Campos et al.

Health and Quality of Life Outcomes 2014, 12:166

[Link]

RESEARCH Open Access

Aging, Gender and Quality of Life (AGEQOL)

study: factors associated with good quality of life
in older Brazilian community-dwelling adults
Ana Cristina Viana Campos1*, Efigênia Ferreira e Ferreira2, Andréa Maria Duarte Vargas2 and Cecilia Albala3

Abstract
Background: In Brazil, a rapidly aging country suffering from large inequalities, the study of the quality of life (QOL)
of aged people is important for the future health. The aim of this study was to examine the associations among
QOL, gender, and physical and psychosocial health in older Brazilian community-dwelling adults to identify factors
that are associated with better QOL.
Methods: The “Aging, Gender and Quality of Life (AGEQOL)” study, which included 2,052 respondents aged 60 or
older, was conducted in Sete Lagoas, Brazil between January and July 2012. The respondents answered questions
regarding their socioeconomic and demographic information, health and social situations, cognitive impairment,
depressive symptoms and family satisfaction. The authors also applied the Brazilian version the World Health
Organization Quality of Life QOL Assessment-Brief Instrument (WHOQOL-BREF) and the World Health Organization
Quality of Life Instrument-Older Adults Module (WHOQOL-Old). Ordinal logistic regression with the Proportional-Odds
and Logit function was used to test the association between QOL and physical and psychosocial health according to
age and socioeconomic status.
Results: Older adults of both genders with five or more years of education, good self-rated health, an absence of
depressive symptoms, and no family dysfunction reported better QOL. Retired men had a better QOL compared
to non-retired men (OR = 2.2; 95% CI = 1.4–3.2), but this association was not observed in females. Men living in
mixed arrangements (OR = 0.5; p = 0.033) and women who did not practice physical activity (OR = 0.7; p = 0.022)
tended to have poorer QOL.
Conclusions: We conclude that there are gender differences related to better QOL in this sample. Women with
good physical and psychosocial health are more likely to have a better QOL. For men, the best QOL was associated
with high socioeconomic conditions and good physical and psychosocial health.
Keywords: Quality of life, Aging, Gender

Background aging”, which was adopted by the World Health

In Brazil, an increase in life expectancy and a decrease Organization (WHO), involves optimizing the opportun-
in the fertility rate have led to a significant aging popula- ities for health, participation and security to improve the
tion. In South America, the aging population/proportion quality of life (QOL) as individuals age [3]. The challenge
of older people is increasing at a more rapid rate than in for aging studies is to understand the conditions associ-
most developed countries [1,2]. ated with aging as a positive process and old age as a stage
Aging is a complex phenomenon that requires increas- of life in which health, well-being, pleasure and QOL can
ing numbers of multidisciplinary studies. The term “active be increased [4-6].
The QOL of older adults could be good, or at least
preserved, provided they have autonomy, independence
* Correspondence: [Link]@[Link]
1
School of Dentistry, Universidade Federal de Minas Gerais, Presidente
and good physical health and provided they fulfill social
Antônio Carlos 6627, Belo Horizonte 31270-901, Minas Gerais, Brazil roles, remain active and enjoy a sense of personal
Full list of author information is available at the end of the article

© 2014 Campos et al.; licensee BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative
Commons Attribution License ([Link] which permits unrestricted use, distribution, and
reproduction in any medium, provided the original work is properly credited. The Creative Commons Public Domain
Dedication waiver ([Link] applies to the data made available in this article,
unless otherwise stated.
Campos et al. Health and Quality of Life Outcomes 2014, 12:166 Page 2 of 11
[Link]

meaning [7]. Epidemiological population-based studies The sampling process was conducted in two stages.
are important for identifying the determinants and The census tracts were first selected, and the households
etiological factors associated with aging. To investigate within each sector were then selected [10]. In each
the determinants of aging, questions must be answered household, all residents aged 60 years or older were
using longitudinal surveys [8]. Longitudinal studies specif- interviewed, regardless of marital status or kinship.
ically designed to assess health, QOL and associated risk
factors are not abundant in the literature, particularly Data collection
those performed in underdeveloped countries in which A pilot study including 107 older adults (approximately
poverty and a low educational level might lead to a differ- 10% of the sample) was conducted prior to data collection.
ent set of variables that affect the aging process [9]. All of the instruments were validated for Portuguese in
In Brazil, a country that is rapidly aging and that suf- Brazil, and the test/retest method was used to assess
fers from large inequalities, the study of the QOL among reliability and concordance. Coefficients greater than
aged people is important for the future health. This 0.80 were obtained (p < 0.001) and included a weighted
study sought to examine the association between QOL, Kappa (95%) value of 0.81 (0.71 to 0.91) and an adjusted
gender and physical and psychosocial health among Kappa value of 0.86.
older Brazilian community-dwelling adults, with the aim The data collection was conducted in the homes of
to identify potential factors associated with better QOL. the older adults between January and July 2012 and in-
volved household interviews and examinations con-
Methods ducted by three examiners and three annotators.
The Aging, Gender and Quality of Life (AGEQOL) study All persons 60+ years in the selected households were
is an observational, cohort study of a community-dwelling informed of the study and were asked to sign an in-
population aged 60 years and older. The sample is repre- formed consent form that had been previously approved
sentative of the city of Sete Lagoas in the state of Minas by the Ethical Committee of the Federal University of
Gerais, Brazil, which has a population of approximately Minas Gerais. The interviews lasted 40 to 60 minutes.
21,000 older adults (10.2% of the population) [10]. This At the end of the interviews, each subject in the city
city is divided into 17 administrative regions, one district received guidance regarding health care and activity
and four rural areas [11]. options as well as the personal contact information of
the researcher responsible for the questionnaire.
Sample
A complex sampling design was adopted for this study Measures
and consisted of a combination of probabilistic sampling The socioeconomic and demographic data included age,
methods for selecting a representative sample of the gender, marital status, income categorized by the median
population [12]. For this sampling, the following two value, years of education, residence and occupation.
calculations were performed: an estimation of the number Most independent variables were dichotomized to enhance
of older adults and an estimation of the number of house- the interpretability of the logistic regression coefficients.
holds to be visited. Physical activity and social participation were measured
The sample size calculation was performed to compare using a single question with a dichotomous answer (yes or
genders by considering the prevalence of functional im- no). The health-related component included self-reported
pairment in instrumental activities for males (86.6%) and health conditions, which were assessed using a Likert
females (72.9%) [13]. The estimated error was up to 5%, scale, and access to and utilization of health services. For
with a power of 80% at 95% confidence intervals (95% CI) this study, the categories were grouped into poor (very
when considering a design effect of two. An estimated poor and poor), regular and good (good and very good).
additional 20% of the sample size was added to compen- With regard to the chronic diseases previously reported
sate for refusals. The samples from each group (men and to be most relevant to the loss of functionality in aging
women) were stratified by age in relation to the population subjects (hypertension, diabetes, cardiovascular disease,
and were corrected based on the probability of dying. musculoskeletal disorders and respiratory diseases), the
Of the total potential participants living in the selected number of diseases was recorded as 0, 1 or ≥2.
dwellings, 25 (1.2%) were excluded because they could Functional limitations were evaluated by combining
not answer the questionnaire or because of cognitive the participants’ responses to questions about six basic
impairment/dementia or difficulty speaking. One hundred activities of daily living (eating, dressing and undressing,
and twenty-five subjects (5.8%) refused to participate in grooming, walking, getting in and out of bed, bathing
the study, and 100 (4.8%) could not be located or had died. and continence) [14] and seven instrumental activities
The final sample consisted of 2,052 individuals, of whom (using the telephone, travel, shopping, meal preparation,
59.7% were female. housework, taking medicine and management of finances)
Campos et al. Health and Quality of Life Outcomes 2014, 12:166 Page 3 of 11
[Link]

