care interventions have been targeted towards serious illness and medical clinicians in settings such as oncology, internal medicine, and critical care. 26
As with palliative care subspecialists, there are significant deficits in the geriatric psychiatry workforce; there are currently approximately 1300 practicing board-certified geriatric psychiatrists nationally.27-29 These deficits are expected to expound given recruitment challenges to geriatric psychiatry and the growth of the older adult population. 30 Such workforce challenges have significant implications on the role of geriatric psychiatrists as primary palliative care experts. First, the growing deficit in geriatric psychiatrists will necessitate that subspecialty trained geriatric psychiatrists provide care to the sickest and most complex older adults—likely the population with the highest palliative care needs. Second, geriatric psychiatrists will be increasingly situated in training roles where they will need to prepare general psychiatrists and other mental health clinicians to provide care to medically complex older adults. In such roles, geriatric psychiatrists may become the de facto psychiatry-palliative care experts, preparing psychiatry trainees to provide primary palliative care and psychiatry care to older adults with mental health and medical comorbidities.
Despite recognition of the importance of integrating palliative care into geriatric psychiatry, there are a lack of data characterizing the perspectives on and practice of primary palliative care among geriatric psychiatrists. To address this knowledge gap, we conducted a cross-sectional study to: 1) characterize the palliative care needs of older adults accessing psychiatric care from the perspective of geriatric psychiatrists; 2) describe the comfort with and utilization of primary palliative care skills by geriatric psychiatrists; and 3) characterize geriatric psychiatrists’ referral patterns regarding specialty palliative care. We hypothesized that geriatric psychiatrists would perceive their patients as having substantial palliative care needs. We further hypothesized that respondents would feel comfortable with and commonly utilize psychosocial and spiritual primary palliative care skills but would be less comfortable with and make less use of physical symptom management skills. Finally, we hypothesized that respondents would refer to specialty palliative care infrequently and experience barriers when doing so.
Methods
Survey Design
Given the lack of any existing instruments validated to answer our research questions, we designed a 22-question de novo survey tool aimed at board-certified/board-eligible geriatric psychiatrists or adult psychiatrists self-identifying as practicing in predominantly geriatric settings.
Domains of Palliative Care Utilized in Survey Instrument and Corresponding Clinical Practice Guideline Domains.
The final survey (see supplemental materials) included personal demographic questions about respondents’ age, gender, and race/ethnicity and professional demographic questions regarding respondents’ board-certification status (geriatric or adult psychiatry), years of practice experience, and administrative roles held. Respondents’ practice characteristics were also collected, including geographic information, size, and setting (inpatient or outpatient, medical or psychiatric setting), and proportion of patients over age 60.
Respondents were queried about perceived palliative care needs of their patients, about how frequently they provided care in each of these domains, and about their comfort level providing care in each domain using 5-point Likert response scales. We also asked respondents about their patients’ need for specialty palliative care and hospice referral, their referral practices, and their success obtaining referral.
Sample Assembly
We sought to survey board-certified geriatric psychiatrists and adult psychiatrists providing geriatric psychiatric care. We built a sample of 1,780 potential participants by searching publicly available professional organization directors, the American Board of Psychiatry and Neurology’s Board Certified Physician search feature (VerifyCERT), 34 and faculty rosters of institutions with geriatric psychiatry fellowships listed on the Accreditation Council of Graduate Medical Education database. Where possible, we matched names to emails via web searches for institutional directories.
Survey Distribution and Administration
We utilized RedCap’s distribution system to anonymously distribute the survey to our list on April 2022. Non-respondents were sent 2 automatic reminders at weeks 2 and 3 following the initial survey invitation.
