Informant Perspectives on the Quality of Life of Healthy Older People, and People With Dementia Living in the Community and Care Facilities

ABSTRACT

Objectives

Quality of life (QoL) ratings for people living with dementia appear to be influenced by several factors. This study examined how informant perspectives and place of residence influence agreement between self‐ and informant‐rated QoL of people living with dementia.

Methods

This cross‐sectional observational study included 74 healthy older controls, 80 community‐dwelling people living with dementia and 351 residing in residential care facilities (RCFs), along with their informants. Quality of life was assessed using the Quality of Life in Alzheimer's Disease (QoL‐AD) scale. Carers provided two ratings: (i) proxy (their estimate of the person's self‐rating) and (ii) opinion (their own view of the person's QoL).

Results

In the control group, self‐ and informant‐rated QoL did not differ significantly. Among community‐dwelling people living with dementia, self‐rated QoL was higher than both proxy (mean difference 2.7, 95% CI 1.3–4.1) and opinion ratings (mean difference 5.3, 95% CI 3.8–6.8). Similarly, those in RCFs rated their QoL higher than informants. The mean difference between self and opinion rating was greater than the mean difference between self and proxy rating (7.2, 95% CI 6.0, 8.5 for opinion rating, and 5.1, 95% CI 3.8, 6.3 for proxy rating). Discrepancies were greater for people living with dementia residing in RCF than in the community.

Conclusions

Informants tend to underestimate QoL in people living with dementia, particularly in residential care. Informant perspective and dementia severity may influence ratings. Researchers and service providers should consider both self‐ and informant perspectives when evaluating QoL in people living with dementia.

Practice Impact

This study highlights that people living with dementia often rate their quality of life higher than informants, particularly in residential care settings. Differences may be influenced by dementia severity and care context. Using both self and proxy perspectives can support more person‐centred practice.

1. Introduction

The global prevalence of dementia is rising due to increased life expectancy and the progressive ageing of the population. In 2019, an estimated 57 million people were living with dementia, with this number forecasted to increase to 153 million by 2050 [1]. A person‐centred approach to the care of people living with dementia is recommended [2]. In this context, measuring quality of life (QoL) is important in assessing care outcomes for people living with dementia [3, 4], given that it is an important outcome from the perspective of both people living with dementia and carers [5].

Although the point of view of people living with dementia themselves has been recognised as important, little is known about the subjective experiences of people living with dementia. Available data suggest that people living with moderate‐to‐severe dementia frequently experience distressing emotions and retain awareness of some of their deficits and needs to a greater degree than has often been thought [6, 7]. In addition, there is evidence that even people whose dementia is moderate to severe can rate their own QoL reliably [8, 9]. Despite this evidence, reliance on informants to rate the QoL of people living with dementia is common, particularly for those living in residential care facilities (RCFs) who tend to have more severe cognitive impairments than those living in the community.

Proxy rating reflects the appraisal that the proxy thinks people living with dementia would make of their own QoL, distinct from opinion rating (i.e., the informant's own opinion) [10]. Discordance between self, opinion and proxy ratings is common in various clinical conditions and settings [11]. There are several potential problems with the use of informant (both opinion or proxy) ratings of QoL of people living with dementia. We have found that informant or proxy ratings underestimate the self‐ratings of QoL of people living with dementia residing in the community [12] or residential care facilities [8] and are associated with factors (such as the person's severity of cognitive impairment) that are not associated with self‐ratings. A subsequent systematic review and meta‐analysis of the available data did not suggest significant differences between staff and family proxy ratings of QoL, and both were associated with rater‐specific factors (such as staff burnout and the experience of family carers) [13]. Other authors have also found that rater‐specific factors (such as age and financial status) are associated with proxy ratings [14]. In RCFs, carer burden and distress contributed to lower scores in informant‐rated QoL as opposed to self‐rated QoL [10]. Similarly, staff who were more distressed rated residents' QoL less favourably. Additionally, relatives rated QoL less positively when the resident had stayed in the RCF for a longer duration, when they observed more restraint, or contributed more to fees [13].