[15]. To evaluate the cognitive status of the older people, The F test was used to analyze the differences and
we used the Mini Mental State Examination, which has characterize the groups with a significance level of 5%.
been validated in Brazil [16] and has a cut-off of 21/22 This type of analysis is an analytical statistical tool that
points [17]. A score ≤21 indicated cognitive impairment. is used to define the development of mutually exclusive,
The presence or absence of a functional limitation was significant subgroups based on the similarities among
determined depending on the type of daily living activity individuals, without prior knowledge of the allocation
and cognitive status, as adapted from Albala [18]. The within the groups. In cases in which the grouping of the
subjects were classified as restricted if they had one or data is successful, the groups are internally homoge-
more limitation in basic or instrumental activities or if neous but have high external heterogeneity [26].
they had cognitive impairment. Canonical discriminant analysis was used posteriorly to
The presence of depressive symptoms was assessed validate the cluster analysis described by two functions.
using the short version of the Geriatric Depression Scale The objective of discrimination is to maximize the variance
(GDS-15) [19], with a cutoff of 5/6; a score ≥6 indicated between and within groups and to verify the efficiency of
suspected depression. Family functioning was assessed the overall correct classification of the model [26].
using the five-item Family Adaptability, Partnership, The QOL level among the clusters was adapted from
Growth, Affection, and Resolve (APGAR) scale, which Oliveira et al. [27]; for all the WHOQOL domains, there
measures the satisfaction of older adults in relation to was a group with good QOL scores, a group with inter-
various aspects of family life [20]. The responses consist mediate QOL scores and a group with worse QOL scores.
of values between 1 (hardly) and 3 (but not always), and Ordinal logistic regression was used to test the associ-
the total score ranges from 5 to 15. A score ≥10 indi- ation between QOL and physical and psychosocial health
cates family satisfaction [21]. after controlling for age and socioeconomic status. All
analyses were performed separately for each gender. In
QOL this study, we applied the Polytomous Universal Model
We used the World Health Organization Quality of Life (PLUM), which incorporates the ordinal nature of the
Assessment-Brief Instrument (WHOQOL-BREF) [22] and dependent variable in the analysis; thus, a logistic regres-
the World Health Organization Quality of Life Instrument- sion model with proportional-odds and Logit function
Older Adults Module (WHOQOL-Old) to evaluate QOL [28] was performed. The odds between the categories of
[23]. The first instrument is composed of 24 facets that the dependent variable were compared by calculating the
are grouped into four domains that focus on physical, crude and adjusted odds ratio (OR), and tests evaluating
psychological, social and environmental aspects. There the homogeneity of slopes and multicollinearity were con-
is no total score for this instrument, and each item ducted using Pearson's adjustment to analyze the validity
contains five Likert response options that are recorded of the model. To ascertain the possible interference of a
as scores of 1–5. The WHOQOL-Old module consists small number of observations, we used residual analysis
of 24 items that are divided into the following six for ordinal data, as proposed by McCullagh [29]. All of
domains: sensory abilities (SAB); autonomy (AUT); these tests showed that the model satisfied all of the
past, present and future activities (PPF); social partici- assumptions, and the effect of the complex sample design
pation (SOP); death and dying (DAD); and intimacy was considered in all of the analyses.
(INT). The scores of all domains are combined to
produce an overall score for QOL in older adults, with Results
higher scores indicating good QOL. The instruments The age of the total sample at baseline ranged from 60
were previously validated by Fleck et al. [24,25] and to 106 years old, and the mean age of all participants at
showed good reliability and validity in the assessment of baseline was 70.89 ± 8.14 years (71.03 ± 8.35 for women
QOL of Brazilian older adults (the Cronbach’s alpha and 70.69 ± 7.83 for men). Table 1 shows the descriptive
score ranged from 0.7 to 0.8 for the WHOQOL-Bref statistics of the socioeconomic and health conditions of
and from 0.7 to 0.9 for the WHOQOL-Old). the participants according to gender. Thirty percent
(625) of the participants were more than 74 years old,
Statistical analysis and 317 (15.4%) older adults were octogenarians. Most
SPSS software (SPSS Institute, Chicago, IL, USA) version men (70.8%) and women (68.7%) were between 60 and
19.0 was used for the analysis and included χ2 tests and 74 years old, and there was no difference in age distribu-
ordinal logistic regression. tion between genders.
K-means clustering analysis was used to obtain three Forty-eight types of living arrangements were identi-
groups by considering the better distance between the fied among older adults in the city under study. When
mean scores of the four dimensions in WHOQOL-BREF taking the three groups of living arrangements that were
and the mean of the total WHOQOL-Old score (Figure 1). established in this study into account, it was observed
Campos et al. Health and Quality of Life Outcomes 2014, 12:166 Page 4 of 11
[Link]

