Quality of life of older persons in nursing homes after the implementation of a knowledge‐based palliative care intervention

Abstract Background The goals of palliative care are to relieve suffering and promote quality of life. Palliative care for older persons has been less prioritised than palliative care for younger people with cancer, which may lead to unnecessary suffering and decreased quality of life at the final stage of life. Aim To evaluate whether a palliative care intervention had any influence on the perceived quality of life of older persons (≥65 years). Methods This study was conducted as a complex intervention performed with an experimental crossover design. The intervention was implemented in 20 nursing homes, with a six‐month intervention period in each nursing home. Twenty‐three older persons (≥65 years) in the intervention group and 29 in the control group were interviewed using the WHOQOL‐BREF and WHOQOL‐OLD questionnaires at both baseline and follow‐up. The collected data were analysed using the Wilcoxon signed‐rank test to compare paired data between baseline and follow‐up. Results In the intervention group, no statistically significant increases in quality of life were found. This result contrasted with the control group, which revealed statistically significant declines in quality of life at both the dimension and item levels. Accordingly, this study showed a trend of decreased health after nine months in both the intervention and control groups. Conclusion It is reasonable to believe that quality of life decreases with age as part of the natural course of the ageing process. However, it seems that the palliative care approach of the intervention prevented unnecessary quality of life decline by supporting sensory abilities, autonomy and social participation among older persons in nursing homes. From the ageing perspective, it may not be realistic to strive for an increased quality of life in older people living in nursing homes; maybe the goal should be to delay or prevent reduced quality of life. Based on this perspective, the intervention prevented decline in quality of life in nursing home residents. Implications for practice The high number of deaths shows the importance to identify palliative care needs in older persons at an early stage to prevent or delay deterioration of quality of life.

These diseases usually progress over a period of time, which makes it challenging to determine when an older person's end-of-life stage begins (Goddard, Stewart, Thompson, & Hall, 2013); therefore, early signs preceding death could be difficult for the nursing home staff to detect (Åvik, Sandgren, Fürst, Ahlström, & Behm, 2018). The endof-life process of older persons has been described as a dynamic individual process (Krishnan, 2017), and the final year of life is often associated with symptoms such as pain, depression, confusion and distress (Fleming et al., 2017), leading to decreased QoL. A high proportion of older people die shortly after moving into a nursing home (Schon, Lagergren, & Karleholt, 2016), and almost one-third of older persons who move into a nursing home die within six weeks (Smedbäck et al., 2017). This situation implies that most of the persons living in nursing homes are in the final stage of life and would benefit from a palliative care approach.
Today, QoL is considered a crucial outcome in healthcare research, and the underlying reason to measure QoL in health care is to ensure that the evaluations focus on the person rather than the disease (Leegaard et al., 2018). To understand QoL in old age, not only do the distresses and impairments resulting from poor health need to be considered but non-health-related aspects must be considered as well (Lawton, 1999). Even if there is no consensus regarding the concept of QoL, it must address the individuals' perceptions of both positive and negative dimensions (Bowling et al., 2015;World Health Organization, 1995, 1997 and include different domains (emotional, physical, social, independence, spirituality and environmental) of a person's well-being (World Health Organization, 1995, 1997. These What does this research add to the existing knowledge in gerontology?
• It seems that the palliative care approach in this intervention prevented unnecessary QoL decline by supporting sensory abilities, autonomy and social participation among older persons living in nursing homes.
• Frail older persons are often excluded from research; this study is one of few intervention studies on palliative care in nursing homes reporting outcomes of the residents.
• There is an urgent need to continue testing different models to identify effective implementation strategies for palliative care due to the increased number of older persons in many countries.
What are the implications of this new knowledge for nursing care with older people?
• This intervention model reduces the gap in the limited knowledge regarding how palliative care can be implemented in nursing homes.
• The high number of deaths shows that it is important that staff are able to identify older persons in nursing homes who are in need of palliative care in the early stage.
• Due to the progression of multi-morbidity in frail older persons living in nursing homes, a realistic QoL goal is to delay or prevent its deterioration.
How could the findings be used to influence policy, practice, research or education?
• This intervention recommends that staff and manager seminars are held more frequently than every five weeks.
• The dropout rate due to death or insufficient energy was considerably high, so a narrower follow-up period is recommended when frail older persons are included in experimental studies.
• Education regarding evidence-based high-quality palliative care is a promising model that can be performed in nursing homes; therefore, this palliative care approach can be easily disseminated to all staff.

