Psychiatry Research 244 (2016) 37–44

Contents lists available at ScienceDirect

Psychiatry Research
journal homepage: www.elsevier.com/locate/psychres

Toward an understanding of the quality of life construct: Validity and

reliability of the WHOQOL-Bref in a psychiatric sample
Sandra E.H. Oliveira a,n, Helena Carvalho b, Francisco Esteves c
a
Instituto Universitário de Lisboa (ISCTE-IUL), CIS-IUL, Lisboa, Portugal
b
Instituto Universitário de Lisboa (ISCTE-IUL), CIES-IUL, Lisboa, Portugal
c
Mid Sweden University, Östersund, Sweden, & CIS-IUL, Lisboa, Portugal

art ic l e i nf o a b s t r a c t

Article history: This study tested the psychometric properties of the WHOQOL-Bref by examining its construct validity,
Received 13 December 2015 predictive validity and reliability in a psychiatric sample. The sample consisted of 403 participants re-
Received in revised form cruited from mental health care facilities. Construct validity was assessed through confirmatory factor
18 June 2016
analysis (CFA) and item-domains correlations. Predictive validity was evaluated via multiple regressions.
Accepted 5 July 2016
Internal consistency was analyzed by using Cronbach's alpha. Results from CFA second-order hierarchical
Available online 8 July 2016
model and item-domain correlational analyses supported the construct validity of the WHOQOL-Bref. A
Keywords: 5-domain model (psychological, physical, social relationships, environment and level of independence)
WHOQOL-Bref demonstrated good-fit and adequate internal consistency. Multiple regression analyses of the domains
Quality of life
with overall quality of life (QOL), general health and general QOL were supportive of predictive validity.
Mental illness
This study found support for the multidimensionality of the WHOQOL-Bref which demonstrated ap-
Confirmatory factor analysis
Validity propriate properties for the assessment of QOL in psychiatric inpatients and outpatients. Thus, a valuable
tool to be incorporated as part of the routine clinical evaluation, monitoring and an important indicator
of treatment outcome and research. Our findings suggest a conceptual distinction between the physical
domain and level of independence domain in this short version of the WHOQOL, as proposed by the
WHOQOL-100.
& 2016 Elsevier Ireland Ltd. All rights reserved.

1. Introduction psychiatric disorder for everyday life, as well as evaluating patient

Improving the quality of life (QOL) of people with severe
mental illness has become a major goal in the context of dein-
stitutionalization (Sartorius, 2006) and increasingly acknowledged
as an important measure of treatment outcome. In fact, since the
provision of mental health care has shifted from long-stay re-
sidence in psychiatric institutions to community-based services,
there has been a growing concern with improving the patients’
QOL, as well as evaluating the impact of healthcare interventions
on patients’ well-being, rather than focusing solely on symptom
reduction.
Besides taking into account the psychosocial implications of
diseases, current conceptualizations of QOL highlight the need to
take into account the subjective experience of the individual's
satisfaction with life (Katschnig, 2006). According to this rationale,
assessing the QOL of people with mental illness may contribute to
gaining a better understanding of the consequences of a
n
Correspondence to: CIS-IUL, Instituto Universitário de Lisboa, Av. das Forças
Armadas, 1649–026 Lisboa, Portugal. Tel.: +351 217903000; fax: +351 217964710.
E-mail address: seholiveira@gmail.com (S.E.H. Oliveira).
outcome and change in QOL over time. This perspective has be-
come central along with the recognition of the requirement to
achieve a cross-cultural QOL self-report measure, allowing valid
comparisons of results from different populations, cultural settings
and countries as an influential factor of subjective well-being
(Hawthorne et al., 2006).
The WHOQOL Group defined QOL as an individuals’ perception
of their position in life in the context of the culture and value systems
in which they live and in relation to their goals, expectations, stan-
dards and concerns (WHOQOL Group, 1995, p. 1405).
The World Health Organization Quality of Life assessment
(WHOQOL-100) resulted in a multilingual, multicultural, cross-
culturally sensitive and generic QOL instrument enabling to assess
variations in QOL across cultures and to compare groups within
the same culture (WHOQOL Group, 1998). Using the same ratio-
nale as the WHOQOL-100, the WHOQOL-Bref, is an abbreviated
version, which was developed simultaneously in 15 international
centers. Additionally it was found to be an adequate alternative
and particularly useful in situations where there is a need to
minimize respondent burden, the facet-level detail is dispensable
and, when time is limited (Skevington et al., 2004; WHOQOL

