Components and determinants of quality of life in community-dwelling older adults

Abstract

The objectives of this study are to detect the main components of global quality of life (QoL) of community-dwelling older adults from their own perspective and to identify determinants of health-related and global QoL in the same population. This is a cross-sectional study covering a representative sample of 1,106 community-dwelling adults aged 60 years and older residing in Spain. The survey collected information on QoL through a face-to-face interview asking for QoL components in free-format, as well as the completion of two QoL measures, the EQ-5D and the Personal Wellbeing Index. The most important QoL dimensions, according to the participants of this study, were health, family, and finances. Depression was the main determinant of both QoL indices, while functional independence and social support specifically influenced health-related and global QoL, respectively. Based on the perspective of the older adults as well as on statistical analysis, this work emphasizes the importance of health, family, and social support as areas of special interest in aging. There was a discrepancy when comparing findings related to the importance of financial status. Results also support that global and health-related QoL share some common determinants, but with different weights for functional independence and social support.

Introduction

The rapid increase of population aging in Europe has attracted great interest in actions targeted at widening knowledge and, subsequently, developing policies and interventions in the promotion of health and quality of life (QoL) in old age (European Research Area in Ageing ERA-AGE 2010; Walker et al. 2004).

The use of multimethod approaches helps to measure not only the overlapping, but also distinct aspects of QoL (Schalock 2004). Therefore, the comparison of standardized measures with qualitative information is useful to better understand the QoL in old age. In this study, we use both approaches to explore older adults’ wellbeing through the multidimensional constructs of health-related and global QoL. The field of health-related QoL especially focuses on the effects of health, illness, and treatment on QoL (Ferrans et al. 2005), making it possible to capture the individuals’ perceptions of those effects in their lives (Martínez Martín 2006). Consequently, attention has been drawn to health-related QoL as an outcome to assess health programs and interventions (Badía Llach et al. 2002). The concept of health-related QoL is controversial. Currently, however, agreement exists about the three broad domains of physical, psychological, and social wellbeing that make up the basis for most approaches to health-related QoL assessment (Phillips 2006). These three domains are consistent with the World Health Organization's definition of health (O’Boyle 1997; Phillips 2006). Another important aspect of the concept of health-related QoL is connected to the subjective perception of health status. It refers to the assessment of health in general, thus its measures do not focus on a specific dimension of health (Baron-Epel and Kaplan 2001).

On the other hand, global QoL is a wider and more comprehensive concept defined as a dynamic interaction between the external conditions of the individuals’ life and the perceptions they have of these conditions (Browne et al. 1994). It is formed by a wide spectrum of life domains (health, family relationships, social support, residential environment, and economic situation, among others) and positive and negative experiences and feelings (Brown et al. 2004). In this sense, the importance and usefulness of subjective evaluations of QoL have been widely demonstrated (Borglin et al. 2005; Browne et al. 1994; Cummins 2005). Specifically, satisfaction scales are extensively used as they represent judgments of an individual’s life as a whole or with respect to specific domains (Bowling 2005; Diener 1994). Accordingly, QoL is largely operationalized through subjective measures such as the Personal Wellbeing Index (PWI), which has been used in many different contexts (International Wellbeing Group 2006; Lau et al. 2005; Rodriguez-Blazquez et al. 2011).

Among the main determinants of QoL, a general agreement exists on the key role of health (Fernández Ballesteros et al. 1996; Michalos et al. 2000). However, its contribution to QoL is variable with discrepancies between health status and QoL: many older adults report a satisfactory QoL despite having several health problems, a finding that may be due to other factors that help to compensate health decline (Fernandez-Mayoralas and Rojo Pérez 2005; Grimby and Svanborg 1997; Grimby and Wiklund 1994; Prieto-Flores et al. 2010). In addition, the process of adaptation to health deterioration in old age contributes to the reduction of the relative importance of health in QoL.

In spite of the key role of health as a QoL determinant, health is not always ranked in the first place by older adults (Bowling 1995; Xavier et al. 2003). Determinants of QoL in old age may vary in type and importance, when compared between groups according to age, health characteristics, and living circumstances (Bowling and Zahava 2004). Different studies have contributed to the identification of QoL domains from the perspective of aging people. Health, family, social relationships and finances (Seymour et al. 2008), autonomy and attitude toward life (Richard et al. 2005), and health of significant others are some of the most relevant dimensions according to older adults’ point of view (Bowling 1995; Prieto-Flores et al. 2010).

