Abstract
This article describes a pilot project involving training, case consultations, and administrative coaching over a period of 1 year aimed at introducing palliative care in 2 nursing homes among 31 residents with advanced dementia. Resident outcomes that examined numerous clinical measures were assessed at 3 points in time. Changes in the knowledge and attitudes of 80 staff members and 33 family members who participated in the multimodal intervention were also assessed at 3 points in time. Limited improvements were demonstrated on measures for residents, staff members, and family members at the first nursing home (site 1) and significant improvements were demonstrated at the other nursing home (site 2). Top leadership turned over 3 times at site 1 which limited the integration of palliative care, whereas leadership of site 2 remained stable. Implications for implementing a program of palliative care in nursing homes are discussed.
Introduction
The advanced stages of Alzheimer’s disease and other dementias are characterized by severe memory and language impairments, incontinence of bowel and bladder, immobility or loss of balance and coordination, and complete dependence on caregivers for instrumental and personal activities of daily living. In addition to being a progressive and chronic illness, dementia may also be considered a terminal condition with a fairly well-defined clinical course in the advanced stages. 1 Dementia is now the sixth leading cause of all deaths in the United States and the fifth leading cause of death in Americans aged 65 years and older. 2 It is estimated that 90% of persons who survive to the advanced stages of dementia live in residential care facilities, primarily nursing homes. 3 Mitchell and colleagues 4 reported that 67% of dementia-related deaths occur in nursing homes, and less than one quarter of residents with dementia are referred to hospice. 5 –7 The Center to Advance Palliative Care estimates that by 2030, more than 3 million Americans will reside in nursing homes and 40% of older adults will die there, most of them with some type of dementia. 8
Nursing home residents with advanced dementia often undergo aggressive medical care with limited or no benefits. In a comprehensive review of literature, Volicer noted numerous negative outcomes associated with this type of care for nursing home residents with advanced dementia. 9 For example, intercurrent infections are a common and almost inevitable consequence of advanced dementia but antibiotic therapy does not prolong survival and is unnecessary for symptom control. In a study of 22 Boston area nursing homes, D’Agata and Mitchell found that 66.4% of residents with advanced dementia were treated with antibiotics over a period of 10 months. 10 Mitchell and colleagues also reported aggressive care practices among nursing home residents with advanced dementia including the use of a tube feedings in 25%, laboratory tests in 49%, restraints in 11%, and intravenous therapy 1,4 in 10%. When aggressive interventions are started, nursing home residents can experience distressing consequences such as pain, pressure ulcers, agitation, and shortness of breath. Family members who face aggressive medical care for relatives with advanced dementia are often conflicted about making care decisions and dissatisfied with results. 11 Moreover, a large proportion of Medicare expenditures for nursing home residents with advanced dementia are spent on aggressive treatments that may be avoidable and have limited clinical benefit. 12
Dementia is conventionally treated using a curative model of medicine that focuses on diagnosing the illness, and aggressively treating all symptoms. This model of care strongly influences nursing homes that traditionally have focused on rehabilitation and restorative care. Palliative care is a more realistic and less burdensome approach by focusing on enhancement of quality of life and holistic comfort measures for residents. For example, simple strategies involving care by well-trained caregivers—such as keeping the mouth clean, using flavorful foods, and hand feeding in small portions or tastes—can prevent infection and manage eating and swallowing problems without resorting to artificial nutrition. 9
Background
The Alzheimer’s Association identifies and recommends best practices to meet the emotional, physical, and spiritual needs of nursing home residents with advanced dementia. 13 These best practices focus on enhancing remaining quality of life through a comfort-related approach for all care-related interactions. At present, there is a dearth of research and specialized programs related to palliative care for advanced dementia in nursing homes. 14 –16 Nursing homes are highly complex and heavily regulated organizations characterized by high turnover at every level: residents, families, and staff. Because of their highly unstable environments, conducting research in nursing homes, from pilot programs to randomized-controlled intervention studies, is well known to be challenging. 17 –19 Furman and colleagues have described numerous barriers to the implementation of palliative care in nursing homes including poor communication, lack of coordination of care, inadequate pain and symptom control, and the perceived insignificant value of palliative care research. 20 Ersek and Wilson have cited similar barriers as well as several regulatory and reimbursement issues affecting nursing homes. 21
The overall goal of this pilot program was to implement of palliative care education, training, consultations, and administrative coaching in 2 Chicago area nursing homes caring for residents with advanced dementia. The pilot program was based upon a similar multimodal program, Palliative Care for Advanced Dementia, developed in Phoenix, Arizona in 2005 through collaboration with Hospice of the Valley, a large, nonprofit hospice organization, and the Beatitudes Campus, a nonprofit, continuing care retirement community. The program in Arizona guiding principle is “comfort at all times” for each resident with advanced dementia. 22 This is accomplished through a variety of practical steps such as incorporating a resident’s life story into care practices and care planning, creating a homelike environment, anticipating each resident’s needs, and empowering frontline caregivers to do whatever is needed to make a resident comfortable.