http://dx.doi.org/10.1016/j.psychres.2016.07.007
0165-1781/& 2016 Elsevier Ireland Ltd. All rights reserved.
38 S.E.H. Oliveira et al. / Psychiatry Research 244 (2016) 37–44
Group, 1995).
The WHOQOL-Bref has been widely field-tested in various
countries and its psychometric properties have demonstrated to
be adequate for use in different cultures and with a variety of
population groups including young people (Li et al., 2009), adults
(Fleck et al., 2000) and the elderly (von Steinbüchel, et al., 2006). It
has been also used in groups with particular medical conditions
such as patients with cancer (Michelone and Santos, 2004), epi-
lepsy (Liou et al., 2005), and mental disorders (Trompenaars et al.,
2005) such as depression (Berlim et al., 2005), bipolar disorders
(Chand et al., 2004), psychosis (Herrman et al., 2002), schizo-
phrenia (Mas-Expósito et al., 2011), and alcohol abuse (Da Silva
Lima et al., 2005). It is a 4-domain model derived from the 6-do-
main model of the WHOQOL-100, in which the level of in-
dependence domain was included in the physical domain, while
the spirituality domain has been merged with the psychological
domain.
Although the study of the construct validity and the model fit
of the WHOQOL-Bref among psychiatric samples demonstrate to
be highly relevant, research on this particular group remains
scarce. Simultaneously, while most of research has not explored
alternative factorial solutions of the 4-domain model proposed by
the WHO, the few existing studies investigating the factor struc-
ture of the WHOQOL-Bref have found some inconsistencies for the
support of its dimensionality. For instance, in a Nigerian study, an
eight-domain factor structure provided a better explanation of the
data than the WHOQOL-Brief's four and six-domain models
(Ohaeri et al., 2004). In another study with patients with tu-
berculosis and healthy referents in Taiwan, Chung and colleagues
(2012) found that while results from exploratory factor analysis
(EFA) on the healthy referents displayed a 4-domain model, the
factor structure generated a 6-domain model for the patient group.
Differences regarding the dimensionality of the WHOQOL-Bref
were also reported by Ohaeri and collaborators (2007) in a factor
analytical study with general population and psychiatric samples.
As for the European Portuguese versions of the WHOQOL-100
and WHQOL-Bref, both instruments have shown good psycho-
metric properties (internal consistency, test–retest stability, con-
vergent validity, discriminant validity). Interestingly, in the na-
tional version psychiatric patients (around 20.4% of the sample,
N ¼604) reported the worst results in QOL scores, except in the
physical domain in the 4-domain WHOQOL-Bref (Canavarro et al.,
2007; Vaz Serra el al, 2006) and in the physical and the level of
independence domains in the six-domain WHOQOL-100 (Cana-
varro et al., 2009). Also, as in the original validation study of the
WHOQOL-Bref (Skevington et al., 2004), while in the Portuguese
sample higher correlations (40.50) were found not just in the
intended domain, but also for some items within other domains.
For instance, two items of the physical domain belonging to the
level of independent domain on the WHOQOL-100 (activities re-
lated to daily living and work capacity) had strong correlations
with the psychological domain (Canavarro et al., 2007). Besides the
‘cross-domain’ correlations, in the European Portuguese validation
studies discriminant validity was best demonstrated in the phy-
sical domain in the WHOQOL-Bref and in the physical and in the
level of independence domains in the WHOQOL-100. These results
suggest that associations between the WHOQOL-Bref facets might
differ between the different populations whereby reducing the 24
items to four domain scores might lead to a loss of relevant in-
