Health‐related quality of life and related factors among a sample of older people with cognitive impairment

Abstract Aim This study aimed to identify factors affecting health‐related quality of life (HRQoL) of older adults with cognitive impairment and to describe the association of these factors with different components of HRQoL. Design A cross‐sectional, descriptive research design was used. Methods Data were collected from 247 individuals aged 60 years and older from a Swedish longitudinal cohort study. The Short‐Form Health Survey‐12 (SF‐12) and EuroQol (EQ‐5D) were used to assess HRQoL. The data were analysed using descriptive and comparative statistics. Results The present study identified several factors that influenced HRQoL of older adults with cognitive impairment. The results of a multiple logistic regression analysis revealed that the following factors were associated with physical and mental HRQoL: dependency in activities of daily living (ADL), receiving informal care and feelings of loneliness and pain.


| INTRODUC TI ON
In an ageing population, the risk of poor health and cognitive impairment increases with age (Teeters et al., 2016). According to Murman (2015), cognition is critical for functional independence as people age. Although many older adults retain good health in old age, cognitive impairment can lead to increased dependency in daily life (Ahn et al., 2009;Reppermund et al., 2013). Such dependency can affect different aspects of health and overall quality of life (QoL) (Ahn et al., 2009;Reppermund et al., 2013). To support health and maintain independent living in old age, it is important to understand the challenges that older individuals face, including the most important factors that influence health and QoL. There is growing interest in measuring these challenges and the factors affecting health-related quality of life (HRQoL) as outcome measures in the field of health research (Leow et al., 2013).
According to its definition, health is "a state of complete physical, mental and social well-being and not merely the absence of disease and infirmity" (p. 100). Despite the many definitions, a measure of health status is difficult to establish using a single measure. Mainly because health varies greatly among individuals and is influenced both by culture and a person's age (Hernandez, Blazer & Institute of Medicine (US) Committee on Assessing Interactions Among Social, Behavioral, and Genetic Factors in Health 2006). But also, because health comprises many interrelated components, such as physical functioning, cognitive functioning, mental well-being and selfregulation (Leist, Kulmala, & Nyqvist, 2014). Furthermore, health is often described as a positive concept that has been operationalized in terms of self-reported (mental and physical) health status and broader health status, which includes the ability to continue with daily social role functioning (Huber et al., 2011;Singh & Dixit, 2010).
Thus, the establishment of health status should be based on definitions of the concept of health that are multidimensional (Bowling, 2017).
Ageing is associated with both biological and psychological changes, including cognitive impairment, such as dementia (Glisky, 2007). Dementia is a multifactorial disorder, which is characterized by progressive deterioration in memory and other cognitive domains (Winblad et al., 2016). Both the prevalence and incidence of dementia increase significantly with advancing age, and it is a major cause of functional dependence, institutionalization, poor QoL and mortality in older adults (Prince et al., 2015;Qiu, von Strauss, Bäckman, Winblad, & Fratiglioni, 2013). The negative impacts of dementia on health and QoL of older adults are well established (Cooper et al., 2012). The effects of dementia on QoL encompass cognitive function, activities of daily living (ADL), social interactions and mental health (Beerens et al., 2014). Previous research demonstrates that even early stages of dementia (i.e. mild dementia or mild cognitive impairment) adversely affect QoL (Bárrios et al., 2013). Research also found a strong association between health and QoL among older adults (Baernholdt, Hinton, Yan, Rose, & Mattos, 2012). Furthermore, self-rated health is shown to be strongly associated with subjective views of QoL in people with dementia (Orgeta, Edwards, Hounsome, Orrell, & Woods, 2015).
Despite the aforementioned findings, the relationship between health and QoL is complex (Borglin, 2005). To measure the health status of older adults with cognitive impairment, their health needs to be based on a definition that encompasses QoL. The application of the concept of QoL in the area of health and illness is referred to as HRQoL. HRQoL is a multidimensional concept that has evolved since the 1980s to encompass components of overall QoL that can be clearly shown to affect health (Bowling, 2001). These components include physical, mental, emotional and social functioning (Bowling, 2001). HRQoL is often used in scientific research to measure health status and population health outcomes since it provides a broad summary measure of perceived health (Kindig, Asada, & Booske, 2008). Previous research of older populations shows that the negative impact of cognitive dysfunction on HRQoL increased with the severity of cognitive dysfunction and is associated with a lower physical and mental health (Pan et al., 2015). In addition, poor mental health is associated with cognitive impairment in old age (Bauermeister & Bunce, 2015).
Many older adults with cognitive decline can continue to live at home in the early stages of the disorder, with support from a family member or relative, who usually serves as an informal caregiver (Lethin, 2016). However, as the condition progresses, the individual may suffer a loss of independence and have an increased need for care assistance with both basic ADL, such as functional mobility, dressing, bathing and toileting and instrumental activities of daily living (IADL) (e.g. preparing meals, cleaning, cooking and handling finances) (DeFries, McGuire, Andresen, Brumback, & Anderson, 2009;Schulz & Martire, 2004). A study of functional ability in older adults demonstrated that individuals with mild cognitive impairment had poorer IADL functioning as compared with that of cognitively intact older individuals (Johansson, Marcusson, & Wressle, ). The same study found that requiring help with ADL had a significant negative effect on older adults HRQoL. Therefore, measuring HRQoL has become an important objective in dementia research, where the importance of valuing the perspective of the person with dementia is emphasized (Hounsome, Orrell, & Edwards, 2011). There has been a recent increase in studies investigating the relationship between cognition and HRQoL. However, research about factors of significance for the HRQoL among this group of older adults is still sparse.
Hence, the purpose of this study was to identify factors affecting HRQoL among older adults with cognitive impairment and to discuss the association of these factors in relation to different components of HRQoL.