Figure 1 Canonical discriminant functions based on the QOL level of older people.

that the majority of older adults who lived alone were (p = 0.001) in functional limitations between men (32.6%)
women (71.5%), whereas 75.5% of men lived with their and women (39.6%) (Table 1).
partners (p < 0.001). There were no differences in the Cluster analysis (k-means) resulted in the formation of
years of education between the different genders; how- the following three groups of older adults in relation to
ever, 10.4% of men and 8.6% of women had completed QOL (Table 2): subjects with poor, fair and good QOL.
over 4 years of study (Table 1). The majority of the older adults were included in the fair
Additionally, there were significant differences re- QOL group (51.4%), which corresponded to the average
lated to marital status, income, retirement, and living level of scores in the WHOQOL. The group with worse
arrangement between genders. The majority of men in the QOL included 371 people (18.1%), whereas the good
sample were married (74.5%), while 61.7% of women QOL group included 627 subjects (30.6%). The results
were single, separated or widowed. Most older adults of the test for equality of the group means between the
had low monthly income (66.1%), and this percentage groups were significant, indicating that the groups
was higher for females (71.5%) compared to males differed in all QOL domains. The overall correct classifi-
(58.1%) (Table 1). cation of the canonical discriminant functions was
The self-perceived health status was different between 97.9%, with a correlation coefficient of 0.89.
men and women (p <0.001). While 50.8% of men rated Differences were observed in all of the QOL variables,
their health as good, most women rated their health as except for those of retired individuals. In this case, the
fair (37.8%) or good (41.8%). Only 15.9% of the older socioeconomic distribution between the genders was
adults did not have chronic diseases; however, the reversed, with 47.7% of the older adults in the higher
percentage of women (59.6%) with more than two QOL group being male and 67.7% of those in the lower
diseases was statistically higher (p < 0.001) than that of QOL group being female. Additionally, there was a
men (44.6%). The prevalence of cognitive impairment gradient association between low QOL and worse health
was 35.3%, with a slightly larger proportion of women perception, cognitive impairment, depressive symptoms,
(36.0%) than men (34.3%) reporting this condition. In family dysfunction and functional limitation. Most of the
relation to depression, there was a 30.2% prevalence older adults who reported two or more chronic diseases
of depressive symptoms and a statistically significant (70.4%) were allocated to the low QOL group (Table 3).
(p < 0.001) difference between genders (23.8% for men The results of the ordinal regression model, which
and 34.4% for women). There was a high prevalence of estimates the OR of good QOL by gender, are shown in
functional limitations (36.7%) and a significant difference Table 4. Age, marital status, income, and cognitive
Campos et al. Health and Quality of Life Outcomes 2014, 12:166 Page 5 of 11
[Link]

Table 1 Characteristics of the sample by gender (N =2052)

Variables Total (n = 2052) Men (n = 826) Women (n = 1226) P*
n (%) n (%) n (%)
Age
60-74 years old 1427 (69.5) 585 (70.8) 842 (68.7) 0.162
≥75 years old 625 (30.5) 241 (29.2) 384 (31.3)
Years of education
0 579 (28.2) 240 (29.1) 339 (27.7) 0.173
1–4 1282 (62.5) 500 (60.5) 782 (63.8)
5–7 130 (6.3) 63 (7.6) 67 (5.5)
>8 61 (3.0) 23 (2.8) 38 (3.1)
Income**
≤R$622.00 1357 (66.1) 480 (58.1) 877 (71.5) <0.001
>R$622.00 695 (33.9) 346 (41.9) 349 (28.5)
Marital status
Married 1084 (52.9) 615 (74.5) 469 (38.3) <0.001
Single/Divorced/Widower 965 (47.1) 211 (25.5) 754 (61.7)
Living arrangements
Living with couple 1065 (53.0) 616 (75.5) 449 (37.6) <0.001
Mixed arrangements 668 (33.2) 121 (14.8) 547 (45.8)
Living alone 277 (13.8) 79 (9.7) 198 (16.6)
Functional limitation 754 (36.7) 269 (32.6) 485 (39.6) 0.001
Self-rated health
Poor 277 (13.5) 105 (12.7) 172 (14.0) <0.001
Fair 705 (34.4) 242 (29.3) 463 (37.8)
Good 1070 (52.1) 479 (58.0) 591 (48.2)
Chronic diseases
0 327 (15.9) 184 (22.3) 143 (11.7) <0.001
1 626 (30.5) 274 (33.2) 352 (28.7)
≥2 1099 (53.6) 368 (44.6) 731 (59.6)
Physical activity 545 (26.6) 216 (26.2) 329 (26.8) 0.385
Symptoms of depression 619 (30.2) 197 (23.8) 422 (34.4) <0.001
Cognitive impairment 264 (12.9) 98 (11.9) 166 (13.5) 0.148
Social participation 157 (7.7) 41 (5.0) 116 (9.5) <0.001
Family satisfaction 1565 (76.3) 630 (76.3) 936 (76.3) 0.520
*χ2 test. ** Brazilian minimum salary = R$622.00 ≈ US$300.

Table 2 Means and standard deviations of QOL clusters of older adults

Domains WHOQOL QOL* Z Wilks’ P
lambda test
Poor (n = 371) Fair (n = 1054) Good (n = 627)
WHOQOL-BREF Mean (±sd) Mean (±sd) Mean (±sd)
Physical 41.0 (±14.0) 61.4 (±11.0) 78.3 (±10.1) 1282.4 0.4 <0.001
Psychological 47.7 (±12.4) 68.2 (±9.0) 80.3 (±8.8) 1335.3 0.4 <0.001
Social Relations 52.0 (±16.1) 69.1 (±10.5) 78.8 (±12.5) 555.1 0.6 <0.001
Environmental 47.3 (±10.3) 58.9 (±8.3) 72.9 (±10.5) 913.5 0.5 <0.001
WHOQOL-Old 53.1 (±9.4) 65.4 (±7.7) 77.0 (±8.4) 1007.2 0.5 <0.001
*Final central cluster for QOL measure. Sd = standard deviation.
Campos et al. Health and Quality of Life Outcomes 2014, 12:166 Page 6 of 11
[Link]