formation (von Steinbüchel, et al., 2006). Therefore, due to the
scarcity of studies with psychiatric samples, the fact that some
findings with this population group suggest structural differences
in the short version of the WHOQOL instrument (which may be
partly due to group-specific influences on some items) and also
because psychiatric patients may reveal some difficulties in com-
pleting the long form (WHOQOL-100), addressing the
psychometric properties of the WHOQOL-Bref may open up new
avenues for its rationale.
The purpose of this study was to examine the psychometric
properties of the WHOQOL-Bref by testing its dimensionality,
construct validity, predictive validity and reliability in a Portu-
guese psychiatric sample of inpatients and outpatients.
Information on the WHOQOL-Bref factor structure in psychia-
tric samples may contribute to improve knowledge on the sub-
jective experience of people with mental illness regarding their
QOL, which, in turn, may stimulate the use of standardized mea-
sures as a routine multidimensional assessment aimed at im-
proving QOL outcomes in the context of the mental health care
system.
2. Methods
2.1. Participants and procedure
Data was collected after approval by the ethical review boards
of the institutions. The participants were recruited from inpatient
and outpatient Portuguese mental health facilities: three general
hospitals, two community-based facilities, a psychiatric hospital,
and a psychiatric institution run by a religious order. The aim of
the study was explained by the researcher and all participants
signed informed consent. Participants were referred to the study
by the psychiatrist according the following inclusion criteria:
(1) adults with a clinical diagnosis based on the Diagnostic and
Statistical Manual of Mental Disorders (DSM-IV, American Psy-
chiatric Association, 1994); (2) without neurological disorders or
severe cognitive impairments or intellectual disabilities, and
(3) medically stabilized while under psychiatric treatment. Based
on these criteria, 15 individuals were excluded because they failed
to complete all the questionnaires and twenty-eight were ex-
cluded owing to low comprehension skills. The final sample con-
sisted of 403 adults, 231 males (57.3%) aged between 19 and 79
years (M ¼43.15, SD ¼12.38 and M ¼45.08, SD ¼13.67, for male and
female participants respectively). Most participants were single
(58.3%), on a disability retirement (49.6%) and diagnosed with
schizophrenia (36.5%). Socio-demographic and clinical character-
istics of participants are presented in Table 1.
2.2. Measure
The participants were administered the WHOQOL-Bref (WHO-
QOL Group, 1998) a 26-item questionnaire rated on a five-point
Likert scale with four domains measuring: psychological health,
physical health, social relationships and environment, plus 2 items
representing the general QOL (overall QOL and general health).
Since items 3, 4 and 26 were negatively formulated, they were
reversed before the analysis so that higher scores refer to high
QOL. For the purpose of our study, we used the Portuguese version
of the WHOQOL-Bref that has demonstrated adequate psycho-
metric properties, discriminating between healthy subjects and
patients with different medical conditions (Vaz Serra et al., 2006).
We followed the WHOQOL-Bref guidelines regarding administra-
tion, scoring, analyses and interpretation of the results (WHO,
1996).
2.3. Statistical analysis
Descriptive statistical analysis (Mean, SD and range) was car-
ried out through the examination of item-response distribution.
Skewness and kurtosis coefficients and respective standard errors
were examined and box-plots checked. In this study with a psy-
chiatric sample we started with exploratory factor analyses (EFA)
 S.E.H. Oliveira et al. / Psychiatry Research 244 (2016) 37–44 39

