Mind the gap: is the Canadian long-term care workforce ready for a palliative care mandate?

Abstract The average expected lifespan in Canadian long-term care (LTC) homes is now less than two years post-admission, making LTC a palliative care setting. As little is known about the readiness of LTC staff in Canada to embrace a palliative care mandate, the main objective of this study was to assess qualities relevant to palliative care, including personal emotional wellbeing, palliative care self-efficacy and person-centred practices (e.g. knowing the person, comfort care). A convenience sample of 228 professional and non-professional staff (e.g. nurses and nursing assistants) across four Canadian LTC homes participated in a survey. Burnout, secondary traumatic stress and poor job satisfaction were well below accepted thresholds, e.g. burnout: mean = 20.49 (standard deviation (SD) = 5.39) for professionals; mean = 22.09 (SD = 4.98) for non-professionals; cut score = 42. Furthermore, only 0–1 per cent of each group showed a score above cut-off for any of these variables. Reported self-efficacy was moderate, e.g. efficacy in delivery: mean = 18.63 (SD = 6.29) for professionals; mean = 15.33 (SD = 7.52) for non-professionals; maximum = 32. The same was true of self-reported person-centred care, e.g. knowing the person; mean = 22.05 (SD = 6.55) for professionals; mean = 22.91 (SD = 6.16) for non-professionals; maximum = 35. t-Tests showed that non-professional staff reported relatively higher levels of burnout, while professional staff reported greater job satisfaction and self-efficacy (p < 0.05). There was no difference in secondary traumatic stress or person-centred care (p > 0.05). Overall, these results suggest that the emotional wellbeing of the Canadian LTC workforce is unlikely to impede effective palliative care. However, palliative care self-efficacy and person-centred care can be further cultivated in this context.


Introduction
The average expected lifespan in many Canadian long-term care (LTC) homes 1 is now less than two years post-admission (Frohlich et al., 2002;Palliative Alliance, 2010;Jayaraman and Joseph, 2013). Within the Canadian LTC context, nearly one in five residents is expected to die in a given year (Canadian Institute for Health Information, 2017), with many unnecessarily visiting the emergency department in the months leading up to death (Gruneir et al., 2010). More ideally, LTC homes would manage end-of-life care and avoid unnecessary visits to acute care settings (Lievesley et al., 2011;World Health Organization, 2011;Canadian Institute for Health Information, 2016). Yet, adoption of a palliative approach remains uncommon in Canada's LTC sector.
Like other nations with a relatively large older adult population, Canada has experienced a gradual change in the level of care acuity in LTC. A previous focus on supported living or convalescence in LTC (Phillips et al., 2006) has been supplanted by more complex care needs, including demand for palliative care (Frohlich et al., 2002;Palliative Alliance, 2010). Moreover, there are some unique palliative care considerations in this context. For instance, a relatively low number of people with a primary diagnosis of end-stage organ failure or cancerwho fall within the traditional domain of palliative carelive in LTC. In contrast, a high number of LTC residents are nearing the end of life with conditions in which the trajectory to death can be more difficult to recognise, including frailty, Alzheimer's disease and other dementias (Murray et al., 2005;van der Steen et al., 2014). For instance, in a large cohort of LTC residents with advanced dementia who died less than one year post-admission, only 4.1 per cent were recognised as having less than six months to live, yet 71 per cent died within that period (Mitchell et al., 2004).
The World Health Organization (2011) has recognised a global need to improve palliative care for older people living in LTC settings, calling for additional research on appropriate and effective service in this sector. Currently available guidelines for a palliative approach within LTC suggest that in addition to managing residents' disease symptoms and physical needs, palliative care in LTC should actively address a range of other important needs, including quality of life, psycho-social and spiritual support, family support, and respect for personhood and dignity (e.g. Australian Palliative Residential Aged Care (APRAC) Project, 2004;van der Steen et al., 2014). These goals rely on the cultivation of relevant knowledge, skills and personal qualities in the LTC sector. At this time, relatively little is known about the status of attributes such as these within the LTC workforce. 2 The main objective of this research is to describe the capacity of employees working within Canada's LTC sector to embrace a palliative care mandate, with particular attention to some of the personal qualities that are necessary for this work. Given evidence that some personal qualities vary by occupational role , an additional goal is to describe strengths and weaknesses in readiness for palliative care across occupational groups. This work builds on previous qualitative work within our team that identified challenges implementing a palliative approach to LTC (Kaasalainen et al., 2007;Sussman et al., 2017).