Table 3 Characteristics of the sample by QOL

Variables QOL P*
Poor (n = 371) Fair (n = 1054) Good (n = 627)
Age n (%) n (%) n (%)
60-74 years old 219 (59.0) 744 (70.6) 464 (74.0) <0.001
≥75 years old 152 (41.0) 310 (29.4) 163 (26.0)
Gender
Male 120 (32.3) 407 (38.6) 299 (47.7) <0.001
Female 251 (67.7) 64.7 (61.4) 328 (52.3)
Years of study
0 159 (42.9) 309 (29.3) 111 (17.7) <0.001
1–4 195 (52.6) 668 (63.4) 419 (66.8)
>5 17 (4.5) 77(7.3) 97 (15.4)
Income**
≤R$622.00 265 (71.4) 730 (69.3) 362 (57.7) <0.001
>R$622.00 106 (28.6) 324 (30.7) 265 (42.3)
Retired 266 (71.7) 778 (73.8) 474 (75.6) 0.392
Marital status
Married 148 (39.9) 554 (52.7) 382 (61.0) <0.001
Single/Divorced/Widower 223 (23.1) 498 (51.6) 244 (25.3)
Living arrangements
Living with spouse 145 (39.9) 544 (52.8) 376 (60.9) <0.001
Mixed arrangements 162 (44.6) 351 (34.1) 155 (25.1)
Living alone 56 (15.4) 135 (13.1) 86 (13.9)
Functional limitation 210 (27.9) 377 (50.0) 167 (22.1) <0.001
Self-rated health
Poor 129 (34.8) 113 (10.7) 35 (5.6) <0.001
Fair 179 (48.2) 407 (38.6) 119 (19.0)
Good 63 (17.0) 534 (50.7) 473 (75.4)
Chronic diseases
0 29 (7.8) 150 (14.2) 148 (23.6) <0.001
1 81 (21.8) 306 (29.0) 239 (38.1)
≥2 261 (70.4) 598 (56.7) 240 (38.3)
Physical activity 59 (15.9) 250 (23.7) 236 (37.6) <0.001
Cognitive impairment 88 (23.7) 115 (10.9) 61 (9.7) <0.001
Depressive symptoms 247 (66.6) 300 (28.5) 72 (11.5) <0.001
Social participation 20 (5.4) 77 (7.3) 60 (9.6) 0.047
Family satisfaction 200 (53.9) 831 (78.8) 534 (85.2) <0.001
*χ test. **Brazilian minimum salary = R$622.00 ≈ US$300.
2

impairment did not remain associated with QOL in the Retired men had better QOL when compared to non-
final model. retired men (OR = 2.2; 95% CI = 1.4–3.2), but this associ-
There was an education gradient for the QOL of men. ation was not observed in females. Men living in mixed
Men with 1–4 and >5 years of education were 2.2 and arrangements (OR = 0.5; p = 0.033) and women who did
4.2 times more likely to have a better QOL than not practice physical activity (OR = 0.7; p = 0.022) tended
illiterate men. Similarly, women with five or more years to have a poorer QOL (Table 4).
of education were associated with good QOL (OR = 2.2; As shown in Table 4, there was an increase in the OR
p < 0.001) (Table 4). for the association between QOL and self-rated health
Campos et al. Health and Quality of Life Outcomes 2014, 12:166 Page 7 of 11
[Link]

Table 4 Results of ordinal logistic regression that explain Table 4 Results of ordinal logistic regression that explain
better QOL in older adults, separated by gender better QOL in older adults, separated by gender
Variables Final model (Continued)
Male OR 95% confidence interval P Depressive symptoms
Years of education No 2.2 1.7–2.9 <0.001
>5 4.2 2.4–7.4 <0.001 Yes 1.0
1–4 2.2 1.3–3.7 0.003 Family dysfunction
0 1.0 No 3.0 2.3–4.0 <0.001
Living arrangements Yes 1.0
Living with spouse 0.9 0.4–2.0 0.809
Mixed arrangements 0.5 0.3–0.9 0.033 for both genders once the model was adjusted for
Living alone 1.0
demographic variables and psychosocial health. Men
with fair health (OR = 3.0; 95% CI = 2.2–4.3) and, in
Retired
particular, good health (OR = 5.0; 95% CI = 3.5–9.4)
No 0.4 0.3–0.7 <0.001 were associated with good QOL. Women with good and
Yes 1.0 fair health were 4.2 (OR = 4.2; 95% CI = 2.8–6.2) and 3.0
Self-rated health (OR = 3.0; 95% CI = 2.3–4.0) times more likely to have a
Good 5.7 3.5–9.4 <0.001 good QOL, respectively.
Fair 3.0 2.2–4.3 <0.001
For both genders, there was a robust association be-
tween QOL and all psychosocial variables, except cog-
Bad 1.0
nitive impairment. Men without depressive symptoms
Chronic diseases and women without family dysfunction were 3.6 (OR =
>2 0.6 0.4–0.9 <0.001 3.6; 95% CI = 2.5–5.2) and 3.0 (OR = 3.0; 95% CI = 2.3–
1 0.6 0.4–0.8 <0.001 4.0) times more likely to have good QOL, respectively
0 1.0 (Table 4).
Depressive symptoms
Discussion
No 3.6 2.5–5.2 <0.001
The physical and psychosocial health and socio-
Yes 1.0 demographic variables examined in this study were
Family dysfunction evaluated using ordinal logistic regression, which
No 1.8 1.3–2.6 0.001 resulted in the following five variables being associated
Yes 1.0 with good QOL for both genders: self-rated health,
Female OR 95% confidence interval P
depressive symptoms, years of education, chronic diseases,
and family dysfunction. Additionally, good QOL for men
Years of education
was associated with retirement, mixed living arrange-
>5 2.2 1.3–3.6 <0.001 ments, and physical activity, whereas good QOL for
1–4 1.3 0.9–2.1 0.188 women was associated with physical activity; these results
0 1.0 are similar to those of other studies [27,30,31]. These
Self-rated health factors represent targets for policy action because they
Good 4.2 2.8–6.2 <0.001
have the potential to affect the health of older individuals
in the general population.
Fair 3.0 2.3–4.0 <0.001
A number of studies have been performed on QOL in
Bad 1.0 older adults. This study is original and innovative because
Chronic diseases it used a representative sample to provide information
>2 0.5 0.3–0.7 <0.001 regarding an ordinal positive relationship between QOL
1 0.7 0.5–0.9 0.010 and self-rated health. Furthermore, our results indicate
0 1.0
that the most important factors for a good QOL for both
genders is a good health perception and a lack of depres-
Physical activity
sion, even when the model was adjusted for socioeco-
No 0.7 0.5–1.0 0.022 nomic conditions.
Yes 1.0 We observed a significant difference of 4.4% when
comparing good self-rated health between the low and
high QOL groups. In the ordinal regression, the men
Campos et al. Health and Quality of Life Outcomes 2014, 12:166 Page 8 of 11
[Link]