K E Y W O R D S
ageing, education, nursing homes, palliative care, quality of life A growing number of older people will require palliative care in nursing home settings (Froggatt et al., 2017;Gott & Ingleton, 2011;Murtagh et al., 2014). However, previous research has noted the need for improved palliative care for older people in these settings (De Gendt, Bilsen, Vander, & Deliens, 2013;Gott & Ingleton, 2011;Schon et al., 2016). Not only are symptom and suffering relief lacking (Fleming et al., 2017;Smedbäck et al., 2017) but the possibilities for older persons to share their thoughts of life and death or talk about passing away with the staff are also insufficient (Österlind, Ternestedt, Hansebo, & Hellström, 2017). Additionally, requests from staff, that is., nurse assistants, for more education on palliative care as well as for support from Registered Nurses and managers have been described earlier (Beck, Törnquist, Broström, & Edberg, 2012). A study of 322 nursing homes in six European countries found that knowledge concerning basic palliative care was suboptimal in all participating countries (Smets et al., 2018). With to the aim of improving palliative care and thereby the QoL of older persons in nursing homes, a Swedish research group implemented a knowledge-based palliative care intervention directed at staff and managers in 20 nursing homes in Sweden .
This study aimed to evaluate whether the palliative care intervention had any influence on the perceived QoL of older persons (≥65 years).

| Research setting and design
This study is a part of the "KUnskapsbaserad PAlliativ vård" [in English: "Knowledge-based Palliative Care"] project, which is abbreviated as the KUPA project. The KUPA project was conducted from 2015 to 2017 as a complex intervention performed with an experimental crossover design . The KUPA project is designed from an implementation perspective, described in the previous study protocol for the project .
In this research approach, a successful change in practice depends on the interplay between several aspects: (a) effectiveness of the implementation strategies; (b) characteristics of the "implementation object"; (c) characteristics of the implementers; (d) the target population; and (e) the context of the implementation Damschroder et al., 2009;Nilsen, 2015). Therefore, the evaluation of the whole KUPA project addresses several different perspectives which will be reported in future papers; staff, managers, context (Alftberg et al., 2018) the older persons and their next of kin and also various barriers and facilitating aspects of the implementation .
The knowledge-based palliative care intervention was implemented with various staff members (Registered Nurses, assistant nurses, social workers, occupational therapists and physiotherapists) and front-line leaders over a 6-month period in 20 nursing homes in two different counties in the south of Sweden (County A and County B). Firstly, the intervention was implemented in ten nursing homes in County A, while ten nursing homes in County B served as a control group. Then, the ten control nursing homes in County B implemented the intervention, and ten new nursing homes in County A, which had not received the intervention, were chosen as a control group .

| The knowledge-based palliative care intervention
The intervention consisted of five 2-hr seminars and was based on two Swedish national documents describing the key principles of palliative care: a care programme designed by the Regional Co-operative

| Sampling and participants
The nursing homes in this study were selected through voluntary participation. This resulted in a mixture of larger and smaller nursing homes from both urban and rural areas .
The inclusion criteria for the older persons were that they were 65 years or older, spoke Swedish, did not have dementia and had enough energy to participate in a structured interview for up to one hour. The recruitment of the older persons was conducted consecutively in equal numbers from both the intervention nursing homes and the control nursing homes. In total, 90 older persons were included in the study. Thirty-eight (42%) dropouts were related to death, not enough energy, and/or lack of interest in participating in the follow-up interview. Altogether, 23 older persons in the intervention group and 29 in the control group were interviewed at both baseline and follow-up and were included in this study. The demographic variables of the two groups were similar, with a median age TA B L E 1 Characteristics of the older persons in the intervention and control groups at baseline Single 3/13 2/6.5 of 87 years, and most of the included older persons were female, widowers or widows (Table 1).

