CHANGES IN THE QUALITY OF LIFE OF PEOPLE WITH DEMENTIA LIVING IN CARE HOMES

Abstract

Quality of Life (QoL) is now an established outcome measure for people with dementia. There is a need to understand if measures are sensitive to change and what factors are associated with change in QoL in dementia in order to develop interventions to improve QoL and identify who may be most likely to benefit. This study aimed to assess change in QoL in people living in 24 hour care homes using the Quality of Life - Alzheimer’s Disease (QOL-AD) scale and investigated which clinical factors predicted changes in QoL in dementia. We used the QOL-AD scale to rate individual and staff perceptions of residents QoL, for 238 people with dementia, recruited from 24 care homes in the UK. A follow up interview was undertaken at 20 weeks. 192(80.7%) of the 238 residents interviewed at baseline were followed up. A reduction in residents’ QoL was predicted by lower baseline depression and anxiety symptoms, higher baseline QoL ratings, and an increase in depressive symptoms and cognitive deterioration at follow up. While QoL does not necessarily diminish as dementia progresses, it is strongly influenced by the person with dementia’s mood. Improvement in cognition and mood may lead to increased QoL.

Introduction

In dementia quality of life (QoL) does not necessarily decrease as dementia progresses^1,2,3. Nevertheless, measures of QoL in dementia should be able to detect any changes in QoL in response to both interventions and the progression of the disease, so that they can be used to establish benefits for patients and caregivers. Dementia is a degenerative disease process characterised by worsening cognition, behavioural disturbance, mood disorder, increased dependency and physical frailty which may lead to institutionalisation⁴. However, little is known about what factors are associated with change in QoL in dementia and studies that examined QoL over time have shown varying results^1,2,3. Lyketsos² detected an overall decline in mean QoL scores at 2 years, although nearly half the QoL scores remained unchanged or were higher at follow up, suggesting QoL is preserved in some people with advanced dementia despite progression of the disease. Selwood¹ found no difference in QoL scores in people with dementia at 1 year follow up, further suggesting that QoL does not necessarily decline over time and baseline QoL was the only predictor of future QoL. Misotten³ found improved QoL scores at 1 year, but no difference in QoL scores at 2 year follow up, despite worsening mental and physical frailty and speculated that there is no relationship between clinical state and QoL. Nevertheless, there is limited research suggesting that QoL in dementia may be improved through enhancing mood and cognition^5,6, increased occupation and maintaining relationships^7,8. Specifically, Woods⁵ showed improved QoL scores for participants of cognitive stimulation therapy were associated with better mood which was mediated by improvements in cognition. It is therefore reasonable to assume that change in QoL is possible and further longitudinal studies with larger samples are needed to further investigate the impact of change in QoL in dementia over time. The aim of this study was to identify clinical and demographic factors predicting QoL change in dementia. We predicted that there would be no significant relationship between change in QoL and change in any clinical or demographic factors.

Method

Sample

This investigation was part of a larger project examining the needs and QoL of older people with dementia living in care homes^9,10,11. 238 people with dementia, over 60 years of age, were recruited from 24 care homes in London, Manchester and North Wales. Those residents, who were permanently placed and had lived in the home for at least one month and had possible or definite memory problems, had a casenote review and diagnostic interview to identify dementia¹². All participants were asked for written consent, or assent depending on their level of cognitive abilities. Trained mental health researchers conducted the interviews, which were stopped if the resident asked to withdraw or showed distress.

Procedure

The residents of the home were interviewed and further information was obtained through staff interviews, observation and the care home documentation. The residents and staff members were interviewed separately. The main study design was a randomised controlled trial of a liaison intervention package to meet the unmet needs of residents with dementia in a group of homes compared to care as usual in the control group homes¹¹. Needs assessment was the primary outcome measure, with QoL used as a secondary outcome to further assess the impact of the study and has previously been described in detail^9,11. The follow up interview was completed at 20 weeks by a second investigator blind to the allocation of homes to intervention and control groups, who reassessed all the residents involved in the study following the procedure already outlined. Participants were followed up where possible even if they had moved out of the home.

Instruments

Instruments completed with residents and staff

Instruments completed with residents and staff: this measures QoL in dementia and is completed by both patient and caregivers¹³. There are 13-items that include: physical health, energy, mood, living situation, memory, family, marriage or significant other, friends, self as a whole, ability to do chores around the house, ability to do things for fun, money, and life as a whole. Possible totals range from 13-52, with higher scores indicating better QoL. Where one or two items are missing the mean QOL-AD item score is calculated and this value is substituted for the missing data¹⁴. QOL-AD scales with 3 or more items missing were excluded. The QOL-AD has been found to have good reliability and validity throughout the range of scores and can be used with people with all severities of dementia^7,15.