| Design
A cross-sectional, descriptive research design was chosen. The Short-Form Health Survey-12 (SF-12) and EuroQol (EQ-5D) were used to assess HRQoL among older adults with cognitive impairment. The data were analysed using descriptive and comparative statistics.

| Study population
The study population consisted of individuals who took part in the Swedish National Study on Aging and Care (SNAC). SNAC is a longitudinal, multicentre, cohort study, which commenced in the county of Blekinge, Sweden in 2001-2003(Lagergren et al., 2004

| Measures
The outcome variable in this study was HRQoL. The study employed Ware's (1987) multidimensional perspective of HRQoL, where there are five inherent health concepts: physical health, mental health, social functioning, role functioning and general well-being. In Ware's definition of HRQoL, the goal of health care is to maximize the health components of QoL. In this study, we applied this definition in the application of two health measurement scales: the SF-12 and EuroQol (EQ-5D). Since both scales aim to measure functioning, well-being and general health status they are considered appropriate to answer the purpose of this study.
With regard to the measure of HRQoL and the health measurement scales used in this study, there were some internal dropouts. A comparison of the study population that completed  and the population that did not complete SF-12 (N = 68) showed that the non-participants/non-responders belonged to the oldest participants (81-87 and 90-105 years). In contrast, a comparison of participants who completed the EQ-5D (N = 237) and those that did not (N = 10) revealed that most the non-participants/non-responders belonged to the youngest participants (60-66 years). Thus, the final study consisted of 247 participants who were included in the health measurements analysis.