Table 1 the relative chi-square [χ2/df r 2, Schreiber et al., 2006), Com-

Sociodemographic and clinical characteristics of the sample (n ¼ 403). parative Fit Index (CFI) Z0.95, Goodness of Fit Index (GFI) Z0.95,
Tucker-Lewis Index (TLI) Z0.95, Root Mean Square Error of Ap-
Characteristic N (%)
proximation (RMSEA) r0.05 and standardized root mean square
Gender Male 231 (57.3) residual (SRMR) r0.08. The Akaike Information Criterion (AIC)
Female 172 (42.7) was also used to compare the models, whereas a lower AIC value
Age 18–35 122 (30.2) represents the better-fitting model.
36–45 102 (25.3) A non-parametric method (bootstrap) was also performed in
46–55 99 (24.6) order to validate the results obtained by maximum likelihood.
þ 56 80 (19.9)
Construct validity was also assessed through the estimation of
Marital Status Single 235 (58.3) intercorrelations of the WHOQOL-Bref items and domains, by
Married/Civil Union 98 (24.3)
using Pearson's correlation coefficient.
Divorced/Separated 57 (14.1)
Widowed 13 (3.2)
Finally, predictive validity was examined by means of multiple
regressions. Overall QOL, general health and general QOL were
Education No former education 8 (2.0)
entered as dependent variables while the five QOL-domains were
Elementary 84 (20.9)
Post-elementary 158 (39.2) used as predictive variables. It was predicted that the five domain-
Secondary 94 (23.3) specific QOL should show a strong and significant correlation with
College/University 59 (14.6) overall QOL, general health and general QOL. Standardized coeffi-
Work Status Student 5 (1.2) cient and Adjusted R2 were reported. Higher R2 indicates better
Employed 85 (21.1) predictive validity. Statistical analyzes were carried out using SPSS
Unemployed/Unoccupied 104 (25.8) (version 20.0) and AMOS (version 20.0).
Disability retired 200 (49.6)
Age retired 9 (2.2)

DSM – IV Diagnoses Schizophrenia 147 (36.5)

Psychotic disorder
Depressive disorder
Bipolar disorder
Anxiety disorder
Personality disorders
Alcohol use disorder
Treatment Context Outpatients
Inpatients
to assess the underlying factor structure and then, we followed
similar analytical procedures reported by the WHOQOL assess-
ment groups by testing construct validity through confirmatory
factor analyses (CFA) and by item-domain correlations. CFA should
be conducted with a different sample, thus, the total sample was
divided into two by random assignment (Cabrera-Nguyen, 2010).
The first subsample comprised about half of the sample
(n ¼191) and was used to perform EFA analyses (with varimax
rotation). The scree plots were also inspected to validate Kaiser's
criterion and the percentage of explained variance analyzed (since
using more than one criterion helps to better support the number
of latent factors underlying factorial structure). Factor loadings
Z0.5 were considered and unidimensionality of each factor was
also validated. A reliability assessment was included and internal
consistency reliability was measured by the Cronbach’s alpha.
Afterward, CFA was performed using maximum likelihood (ML)
estimation with the second subsample (n ¼212). CFA was used to
test the construct validity of the identified EFA structure and to
test the fit of the WHOQOL-Bref standard structure in this psy-
chiatric sample.
For both final EFA and CFA the ratio subject/item was 1:11,
which reduces the average of number of items misclassified on a
wrong factor and therefore, tends to produce more accurate so-
lutions (Costello and Osborne, 2005).
In CFA, each domain represented a latent variable and each
item was used as an observed variable. The existence of outliers
was analyzed by the square Mahalanobis distance (D2). The ana-
lysis of univariate and multivariate normality was observed by
analyzing the values of critical ratio. To guarantee for multivariate
normality in CFA, 7 observations were eliminated from the ana-
lysis. In order to determine how well the two models fit the
sample data (Hooper et al., 2008) the following recommended
“goodness of fit” indices and respective cut-off values were used:
3. Results
42 (10.4)
81 (20.1)
33 (8.2) 3.1. Factor structure
15 (3.7)
57 (14.1) An EFA was performed using all the items of the WHOQOL-Bref.
28 (6.9)
Similarly to other studies, items from the general facet of overall
266 (66.0) QOL and general health were excluded from the factor analysis
137 (34.0)
(Jaracz et al., 2006; Von Steinbuchel et al., 2006). The Kaiser-Meyer
Olkin (KMO) measure of sampling adequacy and the Bartlett test of
sphericity assessed the appropriateness of the data for applying
EFA. KMO was.810 and the Bartlett test of sphericity was sig-
nificant (p o0.001).
The best factor solution consisted of five domains, four of
which consistent with the factor structure proposed by the
WHOQOL-Bref: psychological health, physical health, environment
and social relationships; and a fifth factor corresponding to the
level of independence.
An EFA with a priori criterion was also conducted in order to
replicate the four-domain factor structure of the WHQOL-Bref,
which explained 6% less variance (59.51%) than the five-domain
factor structure. This structure explained 65.36% of the total var-
iance, which is not only higher than the majority of the studies
(Usefy et al., 2010; Skevington et al., 2004; Fleck et al., 2000) but is
also a better solution in accordance with the percentage variance
criterion (Hair et al., 2010).
With regard to the item loadings, in the first factor five of the
six items belonging to the psychological health domain had high
loadings, explaining 17.33% of the variance. The second factor ac-
counted for 13.17% of the explained variance, which is con-
ceptually similar to the WHO physical health domain because five
of the seven items belong to physical health. Three of the eight
items of the environment domain presented loadings 40.5 in the
third factor explaining 12.11% of the variance, which seems to re-
present a narrower definition of the environment domain. In fact,
three items had to be dropped out due to low loading (“health and
social care, availability and quality” [0.259] and “transport”) or
cross loading (“opportunities for acquiring new information and
skills”). The other two items representing the facets of “financial
resources” and “participation in, and opportunities for recreation”
loaded into an independent factor. Although the presenting factor
loadings are greater than 0.5 (0.785 and 0.537, respectively), these
items revealed weak internal consistency (α ¼0.342). For this
reason, these items were also excluded from the analysis in order
40 S.E.H. Oliveira et al. / Psychiatry Research 244 (2016) 37–44