Cornell Scale for Depression in Dementia: a 19 item scale with scores of ≥ 8 indicating significant depressive symptoms¹⁶.

Rating of Anxiety in Dementia (RAID): an 18 item scale with scores ≥11 indicating significant anxiety symptoms¹⁷.

Camberwell Assessment of Need for the Elderly (CANE): a comprehensive, valid and reliable measure of need in older people^18,19. It includes mental and physical health, social and environmental needs, and identifies whether needs are met or unmet using information from patients, carers and professionals. The investigator makes an overall rating of need.

Instruments completed with residents

Mini Mental State Examination (MMSE): a brief, widely used test of cognitive function²⁰.

Instruments completed with staff

Challenging Behaviour Scale (CBS): a 25-item checklist that measures and rates the frequency and severity of challenging behaviour in dementia²¹.

Clifton Assessment Procedures for the Elderly-Behaviour Rating Scale (CAPE-BRS): measures behaviour problems and functional ability and can be used to rate dependency²².

Barthel Scale of Activities of Daily Living Scale: designed to measure the individual’s ability to complete various activities of daily living²³, higher scores indicating better function.

Instruments completed by the investigator

Clinical Dementia Rating (CDR): an investigator rated global score of severity of dementia graded from 0 for no dementia to 3 for severe dementia²⁴.

Power Calculation

The sample size was estimated using the data from Selwood¹ study of QoL in dementia which found that change in QoL was predicted by QoL at baseline. We judged that a significant change would be equal to half a standard deviation and in Selwood’s study (of care-home residents) this was 3 points. This value was taken as the expected and predicted increase or decrease of QoL. For an 80% (power) chance of finding a significant difference at the 1% level (p<0.01), the estimated sample size was 21 participants per quartile (approximately 84 people in the total study population).

Analysis

The data were analysed using SPSS 12.0²⁵. As the data for the QOL-AD were normally distributed t-tests were used to compare mean scores and Pearson’s r test was used for correlations. As there were multiple univariate analyses, p<0.01 was used as the level for significance and correlations of 0.4 and above were considered clinically significant²⁶. An improvement or deterioration in QoL was considered if scores changed by 3 or more points¹. Change in individual QoL scores were correlated with baseline and change scores for pathological measures. A stepwise linear regression analysis was undertaken to determine predictors of change in QoL, (on the QOL-AD) using the resident’s and staff ratings of QoL. We entered: in step 1, the demographic data - age, gender, and length of stay; in step 2, baseline and change scores for the clinical variables including cognition, mood, dependency and behavioural disturbance; in step 3, variables which were significantly associated with change in QoL scores to identify which predicted change in QoL.

Results

Of the 238 residents interviewed at baseline, 192 (80.7%) were followed up. The majority 40(16.8%) of those not followed up were deceased, 1(0.4%) withdrew consent and 5(2.1%) were transferred to another care setting. Those followed up were predominantly female (n=152, 79.2%), white (n=166, 86.5%) and widowed or unmarried (n=181, 94.3%). Other sociodemographic and pathological factors are described in Table 1. There were 96(50%) residents and 190(99%) staff completed QOL-AD scales. Mean QOL-AD scores were imputed for 35(18.2%) resident and 54(28.1%) staff QOL-AD scales¹⁴. A large number of residents had difficulty completing the QOL-AD and 94 residents (48.9%) had an unrecorded score, this is unsurprising when 50% of the sample had MMSE <3¹⁵. There was 21(10.9%) resident (PQoL) completed and 10(10.9%) staff (SQoL) completed QOL-AD scales at follow up but not at baseline. The QOL-AD was completed at both baseline and follow up by 70(36.5%) residents and 182(94.8%) staff.

Table 1.