| The Short-Form Health Survey-12
The SF-12 is a generic, standardized questionnaire used to measure two components of HRQoL: physical and mental health. It is a shorter, yet valid alternative to the most widely used health status scale: the Short-Form-36 Health Survey Questionnaire (SF-36). The SF-12 reproduces the physical and mental component summary scores (PCS/MCS) of the SF-36. As it has a broad generic approach and does not target a specific age or disease group, it is suitable for use with larger samples where broader health outcomes are monitored (Ware, Kosinski, & Keller, 1996). In addition, as the SF-12 contains fewer questions and takes less time to complete than the SF-36, it is more appropriate for use with older adults with common geriatric diseases, such as dementia (Jenkinson & Layte, 1997;Ware et al., 1996). The SF-12 is weighted and summed to provide easily interpretable scales for the two health components.
In this study, the results are presented as PCS-12 and MCS-12, where scores range from zero (lowest HRQoL)-100 (highest HRQoL). In a previous study of the general Swedish population aged 75 years and older, scores of 40.3 (SD 11.6) and 51.5 (SD 11.0) on the PCS-12 and MCS-12, respectively, were considered normal (Sullivan, Karlsson, & Taft, 1997). In this study, a cut-point on the 25th percentile was used to establish the limit for low and high HRQoL since no validated cut-points calculated for an older adult population with cognitive impairment was described in the literature. When establishing cut-points from a continuous scale, distribution-based or anchor-based strategies are usually used in the area of health measurement (Streiner, 2014). The present study applied a distribution-based strategy. A value of 27 or less denoted low physical HRQoL, and a value of 46 or less denoted low mental HRQoL.

| The EuroQoL
The EQ-5D is a generic, multidimensional health self-report questionnaire, designed to generate a single index value for each health state (Bowling, 2017). The questionnaire contains five health-related items: mobility, self-care, usual activities, pain/discomfort and anxiety/depression. The respondents classify their own health status in three levels per item. Utility scores quantify HRQoL along a continuum that ranges from −0.59 (worst health) -1.00 (perfect health).
Respondents are also asked to rate their current health status on a 100-point visual analogue scale (VAS), where higher points represent the best imaginable health state (Bowling, 2017).
The analysis of the EQ-5D index is based on a time trade-off (TTO) value set and the VAS. As there is no Swedish TTO value set for EQ-5D health states, the UK EQ-5D index tariff was used in this study to obtain values for health status (Dolan, 1996). A cut-off on the 25th percentile was used to establish the limit for low and high HRQoL. Cronbach's alpha for the index was calculated as 0.692.

| Assessment of dependency in functional ability
The ADL and IADL indices developed by Katz, Ford, Moskowitz, Jackson, and Jaffe (1963) and Lawton and Brody (1969), respectively, were used to assess the participant's dependency in functional ability in everyday life. The ADL index describes the functional status of elderly and chronically ill patients. It assesses independence in basic ADL, such as personal hygiene care, toileting and dressing. The index is composed of 6 items, where the individual is graded on an ordinal 3-point scale, with lower scores indicating full functioning (Bowling, 2017). In this study, each item was rated dichotomously (0 = dependent and 1 = independent) and then summarized to create an index. The same procedure was performed for the IADL scale, which is composed of eight items. On the IADL scale, higher scores indicate greater functional ability and the scores are corrected according to the gender of the participant.

| Statistical analysis
All statistical analyses were performed using the Statistical Package for Social Science, version 24.0 (SPSS Inc., NY, USA). The clinical and socio-demographic characteristics of the study sample were analysed using descriptive and comparative statistics. The results are presented as absolute frequency (N) and relative frequency (%). As the data were not distributed normally in non-parametric tests, the Mann-Whitney U test and Kruskal-Wallis test were conducted for comparisons between sex and age. Univariate analyses (i.e. correlations) with Spearman's rho (r s ) and binary logistic regression analysis were performed to assess the importance of

| Ethical consideration
This study was conducted in accordance with the ethical principles for research outlined in the Declaration of Helsinki (WMA, 2013). All the participants provided written informed consent before entering the SNAC study, and Research Ethics Committee approval was obtained from the regional ethics review board in Lund (LU No. 650-00 and No. 744-00).

| The EQ-5D scores
According to the EQ-5D, an almost equal proportion of women and men had a low HRQoL (18.5% vs. 19.6%), with no statistically significant difference (p = 0.730). Concerning the different components of HRQoL, there were no sex-related differences in any of the EQ-5D components (Table 3). When the data were compared according to age groups, the results showed significant differences in the EQ-5D components' mobility (p < 0.01), self-care (p = 0.013) and usual activities (p = 0.006) between the different age groups (Table 4).