Data Analysis
Survey results were summarized with descriptive statistics and are reported in the text as means with standard deviations (SD) (for continuous data with a near-normal distribution) and medians with interquartile ranges (IQR) (for ordinal data, including Likert responses). Frequencies are reported as percentages. Data analysis was completed in SPSS statistical software. 35
Multiple linear regression models were constructed to identify factors associated with greater clinician comfort using palliative care. First, a Pearson correlation coefficient was computed to assess the linear relationship between physician comfort level providing each of the 7 palliative care domains. Positive correlations were discovered between several of the variables, allowing 4 composite items to be used in for multivariate analysis. Pharmacologic pain management and non-pain physical symptom management were grouped as “Symptom Management,” resuscitation/code status discussions (hereafter code status), advance care planning and goals of care discussions other than code status (hereafter goals of care), and hospice referral grouped as “End of Life Care,” and “Serious Illness Coordination” and “Spiritual Care” each did not correlate significantly with any of the others. For each of the 4 composite items, a multiple linear regression model was constructed with clinician comfort as the dependent variable and the following independent variables: frequency of clinician provision of the domain, age, years of experience, practice setting, palliative care referral process, and palliative care referral success. We selected these variables based on a hypothesis that more experienced clinicians and clinicians in settings with access to palliative care specialists (academic settings or settings with accessible referral) would feel more comfortable with palliative care skills.
Clinician patterns of referral to palliative care were analyzed using multiple linear regression, with palliative care referral success as the dependent variable and the same variables listed above serving as independent variables. A Pearson correlation coefficient was computed to assess the linear relationship between each of the following practice setting variables: inpatient geriatric psychiatry unit, inpatient geriatric psychiatry consultation in medical settings, inpatient geriatric psychiatry consultation in psychiatric settings, outpatient geriatric psychiatry clinic, psychiatrist embedded in geriatric medicine clinic, and psychiatric consultation in nursing home, subacute rehabilitation, or other facility settings determined the variables correlated in 2 groups: “Practice in Psychiatric Settings” and “Practice in Medical or Mixed Settings.” Pearson correlation coefficients were then calculated to assess the linear relation between practice in psychiatric settings and proportion of patients that would benefit from and frequency of provision of palliative care.
Ethics
This study was reviewed by the Biomedical Research Alliance of New York (BRANY) Institutional Review Board and deemed exempt from IRB review. This determination was acknowledged by the Weill Cornell Institutional Review Board.
Results
A total of 453 of 1780 invitees completed the survey, 56 respondents did not meet inclusion criteria because they did not treat older adults or were no longer in practice. Thus, 397 (response rate 23%) responses were included in the analytic sample.
Demographics
Respondent Demographics and Practice Characteristics.
Perceived Palliative Care-Related Patient Needs
aOn a Likert scale from 1 to 5 with 1 being never and 5 being very often.
bOn a Likert scale from 1 to 5 with 1 being very uncomfortable and 5 being very comfortable.
Palliative Care Delivery
Respondents reported utilizing a range of palliative care skills. On a Likert scale of 1 to 5 with 1 being never and 5 being very often, respondents reported most often utilizing goals of care discussions (median 4, IQR 3-4) and serious illness care coordination (4, IQR 3-4). In descending order of frequency of use, respondents also reported frequency of the following palliative care domains: non-pain physical symptom management (3, IQR 2-4), code status discussions (3, IQR 2-4), pharmacologic pain management (3, IQR 2-4), and spiritual care (2, IQR 2-3), and hospice referral (2, IQR 2-3). Table 3 compares respondent reported frequency of use and comfort with each of the palliative care domains.
Comfort with Palliative Care
We asked respondents to gauge their comfort level providing care in each of the 6 palliative care domains on a Likert scale from 1 to 5, with 1 being very uncomfortable and 5 being very comfortable. Respondents expressed a high comfort level participating in goals of care discussions (4, IQR 4-5), hospice referral (4, IQR 3-5), code status discussions (4, IQR 3-4), serious illness care coordination (4, IQR 3-4), and non-pain physical symptom management (4, IQR 3-4). Respondents felt less comfortable delivering spiritual care (3, IQR 2-4) and managing pain pharmacologically (3, IQR 2-4).
We used multivariate linear regression modeling to identify factors associated with comfort across 4 skill clusters end-of-life care (code status discussions, goals of care discussions, and hospice referral), symptom management (pain and non-pain symptom management), serious illness coordination, and spiritual care. Independent variables included clinician experience, clinician setting, clinician referral access, and frequency of utilization of the palliative care skill cluster. For each skill cluster, frequency of provision was the only variable that significantly predicted comfort with provision (end-of-life care cluster β 0.57, P < .001; symptom management cluster β 0.62, P < .001; serious illness care coordination β 0.66, P < .001; spiritual care β 0.67, P < .001).