Sociodemographic and pathological variables

Variable	Scores at baseline (n=192)
	Mean	(S.D.)	Range
Age	85.8	(7.6)	60 - 104
Length of stay (months)	33.1	(31.0)	1 - 180
MMSE	7.4	(8.0)	0 - 26
CDR	1.9	(0.8)	0.5 - 3
CORNELL	5.6	(4.9)	0 - 24
RAID	5.8	(5.8)	0 - 34
CBS	26.1	(29.6)	0 - 214
CAPE-BRS	16.2	(5.3)	2 - 32
BARTHEL	66.0	(18.4)	30 - 100
CANE - Met needs	12.0	(2.4)	5 - 17
CANE - unmet needs	4.4	(2.6)	0 - 13

Change in quality of life

Baseline and follow up ratings of residents QoL were examined using the QOL-AD scores. There was no significant difference between the mean resident (PQoL) or staff (SQoL) rated QOL-AD total scores at baseline and follow up, PQoL (n=70): 32.6(s.d. 6.4, range 15-46) vs 32.4(s.d. 6.8, range 15-47); SQOL (n=182): 30.6(s.d. 6.0, range 15-49) vs 29.8(s.d. 5.9, range 15-47). Considering individual scores there was no change (±0-2) in PQoL scores for 19(27.1%) residents; 24(34.3%) residents improved in PQoL, with 16(22.9%) of these having an increase of 6 or more points (range 6-16) on the QOL-AD. Similarly, 27(38.6%) residents had decreased PQoL, with 16(22.9%) of these showing a decrease of 6 or more points (range 6-17) on the QOL-AD.

Differences between grouped changes of PQOL

The 70 residents were categorised into three groups

increase, no change and decrease in PQoL. The group with an increase in PQoL (n=24) had a mean QOL-AD score of 29.8(s.d. 4.7) at baseline and 36.8(s.d. 4.8) at follow up, whilst the group with a decrease in PQoL (n=27) had a mean QOL-AD score of 36.0(5.1) at baseline compared to 28.9(s.d. 5.6) at follow up.

Change in PQOL scores (see Table 2)

Table 2.

Correlations with change in resident (PQoL) and staff (SQoL) rated QoL versus baseline scores and change scores for the clinical variables

Clinical variables	Baseline variable v PQoL change (n=70)		Variable change v PQoL change (n=70)		Baseline variable v SQoL change (n=182)		Variable change v SQoL change (n=182)
	r	sig	r	sig	r	sig	r	sig
Barthel	.055	NS	.144	NS	-.007	NS	.078	NS
CAPE-BRS	-.010	NS	-.196	NS	.154	NS	-.412	***
CDR	.022	NS	-.078	NS	.056	NS	-.064	NS
Cornell	.138	NS	-.360	**	.202	*	-.350	***
CBS	.017	NS	-.240	NS	.057	NS	-.340	***
MMSE	-.122	NS	.421	***	-.196	*	.201	*
RAID	.341	**	-.404	***	.120	NS	-.245	***
Total Met Need	.131	NS	-.005	NS	-.103	NS	.154	NS
Total Unmet Need	-.138	NS	-.030	NS	.129	NS	-.204	*
PQoL	-.492	***			-.544	***

NS = Non Significant

^*= p<0.01

^**= p<0.005

^***= p<0.001

Increased PQoL scores at follow up were associated with fewer baseline anxiety symptoms (RAID: p<0.005) and lower QoL at baseline (p<0.001), and with a decrease in depression (Cornell: p<0.005) and anxiety (RAID: p<0.001) symptoms at follow up. A decline in cognition (MMSE: p<0.001) was linked to reduced QoL at follow up.

Change in SQOL scores (see Table 2)

Increased SQoL scores at follow up were associated with fewer depressive symptoms (Cornell: p<0.01), less cognitive impairment (MMSE: p<0.01) and lower SQoL at baseline (p<0.001) and with a decrease in depression (Cornell p<0.001) and anxiety symptoms (RAID p<0.001) at follow up, and fewer unmet needs (CANE: p<0.006). Reduced SQoL at follow up was linked to increased dependency (CAPE-BRS: p<0.001), cognitive decline (MMSE: p<0.01) as well as increased behaviour problems (CBS: p<0.001).

Predictors of change in PQOL between baseline and follow up

We used stepwise linear regression analysis to consider independent predictors of change in PQoL. No predictors of change in PQoL were identified in step 1 (socio-demographic variables). In step 2 (baseline and change in clinical ratings) a decrease in PQoL was predicted by higher baseline QoL (PQoL, p≤0.001), and fewer depressive (Cornell, p≤0.001) and anxiety (RAID, p≤0.01) symptoms at baseline, and by a decline in cognition (MMSE, p≤0.001). This model accounted for 55% of the variance (F=4.3, adjusted R² =.42, p<0.001).