| Dependency in functional ability
When the ability to perform ADL was analysed, the results showed that 89.9% (N = 222) of the participants had full functioning, 6.4% (N = 16) had moderate functioning and 3.6% (N = 9) had severe functional impairment. Regarding the ability to perform ADL, there were no significant sex-related differences in dependency in functional Severe pain 4 (4.6) 6 (5.6) 10 (5.2) Note. The Mann-Whitney U test was used in the comparisons between sex. * p < 0.05 significance level for the difference between sex.
TA B L E 1 (Continued) ability (p = 0.297), although more women than men (11.9% vs. 7.7%) were dependent on support to accomplish ADL (Table 1). In two ADL functions (i.e. bathing and continence), there were significant differences between women and men (p = 0.036 and p = 0.009, respectively). In total, 86.6% (N = 214) of the participants were able to independently perform bathing, whereas 13.4% (N = 33) of the population was dependent on support. Concerning the ADL function "continence," 72.9% (N = 180) of the participants were able to independently exercise complete self-control over urination and defecation, whereas 27.1% (N = 67) of the participants were partially or totally dependent on support. In both cases, women were more dependent than men.
Regarding the ability to perform IADL, the comparison between the means for women and men revealed no significant difference in overall functional ability (p = 0.731), although more men than women were dependent on support to accomplish IADL (22.8% vs. 20.9%) ( Table 1). There was a significant difference between women and men in one IADL function: mode of transportation (p = 0.008).
Women were less likely than men to travel independently, either in their own car or on public transportation and they were less likely than men to travel on their own (38.7% vs. 22.1%).
There were significant sex-related differences in the relationship between physical HRQoL (PCS-12) and functional ability in ADL (p = 0.009 for women; p = 0.003 for men) and IADL (p = 0.012 for women; p = 0.007 for men).
The multiple logistic regression analysis resulted in two models including the outcome variables physical HRQoL (PCS-12) and mental HRQoL (MCS-12). The first model ( Participants who sometimes experienced loneliness were more likely to have a low mental HRQoL than those who never felt lonely. More women than men reported that they sometimes felt lonely (52.4% vs. 43%, p = 0.244). In addition, there was a weak correlation between loneliness and living alone (r s = 0.390, p < 0.01).