Personal qualities and palliative care
Despite the need for additional research in long-term care settings, the personal qualities required for palliative care work have a relatively long history of study in other settings, such as hospice care, oncology and community palliative care. Among the qualities currently receiving greatest attention are employee wellbeing, self-efficacy and person-centredness.

Employee wellbeing
Employee wellbeing is a measure of an employee's emotional response to work. In palliative care, a whole-person approach to care is encouraged, with balanced attention to physical, psychological, social, and spiritual needs (APRAC, 2004;Dobrina et al., 2014;van der Steen et al., 2014). This requires a personal, emotional engagement in work (Meier et al., 2001). At the same time, over the years, there has been concern that ongoing requirements to emotionally invest in work might undermine employee wellbeing (Maslach et al., 2001). Burnout is the most widely measured aspect of employee wellbeing; however, more recently, other negative experiences, such as secondary traumatic stress, and other positive experiences, such as compassion satisfaction, have also been evaluated. Burnout Burnout has been described as a 'progressive loss of idealism, energy, and purpose experienced by people in the helping professions as a result of the conditions of their work' (Edelwich and Brodsky, 1980: 14). Although evidence following decades of research seems to point to a risk of burnout in human service professions (Maslach et al., 2001), it is also clear that there is variability, with some providers remaining resilient to burnout while others succumb (Vachon, 1995;Ramirez et al., 1998). While burnout need not be a fact of palliative care work, some setting and occupational factors are known to elevate risk. For instance, although palliative care physicians are less likely to experience burnout than other physicians, palliative care nurses experience a rate of burnout similar to nurses working in other settings. Yet, both physicians and nurses who work in oncology care experience more burnout than those who work in other health-care settings (Martins Pereira et al., 2011). It is important to understand how occupational roles and contexts contribute to burnout, since burnout is associated with negative work experiences, absenteeism, intention to leave, high turnover (Ablett and Jones, 2007;Stamm, 2010) and poorer quality of care (Maslach, 1976;Astrom et al., 1991;Jenkins and Allen, 1998;Todd and Watts, 2005;Shinan-Altman and Cohen, 2009).
The potential for burnout has been well recognised in LTC contexts (Hare and Pratt, 1988;Kennedy, 2005). Several studies draw attention to a high workload in LTC (e.g. Bowers et al., 2000;Lopez, 2006;Mallidou et al., 2013;Morgan et al., 2008), emotional demands of LTC work (e.g. Abrahamson et al., 2009), and challenges associated with dementia care (e.g. Opie et al., 2002;Fuchs-Lacelle et al., 2008), all of which operate as risk factors for burnout (Maslach et al., 2001;Fuchs-Lacelle et al., 2008;Abrahamson et al., 2009). Beyond the general relevance of burnout for employee's work-related quality of life and for resident care, unrecognised burnout in LTC may also have implications for the acceptance of a palliative care mandate. For example, in one study, higher levels of burnout among LTC staff were associated with significantly lower interest in participating in palliative care education (Frey et al., 2015).

Secondary traumatic stress
In both palliative care and LTC contexts, staff members regularly confront death (Vachon, 1998), and thus their own mortality (Ablett and Jones, 2007). The cumulative effects of such work, and the indirect trauma associated with 'bad deaths' (Costello, 2006), can operate as significant stressors (Perez et al., 2015). The concept of secondary traumatic stresspost-traumatic stress disorder-like effects resulting from some types of care workis relatively new and somewhat controversial (Baird and Kracen, 2006;Devilly et al., 2009). Nevertheless, it has become a mainstay of recent research on the emotional wellbeing of the health-care workforce (e.g. Beck, 2011;Elwood et al., 2011). In studies of hospice care workers, prevalence of post-traumatic stress disorder-like symptoms above the threshold for concern is as high as 37 per cent (Quinal et al., 2009). Like burnout, secondary traumatic stress reduces work-related quality of life (Stamm, 2010). Although its relevance to patient care has not been established, downstream effects seem likely on the basis of construct overlap with burnout (Devilly et al., 2009). As LTC increasingly becomes a palliative care context, it is important to understand the rate of secondary traumatic stress, and predictors of this experience, so appropriate strategies can be introduced to manage risks to mental health, work-related quality of life and care quality.