| Instruments
The data collection was based on two questionnaires that are commonly used together to assess QoL in older persons (Power, Quinn, & Schmidt, 2005), the WHOQOL-BREF and the WHOQOL-OLD. in older adults (Chachamovich, Trentini, & Fleck, 2007;Steinbüchel, Lischetzke, Gurny, & Eid, 2006). Due to the Regional Ethics Review Board in Lund (reference number: 2015/4), the item "sexual activity" in the social relationships' domain needed to be removed. The WHO gave permission to exclude the item to (GA), the project leader of KUPA. Thus, the total score in this study ranged from 25 to 125. 0.66, 0.73, 0.38, 0.60.
The WHOQOL-OLD questionnaire consists of a total of 24 questions and is used to assess QoL in older adults.

| Statistical analyses
The quantitative data collected from questionnaires were analysed using within-group comparisons (the data from after the intervention period were compared with the data from before the intervention period in the intervention group and the control group). The selection of the statistical tests to analyse data other than the descriptive statistics was based on whether the data were distributed normally and the scale level of the instruments. A Wilcoxon signed-rank test was used to compare paired data from after the intervention period to the data from before the intervention period. Subgroup analysis, using the Mann-Whitney U test and the Kruskal-Wallis test, was applied to compare baseline characteristics of the intervention group and the control group. The same analyses were applied for the dropouts in the intervention group with the participants in the intervention group and the dropouts within the control group with the participants in the control group. Either Pearson's chi-square test or Fisher's exact test was used when the expected value was <5.
Analyses were performed using IBM SPSS Statistics version 24. A two-tailed P-value of ≤0.05 was regarded as statistically significant.
Missing data for single items were replaced by the mean score for that item (World Health Organization, 1996.

| Ethical considerations
The To enable the older persons to make an autonomous decision regarding whether or not to participate in the study, both verbal information and written information about the study were provided to the older persons prior to the interviews. Before the written consent was signed, the older persons were informed that participating in the interviews was voluntary and that they could interrupt at any time without having to give a reason and without any consequences. The interviews were conducted by researchers, that is, Registered Nurses trained for the task and with experience approaching older persons to minimise risks of inflicting harm. The principle of non-maleficence was maintained since the older persons were guaranteed confidentiality; that is, individuals cannot be identified as the older person's personal data were encoded. Only the codes were used during the data analysis, and the findings are reported on the group level. The code lists are stored away from the questionnaire forms, which are in locked cabinets. To create as beneficial of an environment as possible for the older persons, the older persons themselves chose the time and place for the interviews and breaks were taken if the older person got tired.

| RE SULTS
No statistically significant differences were detected when compar-  Table 2).
The median score for "general QoL" was three in both groups at baseline, with no statistically significant change at follow-up. The median score for "general health" was four in both groups at baseline, with a non-significant decrease to a median score of three at follow-up. It should be noted that both groups reported a decline in the domain of Environment even though the change was not statistically significant (Table 2).
When comparing the data at the item level, the results revealed that the control group demonstrated a statistically significant decline in the psychological dimension for both "bodily image and appearance" (p-value 0.047) and "negative feelings" (p-value 0.013) ( Table 2). The questions for those items were "Are you able to accept your bodily appearance?" and "How often do you have negative feelings such as sadness, despair, anxiety, depression?" No further itemlevel results from the QOL-BREF are presented since the analyses did not reveal any statistically significant changes.
When comparing the results from the QOL-OLD questionnaire, statistically significant decreases at both the dimension and item levels were revealed in the control group. The median score in the sensory abilities dimension decreased from 16 to 13 (p-value 0.001), the median score in the autonomy dimension decreased from 15 to 12 (pvalue 0.000), and the median score in the social participation dimension decreased from 15 to 13 (p-value 0.000). In the intervention group, no statistically significant changes were found for any of the dimensions (Table 3).
In the control group, statistically significant changes were found for three items. The item "able to do things you like" decreased from a median score of 3 at baseline to a median score of 2 at follow-up (p-value 0.019) and the item "received the recognition you deserve in life" decreased from a median score of 4 at baseline to a median score of 3 at follow-up (p-value 0.040). Furthermore, an increase was found in the control group for "loss of sensory abilities that affect participation in activities" from a median score of 3 to a median score of 4 (p-value 0.005) ( Table 3). No further item-level results from the QOL-OLD are presented since the analyses did not reveal any other statistically significant changes.  ple at a nursing home, these impairments will be untreated, which negatively affects physical activities and communication and contributes to social isolation and loss of autonomy (Evans & Rowlands, 2004;Weinstein, 2015;World Health Organization, 2015). This may explain the significant decrease in this study regarding autonomy and social participation observed in the control group. A contradictory result in the control group was the significant improvement in the loss of sensory abilities affect participation in activities, which indicates that staff recognise these sensory impairments among the older people. Annual health checks for older people in nursing homes may prevent adverse hearing and visual impairments.