Previous research has shown a poor inter‐rater agreement between self and informant ratings of QoL, with informants consistently underestimating the self‐rating by people living with dementia [15, 16]. In one study, informants for people living with dementia receiving home care rated QoL 4.7 points lower than people living with dementia, whereas informants for those living in RCFs rated their QoL one point lower than their carers [16]. However, this study included only people living with dementia across two care settings. There is also a paucity of studies reporting the inter‐rater agreement for people living with dementia living in RCFs and community [15, 17, 18] using the recommended methods of assessing agreement, such as inspection of Bland–Altman plots [19, 20]. Currently available studies have used correlation to analyse agreement, thereby assuming a linear or ranked relationship of QoL scores according to their source [16].

We designed this study to examine whether (i) the presence of dementia and (ii) the setting of residence of people living with dementia affected the agreement in QoL rating of patients and their informants.

2. Methods

2.1. Study Design

Cross‐sectional observational study. This is a post hoc analysis of data from prior studies that had ethics approval (Human Research Ethics Committee at the University of Western Australia [RA 4/1/1685]). All participants provided informed consent.

2.2. Setting

Participants were recruited from the community and RCFs in Perth, Western Australia.

2.3. Participants

Seventy‐four healthy older controls, 80 people living with dementia residing in the community, 351 people living with dementia residing in residential care facilities and their informants participated. Participants for the healthy older control group were aged 65 years or older and were recruited together with their next‐of‐kin. We only included cognitively intact control participants, that is, volunteers who did not have cognitive complaints, with a Mini‐Mental State Exam (MMSE) of ≥ 26, and who did not meet criteria for dementia or mild neurocognitive disorder. We included 74 control pairs in this study. Participants living with dementia were largely recruited for two studies of QoL. Community participants were recruited to QoLCog [21], a prospective observational study to determine the association of cognitive decline with QoL among people living with mild‐to‐moderate Alzheimer's disease (AD) (according to NINCDS‐ADRDA criteria [22]) over time. We excluded community‐dwelling participants with a history of alcohol or substance abuse [23], and those with a medically unstable illness that could compromise survival (such as metastatic cancer). Participants living with AD could be taking cholinesterase inhibitors or memantine but could not be participating concurrently in an experimental study of medications for AD. All participants were competent in spoken English. Participants from RCFs were recruited to the DIRECT study [24], a randomised controlled trial of educational interventions, aiming to improve the QoL of people living with dementia residing in RCFs. Recruitment procedures have been described in detail elsewhere [25]. In brief, RCF participants were all permanent residents of aged care facilities aged over 65 years and with a Mini‐Mental State Examination (MMSE) score of < 24. Residents were excluded if the facility staff identified them as being acutely medically unstable or suffering from delirium, or in the terminal stages of a co‐morbid illness. Informants for people living with dementia residing in RCFs participating in this study were required to have visited the person living with dementia on average at least once per week over the previous year.

2.4. Measurements

The Quality of Life in Alzheimer's Disease (QoL‐AD) scale was used to measure QoL by self, opinion and proxy reports [3, 26]. The 13‐item QoL‐AD scale is a widely used, brief (generally administered in ~10 min) dementia‐specific instrument assessing QoL across multiple domains, including physical, psychological and social aspects [3, 26]. The instrument comprises the following 13 items: physical health, energy, mood, living situation, memory, family, marriage, friends, self as a whole, ability to do chores around the house, ability to do things for fun, money and life as a whole. The available data suggest that the QoL‐AD has good inter‐rater and test–retest reliability [26, 27] and can be used to reliably assess the QoL of people living with severe dementia [9]. The QoL‐AD has been modified to produce a 15‐item scale for assessing the QoL of people living with dementia residing in RCFs to improve contextual relevance for people in this care setting [3, 28, 29]. The adapted 15‐item scale removes two items (money and marriage) from the 13‐item scale and adds four items (people who work here, ability to take care of oneself, ability to live with others and ability to make choices in one's life). In addition, the wording of three items has been changed, for example, from ‘ability to do chores around the house’ to ‘ability to keep busy’ to better represent people living in RCF [28]. The items are rated using the same 4‐point scale (poor, fair, good or excellent).