Compassion satisfaction
Despite the risk of burnout and secondary traumatic stress in health-care work, the labour force is generally resilient (Devilly et al., 2009), and many health-care employees derive considerable satisfaction from their work (e.g. Nolan et al., 1994). Palliative care employees also speak about positive outcomes from regular exposure to death, saying that their experience helps them to cultivate a more person-centred focus and pay attention to their spiritual development (Ablett and Jones, 2007;Sinclair, 2011). Work founded in compassion, such as palliative care work, is known to increase job satisfaction and reduce employee turnover (Sinclair et al., 2016). In LTC, employees report enjoying the relationships they cultivate in their work, and the fulfilment that comes from helping meet the needs of residents (Eldh et al., 2016;Hunter et al., 2016a).
To summarise, both palliative care and LTC work rely on emotional engagement with patients, and this can operate both as a satisfier (Nolan et al., 1994;Ablett and Jones, 2007;Sinclair, 2011) and an occupational hazard (Wakefield, 2000;Vis et al., 2016). Investigations of the emotional wellbeing of the LTC workforce have included a consistent focus on burnout (e.g. Westerman et al., 2014), and a lessconsistent focus on other aspects of emotional wellbeing, such as compassion satisfaction and secondary traumatic stress (Stamm, 2002). A greater understanding of the emotional readiness of the LTC workforce would aid recognition of strengths and limitations in the capacity of the LTC workforce to embrace a palliative care mandate.

Self-efficacy
Like emotional aspects of work, health providers' self-efficacy, or self-perceived confidence in abilities, influences health-care quality (Bandura, 1986). Self-efficacy can apply to very circumscribed skills, such as inserting catheters (Ngo and Murphy, 2005), or to more complex social behaviours, such as interacting with cognitively impaired older adults (Fry et al., 2015) and delivering palliative care (Desbiens et al., 2012). Although measures of self-perceived competence and efficacy were recently developed for palliative care practice (e.g. Desbiens and Fillion, 2011;Lazenby et al., 2012), as yet, little is known about self-efficacy in the palliative care and LTC labour forces. One study suggests that palliative care knowledge and self-efficacy are at lower than desirable levels in the German LTC workforce (Pfister et al., 2013). In the Canadian LTC context, comfort when working with dying residents appears to vary by occupational group .
Meta-analytic research from diverse occupational contexts suggests that selfefficacy is generally positively and strongly related to work performance outcomes (Stajkovic and Luthans, 1998). Self-efficacy also has important implications for employee wellbeing; for instance, it is known to moderate the relationship between work stress and burnout (Smolen-Hetzel, 2010). Although self-efficacy often grows through experience (Simons et al., 2016), it can also be promoted through management interventions that focus on staff empowerment (Manojlovich, 2005a(Manojlovich, , 2005b and education (Ngo and Murphy, 2005;Phillips et al., 2011;Moir et al., 2015). As a result, assessing current self-efficacy for palliative care in LTC is vital to planning for a palliative care mandate.

Person-centredness
Common to palliative care models is an emphasis on the unique needs of each patient, attention to the patient as a whole person and concern for patient autonomy (Dobrina et al., 2014). These features of palliative care models are consistent with a humanistic or person-centred approach to palliative care; in fact, some consider person-centred care to be part of the very definition of palliative care (Higginson, 1999;World Health Organization, 2011). Resources are currently being developed to support person-centred palliative care (e.g. Ewing et al., 2015) and the effects of person-centred palliative care are beginning to be documented in randomised controlled trials (e.g. Brännström and Boman, 2014). Person-centred palliative care has benefits for health providers as well as patients. For example, job satisfaction is higher among palliative care nurses working in person-centred organisations (i.e. organisations that consult actively with employees; Fillion et al., 2007) In LTC contexts, person-centred care is accepted as integral to quality of care, and available evidence, including RCT evidence, generally supports this (Brownie and Nancarrow, 2013). Person-centred care is of particular importance to individuals with cognitive impairment (Hunter et al., 2013(Hunter et al., , 2016a(Hunter et al., , 2016b, who represent the majority of LTC residents, since beliefs about cognitive impairment can diminish respect for human dignity (Kitwood, 1997) and potentially influence treatment plans (Hunter et al., 2013). Positive outcomes of person-centred dementia care have been observed in randomised controlled trials (RCTs; Chenoweth et al., 2009). One RCT suggests that the benefits of person-centred care extend to LTC staff; in particular, emotional exhaustion decreased after implementing person-centred care strategies (Jeon et al., 2012). Organisational factors (including the general environment of the LTC home, collaboration and staff empowerment, including supervisory or organisational support) are known to influence the extent to which staff report providing person-centred care, suggesting that this quality may be responsive to intervention (Caspar and O'Rourke, 2008;Hunter et al., 2016b). Given its importance to people approaching the end of life, their family members and health providers, the degree to which care is person-centred is an important consideration in planning for a palliative approach to LTC.