The combination of standardized and free-format approaches through the comparison of main QoL components and determinants identified might provide an inclusive view in aging research. Ultimately, this approach has potential to inform health and social policies targeting the population of older adults.

The current study is based on a multimethod data analysis that includes, in a single study, qualitative information on the main QoL domains identified by community-dwelling older adults and quantitative data regarding health-related and global QoL. In this context, the objectives of the present study are: first, to detect the main components of QoL of older adults from their own perspective; second, to identify determinants of health-related and global QoL in the same population.

Methods

Participants

We used a cross-sectional QoL survey conducted in 2008 on a representative sample of the community-dwelling population, aged 60 years or older in Spain. A description of the sample has been published previously (Fernandez-Mayoralas et al. 2012; Forjaz et al. 2011; Prieto-Flores et al. 2011; Rodriguez-Blazquez et al. 2011). A representative sample was drawn from the total population in Spain, aged 60 years or older: 9,812,307 individuals (INE 2007). To improve representativeness, we used geodemographically based proportional stratified sampling. The strata consisted of Autonomous Regions (Comunidades Autónomas; 14 groups), town population size (7 groups), age (3 groups), and sex. For an allowable sampling error or level of precision of ±3.5 %, a design effect of 1.4, a 95 % confidence level, and the most unfavorable probability (50 % chance of any event), the sample size was n = 1106. A face-to-face survey at home was undertaken by experienced interviewers, from a consulting company specialized in social and health surveys, who were purposely trained for this study. The interviewers had experience in using the Schedule for the Evaluation of Individual Quality of Life (SEIQoL) as described in a previous study (Fernández-Mayoralas et al. 2007).

The inclusion criteria were: older adults aged 60 years and older, living in the community, who gave informed consent, and did not have cognitive decline that could affect the ability to respond to the questionnaire, according to the Pfeiffer’s Short Portable Mental Status Questionnaire (SPMSQ) (Pfeiffer 1975). A percentage of 4.4 % persons of the original sample (51 subjects) with five or more errors in the SPMSQ, indicative of cognitive decline, were excluded. The participants were contacted personally and were informed about the purpose of the study and that their participation was voluntary. Individuals who refused to participate or failed to meet inclusion criteria were replaced by others from the same stratum. The study was approved by the Bioethics Committee of the Carlos III Institute of Health.

Measures

The survey included information on sociodemographics, psychosocial and health indicators, and health-related and global QoL measures.

To assess the main QoL components, interviewees were asked to name the five aspects that contribute most to their QoL. The process to obtain this information was as follows. First, an open-ended question, at the beginning of the questionnaire, was conducted as part of the QoL survey: “We would like you to name the five aspects that most contribute to your QoL, to your satisfaction or dissatisfaction with life.” The interviewers coded the answers following a list of 18 pre-defined codes or specified into an additional box for “other aspects.” Each participant was interviewed by only one interviewer. The coding process was supervised by the researchers; in case of doubt, the interviewer wrote the participants' statement verbatim, which was coded later. Coding assignment difficulties were solved by consensus after discussion. The list of pre-defined codes was based on a previous study of the research group in which the SEIQoL was applied (Fernández-Mayoralas et al. 2007).

The following measures were used as dependent variables in the statistical analysis: EQ-5D index to assess health-related QoL; and the PWI for global QoL. In the selection of the independent variables, the main QoL aspects reported by the participants (health, family, and financial status) in this and a previous study by our group (Fernández-Mayoralas et al. 2007) were taken into account. These have in good measure coincided with earlier works of QoL in old age (Browne et al. 1994; Seymour et al. 2008). In addition, we selected sense of coherence and social support variables, all the importance of which is reflected in the literature on the multidimensional concept of QoL (Cummins 1996; Schalock 2004).

The EQ-5D index is part of the EQ-5D instrument (Badía et al. 1999; The EuroQol Group 1990), a generic self-applied measure. It includes a descriptive system made up of 5 items: mobility, self-care, usual activities, pain/discomfort, and anxiety/depression, scored from 1 (no problems or symptoms) to 3 (extreme problems or important symptoms). The answers from the descriptive profile were converted by the time trade-off method into an index (the EQ-5D index) with values ranging from 0 (death) to 1 (perfect health), though it can also assume negative values for health situations considered worse than death (Badía et al. 1999). To evaluate the current health status, the Visual Analog Scale (EQ-VAS) was employed. It ranges from 0 (worst imaginable health status) to 100 (best imaginable health state).