and women who reported having good health were 5.7 Considering the importance of the physical and psy-
and 4.2 times more likely to have good QOL, respectively. chosocial aspects of active aging and of QOL in older
Previous studies on QOL in older adults have also shown adults, other results of this study should be briefly
a direct relationship with self-rated health [27,32,33]. In discussed. In this study, an increase in the number of
particular, older adults who evaluated themselves as having chronic diseases was associated with a decrease in
good health tended to have good QOL [27]. QOL, and statistically significant gender differences were
The perception of health in older adults was generally observed between chronic diseases and QOL. Most women
positive because most of the older adults in this sample who reported two or more diseases were classified as
rated their health as good (52.1%), including 58.0% of having poor (73.3%) or fair (62.3%) QOL (data not
the men and 48.2% of the women (48.2%). However, the shown). In general, the prevalence of chronic disease
percentage of poorer self-rated health was higher in among older people in Brazil is high and differs between
women compared to men. This study provides further genders [11], resulting in negative repercussions on
evidence that QOL can be explained by self-rated health QOL [43]. Preventive actions and the promotion of
and its associated factors among older men and women. policies for controlling the effect on health conditions
In the SABE study (Salud, Bienestar y Envejecimiento) in could result in good QOL in this population [30].
São Paulo, Brazil, 8.9% of women and 7.2% of men demon- Physical activity is a protective factor for QOL and has
strated poor health. In other SABE study countries, the been previously discussed in the literature [44,45]. For
participants reporting good/very good health ranged from women, we observed a significant association between
27.9% of women (Mexico) to 69.0% of men (Uruguay) [2]. QOL and physical activity (OR = 0.7; 95% CI = 0.5–1.0),
A previous study on the components of self-rated health i.e., the choice of good QOL was 1.4 times higher for
among adults suggested that physical health (chronic women who practiced physical activity. However, this
diseases and functional limitations) most likely comprises association was not accurate once the confidence intervals
the majority of an individual’s perception of health status included a value of 1.0. Physical activity was measured
[34], and this result was observed in this study. using a single yes/no question, which is an important limi-
Health perception involves an individual’s evaluation tation of this study because these results assume that any
of his/her body in relation to his/her feelings, including level of physical activity will be associated with health.
feelings regarding health and well-being, and this per- We did not observe an association between marital
ception can be altered by environmental stressors and status and QOL, although we observed an inverse asso-
the social context [35]. For older adults, the concept of ciation between QOL and family dysfunction. Men and
self-rated health remains stable despite significant health women who were satisfied with their family relationship
problems, although over time, there might be a reduc- had 1.8 and 3.0 times higher odds of good QOL, respect-
tion in the standard of good self-rated health [36]. ively. Frequent contacts and visits with friends or family
Self-rated health has been shown to be a reliable have been shown to motivate activity and increase self-
method for measuring health status [37] and to be a rated QOL [46].
consistent predictor of mortality in older adults [38]. It Additionally, we found a high percentage of individuals
is essential to use the association between perceived with poor QOL living in mixed arrangements, i.e., shar-
health and QOL in patients, especially in regards to the ing the household with their sons and frequently with
dual direction of this association. sons and grandchildren (44.6%). This situation, which is
We observed a very strong association between QOL common in other Brazilian regions, is in contrast to the
and depressive symptoms, which corroborates the find- living arrangements in developed countries [11].
ings of other studies [32,33,39,40]. Thus, the choice of As shown in Table 4, men living in mixed arrange-
good QOL was 3.6 and 2.2 times higher for men and ments had worse QOL than those living alone. In our
women without depressive symptoms than for those dataset, most men who lived in mixed arrangements had
experiencing depressive symptoms, respectively. This functional limitations and reported more than two
finding could be explained by the high prevalence of chronic diseases (63.3%). It is possible that men in our
depression (30.2%) in this sample; this figure reached sample could have been living in mixed arrangements
34.4% in women and 70.4% in women with poor QOL. because they had poorer health and therefore needed
These disorders are more prevalent in females, but this daily assistance; however, these results should be inter-
gender vulnerability varies with age [41]. preted with caution, as there was a low percentage of
In a study conducted in Łódź, Poland, 30.9% of older men living alone (9.7%). Size, sample stratification and
adults (56.5% females) were found to suffer from depres- corrections minimized these effects, thus permitting
sion. According to the authors, the chances of good self- comparisons in this study.
rated QOL were 9.9 (95% CI = 5.0–19.6) times higher in A mixed living arrangement could have a negative
older adults without depression [42]. effect on the older population [20,47]. However, living
Campos et al. Health and Quality of Life Outcomes 2014, 12:166 Page 9 of 11
[Link]