Research question
The research question guiding this study is: 'Are employees working within Canada's LTC sector ready for a palliative care mandate?' More specifically, we set out to identify strengths and gaps in palliative care readiness among a sample of professional and non-professional LTC employees by measuring current levels of employee wellbeing, palliative care self-efficacy and self-reported person-centred care behaviours. We then considered the implications of these results for the development of a palliative approach to LTC.

Design
This study is part of a larger mixed-methods project exploring the implementation of a palliative programme in four participating LTC homes using a participatory action approach. At the outset of this larger project, we conducted a baseline survey with LTC employees, and this paper reports on the results of that survey. A follow-up survey is planned post-implementation. The survey focuses on personal qualities of LTC employees as relevant to palliative care.

Setting
The survey was conducted in four separate LTC homes in the Canadian provinces of Alberta, Saskatchewan, Manitoba and Ontario. The homes varied in size, housing 128, 50, 104 and 284 residents, respectively. Three of the homes were privately owned by the same company. One was a non-profit entity. Hereafter, to avoid identifying results as associated with particular homes, we refer to the sites simply as Sites 1-4 in a randomly determined order.

Participants
We recruited a convenience sample of staff employed at LTC facilities. Since all LTC employees are likely to be involved in supporting palliative care in some fashion, the only inclusion criterion was employment at the LTC facility. There were no exclusion criteria. Participants included managers and administrators, clerical staff, licensed professional staff (e.g. physicians, nurses, social workers, occupational therapists), nursing care assistants, and other unregulated staff such as housekeepers, recreation workers, maintenance staff and kitchen staff.

Procedure
We worked with participating LTC homes to recruit participants. All current employees were invited to participate in the survey either personally or through workplace mail. A modified Dillman (1978) approach was used to maximise the response rate. That is, we tracked staff who completed the survey, and followed up with those who did not by extending another invitation. The maximum number of invitations was three. To encourage participation further, we held a draw at each of the participating LTC homes and told staff that they would be entered to win a Can $50 gift card on completing the survey. Surveys were collected in the manner each LTC context identified as most appropriate to their setting, such as in a locked drop box.

Measures
There were three constructs of interest: emotional readiness, palliative care selfefficacy and person-centred care. We used the Professional Quality of Life scale (ProQOL; Stamm, 2010), the End-of-life Professional Caregiver Survey (EOL-PC; Lazenby et al., 2012), the Person-Directed Care scale (PDC; White et al., 2008) and the Environmental Support for Person-Directed Care scale (ES-PDC; White et al., 2008) to measure these constructs. Participants also completed a demographics questionnaire.

Employee wellbeing
The 30-item ProQOL version 5 (Stamm, 2010) measures three aspects of professional quality of life (i.e. employee wellbeing) across helping professions: compassion satisfaction (pleasure derived from doing work well), burnout (feeling overwhelmed by work) and secondary traumatic stress (problems resulting from exposure to others' trauma). Responses are based on a five-point Likert-type scale. In prior studies, the ProQOL sub-scales have shown good internal consistency reliability, with Cronbach alpha coefficient estimates of 0.88, 0.75 and 0.81, respectively (Stamm, 2010). In this study, the estimates were similar, at 0.88, 0.74 and 0.81, respectively. Cut scores have been established for each of the three ProQOL sub-scales. For compassion satisfaction, scores lower than the cut score of 22 indicate a concern, whereas for both burnout and secondary traumatic stress, scores above 42 indicate a concern (Stamm, 2010).