The PWI was created by the International Wellbeing Group (International Wellbeing Group 2006; Lau et al. 2005; Rodriguez-Blazquez et al. 2011) and consists of seven items on satisfaction with the following quality of life domains: standard of living; health; achievement in life; personal relationships; safety and protection; community-connectedness; future security; and protection. Items are scored from 0 (completely dissatisfied) to 10 (completely satisfied) with the midpoint, 5, indicating a neutral feeling. The PWI is represented by the mean value of its items transformed into a percentage of the maximum possible score.

The following sociodemographic data were included: sex, age, marital status (with or without a partner), having children or not, and monthly income (categories: <€600, €601–900, and >€900).

Regarding psychosocial indicators, the following scales were used: the Spanish version of the Duke-UNC Functional Social Support Questionnaire (DUFSS) (Broadhead et al. 1988; De la Revilla et al. 1991); and the Sense of Coherence Scale (Lundberg and Peck 1995). The DUFSS (Broadhead et al. 1988; De la Revilla et al. 1991) evaluates the perceived social support through the dimensions of affective support and a confidant relationship. It is made up of 11 items, and higher scores reflect higher perceived social support. The short version of the Sense of Coherence Scale (Lundberg and Peck 1995) based on Antonovsky’s Sense of Coherence Scale contains three items corresponding to the dimensions of manageability, meaningfulness, and comprehensibility, which reflect a global orientation of viewing the world and the individual environment (Eriksson and Lindström 2005). The total scale score ranges from 0 to 6, such that the higher the value, the lower the sense of coherence.

The following health indicators were used: Barthel Index for functional independence (Mahoney and Barthel 1965); the depression subscale of the Hospital Anxiety and Depression Scale (HADS-D) (Zigmond and Snaith 1983); and a comorbidity scale. The Barthel Index (Mahoney and Barthel 1965) evaluates functional capacity for 10 activities of daily living with a total score ranging from 0 (completely dependent) to 100 (completely independent). The HADS-D (Zigmond and Snaith 1983) contains seven items measuring feelings of anhedonia and slowing down. Higher scores indicate more symptoms of depression. The comorbidity questionnaire was based on the Cumulative Illness Rating Scale-Geriatrics (Fernández-Mayoralas et al. 2007; Miller et al. 1992) and the number of self-reported medical conditions was taken for the current study. The questionnaire explores the presence (yes/no) of 20 medical conditions plus an open question for “others.” The conditions considered were: cardiovascular, arterial hypertension, hypercholesterolemia, respiratory, gastrointestinal, bone and joint, genitourinary, visual problems, hearing problems, insomnia, memory problems, depression, Alzheimer’s and Parkinson’s disease, other mental problems, diabetes, mouth and teeth problems, tumors or cancer, allergy, and headaches.

Data analysis

Once the free-format responses to the question of the main aspects of QoL were coded, we presented the five most frequently named aspects by the order in which they were reported. We also reported the total number of responses given for each component of QoL.

On the other hand, to identify determinants of the health-related and global QoL measures, two multiple linear regression models were applied. For the model of health-related QoL, we used as dependent variable—the EQ-5D index. Likewise, in the analysis of global QoL, the dependent variable was the PWI. Sociodemographic, psychosocial, and health characteristics were included as independent variables in both regression models toward identifying differential factors between health-related and global QoL.

The sociodemographic characteristics included in the analyses were: age and sex as control variables, marital status, children, and income. Sense of coherence and DUFSS were included as psychosocial measures; and number of chronic medical conditions, HADS-D and Barthel Index, as health indicators. The Barthel Index was included as an independent variable in the EQ-5D index model even if some of the items on which the EQ-5D index is based are related to problems performing activities of daily living (Spearman correlation between the Barthel index and EQ-5D index: 0.38, p < 0.001). The reasons behind this decision were: first, the great impact that functional dependence has on the life of older adults; second, to analyze the independent effect of other variables controlled by the level of functional independence; third, to use the same independent variables in health-related and global QoL models to detect differential factors between both models; and fourth, functional ability is a predictor of health-related QoL according to the Wilson and Cleary model (1995).

Statistical analyses were performed by means of the Stata IC/10.1 for Windows computer software program (College Station, Texas, USA).

Results

Of the total number of participants, 56.3 % were women and the mean age (±standard deviation) was 72.1 ± 7.8 (range: 60–96) years.

