End-of-Life Care Discussion for Residents With Dementia in Long-Term Care Facilities

Abstract

Background: 

Healthcare professionals play an important role in end-of-life care decisions. Although the factors affecting end-of-life care in long-term care facilities have been well studied in Western contexts, few studies have examined factors affecting end-of-life care discussions for residents with dementia among multidisciplinary professionals in Asian or Taiwanese contexts.

Purpose: 

The aim of this study was to explore the factors related to the end-of-life care discussions of registered nurses, social workers, and physicians with residents with dementia and their families.

Methods: 

This is a cross-sectional and correlational study. A stratified random sampling was conducted, and self-administered questionnaires were mailed to physicians, registered nurses, and social workers who work in long-term care facilities with accreditation ranks ≥ Grade A in Taiwan.

Results: 

Questionnaires were returned by 478 participants (response rate = 94.1%). Fewer than 10% of the participants reported often discussing end-of-life care with residents with dementia and their families. Physicians were significantly more likely to discuss life-sustaining treatment than nurses and social workers. The frequency of professionals discussing advance directives and life-sustaining treatment for residents with dementia was predicted by higher confidence in end-of-life care skills and greater perceived facility support.

Conclusions/Implications for Practice: 

Most professionals did not frequently discuss end-of-life care with residents with dementia and their caregivers. Discussions of end-of-life care in long-term care facilities should be promoted by enhancing the confidence of these professionals in providing end-of-life care and creating a supportive environment for end-of-life discussions and care.

Introduction

Dementia is not only a major determinant of nursing home placement (^{Lithgow, Jackson, & Browne, 2012}) but also the most common cause of death in nursing homes (^{Gjerberg, Førde, & Bjørndal, 2011}). Indeed, approximately 70% of dementia-related deaths in the United States (^{Hicks, Rabins, & Black, 2010}) and 50%–92% of those in European countries occur in nursing homes (^{Houttekier et al., 2010}). However, end-of-life (EOL) care in these settings has been shown to be poor (^{Vandervoort et al., 2013}). Many nursing home residents with dementia (RWDs) are transferred to hospitals and receive intrusive medical intervention that has no demonstrated benefit to their quality of life or quality of death (^{Givens, Selby, Goldfeld, & Mitchell, 2012}).

People with advanced dementia at the end of their lives typically experience acute conditions that require family members and healthcare professionals to make decisions about life-sustaining or palliative care, particularly in long-term care facilities. The quality of EOL care for these RWDs and the satisfaction of their family caregivers with EOL care have been shown to improve significantly with advance care planning (ACP; ^{Givens, Jones, Shaffer, Kiely, & Mitchell, 2010}; ^{Liu, Guarino, & Lopez, 2012}). These findings highlight the need to improve EOL care and discuss ACP in long-term care facilities (^{Givens et al., 2010}). Nevertheless, most studies on this topic have been conducted in Western countries, and their outcomes have limited applicability in Asian contexts because of cultural differences and dissimilar medical systems. Taiwan is a newly developed Asian country in which the aging population and prevalence of dementia are increasing faster than those in most other developed countries (^{Wu et al., 2013}). As in Western countries, a high proportion of nursing home residents in Taiwan (64.5%) experiences dementia (^{Chen et al., 2007}). Therefore, providing EOL care to RWDs is an inevitable challenge for long-term care facility staff members in Taiwan and is a major component of their work.

EOL care involves many complex treatment decisions. However, people with dementia are often unable to participate in their own EOL care decisions. To ensure that the autonomy of residents is respected, the use of advance directives (ADs) is encouraged. Taiwan’s Legislative Yuan passed the Hospice Palliative Care Act in 2000, encouraging people to make ADs that reflect their preferences regarding whether to receive palliative or life-sustaining treatments (LSTs) at EOL (^{Chen, 2009}). Discussions of EOL care and ADs in Taiwan have typically focused on patients with cancer, and patients with dementia have often been neglected in this respect (^{Huang, Weng, & Yeh, 2011}). However, Taiwan’s National Health Insurance program has covered the hospice fees of EOL patients with eight noncancer chronic diseases, including advanced dementia, since 2009 (^{Huang et al., 2011}). This policy has not only raised awareness of dementia-related EOL care but also increased the availability of palliative care for RWDs. Discussion and communication are vital processes in these decisions, irrespective of the ultimate decision regarding ADs and palliative care (^{van der Steen et al., 2014}). However, to date, information is lacking concerning how professionals discuss EOL care with RWDs and their families and what factors influence these discussion processes.