Palliative care self-efficacy
The EOL-PC (Lazenby et al., 2012) is a list of 28 statements describing self-efficacy in providing end of life care. The survey comprises three factors: patient and family centred communication (12 items; e.g. 'I can assist family members through the grieving process'); cultural and ethical values (eight items; e.g. 'I am able to deal with my feelings related to working with dying patients'); and effective care delivery (8 items; e.g. 'I am familiar with palliative care principles and national guidelines').
Responses are based on a five-point Likert-type scale. In prior research, each subscale has shown good internal consistency reliability, with Cronbach alpha coefficient estimates of 0.89-0.95 (Lazenby et al., 2012). In this study, the estimates were similar, at 0.85-0.93.

Person-centred care
The 35-item PDC (White et al., 2008) contains five sub-scales that measure selfreported person-centred care. The three sub-scales chosen for this study have high overlap with palliative care domains (van der Steen et al., 2014): Knowing the Person (e.g. PDCknowing; seven items; e.g. 'Thinking about the people in your care, for how many do you know their feelings about dying?'), Comfort Care (PDCcomfort care; eight items; e.g. 'Thinking about the people in your care, for how many can you minimise or ease pain?') and Support for Relationships (PDCfoster relationships; six items; e.g. 'Thinking about the people in your care, how often are you able to keep them connected to their families?'). Responses are based on a five-point Likert-type scale. The Cronbach alpha coefficient estimates for the five sub-scales reported in the previous research were 0.91, 0.88 and 0.91, respectively (White et al., 2008). In this study, estimates were similar: 0.90, 0.87 and 0.90, respectively.

Demographics questionnaire
We also collected the following demographic information: age, gender identity, education level, occupation, employment details (e.g. full time), and years of experience in LTC and within the current LTC facility.

Analysis
Prior to analysis, participants were identified as belonging to one of two groups. The first group, non-professional staff, was comprised of unregulated direct care staff and other support staff (e.g. including nursing assistants, recreation workers, dietary aides and housekeepers); the majority of this group were nursing assistants. The second group, professional staff, was comprised of clinical/therapeutic staff (e.g. regulated professional staff, managers, spiritual care providers and recreation therapists); the majority were nurses. It was not possible to make further distinctions among occupational groups without compromising statistical power.
To answer the research question, 'What are some strengths and gaps in palliative care readiness in an LTC context, and how do these vary by occupational role?', we calculated the mean, standard deviation (SD) and score range for each outcome variable by occupational group. We compared these mean values to established cut scores when available; in these cases, documenting the proportion of concerning cases. Independent-samples t-tests were then conducted to evaluate whether these outcomes varied by occupational group (i.e. professional or non-professional).
A further series of analyses was used to test the association of demographic variables (gender identity, education level, age and work experience) with outcomes of interest. First, to explore any potential association between years of work experience and the personal qualities under study, we calculated Pearson's correlation coefficients. Next, independent-samples t-tests were conducted to evaluate whether outcomes varied by gender identity (i.e. female versus all other identities) or education level (i.e. university degree or no degree). Since t-tests assume homogeneity of variance, we substituted Welch t-tests whenever variances were not homogeneous. Finally, one-way analysis of variance (ANOVA) and post-hoc Tukey's HSD tests were used to test whether outcomes varied across four age bands (see Table 1). Since ANOVA assumes homogeneity of variance, the Brown-Forsythe test (with post-hoc Games-Howell tests) was used whenever this assumption was not met. We used IBM SPSS Statistics version 24 for all analyses.