Table 1 shows the results for the five dimensions rated by individuals as being most important for their QoL in the order in which these were informed. From the first to the fifth named component, there was a slight increase of missing values. The most reported dimensions were health, children, financial status, spouse/partner, and other family members. Participants named specific family members when reporting the main QoL dimensions, whereas health was mentioned as a unique concept without any specification. Among the main five named aspects, family occupied a major place, not only because each dimension (children, partner, and other family members) was frequently reported, but also because when adding them together family occupies then the first place before the health dimension. Either way, financial status remains in the third place. Other aspects mentioned by the older adults interviewed were mood or emotional state, friends, housing, lifestyle, neighborhood, free time, neighbors, and values and attitudes to life among others.

Table 1.

Aspects that contribute most to quality of life, ordered by total, according to participants’ perspectives

	Order in which reported					Total no. of responses
	1st	2nd	3rd	4th	5th
Health	585	128	158	105	36	1,012
Children	212	298	204	82	72	868
Financial status	34	197	112	184	139	666
Spouse or partner	194	204	116	59	15	588
Other family members	12	96	160	135	79	482
Other QoL domainsa	63	177	347	528	753	1,868
Total of participants	1,100	1,100	1,097	1,093	1,094

^aOther QoL domains named, in the order of number of times named: Mood or emotional state (294), Friends (254), Housing (252), Lifestyle (223), Neighborhood or town of residence (180), Free time available (132), Neighbors (120), Values and attitudes to life (111), Occupational status (110), Religious beliefs and spiritual life (78), Leisure activities pursued (62), Social institutional support (32), Participation in social and community life (30)

Concerning the health characteristics, the participants reported an average of 3.3 ± 2.5 (range: 0–15) medical conditions, 117 individuals (10.6 %) stated that they had no medical condition, and 2 persons (0.2 %) reported up to 15. The five most frequently reported medical conditions (affecting over 25 % of the sample) were: bone and joint problems (53.3 %); arterial hypertension (39.9 %); visual problems (29.0 %); hypercholesterolemia (28.8 %); and cardiovascular problems other than arterial hypertension (26.1 %). Only 31 individuals (2.8 %) had severe or very severe disability according to the Barthel Index, while 837 individuals (77.9 %) had no functional limitation whatsoever.

With respect to EQ-5D items, 22.0 % of the respondents reported problems of mobility, 9.1 % experienced problems in personal care, 17.6 % had problems in daily activities, 50.5 % had pain, and 21.4 % suffered from anxiety/depression. Data on the psychosocial, health, and QoL variables are shown in Table 2. Overall, the sample displayed a relatively good health status (EQ-VAS) though the score range indicates that both extremes were represented in the sample. The participants reported a good satisfaction with the PWI quality of life domains, as indicated by a PWI value of 71.03 ± 13.50 (range: 15.71–100). The EQ-5D index and PWI were moderately correlated (Spearman correlation coefficient r = 0.397, p < 0.01).

Table 2.

Descriptive characteristics of the sample

	Median	Mean	SD	Range
Psychosocial indicators
DUFSS	46	44.95	8.99	11–55
Sense of coherence	2	2.30	1.28	0–6
Health indicators
Barthel Index	100	95.82	11.05	0–100
HADS-D	4	4.92	4.30	0–21
Medical conditions	3	3.29	2.48	0–15
Health-related QoL
EQ-5D index	0.88	0.83	0.25	−0.65–1.00
EQ-VAS	70	66.16	20.88	0–100
Global QoL
Personal wellbeing Index	71.42	71.03	13.50	15.71–100

SD Standard deviation, DUFSS Duke-UNC Functional Social Support Questionnaire, HADS-D Hospital Anxiety and Depression Scale-Depression subscale, QoL Quality of Life, EQ-5D index Index of health-related QoL, EQ-VAS Visual Analog Scale of health status

The regression models used to measure the effect of sociodemographic, health, and psychosocial characteristics on health-related and global QoL are presented in Table 3. Standardized regression coefficients are presented to allow comparisons of the relative impact of the variables. Depression (HADS-D) was the determinant with the greatest influence, showing a negative effect in the model of health-related QoL (−0.335, p < 0.001), and in the model of global QoL (−0.347, p < 0.001). The number of medical conditions also contributed to explain both models (−0.335, p < 0.001 and −0.347, p < 0.001, respectively). Functional independence (Barthel Index) was only significant for the health-related QoL model (0.321, p < 0.001), whereas social support (DUFSS) (0.189, p < 0.001) and sense of coherence (−0.242, p < 0.001) strongly influenced global QoL, but not health-related QoL. Finally, being male was positively associated with health-related QoL (0.063, p = 0.006). According to the regression models, age, income, and having children and partner did not contribute to either health-related or global QoL.