For appropriate EOL care decisions to be reached, RWDs and their caregivers must be informed about their right to make ADs and communicate their concerns regarding palliative care and LSTs. The key professionals in long-term care facilities who communicate with RWDs and their family surrogates regarding EOL care are physicians (^{Ampe, Sevenants, Smets, Declercq, & Van Audenhove, 2016}; ^{van Soest-Poortvliet et al., 2015}), registered nurses (^{Rurup, Onwuteaka-Philipsen, Pasman, Ribbe, & van der Wal, 2006}), social workers (^{Lacey, 2005}), and management staff (^{Ampe et al., 2016}).

Understanding the factors that influence the practice of EOL care by these professionals for RWDs may enable policy makers and managers of long-term care facilities to provide more focused support. Moreover, this understanding may help ensure that these influencing factors are addressed to improve the quality of EOL care for RWDs as well as the satisfaction of family surrogates with EOL care decision-making. Physicians, nurses, and families of RWDs in the Netherlands were reported to have different attitudes toward EOL care decisions for RWDs because of differences in religious beliefs and responsibilities (^{Rurup et al., 2006}). Moreover, social workers in the United States reportedly perceive that discussing ADs with RWDs and their families is vital (^{Lacey, 2005}). A qualitative study (^{Casey et al., 2011}) and a mixed-method study (^{Marshall, Clark, Sheward, & Allan, 2011}) further suggested that the EOL care delivery of long-term care professionals in Ireland (^{Casey et al., 2011}) and New Zealand (^{Marshall et al., 2011}) was influenced by their perceptions of their working environment and organizational culture. Nevertheless, these studies were conducted in Western countries and lacked comprehensive quantitative data on the factors affecting the EOL care discussions of these key professionals for RWDs in long-term care facilities. Moreover, these studies have not addressed the importance of cultural sensitivity in EOL care discussions and the role of multidisciplinary cooperation in EOL care quality. These knowledge gaps necessitate that studies on EOL care discussions be conducted in a diverse range of cultural settings and with different categories of long-term care facility professionals. Therefore, the authors of this study hope that their findings may establish a future reference for developing culturally sensitive interventions to promote EOL care discussions for RWDs.

Purposes

The purpose of this study was to explore the factors related to how healthcare professionals approach EOL care for RWDs in long-term care facilities. Specifically, we aimed to (a) investigate the EOL care discussions for RWDs among healthcare professionals in long-term care facilities; (b) compare the EOL care knowledge, attitudes, confidence, perceived facility support, and practices of physicians, registered nurses, and social workers with regard to their discussing EOL care for RWDs; and (c) explore the factors influencing EOL care discussions for RWDs, particularly discussions about ADs and LSTs.

Methods

Design

A cross-sectional and correlational study was used to describe the current status of EOL care discussions for RWDs among healthcare professionals in long-term care facilities and to explore the factors associated with these EOL care discussions in Taiwan. For the sample to be representative and accessible, stratified random sampling and postal surveys were applied in this study.

Setting and Participants

In Taiwan, people with dementia reside in four types of long-term care facilities, including dementia-specific facilities, freestanding nursing homes, hospital-based nursing homes, and general long-term care facilities. Participants in the this study included physicians, registered nurses, and social workers at accredited long-term care facilities in Taiwan that were rated at Grade A or higher and had capacities of more than 50 beds (i.e., a middle-scale facility or larger). Long-term care facilities in Taiwan are accredited using five grades (from highest to lowest): excellent, A, B, C, and D. Eligible facilities were selected from two government accreditation reports: Nursing Home Accreditations (^{Ministry of Health and Welfare, Taiwan, ROC, 2013}) and Senior Welfare Institution Accreditations (^{Ministry of the Interior, Taiwan, ROC, 2012}). Because EOL care in long-term care facilities is a recent concept, we assumed that middle-scale facilities graded A and higher have more workers and resources to implement this care. The sample size was estimated using G*Power 3 (^{Faul, Erdfelder, Buchner, & Lang, 2009}). Setting the α level at .05, effect size at .2, and power at .90, we estimated the sample size to be at least 207. Assuming the manpower requirement of a middle-scale nursing home (50–100 beds) to be at least four to seven nurses (nurse/residents = 1:15), one part-time social worker, and one contract physician (^{Ministry of Health and Welfare, Taiwan, ROC, 2013}), 36–63 facilities would need to be recruited. Furthermore, we estimated that the number of eligible facilities must be doubled if the rate of facility agreement and participant response was 50%.