alone presents a greater risk of loneliness and isolation WHOQOL-BREF and WHOQOL-Old, which were devel-
because loneliness increases as the social contacts of oped by the WHO, are widely reported in the scientific
older individuals decrease [46]. literature and have been validated in Brazil [24,25]. The
Similar to the results of other studies [33,48,49], we results of this study corroborate those reported by the
found an association between QOL and education. Sete Brazilian WHOQOL group. For older Brazilian adults, a
Lagoas is a Brazilian city with high life expectancy positive QOL includes several aspects such as activity,
(73.9 years) and good social indicators [50]. In addition, income, social life and family relationships, whereas a
illiteracy is high in this sample (28.2%) compared to the negative QOL is related to poor health, which differs
current national data (24%) [51]. These results are often between individuals [53].
found in most Latin American countries [2] and in some WHOQOL-Old is a supplementary module for older
regions in Brazil, where very different educational op- adults and can be added to the existing WHOQOL instru-
portunities are available for the rich and poor. ments [22]. Bowling [7] compared generic QOL scales
A low level of education is an important aspect to be used for older adults and showed that the WHOQOL-Old
considered when developing public policies for older was the most comprehensive instrument; the questions in
adults and a proposed collective action. In our study, the this instrument are based on measuring suffering, but the
illiteracy rate was similar between genders (29.1% for men questionnaire is relatively long, and the Likert scale format
and 27.7% for women). A previous study investigated might be boring to the subjects (although there is no
trends in educational inequalities in terms of old-age evidence that this characteristic has adversely affected
mortality in Norway from 1961 to 2009, and the authors responses to date).
observed that relative educational inequalities in old-age Additionally, Bowling emphasized the need for a gen-
mortality were increased for both genders [52]. eric, truly multidimensional QOL measure with minimal
The association of years of education with QOL was respondent burden for evaluating the outcomes of health
different between the genders. We observed an ordinal and social care in older populations [7]. The reason for
crescent impact of years of education on QOL for men, the existing difficulties in the assessment of QOL that
indicating that education can be a protective factor for limits its inclusion in clinical practice and public health
good QOL among men. The QOL among women with services is relevant [54].
1–4 years of education was no different than that of To minimize these limitations, one specific method of
illiterate women. analysis was conducted in this study. Based on a Brazilian
Our results correspond to the baseline data reported study [27], we used cluster analyses and canonical
for the AGEQOL study. However, the lack of under- discriminant analyses to compile both WHOQOL in-
standing of the ways in which specific levels of education struments into a unique measure for QOL. This ana-
interfere in the association between SES and QOL is the lysis was performed to provide an ordinal variable with
first limitation of this study. A longitudinal follow-up three internally more homogeneous groups that were
study of older adults would permit better comparisons distinct from each other. Additionally, we minimized
of this study with others, although such comparisons the variations between the mean scores of the five
might be hampered by differences in the QOL models dimensions of QOL that were considered. We found a
and measures that are employed across studies. It is not high percentage of correct classification (97.9%) and a
yet possible to determine whether there is a temporal high correlation coefficient (0.89), which indicated the
relationship between the studied variables. likelihood that we had constructed a good measure of
The response rate in this study could be considered QOL for older adults in this sample.
high (98.8%); therefore, this study is one of the few studies In future studies, we suggest replicating this statis-
that have been performed using a probabilistic sample of tical model, considering gender and age stratification
older adult community residents with an adequate variations and including other independent variables
number of participants to perform an ordinal logistic concerning nutrition and lifestyle. Adaptation and
regression. Our results are valid and representative of resilience might also play a role in maintaining good
the population living in the community that lacks QOL [55].
significant cognitive and/or physical deficits. Despite these limitations, this study confirmed that the
In addition to the limitations of this being a cross- QOL of older adults differed between the three clusters
sectional study, it should be emphasized that the evalu- that were formed, with a good QOL being strongly asso-
ation of QOL presupposes the quantification of a construct ciated with good self-rated health, the absence of depres-
that is sensibly marked by the subjectivity of individual sive symptoms, and family satisfaction.
experiences, beliefs, expectations and perceptions [24]. Overall, the results demonstrate that active aging
In this sense, it is necessary to discuss the instruments in Sete Lagoas, Brazil, does not occur evenly across
used to measure QOL in older adults. We used the genders. Better healthcare requires the inclusion of such
Campos et al. Health and Quality of Life Outcomes 2014, 12:166 Page 10 of 11
[Link]