The Standardised MMSE [30] was administered to controls and people living with dementia, and the QoL‐AD was administered to controls, people living with dementia and their informants, as structured interviews by one of the authors or trained research staff. The inter‐rater reliability of the QoL‐AD administered by our research staff has been described previously [31]. Interviewers used a standard set of instructions. Carers of people living with dementia who could not be interviewed face to face were interviewed over the telephone. In face‐to‐face interviews, participants were handed their own copy of the questionnaire, which they could follow if able to. Participants were able to indicate responses verbally or by circling the responses. If the participants were unable to respond to more than two items, they were considered unable to complete the measure and their results were excluded. Carers of people living with dementia rated QoL from two perspectives: (i) their estimation of the person's self‐rating (‘proxy rating’), and (ii) their opinion of the person's QoL (‘opinion rating’).

2.5. Data Handling and Analysis

Because of the use of different (13‐ and 15‐item) versions of the QoL‐AD, scores were converted to a percentage of the score maximum (% SM). Item 7 (marriage) was found to be frequently not applicable to community‐dwelling people living with dementia (36% were widowed or not in a relationship) and was thus excluded from the analysis. Data were handled using the Statistical Package for Social Sciences (SPSS) software. The analysis comprised descriptive statistics. Nonparametric variables were compared using the Mann–Whitney U test. Bland–Altman plots [19] were constructed and paired t‐tests were calculated to determine differences and agreement between self, informant and proxy ratings. The solid horizontal line in the Bland–Altman plot represents the mean bias (i.e., the mean difference); dashed lines indicate the 95% limits of agreement (mean difference ± 1.96 SD). A diagonal regression line was added to assess proportional bias. Missing data were handled using an available‐case approach, whereby participants were included in analyses when relevant data were available. Participants with missing data were not excluded from the analyses. The number of participants with data available at each stage of the analysis is indicated in Table 1.

TABLE 1.

Demographic characteristics and rating of quality of life according to rater and informant's perspective.

	Group
	Control		People living with dementia residing in community		People living with dementia residing in residential care facilities
	Median or mean (% SM)	IQR, Std Dev, or 95% CI	Median or mean (% SM)	IQR, Std Dev, or 95% CI	Median or mean (% SM)	IQR, Std Dev, or 95% CI
Total number, n	74		80		351
Age	76	71–79	78	72–85	86	81–91
MMSE	29	28–30	19	15–22	14	6–19
QoL‐AD rating
Self‐rating, n	74		80		226
QoL‐AD (% SM)	80	10	72	11	69	10
Proxy rating, n	72		80		282
QoL‐AD (% SM)	79	10	67	13	57	16
Informant rating, n	72		79		292
QoL‐AD (% SM)	79	11	61	11	54	14
Mean difference between self and proxy ratings of QoL
Proxy, n	72		80		186
Mean difference	0.4	−0.7, 1.6	2.7	1.3, 4.1	5.1	3.8, 6.3
Mean difference between self and informant ratings of QoL
Informant, n	72		79		189
Mean difference	0.6	−0.7, 1.8	5.3	3.8, 6.8	7.2	6.0, 8.5

Abbreviations: % SM, percentage of the score maximum; IQR, interquartile range; MMSE, Mini‐Mental State Examination; QoL, quality of life; Std Dev, standard deviation.

3. Results

Control participants were aged 76 (IQR 71–79) years with a median MMSE score of 29 (28–30) (Table 1). Compared to controls, people living with dementia residing in the community were not significantly older (78 years, 72–85, p = 0.12) but had lower MMSE scores (19, 15–22, p < 0.001). In turn, people living with dementia residing in RCFs were significantly older (86 years, 81–91, p < 0.001) and had lower MMSE scores (14, 6–19, p < 0.001) than those living in the community. Control participants had a mean self‐rated QoL‐AD score of 80% SM ± 10. People living with dementia residing in the community had significantly lower self‐rated QoL‐AD scores (72% SM ± 11, p < 0.001) than controls. Additionally, self‐rated QoL‐AD scores were lower among people living with dementia residing in RCFs (69% SM ± 10) than those living in the community (p = 0.02).