Eligible facilities were stratified by random sampling and weighted by the number of facilities in different areas of Taiwan (north, central, south, and east). Of the 201 eligible facilities in Taiwan, 63 were selected at random. The chief administrators of these 63 facilities were contacted about participating in the survey, with 15 refusing, yielding a final sample of 48 facilities (agreement rate = 76%). When the chief administrators agreed to take part in this survey and reported the numbers of each type of professional in their facilities, these professionals became the target sample. Five hundred twenty-nine potential participants (410 nurses, 71 social workers, and 48 physicians) were referred by the study facilities in 2011. Of the 529 questionnaires mailed to these potential participants, 498 were returned, yielding a return rate of 94.1%. Valid questionnaires were defined as those questionnaires that were at least 90% complete (n = 478).

Data Collection and Instruments

Data were collected from November 2010 to July 2011 using a standardized questionnaire that was mailed to physicians, registered nurses, and social workers in long-term care facilities in Taiwan. On the basis of information from the consenting chief administrators of targeted facilities, questionnaires were mailed to administrators with cover letters, informed consent forms, and self-addressed, prepaid envelopes. The administrators then sent a questionnaire, an informed consent form, and a self-addressed prepaid envelope to each potential participant. All of the participants were volunteers and remained anonymous, with this explained to them in the cover letter. Anonymity was assured by using no identifying code on the questionnaire. All participants were asked to send back the informed consent form and the completed questionnaire by prepaid mail if they agreed to participate.

Frequency of end-of-life care discussions for residents with dementia

The frequency of the discussion of EOL care for RWDs was assessed using a six-item scale on ACP with RWDs and their family surrogates (^{Lacey, 2005}). Items on this scale involved questions about explaining AD forms and do-not-resuscitate orders; discussing artificial nutrition, artificial hydration, and antibiotics; and helping people clarify their thoughts concerning LSTs. The six-item scale that was used in this study is divided into two subscales: a discussion of ADs (three items) and a discussion of LSTs (three items). Responses are rated on a 5-point Likert-type scale from 1 (never) to 5 (always), with the potential total score for each subscale ranging from 3 to 15 and higher scores indicating that professionals more often discuss ADs and LSTs with RWDs and their families. The internal consistency reliability (Cronbach’s α) of the Frequency of EOL Care Discussions for RWDs Scale was determined to be .71 (^{Lacey, 2005}). Because this scale is an English version and was being used in Taiwan for the first time, translation, back-translation, and content validity index (CVI) were performed to confirm accuracy and cultural sensitivity. The English version was translated into Chinese by the first author, and the Chinese version was then back-translated by an English speaker. The back-translated English version was again confirmed by the second author. The CVI and the pilot study were conducted to test validity and reliability. (The detailed procedures are described in the next paragraph.) In this study, the CVI of the scale was determined to be .97, with the two subscales earning Cronbach’s α values of .86 and .88. For the total scale, the Cronbach’s α value was .88, and the test–retest reliability was .83.

The next four scales were developed from four sources: (a) relevant studies (^{Lacey, 2005}; ^{Rurup et al., 2006}); (b) four focus groups with four to six long-term care facility nurses and social workers in each group; (c) content validity testing by six experts, including one professor of palliative care, one associate professor, one lecturer of gerontological nursing, two master-degree nurses, and one master-degree social worker, all of whom worked in dementia care units; and (d) a pilot study that was performed to test internal consistency reliability (Cronbach’s α or Kuder-Richardson Formula 20 [KR-20]) and 2-week test–retest reliability with 14 nurses and four social workers who worked in two long-term care facilities.