differences as part of the comprehensive evaluation of Received: 24 March 2014 Accepted: 4 November 2014
older adults [56].
The discussions of aging in the different genders
in relation to living conditions and perceived health References
1. Paskulin LMG, Aires M, Gonçales AV, Kottwitz CCB, Morais EP, Brondani MA:
that are presented in this study need to be further Aging, diversity, and health: the Brazilian and the Canadian context.
explored, as there are particularities of each group Acta Paul Enferm 2011, 24:851–856.
that may have been missed during routine analysis. 2. Albala C, Lebrão ML, León Díaz EM, Ham-Chande R, Hennis AJ, Palloni A,
Peláez M, Pratts O: Encuesta Salud, Bienestar y Envejecimiento (SABE):
We believe that this study may contribute to the metodología de la encuesta y perfil de la población estudiada. Rev
formulation of new public health and social care pol- Panam Salud Publica 2005, 17:307–322.
icies for older adults in the medium and long term. 3. World Health Organization: Active Ageing: A Policy Framework. Geneva: 2002
[[Link]
Older adults will benefit from interdisciplinary moni- 4. Lowsky DJ, Olshansky SJ, Bhattacharya J, Goldman DP: Heterogeneity in
toring that focuses on promoting health, improving healthy aging. J Gerontol A Biol Sci Med Sci 2014, 69:640–649.
QOL and active aging. 5. Hekmatpou D, Shamsi M, Zamani M: The effect of a healthy lifestyle
program on the elderly's health in Arak. Indian J Med Sci 2013, 67:70–77.
6. De Luca d’Alessandro E, Bonacci S, Giraldi G: Aging populations: the health
Conclusions and quality of life of the elderly. Clin Ter 2011, 162:e13–e18.
7. Bowling A: Quality of life in older age: what older people say. In Quality
We conclude that there are gender differences related to of Life in Old Age. International and Multi-Disciplinary Perspectives. Edited by
better QOL in this sample cohort. Women with good Mollenkopf H, Walker A. Dordrecht: Springer; 2007.
physical and psychosocial health are more likely to have 8. Ramos LR: Fatores determinantes do envelhecimento saudável em
idosos residentes em centro urbano: Projeto Epidoso, São Paulo.
a better QOL. For men, the best QOL was associated Cad Saude Publica 2003, 19:793–798.
with high socioeconomic conditions and good physical 9. Ramos LR, Toniolo JN, Cendoroglo MS, Garcia JT, Najas MS, Perracini M,
and psychosocial health. We hope that our study con- Paola CR, Santos FC, Bilton T, Ebel SJ, Macedo MBM, Almada CM, Nasri F,
Miranda RD, Gonçalaves M, Santos ALP, Fraietta R, Vivacqua IN, Alves MLM,
tributes to future discussions on the most important Tudisco ES: Two-year follow-up study of elderly residents in S. Paulo, Brazil:
predictors for assessing QOL in older adults and on methodology and preliminary results. Rev Saude Publica 1998, 32:397–407.
long-term changes in the perception of QOL in this 10. Instituto Brasileiro de Geografia e Estatística: Censo 2010.
[[Link]
population. 11. Lei n° 7804 de 14 de Outubro de 2009 de Sete Lagoas, Define as
Regiões Administrativas - RA no Perímetro Urbano do Distrito Sede do
Abbreviations Município de Sete Lagoas. [[Link]
ADL: Activity of daily living; APGAR: Adaptability, partnership, growth, lei-7804-09-sete-lagoas]
affection, and resolve; AUT: Autonomy; DAD: Death and dying; GDS: Geriatric 12. Kish L: Survey Sampling. New York: John Wiley; 1965.
depression scale; IADL: Instrumental activities of daily living; 95% CI: 95% 13. Santos JLF, Lebrão ML, Duarte YAO, Lima FD: Desempenho funcional de
confidence interval; INT: Intimacy; MMSE: Mini-mental state examination; idosos nas atividades instrumentais da vida diária: uma análise no
OR: Odds ratio; PPF: Past, present and future activities; PLUM: Polytomous Município de São Paulo, Brasil. Cad Saúde Pública 2008, 24:879–886.
universal model; QOL: Quality of life; SAB: Sensory abilities; SOP: Social 14. Lino VTS, Pereira SRM, Camacho LAB, Ribeiro Filho ST, Buksman S:
participation; WHO: World Health Organization; WHOQOL-BREF: World Health Adaptação transcultural da Escala de Independência em Atividades da
Organization Quality of Life Assessment-Brief Instrument; WHOQOL-Old: World Vida Diária (Escala de Katz). Cad Saúde Pública 2008, 24:103–112.
Health Organization Quality of Life Instrument-Old. 15. Santos RL, Virtuoso Júnior JS: Confiabilidade da versão brasileira da Escala
de atividades Instrumentais da vida Diária. RBPS 2008, 21:290–296.
Competing interests 16. Brucki SMD, Nitrini R, Caramelli P, Bertolucci PHF, Okamoto IH: Sugestões
The authors declare that they have no competing interests. para o uso do mini-exame do estado mental no Brasil. Arq Neuropsiquiatr
2003, 61:77–81.
17. Quiroga P, Albala C, Klaasen G: Validación de un test de tamizaje para el
Authors’ contributions diagnóstico de demencia asociada a edad, en Chile. Rev Med Chil 2004,
ACVC conducted all data analyses and drafted the manuscript. AMDV and 132:467–478.
EFF contributed to the conception and design of the study. CA participated 18. Albala C, Lera L, García C, Arroyo P, Marin PP, Bunout D: Searching a
in the interpretation and discussion of the data and made critical revisions to common definition for functional limitation in Latin America.
the manuscript. All authors read and approved the final manuscript. Gerontologist 2004, 44:550.
19. Almeida OP, Almeida AS: Confiabilidade da versão brasileira da Escala de
Acknowledgments Depressão em Geriatria (GDS) versão reduzida. Arq Neuropsiquiatr 1999,
The study was conducted with financial help from the National Council for 57:421–426.
Scientific and Technological Development – CNPq (481672/2011-7, 305032/ 20. Duarte YAO, Lebrão ML, Lima FD: Contribuição dos arranjos domiciliares
2012-7 and 141307/2011-0); the Coordination of Improvement of Higher para o suprimento de demandas assistenciais dos idosos com
Education – CAPES (process 8314/13-6); Pró-Reitoria de Pesquisa da Universidade comprometimento funcional em São Paulo, Brasil. Pan Am J Public Health
Federal de Minas Gerais – PRPq/UFMG; and Fundação de Amparo à Pesquisa de 2005, 17:370–378.
Minas Gerais – FAPEMIG. 21. Mengel M: The use of the family APGAR in screening for family
dysfunction in a family practice center. J Fam Practice 1987, 24:394–398.
Author details 22. Harper A, Power M: Development of the World Health Organization
1
School of Dentistry, Universidade Federal de Minas Gerais, Presidente WHOQOL-BREF quality of life assessment. The WHOQOL Group. Psychol
Antônio Carlos 6627, Belo Horizonte 31270-901, Minas Gerais, Brazil. Med 1998, 28:551–558.
2
Department of Community and Preventive Dentistry, School of Dentistry, 23. Power M, Quinn K, Schmidt S: Development of the WHOQOL-old module.
Universidade Federal de Minas Gerais, Presidente Antônio Carlos 6627, Belo Qual Life Res 2005, 14:2197–2214.
Horizonte 31270-901, Minas Gerais, Brazil. 3Unidad Nutrición, Salud Pública y 24. Fleck MPA, Louzada S, Xavier M, Chachamovich E, Vieira G, Santos L, Pinzon V:
Envejecimiento Saludable, INTA, Universidad de Chile, El Líbano 5524 Macul, Application of the Portuguese version of the abbreviated instrument of
Santiago 138-11, Chile. quality life WHOQoL-bref. Rev Saude Publica 2000, 34:178–183.
Campos et al. Health and Quality of Life Outcomes 2014, 12:166 Page 11 of 11
[Link]