The QoL‐AD scores in the control group rated by self, proxies and opinion were not significantly different (mean difference between self and opinion ratings 0.6, 95% CI −0.7, 1.8; and between self and proxy ratings 0.4, 95% CI −0.7, 1.6). However, among people living with dementia residing in the community, the mean self‐rated QoL was higher than proxy rating and opinion rating (mean difference 2.7, 95% CI 1.3, 4.1 for proxy rating, and mean difference 5.3, 95% CI 3.8, 6.8 for opinion rating). People living with dementia residing in RCFs also rated their QoL higher than both proxies and informant opinion. The mean difference between self and opinion rating was greater than the mean difference between self and proxy rating (7.2, 95% CI 6.0, 8.5 for opinion rating, and 5.1, 95% CI 3.8, 6.3 for proxy rating). The differences in ratings were more pronounced for people living with dementia residing in RCFs than for those living in the community.

Inspection of Bland–Altman plots (Figures 1 and 2) showed clustering of the three groups of participants, with a poor agreement between self, informant opinion and proxy rating for both people living with dementia residing in RCFs and the community. The mean difference between self and proxy ratings was 3.5 (95% CI 2.7, 4.4), and the limit of agreement ranged between −11.8 and 18.8 scale points (Figure 1). Additionally, the mean difference between self and opinion ratings of QoL was 5.4 (95% CI 4.5, 6.2), and the limit of agreement ranged between −10.4 and 21.1 scale points (Figure 2).

FIGURE 1.

Bland–Altman plot of agreement between self and proxy ratings of QoL‐AD.

FIGURE 2.

Bland–Altman plot of agreement between self and informant ratings of QoL‐AD.

People living with dementia residing in RCFs tended to cluster towards a positive discrepancy, indicating that their self‐rated QoL‐AD scores were generally higher than informant ratings, and they also had lower mean QoL‐AD scores overall. In contrast, people living with dementia residing in the community tended to cluster centrally on the plots. Controls tended to cluster towards the lower right of the plot, reflecting a negative discrepancy between self and informant ratings and higher mean QoL‐AD scores. The differences appeared more pronounced (a greater difference between the three groups) for informant rating compared to proxy rating. Additionally, variability (the spread of data points) appeared to be greatest for people living with dementia residing in RCFs. As indicated by the regression lines in the Bland–Altman plots, there was a statistically significant proportional bias in the differences between self‐rated and informant (R ² = 0.127, p < 0.001) and between self‐rated and proxy‐rated QoL scores (R ² = 0.120, p < 0.001). This suggests the difference between the self and other ratings (proxy or informant ratings) was higher at lower levels of QoL ratings.

4. Discussion

The findings suggest that informants often underestimate the self‐rated QoL of people living with dementia. Although when explicitly asked to imagine how the person being evaluated would rate their QoL, the ratings appeared to align more closely with the individual's self‐rating. There were greater discrepancies in QoL ratings for people living with dementia residing in RCFs compared to those living in the community. Informant and proxy ratings of QoL appear to be acceptable for cognitively intact people, but may substantially underestimate the self‐rating by people living with dementia of their own QoL, especially those with a more severe cognitive impairment [3, 32].

People living with dementia residing in RCFs generally have high levels of impairment and complex care needs. Variability in ratings of QoL and the discrepancy between self and informant reports are particularly high among these older, more severely impaired people living with dementia residing in RCFs. Our finding is consistent with other studies that identified the place of residence as a factor contributing to the difference between self and informant ratings [16, 33]. This study shows that greater cognitive impairment is associated with increased inter‐rater discrepancy. The severity of cognitive impairment and anosognosia in dementia (defined as impaired awareness of the functional consequences of dementia [34]) is among the most frequently mentioned factors for differences in ratings of QoL [8, 33, 35]. However, this study did not formally test independent predictors of disagreement using multivariable modelling. The variability in perceived QoL ratings may also be influenced by carers' depressive symptoms [33], quality of the relationship between people living with dementia and their carers [33], behavioural and psychological symptoms of dementia [16] and perception of caregiver burden [3, 16, 33].