End-of-life care knowledge

Ten statements were constructed to test the EOL care knowledge of long-term care facility professionals. The scale involves five statements concerning LST, including cardiopulmonary resuscitation side effects, EOL hunger, dehydration, artificial nutrition, and antibiotics (^{Lacey, 2005}; ^{Rurup et al., 2006}); one statement concerning pneumonia prevention; and four statements on the Hospice Palliative Care Act and National Health Insurance payment in Taiwan. For each statement, the participants were required to answer “true,” “false,” or “do not know.” During processing by the statistical software, each “correct” answer was counted as 1; each “incorrect” or “do not know” answer was equated with having incorrect knowledge or a lack of knowledge regarding EOL care and counted as 0. Higher scores indicated a higher level of knowledge. The CVI of the EOL care knowledge scale was determined to be .96, and the internal consistency reliability (KR-20) for this study was .63.

Attitudes toward the end-of-life care decision-making rights of residents with dementia

The five-item Attitudes toward RWDs’ EOL Care Decision-Making Rights Scale was developed to measure the attitudes of professionals toward the right of RWDs to know about ADs, respect for RWDs’ decisions related to EOL care (e.g., regularly informing RWDs about how to fill out ADs when they are admitted to the facility), respect for RWDs’ choice to refuse food, the reliability of RWDs’ decisions to accept or refuse treatments, and the ability of residents with mild dementia to comprehend the meaning of ADs. Responses were rated on a 5-point Likert-type scale ranging from 1 (strongly disagree) to 5 (strongly agree). Total possible scores ranged from 5 to 25, with higher scores indicating greater respect for RWDs’ decision-making rights and belief in the reliability of RWDs’ EOL care decisions. The CVI of attitudes toward RWDs’ EOL Care Decision-Making Rights Scale was determined to be .95, the Cronbach’s α was .68, and the 2-week test–retest reliability was .66.

Confidence in end-of-life care and communication skills

The confidence of professionals in their ability to manage RWDs’ problem behaviors and end-stage symptoms and in communicating with RWDs and their families about EOL care was assessed by applying the researcher-developed, 17-item Confidence in EOL Care and Communication Skills Scale. Responses were rated using a 4-point Likert-type scale ranging from 1 (no confidence) to 4 (highly confident). Total possible scores ranged from 17 to 68, with higher scores indicating greater confidence in management and communication regarding EOL care for RWDs. The CVI of the confidence in EOL care and communication skills scale was .92, the Cronbach’s α was .97, and the 2-week test–retest reliability was .77.

Perceived facility support for end-of-life concerns

The perceptions of the participants regarding the support of their facilities on EOL issues were assessed using a researcher-developed 10-item Perceived Facility EOL Support Scale. This scale assesses professionals’ perceptions of administrators’ support for providing palliative care to RWDs and specifically addressed issues such as whether the facility has an explicit policy to forgo LST; if the facility educates or prepares its workers to provide palliative care; if the facility has a policy of routinely informing RWDs about ADs, a palliative care team, a special room for dying RWDs, and an available or cooperative referral hospice; if the facility has the ability to provide continuity of care to residents who are referred from a hospice unit and to build consensus among workers on palliative care for RWDs; and if the facility provides adequate palliative care to RWDs. Responses were rated using a 5-point Likert-type scale ranging from 1 (strongly disagree) to 5 (strongly agree). Total possible scores ranged from 10 to 50, with higher scores indicating greater perceived support from the facility. The CVI of perceived facility support for EOL concerns scale was .92, the Cronbach’s α was .91, and the 2-week test–retest reliability was .76.

Demographics and related variables

Data were collected on the participants regarding their age, gender, level of education, work experience (years), and participation in continuing education on dementia and EOL care during the previous 2 years. If participants had taken part in continuing education, they were asked to report the number of education hours. The facilities where participants worked were categorized by type: dementia-specific facility, freestanding nursing home, hospital-based nursing home, and general long-term care facility.

Ethical Considerations

Before data collection, this study was approved (99–1005B) by the institutional review board of Chang Gung Memorial Hospital.

Data Analysis

All of the study variables were analyzed using predictive analytics software (IBM SPSS Statistics 22.0). Frequencies, percentages, means, standard deviations, and ranges were included as descriptive statistics to identify the demographic characteristics of participants, discussions about EOL care, and other relevant main variables. Frequency distributions of discussions about EOL care were measured using chi-square tests, mean-difference-related main variables with one-way analysis of variance, and Scheffe’s post hoc analysis to determine any possible differences among the various professional groups. Factors that were identified as significantly associated with the practices of discussing EOL care for RWDs were analyzed using two multiple regression models, and all of the variables were entered into each regression model by forced entry. Model 1 was used to predict the factors that were associated with discussing ADs, and Model 2 was used to predict the factors that were associated with discussing LSTs. Moreover, p values of < .05 were considered statistically significant.