The Chicago-based pilot program aimed to demonstrate the feasibility of replicating the Phoenix multimodal intervention and its positive outcomes that included decreased use of antipsychotics, decreased hospitalizations, decreased laboratory tests, increased pain assessments, increases use of analgesics, and increased hospice utilization. In addition, the pilot study aimed to replicate the Phoenix model’s outcomes for direct care staff and families of nursing home residents that included increased knowledge and attitudes about best practices in dementia care. 23,24
The Project
The pilot project was initiated and led by experts from a local Alzheimer’s Association chapter who had extensive experience in dementia care and end-of-life care. A contractual relationship was formed with a local nonprofit hospice to assist with training and palliative care consultations. Two Chicago area nonprofit nursing homes with dementia special care units with the same parent organization were selected for their willingness to participate in the pilot project. Site 1 is licensed for 253 beds and had 230 residents at the time the project began. Site 2 is licensed for 135 beds and had 120 residents at the time the project began. The parent organization’s Institutional Review Board granted approval for protection of human individuals.
A total of 85 residents from both sites were screened using the following criteria: (a) diagnosis of Alzheimer’s disease or related dementia, (b) Mini-Mental State Examination (MMSE) 25 score of 9 points or less, (c) Functional Assessment Stage Tool (FAST) 26 score of 7a-e, and (d) not currently enrolled in a hospice or palliative care program. A total of 31 residents with advanced dementia-met criteria were enrolled (16 from site 1 and 15 from site 2). Each resident’s Power of Attorney for Healthcare or designated legal representative signed consent forms for enrollment. All proxies plus 2 additional family members were also enrolled for a total of 33 family members.
The 16 residents at site 1 lived on the dementia special care unit and 2 other units of the nursing home. The 15 residents at site 2 lived on the dementia special care unit and 3 other units of the nursing home. Because residents with dementia were living on skilled and special care units, all staff working on these units were invited to participate in the pilot program. A total of 80 staff members from the 2 nursing homes were enrolled in the study, including 39 of 55 staff at site 1 and 41 of 60 staff at site 2.
The intervention consisted of training, consultations, and administrative coaching. Training consisted of 12 hours of interactive sessions in dementia-specific, palliative care philosophy and care strategies, at each nursing home in 6 modules at times convenient for all shifts. The 6 modules covered the following topics: (1) key concepts in palliative care, (2) assessing and addressing pain, (3) daily care, (4) sensory stimulation, (5) challenging behaviors, and (6) end-of-life care and family caregiving. At the request of the administrator at site 2, 41 additional staff members attended a 6-hour, condensed version of the training, but they were excluded from the study.
The consultation component consisted of 1 palliative care consultation by physicians from the hospice for all enrolled residents and their available family members. Consultation involved a chart review, interview, and physical examination with each resident, and a letter of suggested palliative care strategies was sent to the resident’s primary care physician. Case consultations by a project nurse with experience and expertise in dementia and palliative care were provided weekly and as needed for nurses and nursing assistants at the 2 nursing homes. These nursing consultations involved a combination of chart reviews, resident examinations, encouragement of critical thinking, and problem-solving about palliative care strategies.
The administrative coaching component consisted of monthly meetings of “Comfort Care Advisory” committees established at each nursing home that included the administrator and directors of food service, social services, pastoral care, nursing, and activities. These meetings, led by project staff, consisted of 15 leaders of both nursing homes who focused on strategizing and creating procedures for implementing palliative care measures.