Table 2 Table 3
PCA factor loadings and internal consistent of the WHOQOL-Bref 5-domains in a Correlations item/domains of the WHOQOL-Bref.
psychiatric sample (n ¼191).
Items Psychological Physical Social Environment Level of
WHOQOL-Bref items Component and Relations Independence
Domains
I II III IV V
Q3 0.01 0.69***  0.16** 0.08 0.18***
Spirituality/religion/personal beliefs 0.90 0.04 0.20 0.16  0.01 Q4 0.01 0.70***  0.02  0.02 0.23***
Positive feelings 0.90 0.05 0.13 0.19 0.01 Q5 0.88*** 0.22*** 0.37*** 0.38*** 0.38***
Self-esteem 0.71 0.08 0.32 0.17 0.23 Q6 0.89*** 0.22*** 0.41*** 0.38*** 0.37***
Body image and appearance 0.58 0.02  0.04 0.08 0.41 Q7 0.67*** 0.18*** 0.17** 0.34*** 0.36***
Thinking, learning, memory and 0.55 0.04  0.05 0.26 0.29 Q8 0.38*** 0.09 0.26*** 0.80*** 0.27***
concentration Q9 0.32*** 0.06 0.22*** 0.77*** 0.20***
Dependence on medication or  0.11 0.79  0.02  0.07 0.06 Q10 0.41*** 0.62*** 0.19*** 0.36*** 0.43***
treatments Q11 0.71*** 0.22*** 0.19*** 0.31*** 0.41***
Pain and discomfort 0.01 0.70  0.30 0.26  0.14 Q15 0.21*** 0.70*** 0.00  0.04 0.34***
Mobility 0.12 0.68  0.07  0.25 0.20 Q16 0.30*** 0.71*** 0.03 0.19*** 0.46***
Sleep and rest 0.13 0.62  0.05 0.01 0.38 Q17 0.43*** 0.47*** 0.17** 0.27*** 0.89***
Energy and fatigue 0.23 0.53 0.18 0.35 0.33 Q18 0.43*** 0.37*** 0.22*** 0.28*** 0.90***
Sexual activity 0.10  0.07 0.88 0.06  0.02 Q19 0.83*** 0.21*** 0.40*** 0.39*** 0.42***
Personal relationships 0.06  0.09 0.87 0.07 0.06 Q20 0.21***  0.07 0.81*** 0.18*** 0.07
Social Support 0.27  0.07 0.55 0.26 0.07 Q21 0.27***  0.03 0.81*** 0.17** 0.16**
Physical safety and security 0.26  0.04 0.07 0.79 0.13 Q22 0.42*** 0.09 0.69*** 0.32*** 0.28***
Physical environment 0.22  0.01 0.17 0.77  0.01 Q23 0.33*** 0.20*** 0.20*** 0.73*** 0.24***
Home environment 0.09 0.02 0.07 0.62 0.16
Activities of daily living 0.19 0.29 0.05 0.04 0.83 Correlation coefficients item/domain are shown in bold.
Working capacity 0.16 0.11 0.10 0.27 0.77 **
p o 0.01,
Eigenvalues 3.12 2.37 2.18 2.15 1.95 ***
p o0.001
Explained variance (%) 17.35 13.17 12.11 11.95 10.81
Cronbach's α 0.85 0.73 0.73 0.68 0.77