Results
A total of 228 questionnaires were returned. Response rates varied from 29 to 53 per cent across Sites 1-4, representing an overall response rate of 32 per cent. Of the returned questionnaires, 76 (33%) were returned by professional staff and 150 (66%) by other staff. Among the participants, 192 (84%) were female, 21 (9%) were male, one (0.4%) did not identify as either male or female and 14 (6%) did not specify a gender identity. Age was given as a range and more than 50 per cent of participants were 45 years and older (see Table 1). The average participant had worked for 12 years (SD = 9.4) in LTC and nine years (SD = 8.7) in the place they were currently employed. Casual employees represented 11 per cent of respondents, part-time employees 31.3 per cent and full-time employees 57.6 per cent. Professional staff generally had a higher education level (53.9% with university degrees versus 19.7% for non-professional staff). Note: Twenty-four (10.5%) participants did not report an age.
Cut-off scores were available for the three domains of employee wellbeing assessed by the ProQOL. Among professional staff, no one scored below the threshold for compassion satisfaction or above the threshold for burnout; however, one individual scored above the threshold for secondary traumatic stress. Among the nonprofessional staff, one individual had a compassion satisfaction score below threshold and one person had a secondary traumatic stress score above threshold, but no one had a burnout score above threshold. Table 2 shows means for each outcome variable against maximum scores, and, where available, accepted threshold or cut-off values. Results are presented for each occupational group. On average, staff in each occupational group scored well below accepted thresholds for burnout and secondary traumatic stress, and well above a threshold for low compassion satisfaction.
t-Tests confirmed statistically significant ( p < 0.05) differences between professional and non-professional staff for all self-efficacy sub-scales, and on compassion satisfaction and burnout, but not on any of the person-centred care sub-scales (see Table 3). In all cases, results favoured professional staff.
As Table 4 shows, there was no correlation between work experience and any of the outcome variables for either occupational group. To evaluate whether outcomes varied by gender identity (i.e. female versus all other identities) or education level (i.e. any university degree or no degree), we conducted a further series of independent-samples t-tests. Given that previous analyses showed differences between the occupational groups, these tests were conducted separately for professionals and non-professionals. No differences were observed across the two gender identity groupings, and this was true for both occupational groups ( p > 0.05). No statistically significant effects of education were observed within the professional group (all p > 0.05). Within the non-professional group, however, there was an effect of education on one domain of person-centred care, comfort care, Welch's t(72.38) = 3.56, p < 0.05, and the result for burnout was at the threshold of statistical significance, Welch's t(71.62) = 1.99, p = 0.05. Both results favoured those with a university degree. Finally, tests of age differences across the outcomes of interest were conducted separately for professionals and non-professionals. There were no statistically significant differences attributable to age ( p > 0.05).

Discussion
We set out to explore the readiness of Canadian LTC homes for a palliative care mandate by exploring personal qualities that have been identified as relevant to palliative care work. These qualities included palliative care self-efficacy, person-centred care and emotional readiness. We also examined whether these personal qualities varied by occupational role, and other employee characteristics such as age and work experience. On the basis of this work, a number of important observations can be made about strengths and gaps in readiness for palliative care in the Canadian LTC context.