Finally, a 21-page booklet, Encouraging Comfort Care: A Guide for Families of People with Dementia Living in Care Facilities, was disseminated to all participating family members and staff members. 27 This free, Web-based booklet was created for this project and covers basic facts about dementia, medical decisions that family caregivers may face, and practical ideas for providing holistic care to persons with advanced dementia.
Measures
Evaluation consisted of examining a variety of resident, family, and staff outcomes at pretraining, 6-month posttraining, and 12-month posttraining. First, resident data were collected through chart reviews to determine the following: use of antipsychotics, monthly blood draws, antibiotics, physical restraints, artificial nutrition, dietary supplements, diets without restrictions, weekly assessment of pain assessment, daily treatment of pain, average body weights, emergency department visits, hospital admissions, use of “Do Not Resuscitate Orders,” hospice utilization, and monthly visits by pastoral care staff.
Knowledge and attitudes of staff members were evaluated with the Questionnaire of Palliative Care for Advanced Dementia that has shown good reliability and validity. 28 This measure consists of 2 scales about knowledge and attitudes about advanced dementia care issues. A total of 25 questions assess knowledge about advanced dementia, with respondents given 3 choices: true, false, do not know. The attitude scale includes 12 statements to assess attitudes regarding caregiving, job satisfaction, and end-of-life care decisions, with respondents asked to rate their attitudes using a 5-point Likert scale (1 = strongly disagree, 5 = strongly agree). Both the knowledge and attitudes scales have strong internal consistency, with Cronbach's coefficient α of .83 and .87, respectively. 28
Family members completed an abbreviated version of the Questionnaire of Palliative Care for Advanced Dementia consisting of 11 of the 25 knowledge questions. Family members were also asked to rate 10 items created for this study regarding their confidence in staff to provide them and their relatives with good care using a 5-point Likert scale (1 = not at all confident, 5 = extremely confident).
Analysis
Analyses were conducted for resident, staff, family, and data using data collected prior to training, 6 months posttraining, and 12 months posttraining. Analyses of 6-month posttraining data included those with nonmissing data at pretraining and 6 months posttraining. Analyses of 12-month posttraining data included those with nonmissing data at pretraining and 12 months posttraining; 12-month data were compared to the pretraining data, not the 6-month data. Analyses were conducted separately for data collected at sites 1 and 2.
For resident data, paired samples t tests and chi-squares were computed, as appropriate, to examine differences between 6-month posttraining and pretraining, and between 12-month posttraining and pretraining.
Data collected from staff members were divided between the 2 nursing home sites and the hospice organization. For the sake of comparison between the 2 sites, data for hospice staff were excluded from results reported here. The percentage of participants that responded correctly to the item was computed for all knowledge items. Paired samples t tests were used to examine differences in the percentage of participants that responded correctly to each item between 6-month and pretraining and between 12-month and pretraining. Paired samples t tests were also used to examine differences in the mean number of items correct between 6-month and pretraining and between 12-month and pretraining. The mean of each attitude item and the mean across all attitude items were computed. Paired samples t tests were used to examine differences in the mean score for each item, and the mean score across all items, between 6-month and pretraining and between 12-month and pretraining.
In regard to family data, the percentage of participants that responded true, false, or do not know (or left the item missing) was computed for the knowledge items. Differences in the patterns of responses were examined between 6-month and pretraining and between 12- month and pretraining using chi-square tests. For the confidence items, the mean score across participants was computed for all items. Paired samples t tests were computed to determine differences in mean scores between 6-month and pretraining and between 12-month and pretraining.
Results
Summary data regarding the 31 residents enrolled in the study at pretraining, their 33 family caregivers, and 80 staff members from the 2 nursing homes and hospice are outlined in Table 1. White females comprised the vast majority of residents at both sites, with an average age of 85 years for both sites. The average length of stay at both sites was more than 2 years. Residents at site 2 were more impaired overall compared to site 1, as shown by MMSE and FAST scores. Most family caregivers were adult daughters, with a mean age of 61.6 years at site 1 and 66.6 years at site 2. As expected, the vast majority of staff members were female, 76.9% at site 1 and 85.4% at site 2, with an average age of 48.6 and 43.8 years, respectively. Whites and Asians were the major racial groups at both sites.