Results

Participant Characteristics

Of the 478 participants, 81.0% were registered nurses, 14.2% were social workers, and 4.8% were physicians. Their mean age was 35.5 ± 9.5 years. Approximately 93% of the survey participants in the sample were female. Most of the male participants were physicians. More than half of the participants had a bachelor’s degree or higher (51.3%). However, in the nurse group, only 40.3% had a bachelor’s degree or higher. Over half (62.6%) of the participants reported having never been exposed to curricula on dementia care in school, and exactly half (50.0%) had never taken curriculum-based lessons on EOL care in school. Most of the participants reported participating in continuing education on dementia care (77.2%) and EOL care (69.9%) during the previous 2 years. The mean number of hours of continuing education on dementia and EOL care was 10.0 ± 21.7 and 7.0 ± 16.4, respectively. The data on the characteristics of the participants and facilities are summarized in Table 1.

TABLE 1.: Participants’ Demographic Characteristics by Profession (N = 478)

Most participants did not frequently discuss EOL care with RWDs or their families, with fewer than 10% reporting often or always providing documents about ADs, explaining how to fill in related documents, and discussing the benefits and risks of artificial nutrition and antibiotics. Less than 15% of the participants reported that often or always discussing items would facilitate decision-making regarding do-not-resuscitate orders (11.5%) and clarify problems related to life-extending treatments (11.5%). Although the three professional groups did not differ significantly regarding the three items of discussions about ADs with RWDs’ families, the physician and nurse participants differed significantly from social workers in the three items of discussions about LSTs (benefits and risks of artificial nutrition, χ² = 71.21, p < .001; benefits and risks of antibiotics, χ² = 83.20, p < .001; and clarify problems related to life-extending treatments, χ² = 23.28, p < .001; see Table 2).

TABLE 2.: Profile of Discussions About EOL Care for RWDs by Profession (N = 478)

Knowledge of, Attitudes Toward, Perceived Facility Support for, and Confidence in End-of-Life Care of Residents With Dementia

The mean scores for perceived knowledge related to EOL care for RWDs differed significantly among registered nurses, social workers, and physicians (F = 2.96, p = .013); physicians’ scores were significantly higher than those of social workers in a post hoc test. However, these professionals did not differ significantly in their attitudes toward RWDs’ decision-making rights (F = 2.26, p = .106) and perceived support from their facilities (F = 0.03, p = .973). The mean confidence score of social workers was significantly lower than that of registered nurses and physicians (F = 5.85, p = .003), and physicians and nurses’ scores were significantly higher than those of social workers in a post hoc test. Regarding discussing EOL care for RWDs, the three professional groups did not differ significantly in their frequency of discussing ADs (F = 0.40, p = .674), but they did differ significantly in their frequency of discussing LSTs (F = 29.48, p < .001). Physicians’ scores were significantly higher than those of the nurses, and nurses’ scores were also significantly higher than those of the social workers in a post hoc test (Table 3).

TABLE 3.: Main Study Variables by Profession

Factors Related to the Practice of Discussing End-of-Life Care for Residents With Dementia

The associations among all of the variables for discussing ADs and LSTs for RWDs were examined using two multiple regression models. The first model indicated that significant predictors of the frequency of professionals’ discussions about ADs for RWDs included more hours of EOL-care-related continuing education (t = 4.25, 95% CI [0.02, 0.05], p < .001), more positive attitudes toward RWDs’ decision-making rights (t = 2.68, 95% CI [0.03, 0.22], p = .008), more confidence in their EOL care and communication skills (t = 4.67, 95% CI [0.03, 0.08], p < .001), and higher perceived support from their facilities (t = 5.09, 95% CI [0.07, 0.16], p < .001).