Principal component analysis. varimax rotation with Kaiser Normalization.
a
Factor loadings 40.50 are in bold.
b
I ¼ psychological domain; II – physical domain; III – social relationships domain;
IV – environment domain; V ¼ level of independence domain.
to enhance stability of the environment domain. The fourth factor
replicated the original 3-item social relationships domain, and
explained 11.95% of the variance. Finally, a fifth factor with two
higher loadings items (“activities of daily living” and “work capa-
city”) originally belonging to the physical health contributed with
10.81% for the variance explained. It is noteworthy although not
surprising that these items sorted into a conceptually autonomous
factor, since they pertain to the level of independence facets of the
enlarged version - WHOQOL-100. Moreover, in this psychiatric
sample, the two aforementioned items presented cross-loadings,
and almost completely not significant, in the four domain factor
structure. The EFA results with the five-domain factor structure are
displayed in Table 2.
3.2. Reliability
The internal consistency of the domains was assessed through
the Cronbach reliability coefficient. Cronbach's alpha values were
0.85 for psychological health, 0.73 for physical health, 0.73 for the
social relationships, 0.68 for environment and 0.77 for level of
independence domain, demonstrating the adequate reliability of
the WHOQOL-Bref for this sample.
3.3. Construct validity
In the current study construct validity was assessed through
CFA, as in other studies with the WHOQOL-Bref (e.g. Chen et al.,
2009) and by the analysis of item and domains correlations (e.g.
Skevington et al., 2004).
The item-domain correlation analysis showed that the items
correlated significantly within their domains and that no item
correlated more strongly with other domain than with its own
domain (Table 3). Correlation coefficients were high and ranged
from 0.67 to 0.89 for the psychological domain, from 0.62 to 0.71
for the physical domain, from 0.69 to 0.81 for the social
relationships domain, from 0.73 to 0.89 for the environmental
domain, and 0.89 and 0.90 for the level of independence domain.
CFA was used to test the construct validity of the identified
structure with a five-domain factor and to re-evaluate the fit of the
WHOQOL-Bref standard structure (i.e., four-domain factor) in this
psychiatric sample.
As presented in Fig. 1, two-order CFA was performed to ex-
amine the hierarchical factor model of the WHOQOL-Bref with five
domains in the second psychiatric subsample (n¼ 212). Since the
standard chi-square test may not be a reliable indicator to model
adequacy (Hu and Bentler, 1998) the relative chi-square fit index
(χ2/df) was also considered (values less than two have been sug-
gested as representing “good” data-model fit (Ullman, 2001). The
relative chi-square fit index for this model reached the re-
commended cut-off values, (χ2/df ¼1.515). Accepted values were
also found for four other “goodness of fit” indices: (χ2 ¼180.262;
po 0.001; CFI ¼ 0.955; GFI ¼0.914; TLI ¼0.942; RMSEA ¼0.050;
SRMR ¼0.055), suggesting for a relative good fit between the
hypothesized model and the observed data (Hu and Bentler, 1999).
In CFA the fit indices of the 5-domains EFA solution were
compared with the fit indices of the 4-QOL domains solution in
this psychiatric sample. For the AIC values, the 5-QOL model per-
formed better than the WHOQOL-Bref 4-model. Despite the par-
simony of the WHOQOL-Bref structure, the model with the level of
independence domain shows better fit indices (Table 4). A non-
parametric method (bootstrap) was also used in order to validate
the results obtained by a parametric method (ML) the bias be-
tween the two methods was minimum.
3.4. Predictive validity
As can be seen in Table 5, the regression results for predictive
validity of domain scores show that with exception of the physical
domain, all domains had significant predictive effects on overall
QOL, general health and general QOL. However, the predictive ef-
fects of the domains on the dependent variables were different.
The psychological domain contributed most for explaining overall
QOL, general health and general QOL. As remarked above with the
exception of the physical domain (p 40.05), all domains predicted
overall QOL significantly (p o0.05) while only the psychological
 S.E.H. Oliveira et al. / Psychiatry Research 244 (2016) 37–44 41