25. Fleck MPA, Chachamovich E, Trentini CM: Development of the WHOQoL-old dos idosos de um município do Sudeste do Brasil. Ciência & Saúde
module in Brazil. Rev Saude Publica 2006, 40:785–791. Coletiva 2011, 16:2907–2917.
26. Hair JF, Black WC, Babin JB, Anderson RE, Tatham RL: Multivariate Data 50. Pesquisa Nacional por Amostra de Domicílios: Perfil do Município de Sete
Analysis. Upper Saddle River: Prentice-Hall; 2009. Lagoas, MG. [[Link]
27. Oliveira SES, Hohendorff JV, Müller JL, Bandeira DR, Koller SH, Fleck MPA, 51. Instituto Brasileiro de Geografia e Estatística: Síntese de Indicadores Sociais:
Trentini CM: Associations between self-perceived quality of life and uma Análise das Condições de Vida da População Brasileira, 2012. Rio de
socio-demographic, psychosocial, and health variables in a group of Janeiro: 2012 [[Link]
elderly. Cad Saúde Pública 2013, 29:1437–1448. Indicadores_Sociais_2012/SIS_2012.pdf]
28. De Carlo LT: Using the PLUM procedure of SPSS to fit unequal variance 52. Moe JO, Steingrímsdóttir ÓA, Strand BH, Grøholt EK, Næss Ø: Trends in
and generized signal detection models. Behav Res Methods Instrum educational inequalities in old age mortality in Norway 1961–2009: a
Comput 2003, 35:49–56. prospective register based population study. BMC Public Health 2012,
29. McCullagh P: Regression models for ordinal data. J R Stat Soc Series B Stat 27:911.
Methodol 1980, 42:109–142. 53. Xavier FMF, Ferraz MPT, Norton M, Escosteguy NU, Moriguchi EH: Elderly
30. Lera L, Fuentes-Garcia A, Sanchez H, Albala C: Validity and reliability of the people’s definition of quality of life. Rev Bras Psiquiatr 2003, 25:31–39.
SF-36 in Chilean older adults: the ALEXANDROS study. Eur J Ageing 2013, 54. Seidl EMF, Zannon CMLC: Qualidade de vida e saúde: aspectos
10:127–134. conceituais e metodológicos. Cad Saúde Pública 2004, 20:580–588.
31. Estrada A, Cardona D, Segura AM, Chavarriaga LM, Ordóñez J, Osorio JJ: 55. Netuveli G, Blane D: Quality of life in older ages. Brit Med Bull 2008,
Calidad de vida de los adultos mayores de Medellín. Biomedica 2011, 85:113–126.
31:492–502. 56. Samper-Ternent R, Kuo YF, Ray LA, Ottenbacher KJ, Markides KS, Al Snih S:
32. Trentini CM, Chachamovich E, Wagner GP, Müller DH, Hirakata VN, Fleck Prevalence of health conditions and predictors of mortality in oldest old
MPA: Quality of life (QoL) in a Brazilian sample of older adults: the role Mexican Americans and non-Hispanic whites. J Am Med Dir Assoc 2012,
of sociodemographic variables and depression symptoms. Appl Res Qual 13:254–259.
Life 2011, 6:291–309.
33. Burström K, Johannesson M, Diderichsen F: Health-related quality of life by doi:10.1186/s12955-014-0166-4
disease and socio-economic group in the general population in Sweden. Cite this article as: Campos et al.: Aging, Gender and Quality of Life
Health Policy 2001, 55:51–69. (AGEQOL) study: factors associated with good quality of life in older
34. Onadja Y, Bignami S, Rossier C, Zunzunegui MV: The components of self-rated Brazilian community-dwelling adults. Health and Quality of Life Outcomes
health among adults in Ouagadougou, Burkina Faso. Popul Health Metr 2013, 2014 12:166.
11:15.
35. Blazer DG: How do you feel about…? Health outcomes in late life and
self-perceptions of health and well-being. Gerontologist 2008, 48:415–422.
36. Galenkamp H, Huisman M, Braam AW, Deeg DJ: Estimates of prospective
change in self-rated health in older people were biased owing to poten-
tial recalibration response shift. J Clin Epidemiol 2012, 65:978–988.
37. Bailis DS, Segall A, Chipperfield JG: Two views of self-rated general health
status. Soc Sci Med 2003, 56:203–217.
38. Ferraro KF, Wilkinson LR: Alternative measures of self-rated health for pre-
dicting mortality among older people: is past or future orientation more
important? Gerontologist 2013, [doi:10.1093/geront/gnt098].
39. Renaud J, Bédard E: Depression in the elderly with visual impairment and
its association with quality of life. Clin Interv Aging 2013, 8:931–943.
40. Campbell KH, Huang ES, Dale W, Parker MM, John PM, Young BA, Moffet
HH, Laiteerapong N, Karter AJ: Association between estimated GFR,
health-related quality of life, and depression among older adults with
diabetes: the Diabetes and Aging Study. Am J Kidney Dis 2013, 62:541–548.
41. Faravelli C, Alessandra Scarpato M, Castellini G, Lo Sauro C: Gender
differences in depression and anxiety: the role of age. Psychiatry Res 2013,
210:1301–1303.
42. Bryła M, Burzyńska M, Maniecka-Bryła I: Self-rated quality of life of
city-dwelling elderly people benefitting from social help: results of a
cross-sectional study. Health Qual Life Outcomes 2013, 11:181.
43. Lima MG, Barros MB, César CL, Goldbaum M, Carandina L, Ciconelli RM:
Impact of chronic disease on quality of life among the elderly in the
state of São Paulo, Brazil: a population-based study. Rev Panam Salud
Publica 2009, 25:314–321.
44. Sampaio PY, Ito E: Activities with higher influence on quality of life in
older adults in Japan. Occup Ther Int 2013, 20:1–10.
45. Bowling A, Banister D, Sutton S, Evans O, Windsor J: A multidimensional Submit your next manuscript to BioMed Central
model of the quality of life in older age. Aging Ment Health 2002, and take full advantage of:
6:355–371.
46. Noro A, Aro S: Health-related quality of life among the least dependent
• Convenient online submission
institutional elderly compared with the non-institutional elderly
population. Qual Life Res 1996, 5:355–366. • Thorough peer review
47. Wang H, Chen K, Pan Y, Jing F, Liu H: Associations and impact factors • No space constraints or color figure charges
between living arrangements and functional disability among older
Chinese adults. PLoS One 2013, 8:e53879. • Immediate publication on acceptance
48. Molzahn AE, Kalfoss M, Schick Makaroff K, Skevington SM: Comparing the • Inclusion in PubMed, CAS, Scopus and Google Scholar
importance of different aspects of quality of life to older adults across • Research which is freely available for redistribution
diverse cultures. Age Ageing 2011, 40:192–199.
49. Pereira RJ, Cotta RM, Franceschini Sdo C, Ribeiro Rde C, Sampaio RF, Priore
SE, Cecon PR: Influência de fatores sociossanitários na qualidade de vida Submit your manuscript at
[Link]/submit