Demographic and Clinical Data of Residents, Family Caregivers, and Staff at Pretraining
Abbreviation: SD, standard deviation.
Resident Data
Table 2 summarizes a variety of outcomes regarding residents over time. At 12 months posttraining, 20 of 31 residents remained in the study, with 10 residents deceased and 1 resident withdrawn by the proxy/family caregiver in order to limit the interaction of new staff with the relative with advanced dementia. All MMSE and FAST scores of surviving residents declined as expected over 12 months.
Nursing Home Resident Outcomes
The most pronounced improvement was that staff at both sites began to evaluate residents for pain more routinely—at a minimum on a weekly basis—from 45.5% to 81.8% at site 1 and from 75% to 100% at site 2. As both sites aimed to reduce special diets and offer liberalize diets, as well as provide snacks at any time, reductions were demonstrated in the use of nutritional supplements with minimal weight loss of just 6 pounds on average over 12 months. The use of 2 or more antipsychotic medications decreased primarily in the last 6 months at site 2. A striking difference between the 2 sites was the number of hospice referrals. All residents met Medicare criteria for hospice at 12 months posttraining. 29 However, site 1 made no referrals to any hospice program but site 2 referred 67% of residents to hospice programs.
Staff Data
Improvements over time in knowledge of staff were shown at site 2 but not at Site One. Improvement in attitudes of staff improved at both sites. However, the small sample size due to attrition over time limited the ability to detect significant change. Table 3 shows overall results regarding knowledge and attitudes of staff at the two sites and includes selected items that were shown to improve significantly over time.
Staff Knowledge and Attitudes
At site 1, the percentage of correct items on the 25-item knowledge section of the Questionnaire of Palliative Care for Advanced Dementia was73.6% at pretraining (n = 38), 64.4% at 6-month posttraining (n = 24), and 72% at 12-month posttraining (n = 8). At site 2, the percentage of correct items on the 25-item knowledge section of the Questionnaire of Palliative Care for Advanced Dementia was 67.6% at pretraining (n = 41), 75.2% at 6-month posttraining (n = 17), and 68.8% at 12-month posttraining (n = 15).
In terms of the 12 items on the attitude section of the Questionnaire of Palliative Care for Advanced Dementia, staff at site 1 at preintervention rated an average of 3.5/5(n = 38), 3.9/5 at 6-month posttraining (n = 24), and 4/5 at 12-month posttraining (n = 8). At site 2, attitudes improved from a mean rating of 3.2 at pretraining (n = 41), to 3.7/5 at 6-month posttraining (n = 17), and 3.5 at 12-month posttraining (n = 15).
Family Data
Improvements over time in knowledge of family members were shown at Site Two but not at Site One. Again, the small sample size due to attrition limited the ability to detect significant change. Table 4 shows overall results regarding knowledge and attitudes of family members at the 2 sites and includes selected items that were shown to improve significantly over time.
Knowledge and Attitudes of Family Members
At site 1, the percentage of correct items on the 11-item knowledge section of the Questionnaire of Palliative Care for Advanced Dementia was 75% at pretraining (n = 14), 83.3% at 6-month posttraining (n = 9), and 66.6 at 12-month posttraining (n = 9). At site 2, the percentage of correct knowledge items was 49.1% at pretraining (n = 17), 45.9% at 6-month posttraining (n = 7), and 69.9% at 12-month posttraining (n = 5).
In terms of 10 items pertaining to family’s confidence in the ability of staff to care for them and their relative with dementia, family members at site 1 gave a mean rating of 3.4/5 at pretraining (n = 14), 3.8/5 at 6-month posttraining (n = 9), and 3.4/5 at 12-month posttraining (n = 4). At site 2, family attitudes improved from a mean rating of 2.8 at pretraining (n = 17), to 2.8/5 at 6-month posttraining (n = 7), and 4/5 at 12-month posttraining (n = 5).
Discussion
There are numerous limitations that are to be expected in a pilot study. First, only 2 nursing homes in the Chicago area were selected for the sake of convenience, and there was no comparison group. The sample sizes of both residents and families were small and the vast majority of residents and family members were Caucasians. Attrition of residents mainly due to death was to be expected in this fragile population. There were also diminishing numbers of family and staff responses to questionnaires at 6 and 12 months posttraining. In addition, it was not feasible to train all staff members on the units where the residents lived due to staff shortages.