Table 3.

Determinants of health-related and global quality of life

	Multiple linear regression model of health-related QoL (EQ-5D index)		Multiple linear regression model of global QoL (Personal Wellbeing Index)
	Standardized coefficient	P value	Standardized coefficient	P value
Age	0.021	0.398	0.015	0.576
Men	0.063	0.006	0.031	0.220
No. of medical conditions	−0.335	p < 0.001	−0.134	p < 0.001
Barthel Index	0.321	p < 0.001	0.008	0.757
DUFSS	0.017	0.504	0.189	p < 0.001
Sense of coherence	0.012	0.650	−0.242	p < 0.001
HADS-depression	−0.335	p < 0.001	−0.347	p < 0.001
Income (€601–900 vs. <€600)	0.003	0.906	0.004	0.908
Income (>€900 vs. <€600)	−0.009	0.749	0.011	0.722
Children	0.004	0.877	0.025	0.341
Partner	0.017	0.490	−0.007	0.802
Constant	–	0.024	–	<.000

Note: Health-related QoL model: Adjusted R ² = 0.56. Global QoL model: Adjusted R ² = 0.46

Discussion

This study furnishes information on the QoL of community-dwelling older adults. Specifically, it aimed at identifying the most important global QoL dimensions for this group and it also sought to analyze the main predictors of health-related and global QoL. While the first objective was to seek the main QoL dimensions according to the participants’ point of view, the second objective was to assess its determinants based on regression analysis. This approach allows multimethod data to be complemented so as to better understand the most important aspects of QoL in old age. This study also helps to identify differential factors associated with health-related and global QoL.

Health was a central dimension for global QoL, since it was a frequently mentioned component, in a spontaneous way, and was ranked first by the population surveyed, thereby demonstrating the importance of health in the global QoL of older individuals. Previous studies coincide in stressing the marked weight of health among the various QoL domains from the standpoint of older adults (Bowling et al. 2001; Fernández-Mayoralas et al. 2007; Seymour et al. 2008).

The participants of this study also gave a great importance to family, specifically to children and partners, when asked to name the five aspects that most contributed to their QoL. When considering children, partners, and other relatives together, the importance of family dimension overrides health. Our results are consistent with previous studies on QoL in old age. Bowling (1995) found that the relationships with family and health were the main dimensions of QoL from the individuals’ perspective. In the same way, health and family, particularly in relation to partners, children and grandchildren were the central QoL aspects based on grounded theory research (Prieto-Flores et al. 2010).

Another key dimension reported by the participants was financial status. This was also one of the most often named components of QoL in the longitudinal study by Seymour et al. (2008). In addition, they found an increasing number of older adults mentioning finances over a period of 3 years.

According to the regression models, depression was the most influential predictor in health-related and global QoL, in line with other studies that show the important role of depression in the QoL of older adults (Brown et al. 2004). On one side, depression negatively affects several dimensions of QoL, nevertheless, a perception bias might also occur, so that depressed people tend to see things more negatively, and also rate their QoL as negative. Other authors have demonstrated the association of depressive disorders and anxiety with worse perceived health (Azpiazu Garrido et al. 2002).

The number of medical conditions emerged as a determinant of health-related and global QoL of older adults, a finding pointing in a similar direction as previous research like the study conducted by Smith et al. (2002) who observed this effect on the subjective wellbeing of the oldest old. Other studies have reported a close link of various medical conditions (such as presence of chronic disease) with perceived health and health-related QoL (Jayasinghe et al. 2009; Zahran et al. 2005). Nevertheless, the number of medical conditions showed a stronger association for health-related QoL than for global QoL in our study. This could be explained by the items included in the EQ-5D and PWI, respectively. The first instrument is formed by five dimensions that can be directly affected by medical conditions (mobility, self-care, usual activities, pain/discomfort, and anxiety/depression), whereas the number of the medical conditions may have a lesser effect on the PWI, which includes other QoL domains besides health.