The second model showed that nurses (t = −2.78, 95% CI [−2.47, −0.42], p = .006) and social workers (t = −5.51, 95% CI [−4.48, −2.12], p < .001) were significantly less likely to discuss LSTs for RWDs than physicians. In addition, the significant predictors of the frequency of discussing LSTs for RWDs were more hours of EOL care continuing education (t = 2.54, 95% CI [0.01, 0.03], p = .012), more confidence in EOL care and communication skills (t = 6.42, 95% CI [0.05, 0.09], p < .001), and greater perceived support from their facilities (t = 2.87, 95% CI [0.02, 0.10], p = .004; Table 4).

TABLE 4.: Factors Associated With Frequency of Discussing EOL Care for RWDs

Discussion

On the basis of a review of the literature, this study is the first to comprehensively examine the self-reported knowledge, attitudes, confidence, and frequency of EOL care discussion among registered nurses, social workers, and physicians working in a random, stratified sample of long-term care facilities.

The findings reveal that a large proportion of Taiwanese long-term care professionals never or rarely discuss EOL care for RWDs, which are similar to the results reported regarding managers of Norwegian nursing homes (^{Gjerberg et al., 2011}). The low proportion of healthcare professionals discussing EOL care with RWDs may be attributed to various reasons. One reason may be the passage by Taiwan’s Legislative Congress of the Hospice Palliative Care Act in 2000. However, the Hospice Palliative Care Act has neither a regulation component requiring that ADs be introduced to nursing home residents nor an incentive strategy to promote ADs or ACP (^{Lo, Wang, Liu, & Wang, 2010}). Another reason may be that, in general, EOL care is discussed at the patient’s EOL or at the diagnosis of a severe illness. However, the disease trajectory of dementia differs from that of cancer. Prognostications in relation to dementia are difficult. Thus, dementia is not usually considered as a life-limiting disease (^{Huang et al., 2011}). A third reason may be that EOL care is still a taboo subject for Taiwanese people because they believe it will result in bad luck and unpleasant thoughts or that it means forgoing all treatment (^{Lo et al., 2010}). Therefore, if professionals perceive that they have insufficient communication skills or confidence to discuss EOL care with RWDs or their families, they will be inclined to avoid talking about this issue.

This study found that more EOL-care-related continuing education and greater perceived confidence in EOL care planning were associated with an increased probability of participants discussing ADs and LSTs with RWDs and their families. However, our results show that nearly one third of Taiwanese professionals working in long-term care have never had EOL-care-related continuing education. This finding corroborates earlier findings showing that lack of EOL care education was common among healthcare professionals in Western Europe (^{Lynch et al., 2010}; ^{Marshall et al., 2011}). We observed that professionals who perceived higher levels of support from their facilities conducted more EOL care discussions for RWDs. This result is consistent with previous reports that the provision of effective EOL care is greatly impacted by organizational culture (^{Casey et al., 2011}), work environment (^{Marshall et al., 2011}), and ACP policies (^{Ampe et al., 2016}; ^{van Soest-Poortvliet et al., 2015}). These findings suggest that nursing home managers should put EOL care into a continuing education routine and develop a supportive culture and strategies that guide professionals in initiating quality discussions on EOL care.

Furthermore, the findings of this study show that professionals with more positive attitudes toward RWDs’ decision-making rights more frequently discuss ADs with RWDs and their families. This finding is noteworthy given the debate over the validity of ADs in the population with dementia (^{Hegde & Ellajosyula, 2016}). Cognitive decline is often assumed to be accompanied by decreasing awareness. Indeed, persons with dementia are usually identified as lacking awareness, assigned the status of a “nonperson,” and judged as being incapable of making or contributing to decisions about their own lives (^{Hegde & Ellajosyula, 2016}). People who perceive that RWDs lack insight tend to show a lack of respect toward them, overrule them, and ignore their intentions and autonomy. A recent study showed that professionals who assume that people with dementia are incapable of making reliable decisions for themselves are not informed of RWDs’ rights to ADs (^{De Vleminck et al., 2014}). However, a diagnosis of dementia or a low score on the Mini-Mental State Examination is not a criterion for determining incapacity (^{Dening, Jones, & Sampson, 2011}). Professionals who communicate effectively with RWDs are inclined to believe that RWDs can make valid ADs and express their wishes even at the advanced stages of dementia (^{Hegde & Ellajosyula, 2016}; ^{Rurup et al., 2006}).