Informant opinion and proxy ratings need to be considered with caution in this population. Many (well‐meaning) caregiving situations impose limitations on people living with cognitive impairments, restricting the choice and freedom of people living with dementia [36]. Thus, the discrepancies in opinion and proxy perspectives that we identified may be potentially clinically important, as they could influence care decisions in ways that inadvertently restrict autonomy or result in interventions misaligned with the person's preferences. For instance, limiting activities or interventions that a person living with dementia would otherwise value. Discrepancies could be minimised by asking informants to explicitly consider the perspective of the person living with dementia, and by carefully triangulating input from both people living with dementia and informants, wherever possible, to ensure care decisions align with the patient's preferences and best interests.

Given that the self‐ratings of QoL by people living with dementia residing in RCFs were lower than those living in the community and healthy controls, the data appear to hold face validity. This is important in the context of the measurement of an essentially subjective construct. Quality of life is a subjective state best measured by the person of interest [37]. In this respect, our data accord with other studies suggesting that people living with moderate‐to‐severe dementia can validly and reliably rate their own QoL [9, 27]. Our data are also consistent with that of other authors, confirming that informant ratings tend to be lower than self‐ratings and are thus not interchangeable but complementary. Future studies may consider further investigating the impact of place of residence on QoL ratings and the underlying factors contributing to the differences in ratings.

Apart from clinical contexts, these findings also have important implications for research and policy. From a research perspective, understanding the potential differences in self‐ratings, informant ratings and proxy ratings can inform study design and data interpretation. For example, by indicating which measures are appropriate for participants living with varying levels of cognitive impairment and highlighting the need for sensitivity analyses to account for inter‐rater discrepancies. In terms of policy and care delivery in RCF settings, there is an urgent need for guidelines that support staff in capturing and interpreting QoL from multiple perspectives. Standards and procedures that encourage the routine inclusion of residents' self‐reported QoL, combined with other sources as needed, may improve the alignment of care with resident priorities and guide resource allocation. Furthermore, using proxy ratings for quality indicators (e.g., Star Ratings for residential aged care in Australia [38]) may misrepresent the quality of a facility if not taken into consideration for the proportion of residents with cognitive impairment, as proxies may rate QoL lower than residents do.

The primary strength of this study lies in its inclusion of a control arm (people without dementia) that provides a baseline for comparison. However, several limitations should be acknowledged. The data are limited by their descriptive nature, which affects the ability to infer causation and limits the conclusions that can be reached. Future studies could consider regression‐based analyses to explore independent predictors of disagreement, such as cognitive impairment, place of residence and informant‐related factors, as well as intraclass correlation coefficients to further quantify agreement. Second, the post hoc nature of the study should be considered when interpreting the findings. Participants drawn from parent studies may introduce potential heterogeneity in participant characteristics and influence QoL assessments. For instance, the residential care cohort included people living with dementia without requiring confirmation of a specific dementia subtype, potentially resulting in a population with variation in dementia subtypes and disease severity. Differences in participant characteristics and assessment contexts across the parent studies may therefore have influenced QoL ratings and should be considered when interpreting the findings. Finally, the use of both the 13‐ and 15‐item versions of the QoL‐AD and the conversion to percentage of the score maximum were carefully considered to enable comparability across care settings. However, it is possible that it may introduce some measurement imprecision due to differences in item content.

5. Conclusions

Informant ratings of the QoL of people living with dementia systematically underestimate self‐ratings, with a greater inter‐rater discrepancy for those living in RCFs than in the community. The perspective of informants and the severity of dementia may influence the ratings of QoL. Researchers and service agencies may need to gather QoL ratings from multiple sources and explicitly consider the perspective of people living with dementia as well as informants.

Funding

The authors have nothing to report.

Ethics Statement

This is a post hoc analysis of data from prior studies that had ethics approval (Human Research Ethics Committee at the University of Western Australia [RA 4/1/1685]). All participants provided informed consent. For people with dementia, a structured consent procedure was utilised, comprising informed written or verbal consent, as well as the agreement of next of kin and, when the severity of dementia precluded the resident from providing consent that was clearly informed, agreement from the resident's ‘next of kin’ only.

Conflicts of Interest

Professor Leon Flicker is a member of the Editorial Board of the Australasian Journal on Ageing. No other conflicts of interest declared.

Data Availability Statement

The data that support the findings of this study are available from the corresponding author upon reasonable request.