Emotional readiness for a palliative care mandate
Our data gave no indication that employee wellbeing is lacking in the Canadian LTC context. Scores below accepted thresholds were very rare in our sample, and average scores differed substantially from threshold values. This is encouraging, Notes: PDC: Person-Directed Care scale. ProQOL: Professional Quality of Life scale. Max: maximum scale score. ET: established threshold (for compassion satisfaction, scores lower than threshold indicate a concern; for burnout and secondary traumatic stress, scores above the threshold indicate a concern). SD: standard deviation.
given other reports that seem to suggest a high potential for burnout among human service workers (Maslach et al., 2001), including LTC staff (Woodhead et al., 2016).
Although it is possible that movement towards a palliative care mandate could change this fact and contribute to increased burnout, this is very unlikely, given that LTC employees are already providing terminal care. Moreover, in the palliative care field, initial concern about burnout was followed by evidence that burnout actually can be lower than in other professional settings (although this is not universally true) (Vachon, 1995).  Within palliative care contexts, where employee wellbeing has been more robustly studied than in long-term care contexts, social support, a realistic workload and involvement in workplace decision making have been identified as resources that can help to protect employees' emotional wellbeing (Vachon, 1995). Despite the robust levels of LTC employee wellbeing suggested by the current results, it seems important to explore whether resources such as social support, a realistic workload and involvement in workplace decision making might assist LTC employees in transitioning to a palliative care mandate. Some researchers have documented unrealistic workloads in LTC (Bowers et al., 2000;Lopez, 2006;Dellefield et al., 2012;Qian et al., 2012) and high levels of turnover, a variable known to be related to workload (Ablett and Jones, 2007). Others have shown that non-professional staff are excluded from decision making (Ribbe et al., 1997;Daly and Szebehely, 2012;Kaasalainen et al., 2017). Furthermore, qualitative work suggests that direct care staff may need support for grief as they begin to embrace a palliative care mandate , perhaps especially in the first few years of work (Zimmerman et al., 2005). Thus, continued attention to workforce wellbeing seems essential to guarantee the mental health of the LTC workforce over time, and particularly in the context of a new palliative care mandate.
Personal readiness for palliative care by age and work experience Neither age nor work experience bore significant correlations with any of the personal qualities we studied. All observed correlations were small and not statistically significant, with many approaching zero. Since acuity of care has increased over time in Canadian LTC homes, which are increasingly a place for end-of-life care (Frohlich et al., 2002;Palliative Alliance, 2010), it is very encouraging that those with more work experience are not showing any sign of an adverse reaction (such as burnout) in response to evolving changes in the LTC sector or to the accumulation of experience with care provision and loss.
Given that confidence and skill generally grows with work experience (Simons et al., 2016), it is somewhat surprising that work experience was not associated with self-efficacy and person-centred care. If this result is replicated in other settings, self-efficacy and skill in person-centred palliative care may need to be supported more directly; for instance, through continuing education and other workplace interventions (see e.g. Pan et al., 2016;Thompson et al., 2016), as LTC embraces a palliative care mandate (Venturato and Drew, 2010). In this study, the finding that non-professional staff with higher levels of education endorsed higher levels of comfort care lends support to this interpretation.

Personal readiness for palliative care by profession
In line with prior research suggesting that occupational characteristics influence personal qualities such as burnout (Martins Pereira et al., 2011), we also observed differences among professional and non-professional staff in efficacy and burnout, but not in person-centred care.
With regard to palliative care efficacy, professionals had a clear advantage over non-professionals, but both groups showed potential to grow. This corresponds well to previous survey results documenting palliative care education needs among staff in LTC, which showed that professionals had greater knowledge than support staff, but nevertheless had knowledge gaps . With appropriate training and support in place to improve skill and efficacy, professional staff may be well-positioned, by virtue of their greater efficacy, to mentor other staff to increase skill and confidence in palliative care. Nevertheless, given their close involvement with dying residents, it is critical to ensure that non-professional staff in direct care roles are included in initiatives to cultivate palliative care-related skills and efficacy in LTC. Promising initiatives are under development, but need further evaluation. These include job shadowing in hospice settings , attending inter-disciplinary comfort care rounds to promote shared communication about palliative care in LTC (Wickson-Griffiths et al., 2014), professional leadership (Manojlovich, 2005c) and capacity development (Kelley and McKee, 2013).
With respect to person-centred care, while both groups evidenced room to improve, there was no advantage for professional staff over non-professional staff. This mirrors other research suggesting that non-professional staff members value person-centred care, although they experience some challenges in providing it (Hunter et al., 2016a). Programmatically cultivating an organisation-wide person-centred approach in LTC is possible, and is associated with satisfaction and intention to remain in the work setting (McCormack et al., 2010). More specific approaches to cultivating humanistic care (e.g. assurance of adequate staffing, availability of educational opportunities that match employees' current needs, promoting effective inter-professional communication) are also associated with positive outcomes in health-care organisations (Gunnarsdottir et al., 2007;Manojlovich and Laschinger, 2007;McCormack et al., 2010;Spreitzer et al., 2010). Overall, realising a greater potential for person-centred approaches to LTC is likely to have benefits for residents, staff and LTC organisations.
Burnout was low and compassion satisfaction high in this sample. For a long time, both research and clinical lore have emphasised disadvantageous aspects of work in health care and LTC, such as the potential for stress and burnout. More recent research suggests that despite a high workload and the presence of ongoing stressors, LTC employees are generally satisfied with their work, and have low levels of burnout (McCormack et al., 2010;Lehuluante et al., 2012). Some have noted that these and other rewards of working in LTCsuch as the clinical challenges of working with residents with complex needs, or the personal meaning associated with promoting comfort and relationship at the end of lifehave not been sufficiently explained to graduating nursing students, very few of whom have been directly exposed to LTC environments during their training (Canadian Nurse, 2011). Our results and others suggesting generally positive experiences of work in LTC contexts might be used in ongoing efforts to address LTC workforce recruitment issues.
Nevertheless, some of our results suggest that more work can be done to promote positive outcomes among non-professional staff. For instance, nonprofessionals had significantly more symptoms of burnout than professionals, and experienced less compassion satisfaction. Other work suggests that promoting positive outcomes among non-professional staff involves the cultivating of effective inter-professional communication, teamwork and support (Eaton, 2001;Harahan et al., 2003). Canadian LTC environments have tended to emphasise a task-oriented, role-differentiated approach to work (Cott, 1997;Daly and Szebehely, 2012). Other research also documents the exclusion of non-professional staff from team decisions (Ribbe et al., 1997;Daly and Szebehely, 2012). This is worrisome, given that nursing assistants make up the majority of the LTC workforce and are responsible for most interactions with LTC residents. In contrast to a role-differentiated model of care, palliative care has traditionally relied on more flexible and collaborative inter-professional relationships (Dobrina et al., 2014). Thus, identifying ways to improve team-based collaboration and communication in LTC might be worthwhile in promoting a palliative approach to care in LTC.