Insofar as functional independence is concerned, this study confirms earlier results regarding its impact on health-related QoL (Rojo-Pérez et al. 2009) even though the prevalence of dependence in our sample was low. Specifically, the effect of difficulty in mobility (Groessl et al. 2007) and performing activities of daily living (Netuveli et al. 2006) is of great importance in health-related QoL. In the conceptual model of health-related QoL, Wilson and Cleary identify functional ability as an important predictor (Willson and Cleary 1995). In our study, functional independence did not contribute to explain global QoL. This is congruent with previous research showing that many people perceive their QoL as positive despite their disability (Albrecht and Devlieger 1999). Our results suggest that other life domains could attenuate the effect of functional independence. This could be the case of social support, which presented a protective effect on global QoL. Our results are in line with the results previously found in the older population in Spain (Fernández Ballesteros 2002).

In our study, perceived social support did not appear to influence health-related QoL, even though the positive effect of social support and social relationships on health has been shown elsewhere (Azpiazu Garrido et al. 2002; Otero Puime et al. 2006; Zunzunegui et al. 2001). Sense of coherence likewise exerted a significant influence on the global QoL, but not on health-related QoL, of the study sample, thereby confirming its important positive effect in older adults QoL. This personality resource contributes to coping with health problems and to maintain an optimal QoL (Nesbitt and Heidrich 2000).

When comparing determinants of health-related and global QoL, our results stress the differential contribution of psychosocial and health determinants. Health indicators (depression and medical conditions) contributed to explain both health-related and global QoL. This is congruent with what was expressed by the participants themselves, rating health as a main QoL dimension. On the other hand, psychosocial indicators (social support and sense of coherence) were only significantly associated with global QoL in line with the idea that global QoL is more than health-related QoL. In the current study, psychosocial indicators were not statistically significant in the model of health-related QoL contrary to previous findings. This suggests that the health indicators account for so much of the model variance that the contribution of the other variables is displaced. Our results should not be interpreted as denying the relevance of psychosocial factors on health-related QoL. Instead, they highlight that some determinants are more closely associated with health-related and others with global QoL.

The findings presented, based on a multi-method approach, contribute to capture the multidimensional nature of QoL. In this direction, this study provides information about the aspects that most contribute to the QoL of older adults, to their satisfaction or dissatisfaction with life, from their own perspective. The study also comprises generic, validated scales widely used to measure QoL and health-related QoL, of which associated factors were analyzed. As in other complementarity multi-method studies, the two approaches assess similar and also different aspects of the same phenomenon (Greene et al. 1989; Schalock 2004).

As convergent findings according to our different approaches, we can highlight the importance given by the participants to the family domain, which is in line with the significant contribution of social support in the model of global QoL. In addition, health also displayed an important effect on the QoL of older adults. The significance of personality resources in QoL, such as sense of coherence, was also reflected through the “values and attitudes to life” dimension, named by the study population among the main QoL components. As a divergent finding, financial status was identified by the interviewees as one of the aspects that most contribute to their QoL, but its effect was not significant when combined in the statistical models. A possible explanation could be that although financial status is considered an important component of QoL in the participants’ perspective, its relative importance is attenuated when controlling for health and psychosocial variables.

Besides pointing out the strengths of this study, we also acknowledge some limitations. First, due to the cross-sectional design, causal inferences should not be made and a longitudinal approach would be needed to examine if the associated factors are indeed predictors of health-related and QoL. Second, results cannot be generalized to older adults living in residential care settings (2.83 % of the Spanish population aged 65 and over) (INE 2009) or with cognitive impairments. By replacing non-respondents by others of the same strata, selection bias was attenuated. However, the extent of selection bias was not possible to assess. More studies are needed with representative samples of community-dwelling older adults. Third, the open-ended question of QoL was not designed to obtain in-depth information, such as why a particular dimension was important, but to identify the most important aspects of QoL named by the participants. Finally, it was not possible to obtain inter-rater reliability information from the coding procedure of the open-ended QoL question.

This study provided two important contributions. First, it allowed detecting the main components of QoL according to the community-dwelling older adults. Our results, based on the perspective of older adults and on the statistical models, underscore the importance of health, family, and social support as areas of special interest in research on aging. There was a discrepancy when comparing findings related to the importance of financial status. Another contribution was the identification of explanatory factors of health-related and global QoL in parallel, thereby allowing comparison between potential determinants of these two constructs. Knowledge about the determinants at the intersection of both constructs (depression and medical conditions), those specific to global QoL (social support and sense of coherence) and those to health-related QoL (functional independence), has potential to inform health and social care policies about priority aspects in older adults’ wellbeing.