In step 3, the baseline and change in clinical variables for the MMSE, Cornell, RAID and baseline PQoL were entered. A decrease in PQoL at follow up was predicted by higher baseline QoL (QOL-AD: β=-0.433: p<0.001) and fewer depressive (Cornell; β=-0.870, p<0.001) and anxiety (RAID, β=0.684, p<0.001) symptoms at baseline. Reduced PQoL at follow up was also predicted by worsening depression (Cornell: β=-0.541, p≤0.005) and deteriorating cognition (MMSE: β=0.356, p≤0.001). This model accounted for 52% of the variance (F=9.6, adjusted R² =.47, p<0.001).

Predictors of change in SQOL between baseline and follow up

The stepwise linear regression analysis was then repeated to determine which factors were independent predictors of change in SQoL. No predictors of change in SQoL were identified in step 1 (using socio-demographic variables). In step 2 (baseline and change in clinical ratings) a decrease in SQoL was predicted by higher baseline QoL (SQoL, p≤0.001) and lower dependency (CAPE-BRS, p<0.05) at baseline and by increased dependency (CAPE-BRS, p<0.001) and depressive (Cornell, p≤0.01) symptoms at follow up. This model accounted for 52% of the variance (F=11.8, adjusted R² =.47, p<0.001).

In step 3, just the baseline and change in clinical variables for the CAPE-BRS, Cornell, and baseline SQoL were used as the independent variables. A decrease in SQoL at follow up was predicted by higher baseline QoL (QOL-AD: β=-0.596: p<0.001) and less dependency (CAPE-BRS: β=-0.366, p<0.001) at baseline. Reduced SQoL at follow up was predicted by increasing dependency (CAPE-BRS: β=-0.383, p<0.001) and worsening depression (Cornell: β=-0.256, p<0.001). This model accounted for 49% of the variance (F=33.5, adjusted R² =.47, p<0.001).

Discussion

There are key issues raised by the study which include measuring change in QoL and the nature of QoL itself. This study, examined change in QoL of people with dementia in residential homes, had a high response rate (80.7%)²⁷, used validated instruments and trained raters and included a variety of homes. The results may therefore be generalised to homes throughout the UK. We found that, like previous studies, whilst there was no change in overall mean QoL, changes in QoL were very frequent for individual residents^1,2,3. In this study, individual changes in QoL were apparent for nearly three quarters of the sample population, with just under half of the residents showing large changes in QoL (6 or more points on the QOL-AD scale, equivalent to 1 standard deviation). A reduction in residents’ QoL was predicted by lower baseline depression and anxiety symptoms, higher baseline QoL ratings, and an increase in depressive symptoms and cognitive deterioration at follow up. Other studies have found change predicted by increased functional dependency at 1 year and worsening cognition at 2 years³; and in those treated with cognitive stimulation therapy improvement in QoL scores were mediated by improved cognition^5,6. Furthermore, Livingston²⁸ found that future QoL is predicted by baseline mood and social relationships, and that wellbeing in dementia related adversity remains stable over time. Cognition was found to influence wellbeing in adversity but was mediated through mood²⁸.

Change in QOL

Changes in QoL for people with dementia were associated with baseline and change in clinical factors such as mood and cognition and this is a new and important finding. Higher baseline QoL scores tended to be reduced at follow up, whereas lower baseline QoL scores tended to be increased as would be expected with regression towards the mean over time²⁹. Selwood¹ measured change in QoL from baseline at one year for ratings based on self reports and compared baseline QoL ratings to those at follow up. Comparing total sample statistics at two points in time may be misleading when assessing change, in case underlying changes go undetected. It is important therefore to assess changes using a change score calculated for each participant²⁷. Lyketsos² measured change in QoL between baseline and two year follow up based on proxy reports and compared baseline clinical variables with change in QoL ratings. Higher ratings of QoL are known to be associated with fewer depressive symptoms^13,10 and in this study a decline in individual QoL ratings was associated with an increase in depressive symptoms, suggesting that reducing depressive symptoms may improve residents’ QoL. When the changes in staff QoL ratings were examined, lower ratings of QoL were given to residents whose functional dependency had declined and an analysis of the baseline data also showed that staff ratings of residents’ QoL are strongly influenced by levels of dependency¹⁰.

Ability to detect change using the QOL-AD

The QOL-AD scale was able to detect individual changes in QoL over a 20 week period. The QOL-AD has good reliability^13,7 and the ability to detect change supports the use of the QOL-AD as a valid longitudinal measure of self rated QoL by people with dementia. Improving QoL has now become a key area of focus in the quality of care provided within care homes³⁰. The QoL of residents in care homes may be influenced by different factors compared to those people with dementia living in their own homes or with relatives; such as having a sense of autonomy, security and privacy. QoL has been shown to be higher for those people with dementia living in the community compared to those living in institutions³¹.