Specialist Palliative Care Utilization and Access
Referral to Specialist Palliative Care.
aOn a Likert scale from 1 to 5 with 1 being never and 5 being very often.
bRespondents could select multiple options among these items resulting in response numbers exceeding total respondents.
Respondents felt that across their patient populations, approximately one third (32.7 ± 24%) would benefit from specialist palliative care referral. Working in a psychiatric setting, rather than consultatively in a medical setting, was inversely correlated (albeit weakly) with reported percentage of patients who would benefit from specialist palliative care consultation (r -0.12, P = .007).
Respondents were asked to identify issues for which they would refer to specialist palliative care (respondents could select multiple issues so responses exceed 397/100%). Pharmacologic pain management (270 responses; 68%), advance care planning (258 responses; 65%), and serious illness care coordination (254 responses; 64%) were the most frequent choices. Spiritual care (144 responses; 36%) and code status discussions (176 responses; 44%) were the least cited issues that would prompt specialist referral.
Despite an identified need among a substantial minority of patients, respondents infrequently initiated specialist hospice (2, IQR 2-3) or palliative care referral (median 2, IQR 2-3). However, when respondents did pursue referrals to specialist palliative care, they were frequently able to access such referrals (median 4, IQR 3-5). Most respondents reported referring within their own institution (270, 68%) with few pursuing external referrals (89, 22%). Barriers to referral included lack of clinicians to whom to refer (79, 20%), inadequate knowledge about indications for referral (54, 14%), overly complicated referral process (53, 13%), and referrals not addressing issues of concern (33, 8%). Only a small minority of respondents (25, 6%) were concerned that palliative care physicians would be uncomfortable working with patients with mental health needs.
In a multivariate model predicting referral success academic practice setting (β .17, P = .002) and within-institution referral availability (β .322, P < .001) predicted greater success obtaining palliative care referrals.
Discussion
Older adults with mental illness have high rates of multimorbidity and/or serious medical illness for which palliative care can offer important improvements in quality of life and goal concordant care. As the number of older adults with major neurocognitive disorder and/or medical and psychiatric comorbidity grows, so too will the need for both geriatric psychiatry and palliative care services. This is reflected by a growing literature explicitly identifying the importance of aligning geriatric psychiatry and palliative medicine to improve the care of older adults.20,23,36 However, to the authors’ knowledge, this cross-sectional national survey of geriatric psychiatrists is the first to characterize geriatric psychiatrists’ practice of, referral to, and perceived needs regarding palliative care.
In our study, geriatric psychiatrists identified (1) a high prevalence of palliative care needs among their patients across settings, (2) frequent use of primary palliative care skills in their clinical practice, (3) and variable comfort with specific domains of primary palliative care (4) variable utilization of and success with referrals to specialist palliative care and hospice. As we hypothesized, geriatric psychiatrists were more comfortable with and frequently provided psychosocial domains of palliative care versus pain management (though respondents expressed relative comfort with non-pain symptom management). Respondents perceived physical symptoms as a prompt for referral to specialist palliative care. These data are consistent with fellowship training of geriatric psychiatrists; among fellowships, pain management is taught much less frequently than advance care planning. 25 The frequency of and success with making referrals to specialty palliative care varied among respondents; academic practice environment and availability of within-institution referral were associated with greater success making a palliative care referral.
Importantly, respondents identified that only a subset of patients with palliative care needs required specialist care referral. Our data demonstrate that geriatric psychiatrists frequently utilize primary palliative care skills and might benefit from targeted training opportunities to expand these skills, particularly in the domain of symptom management. Increasing opportunities to utilize a range of palliative care skills in fellowship training may strengthen comfort with a range of palliative care skills in practice. These benefits are likely to extend to clinicians across diverse practice settings as academic versus non-academic setting was not independently associated with comfort practicing palliative care skills. As hypothesized, respondents reported lower levels of comfort managing pain in their patients. This finding provides support for future intervention efforts that seek to empower geriatric psychiatrists to effectively provide primary palliative care in the domains of pain and symptom management.