Research strengths and limitations
As with any research study, there are some limitations. First, given the descriptive nature of this study, the ANOVAs were not corrected. Thus, there is some risk that the number of comparisons we conducted increases the probability of observing at least one statistically significant result by chance (Type I error). Second, any comparisons we made are pseudo-experimental, meaning that any intimation of cause-effect relationships should be very cautiously evaluated. Thirdly, while our response rate borders on that reported on average for other organisational science research (Anseel et al., 2010), our sample included just four Canadian LTC homes and only one-third of potential respondents within those homes; thus, results may not be fully representative of Canadian LTC employees' experiences. Similarly, our sample of homes was too small to analyse meaningfully how differences in organisational structure or culture might influence the outcomes we considered in this study. Finally, we have noted that a university education is relevant to some of our results, yet have left unexplored the potential effects of tailored forms of learning, such as supervision or training in the practice of palliative care. Alongside these limitations, our study has the following strengths: we obtained an inter-provincial sample and our sample size provided adequate power for our planned analyses.

Conclusion
We set out to document strengths and gaps in the readiness of Canadian LTC homes for a palliative care mandate. Our results suggest that some strengths include the emotional readiness of LTC employees (i.e. low levels of burnout and secondary traumatic stress, and robust levels of compassion satisfaction), and engagement in person-centred care among both professionals and non-professionals. Identified weaknesses included the existence of some systematic differences in self-reported person-centred skills and compassion satisfaction across the four participating homes, room to grow in self-efficacy and person-centred care, and greater levels of emotional readiness and self-efficacy among professionals, as compared to nonprofessionals. Supporting the adoption of a palliative care mandate in LTC has the potential to improve palliative care-related outcomes including person-centredness and self-efficacy. Given an emphasis on inter-professional collaboration in palliative approaches to care, a shift towards a palliative approach to care might also contribute to more equitable professional and non-professional staff outcomes.
Author contributions. PVH, LM, SK, TS, GT, LV and AW-G conceptualised the study, co-designed the survey and oversaw data collection. PVH analysed the survey data in consultation with LM, SK and NA-D. PVH wrote the initial draft of the manuscript in consultation with LM. All authors edited the manuscript. PVH formatted the article for publication. PVH is guarantor. All authors read and approved the final manuscript.
Financial support. This work was supported by a Canadian Institutes for Health Research Partnerships for Health Systems Improvement grant.
Conflict of interest. The authors declare no conflicts of interest.
Ethical standards. The study protocol was reviewed and approved by seven university-affiliated or integrated ethics boards (Hamilton Integrated Research Ethics Board, No. 0427;Brock University, No. 15-102;McGill University, No. 281-1214;University of Saskatchewan, No. 15-270;University of Regina, No. 15-190;University of Manitoba, No. H2015:374;University of Calgary, No. 15-2277) and by local health authorities or organisations as required. Study participants reviewed an information form that reviewed all known requirements, risks and benefits associated with participating. Choosing to proceed with the survey implied free, informed consent.