| D ISCUSS I ON
The present study aimed to identify and describe factors associated with HRQoL among older adults with cognitive impairment.
The results revealed that HRQoL of the study population as regards physical and mental health was similar to that of the general Swedish geriatric population (Sullivan et al., 1997) TA B L E 2 HRQoL outcome for gender and age cohorts impairment. In addition, there were hardly any differences in HRQoL between males and females. Earlier research (Franzén, Saveman, & Blomqvist, 2007;Hopman et al., 2009) that studied HRQoL in older adults with chronic diseases showed that chronic heart failure was strongly associated with reduced HRQoL. Further, comorbid conditions, such as diabetes and respiratory diseases, as well as female sex and older age, were associated with poorer physical health.
The results further revealed an association between pain and functional ability in relation to physical health. Having moderate to severe pain or being dependent in ADL was correlated with lower physical health. This is in line with previous research (Fagerström & Borglin, 2013), which showed that functional ability in older adults was associated with both physical and mental health. In the present study, the dependency levels of women and men differed, with women in general more dependent than men on support to perform daily activities. Despite this finding, the results indicated that women who were dependent on support to accomplish ADL had a higher HRQoL than men who were dependent, both in terms of physical and mental health. Previous studies of the functional ability of older adults with cognitive impairment did not clarify the impact of basic ADL functioning on HRQoL (Mlinac & Feng, 2016;Pusswald et al., 2015). According to Bowling (2017), most measures of functioning focused narrowly on mobility, self-care and instrumental tasks and did not always include assessments of more meaningful aspects of a person's functioning (e.g. the time taken to complete a task). Many measures of functional ability have been developed that are suited to a particular age or disease group and validated with people living in institutions or in the community (Comans et al., 2018;Wilson et al., 2011). Combining such measurements with self-rated health measurements, as was done in the present study, may provide a clearer picture of a person's overall functional status. Knowledge about functional status could thereby support HRQoL.
The results of the current study also identified an association between loneliness and mental health. Participants who sometimes experienced loneliness were more likely to have a low mental HRQoL than those who never felt lonely. According to previous research, loneliness and social isolation were associated with mental illness in old age (Loboprabhu & Molinari, 2012;Wenger, Davies, Shahtahmasebi, & Scott, 1996;Wilkes, 1978). Studies also found that the absence of loneliness was related to a good QoL among older adults (Gerino, Rollè, Sechi, & Brustia, 2017;Kang, Park, and Wallace (Hernandez) (2016)). However, according to other research, loneliness was distinct from social isolation, with the former linked to the absence of an intimate relationship (emotional loneliness) or a wider social network (social loneliness) (de Jong-Gierveld, van Tilburg, & Dykstra, 2006). Andersson (1998) and Tornstam (1990) asserted that it was reasonable to assume that both levels of loneliness are related and that inconvenience caused by one type can be alleviated by compensation from the other level. The results of the present study cannot shed light on the levels of loneliness experienced by the participants, as the main focus was to assess whether loneliness, as a one-dimensional concept, influenced HRQoL of older adults with cognitive impairment.
However, the different levels of loneliness could be an interesting topic for further research.
The findings from this study are in line with earlier research that investigated HRQoL among older adults (Fagerström & Borglin, 2013;Pan et al., 2015), which emphasizes the importance of purposeful care for this group. Assessing older adults physical and mental health has become more urgent. Mainly due to the increasing proportion of mental illness. Such research is also necessary because older adults with functional impairments, due to physical and cognitive illnesses, can have difficulty meeting basic-level needs. The latter may affect

| Limitations
The present study has some limitation as regards the clinical importance of the results. Dropouts in the health measurement scales (SF-12 and EQ-5D) may have affected the results as the participants selected for the analysis in this study was based on the specific MMSE score 20-26 points. A possible explanation for the dropout rate is therefore the measurement of HRQoL was based on selfreported data, which can lead to an increased risk of loss of infor- on health measurement scales, cut-off points are used to predict a dichotomous outcome. When establishing cut-off points, although a difference in mortality of 1% may be of little consequence, a difference of 3.1 points on a HRQoL scale is much less comprehensible, as the measurement of health and QoL is subjectively self-rated.
Consequently, there have been attempts to establish minimum standards for clinically important change using cut-off points (Streiner, 2014). We believe that the cut-points used in this study do not change the clinical importance of the results.

| CON CLUS ION
The results of this study provide an overview of important factors that influence HRQoL of older adults with cognitive impairment. Living at home and having the ability to independently perform ADL without receiving informal care, together with the absence of moderate to severe pain and loneliness, were associated with high HRQoL in the study population. Recognizing the significance of the aforementioned factors is important when adopting a person-centred approach to support care needs and enhance independence in self-care activities of older adults with cognitive impairment. It can also increase the possibility of maintaining high HRQoL among older adults with cognitive impairment. The results of this study can be used as a foundation for qualitative research aimed at improving the overall QoL of this group. Further research is needed to explore interventions to support and improve HRQoL among older adults with cognitive impairment.

ACK N OWLED G EM ENTS
This study was supported by Blekinge Institute of Technology, Sweden and The Swedish National study on Aging and Care, SNAC (www.snac.org) which is financially supported by the Ministry of Health and Social Affairs in Sweden. The SMART4MD project, funded by the European Union-Horizon 2020, financed the PhD studies for LC.

CO N FLI C T S O F I NTE R E S T
The authors declare no conflicts of interest with respect to the research, authorship and/or publication of this article.