Despite these major limitations, improvements were demonstrated on all measures for residents, staff members, and family members at site 2 but improvements were not demonstrated at site 1 except for pain assessment. From the outset of the pilot project, the administrator and managers of site 2 embraced palliative care, communicated this approach to direct care staff and families, and took concrete steps to put this approach into practice. In addition to improving clinical outcomes for residents, site 2 aimed to improve the physical environment and to empower direct care staff to improve care practices. Moreover, site 2 examined the manner in which it addressed the active dying process of all residents in order to offer a more comfortable and meaningful experience for all concerned. As previously noted, the administrator at site 2 requested that 41 additional staff members participate in a condensed version of the training program; this too may have influenced the findings.
At site 1, improvements were made in pain assessment but not other measures. There were 3 different administrators during the 1-year pilot program. Despite initial interest, none of these administrators actively promoted palliative care and consequently, efforts to motivate staff to improve outcomes were hindered. In addition, there were 2 Medical Directors and 2 Activity Directors due to turnover during the 1-year pilot program. Additional barriers to implementation of palliative care described in other studies 21,22 were also encountered at site 1 including poor communication, lack of coordination of care, and the perceived insignificant value of palliative care research. Fortunately, at site 1, there was no turnover in the Director of Nursing who took an active interest in assessing and treating pain that may help explain this sole positive finding. Staff turnover in nursing homes is a widespread and serious problem among all levels of staff, resulting in poor resident outcomes. 30 –32 Conversely, studies have shown the benefits of positive and consistent leadership in lowering turnover and improving resident outcomes. 33 –35 The present project lends support to the growing body of evidence linking leadership, staff turnover, and resident outcomes. Finally, this pilot project supports the feasibility of replicating positive outcomes demonstrated in prior work conducted by creators of the palliative care for advanced dementia program in Phoenix.
It is likely that to some extent that this pilot work affected all residents with dementia in both nursing homes, although outcomes were examined among a minority of residents enrolled by their proxies. Including everyone with a primary or secondary diagnosis of dementia, regardless of severity, may contribute to a better understanding of an intervention’s impact. Alternatively, conducting an organizational assessment using aggregate data at several points in time might better capture the effects of an intervention aimed at changing care practices and improving resident outcomes. Also, choosing 1 or 2 clinical measures for improving resident outcomes such as pain assessment and pain treatment might be a more realistic approach than seeking many improvements simultaneously. A more focused approach instead of a multimodal approach may also simplify the intervention and reduce the time involved in staff training and case consultation. Teasing apart different aspects of a multimodal approach would be helpful in identifying the most effective means of improving outcomes
Conclusion
Based upon the positive outcomes demonstrated for residents, their family members, and staff members, this pilot program lends support to introducing a program of palliative care in nursing homes. Specialized staff training, nursing and medical consultation, and administrative coaching are key components. A critical aspect for successful implementation is consistent participation by a nursing home’s leadership and ongoing promotion of steps to improve the comfort of residents. Moreover, collaboration between external researchers and nursing homes staff requires good communication and a firm commitment to work on mutually agreeable goals.
Footnotes
Acknowledgments
The authors would like to express gratitude for the participation of residents, families, and staff members of the 2 collaborating nursing homes, Villa Scalabrini Nursing and Rehabilitation Center and Maryhaven Nursing and Rehabilitation Center, as well as the staff of Rainbow Hospice and Palliative Care, especially Patricia Ahern, RN, MBA, and Amy Frazier, RN, MS. In addition, thanks to Dr Bradley Fulton for his statistical expertise and to Susan Rothas, RN, BSN, and Courtney Bouker, LNHA, for their assistance with data collection. Finally, special thanks to the Palliative Care for Advanced Dementia team in Phoenix, especially Dr Carol Long, Dr Maribeth Gallagher, and Tena Alonzo, MA.
The first author serves on the editorial board of this journal but the reviewers were blinded to authorship. The other authors declared no potential conflicts of interest.
This work was supported by a grant from the Retirement Research Foundation to the Alzheimer’s Association-Greater Illinois Chapter.