| D ISCUSS I ON
In the autonomy domain, the three scores decreased in the control group, while the scores of the intervention group were stable over time. This result is supported by a qualitative study, where a total of 50 staffs were interviewed about autonomy at nursing homes in Sweden. These researchers found that the residents were labelled by the staff as old people who were unable and unwilling to strive for autonomy (Wikström & Emilsson, 2014). Furthermore, a study regarding QoL in Irish nursing homes showed that the more dependent the resident was on the staff, the less likely they were to have a choice (Murphy, Shea, & Cooney, 2007). This fact might explain the signifi-  (ACP) for all nursing home residents. ACP is a discussion with the resident or a representative of the resident about the goals and desired direction of the patient's care. One RCT study using ACP for 309 older persons aged 80 and over found that when using ACP, end-oflife wishes were much more likely to be known (Detering, Hancock, Reade, & Silvester, 2010). Furthermore, a systematic review of 113 studies using ACP in different settings (Brinkman-Stoppelenburg, Rietjens, & van der Heide, 2014) found evidence that ACP positively impacts the quality of care at the end of life.
Earlier evaluations of the care of older persons in nursing homes have shown that the care can often be task-centred and does not correspond to the individual's needs (Beck et al., 2012). By introducing a palliative care approach in which the older person is seen in a holistic view (World Health Organization, 2006), the staff might be open to the idea that the care should be individualised according to each resident's need. A qualitative study by Reitinger, Schuchter, Heimerl, and Wegleitner (2018) exploring relatives' views of the palliative care culture in nursing homes found that the change in culture is highly dependent on management. The relatives concluded that in order to create a palliative care culture, management must support the approach and develop the appropriate structures. However, an qualitative study of managers in the KUPA project  showed that the managers experience limited organisational readiness to develop evidence-based palliative care as a result of variation in the nursing home staff's change efficacy and change commitment as well as restrictions related to many contextual conditions.

| Methodological considerations
Frail older persons are often excluded from research (Lloyd-Williams, Kennedy, Sixsmith, & Sixsmith, 2007;Smedbäck et al., 2017;Ternestedt & Franklin, 2006;Towsley, Hirschman, & Madden, 2015), and this study is one of few intervention studies on palliative care in nursing homes reporting outcomes for residents. However, this study poses some methodological problems that will be discussed below.  (Chachamovich et al., 2007;Power et al., 2005;Steinbüchel et al., 2006). However, one question in the WHOQOL-BREF questionnaire was excluded for ethical reasons based on a decision from the Ethical board, which lowers the validity of the social relationships' domain.

| CON CLUS IONS
The evaluation showed no increase in the QoL dimensions, which could be an expected result of a six-month educational palliative care intervention. The intervention group exhibited no decline in the QoL dimensions at follow-up. However, the control group showed significant declines in the sensory abilities, autonomy and social par-

ACK N OWLED G EM ENTS
We are grateful to the older persons in this study who contributed to their experiences during their last phase of life and for the support from the contact persons in the municipalities. We would like to acknowledge the grants from the Swedish Research Council; the

CO N FLI C T O F I NTE R E S T
The authors declare that they have no conflict of interests.

AUTH O R S CO NTR I B UTI O N
Analysis of the data and drafting the article: CB; The project leader, GA, design of the study, national research grant and recruiting process: GA (the project leader); interviews with the older persons: LB.
All authors contributed to the content of the manuscript text and critically reviewed, discussed, and approved the final version of the manuscript.

DATA ACCE SS I B I LIT Y
The data sets analysed in the current study are not publicly available due to sensitive information from a very vulnerable group, namely frail older persons. Even though the data were anonymised, the study contains sufficient details to enable the identification of individuals. Therefore, before approving the study, the Regional Ethics Review Board in Lund set severe restrictions regarding the accessibility of the data, but the data are available from the project leader (GA) on reasonable request.