Perceptions of Quality of Life for people with dementia

For the person with dementia it is their experience rather than their disability that most influences their QoL, so that mood is the main determinant of QoL at every stage of dementia severity^7,10,,^13,15,31. Whilst all aspects of health are important in assessing the impact of the disease process, it is the individual’s subjective interpretation of the objective experience that truly defines QoL³². Mood relates to psychological well-being, happiness, self-esteem and judgements of overall life satisfaction which are all determinants of QoL^28,33,34. An important feature of psychological well-being is the capacity of individuals to adapt to losses and maintain a positive self-belief. Increased frailty does not necessarily diminish the strength of self³⁴, thus supporting the concept of personhood, in that individuality is maintained throughout the dementia process. An individual’s sense of control is also core to QoL and this incorporates feelings of self-efficacy, perceived control and learned helplessness which relate to people’s personal beliefs and expectations³⁵. Within dementia a person’s sense of control may be diminished through the degenerative and debilitating nature of the disease process. Relationships are a key element in sustaining QoL in that they reinforce a sense of personal worth, a sense of agency, social confidence and hope^28,36. Moreover, relationships remain important even in the more severe stages of dementia and are influenced by the environment^8,15,28,31. The significance of maintaining relationships and establishing new ones should be recognised when people with dementia are placed in residential care and every effort should be made to promote integration for new residents and their families.

Limitations

This investigation was part of a larger randomised controlled study which offered interventions for the unmet needs identified for individual residents in the intervention homes and included feedback on resident assessments and evaluations of their individual needs, both met and unmet. The needs for both groups improved, within the intervention and control homes and there was no significant difference at follow up¹¹. The intervention is therefore unlikely to have significantly affected these analyses. The study collected data at two points in time; nevertheless, it is difficult to be sure of the direction of causation. Dementia is a non-linear and complex disorder and has multiple causes and outcomes³⁷; it is therefore difficult to encompass all factors influencing its progression and the QoL experienced by individuals with the disease. We did, however, include a wide range of factors within this study which included mental and neuropsychiatric disorders such as anxiety, depression, and behavioural disturbances, which are known to commonly occur in dementia³⁸. Consistent with earlier studies, half of the respondents in the follow up study could not complete the QOL-AD^15,10, though, in this study the QOL-AD was completed by 21 residents and 10 staff at follow up but not at baseline. This indicates that some of the QoL ratings may have been unnecessarily missed at baseline and that QOL-AD scores could have been provided for these residents and staff members. It is possible that these discrepancies were due to observer bias, as rater competence and confidence may have been increased over the course of the study as the researchers became more familiar with the scales. This may be indicated by the increased response rate for staff members through time. Alternatively, the residents’ ability to complete the QOL-AD may be as a result of fluctuations in cognition, particularly in attention and comprehension, which may have been influenced by underlying physical ill health, or apathy and withdrawal. All participants were resident in a care home, therefore these findings may only be generalised to this population. Perceptions of QoL may also be different for those people unable to complete the QOL-AD and those lost to follow up as they tended to have more severe dementia

Implications

The QOL-AD can be used as part of a package to assess the impact of interventions and treatment in care homes, for research or in clinical practice^5,6. According to a European consensus, the QOL-AD is the measure of choice for evaluating the effects of psychosocial interventions on QoL for people with dementia³⁹. There is a need for further longitudinal studies to investigate naturalistic change in QoL which should allow for QoL in people with dementia to be assessed at regular intervals in time, thus allowing for the impact of dementia on QoL to be assessed throughout the course of the disease. Future studies should include QoL consistently as a measure of outcome in research.

Conclusion

In conclusion, changes in QoL were apparent for individual residents with dementia at follow up and were associated with changes in mood and cognition. QoL does not necessarily diminish as dementia worsens but continues to be strongly influenced by the resident’s mood and interaction with their environment. Improvement in cognition and mood may lead to increased QoL.

The routine use of QoL as an outcome measure for people with dementia in care homes is indicated to identify the benefits of clinical interventions and treatments, and to assess the impact of the disease process on individuals through the course of the illness. Further studies investigating change in QoL in people with dementia should examine the individual scores for participants’ QoL and clinical ratings and not just the sample mean, as this may be misleading.