In this study, the registered nurse and social worker participants discussed LSTs with RWDs and their families significantly less often than the participants who were physicians. Although these professionals have different roles and responsibilities in long-term care facilities, physicians still have the greatest influence in the hierarchy of treatment decision-making (^{Rurup et al., 2006}). Nevertheless, considering the current situation in long-term care facilities in Taiwan, nurses and social workers at these facilities constitute the primary professional workforce. They have the closest and most trusting relationships with RWDs and their families, and because physicians are not always in long-term care facilities, they are expected to assume more responsibility and to expand their professional roles in the EOL care process. Although nurses and social workers have less direct influence during EOL care decision-making, they are well positioned to play a valuable, indirect role in its initiation, functioning as RWDs’ advocates, providing guidance and information, and supporting residents’ families.

Although our findings indicate that the physician participants discussed LSTs more often than either the nurse or social worker participants, the LST discussion frequencies of the physician participants were still low. Approximately 70% of the physicians reported never or rarely discussing LSTs with RWDs and their families. This may be because physicians believe that RWDs or their families were reluctant to discuss EOL issues (^{Cheng et al., 2015}) or because Taiwan lacks a consensus on or widely accepted medical or legal guidelines related to treating patients with advanced dementia at the EOL stage.

The three professional groups that were surveyed in this article did not differ significantly in their attitudes toward the decision-making rights of RWDs and their perceptions of facility support for discussing ADs with RWDs. However, the social worker participants had significantly lower perceived knowledge and confidence in EOL dementia care and discussed LSTs less frequently than their physician and registered nurse peers. Lower confidence in providing EOL care and less frequent discussion of LSTs are attributable to social workers not providing direct nursing care and having insufficient medical care training. Because they face increased demands for EOL care in long-term care facilities, facility administrators likely expand the roles of social workers to include not only discussing ADs with RWDs and their families but also maintaining ongoing discussions about the goals of medical care (^{Lacey, 2005}). EOL care must be multidisciplinary, and promoting the knowledge and confidence of social workers regarding EOL care contributes to equitable cross-disciplinary dialogue.

Limitations

A strength of this study was its use of a randomized, stratified, nationwide sample of long-term care facilities. However, the generalizability of the results is limited to professionals in Grade A or higher long-term care facilities. The findings are also limited by only 23 physicians (4.8%) having responded to the questionnaires, with most of these (65.2%) working at hospital-based nursing homes. In addition, data were not collected on other potentially vital factors that are associated with EOL care practices such as facility staffing characteristics and facility policies regarding ADs. We suggest controlling for these factors in future studies. Furthermore, the instruments used to measure the knowledge of, attitudes toward, perceived facility support for, and confidence in EOL care were developed for this study, and this is the first time that they have been used. The validity and reliability of these measures should be retested. Finally, caution must be taken in interpreting these findings because self-reported information may not reflect the actual practice of discussing EOL care.

Conclusions

This study profiled discussions about ADs and LSTs in long-term care facilities in Taiwan. Healthcare professionals at these facilities did not frequently discuss ADs and LSTs with RWDs and their family caregivers. Greater perceived confidence in EOL care and communication skills, as well as facility support, were associated with more frequent discussions of ADs and LSTs with RWDs and their families. Professionals with more positive attitudes toward the decision-making rights of RWDs more frequently discussed ADs with RWDs and their families.

This study further determined that the frequency of professionals’ discussions about ADs and LSTs was significantly and positively related to higher perceived confidence in EOL care and higher perceived support from the facility. We recommend that training for long-term care facility professionals should focus on promoting their confidence in EOL care and ability to communicate about EOL care with RWDs and their family caregivers. In addition, facility managers should establish explicit EOL care policies to guide professional staff in routinely discussing EOL care for RWDs. Managers should also work with staff members to mobilize the internal resources of their facility by promoting staff knowledge and multidisciplinary communication, organizing palliative care teams, and collaborating with external resources such as cooperative referral hospice units and palliative care providers to make professionals more comfortable and confident to engage in EOL care discussions. Moreover, we determined that the frequency of discussing ADs for RWDs was significantly predicted by the degree of positivity held by professionals toward RWDs’ decision-making rights. This result suggests that the attitudes of professionals toward the awareness and decision-making rights of RWDs must be addressed to improve the ability of professionals to communicate with RWDs.