This study has several strengths. To the authors’ knowledge, this is the first study to systematically characterize comfort with and practice of primary palliative care skills by geriatric psychiatrists. In addition, we were able to address both primary and specialist palliative care needs among older adults receiving psychiatric care by asking respondents both about their own practice and about referral experiences to specialist palliative care. Our study also contributes a clinician-focused perspective to evidence about the high prevalence of palliative care needs among older adults receiving mental health services. Our data were derived from a sizable sample of geriatric psychiatrists and adult psychiatrists practicing with large populations of older adults representing diverse practice settings across the United States suggesting good generalizability.
Limitations of this study include the possibility of response bias as well as a possible sampling bias towards academic clinicians. Perhaps the most striking limitation is that clinicians’ comfort with clinical skills may not align with actual skill or knowledge level. Though we generally consider comfort with palliative care skills as a desirable trait among non-palliative care clinicians, such comfort can be counterproductive in the context of inadequate training. Data from medical education routinely shows incongruity between confidence and skill-level among learners in a range of settings.37,38 A further limitation is that our response rate of 25% suggests that our data are impacted by response bias; it may be that respondents were more attuned to the importance of palliative care than non-respondents. This is particularly true as our initial sample was likely skewed towards academic clinicians by virtue of our sampling approach because academic clinicians have more readily available contact information through institutional websites. Finally, our respondents varied in the extent of their practice dedicated to adults over age 60 and some respondents were working in general adult psychiatry settings, rather than geriatric settings.
Of note, the conceptual framework of our study is rooted in the traditional framework of palliative care for serious medical illness. However, as the interface between psychiatry and palliative care grows, there are nuances in the field of palliative care psychiatry that require further study. Among older adults with serious illness and mental health comorbidity, there may be differences in needs and care utilization between individuals with mental illness (particularly serious mental illness) who develop serious medical illness and individuals who develop new onset psychiatric illness in the setting of coping with a serious medical illness. 39 The latter group in particular may lie at the intersection of subspecialty care between geriatric psychiatry, consult-liaison psychiatry, and palliative care in the emerging field of palliative care psychiatry. 40 In future studies, it may be important to differentiate the role and experience of geriatric psychiatrists caring for these 2 groups. Particularly vis-à-vis models of psychiatric care which are embedded or situated in palliative care settings and which may skew towards addressing new onset mental illness among older adults with serious illness.41,42 Our study also did not address emerging scholarship on application of palliative care approaches to the care of patients with refractory serious mental illness; as the dimensions of such care are concretized and disseminated, it will be important to attend to this additional sub-domain of palliative care likely to pertain to psychiatrists providing psychiatric care to older adults.43-47
In conclusion, older adults with comorbid serious medical and psychiatric conditions constitute a growing population with significant needs who often face disparities in quality of life and access to appropriate care. Patients with neurocognitive disorders like dementia make up an important subgroup of these patients. Geriatric psychiatrists who provide mental health care to these patients recognized the potential benefits of palliative care among these patients, both in psychosocial and physical domains of palliative care. Geriatric psychiatrists may benefit from additional training and supervision opportunities to further develop core skill sets in these areas. Given that geriatric psychiatrists are in a role to reach many patients who may be unable to access specialty palliative care, their use of primary palliative care skills represents an opportunity to improve patients’ quality of life. Our data suggest the possible utility of creating more operationalized palliative care training opportunities for geriatric psychiatrists. Existing opportunities, such as those offered by the Center to Advance Palliative Care and American Academy of Hospice and Palliative Medicine should be assessed for appropriateness for psychiatrists.48,49 Further downstream, there is a need for development and evaluation of care models that integrate specialist palliative care into geriatric psychiatry settings.
Footnotes
Author’s Note
This work was completed at Weill Cornell Medicine in the Division of Geriatrics and Palliative Medicine.
Author Contributions
DS and HE conceptualize the study. HE, DS, and MCR designed and refined the survey instrument. DS and HE developed the distribution plan and collected data. DS, HE, GAR and JM conducted data analysis. HE, DS, GAR, and MCR contributed to interpretation of results. HE, DS, and GAR collaboratively wrote the manuscript with critical feedback from JM and MCR.
Declaration of Conflicting Interests
The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding
The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This work was supported by the grant number UL1 TR 002384 from the National Center for Advancing Translational Sciences (NCATS) of the National Institutes of Health (NIH) and by grant number T32AG049666 of the National Institute on Aging.
Data Availability
The data has not been previously presented orally or by poster at scientific meetings.