Fig. 1. WHOQOL-Bref: 5-domain confirmatory factor model.

Table 4 that multicollinearity was not a concern (all values met the cut-off
Comparison of fit indices of WHOQOL-Bref models. criterion of VIF o5 and the tolerance 40.20).
Fit Indices Structural models

4-domain model 5-domain model 4. Discussion

χ2 212.317 180.262 The purpose of this paper was to investigate the dimension-
df 128 119
χ2/df 1.659 1.515
ality, reliability, construct validity and predictive validity of the
CFI 0.938 0.955 WHOQOL-Bref among psychiatric inpatients and outpatients. Our
GFI 0.902 0.914 results showed that the instrument assesses QoL well for the
TLI 0.926 0.942 psychiatric population group. The instrument's four-factor struc-
RMSEA 0.057 0.050
ture was not fully confirmed in our sample, although international
SRMR 0.078 0.055
AIC 298.317 284.262 (Skevington et al., 2004) and European Portuguese studies (Ca-
navarro et al., 2007; Vaz Serra et al., 2006) have been able to va-
lidate this structure. Findings from our study suggest that the ideal
structure is a five-domain model made up of the four domains
Table 5 proposed by the WHOQOL-Bref (psychological health, physical
Predictive validity of domain scores for Overall QOL, General Health and General health, social relationships, and environment), plus the level of
QOL.
independence domain. This result is not unexpected since two
Dependent variables Predictors Adjusted R2 items that pertain to the level of independence domain in our
study also belong to this domain in the extended version (WHO-
PSY PHY SOC ENV IND QOL-100), which appear to give support for its conceptual au-
tonomy. Furthermore, in the development of a more recent ab-
Overall QOL (Q1) 0.42*** 0.07 0.10* 0.12** 0.11* 0.38
General Health (Q2) 0.51*** 0.02 0.03 0.06 0.18*** 0.44
breviated version of a WHOQOL instrument, the WHOQOL-HIV
General QOL (Q1 þ Q2) 0.52*** 0.05 0.07 0.10* 0.16*** 0.51 Bref, the level of independence domain although often subsumed
within physical health in QOL assessments as ‘functional status’ has
PSY psychological domain; PHY physical domain; SOC social relationships domain; shown that it is distinctive and therefore warrants particular clinical
ENV environmental domain; IND level of independence domain.
attention (O’Connell and Skevington, 2012, p. 459), suggesting that
Standardized coefficients were reported.
* this aspect of QOL might be particularly valued by those with
p o 0.05,
**
po 0.01, clinical conditions. Results on these quantitative assessments are
***
p o 0.001 supported by qualitative research, therefore helping to establish
the domains of QOL that are important to people with mental
illness. For instance, in a recent systematic review of qualitative
and the level of independence domains predicted general health research of the meaning of QOL for people with mental illness,
(p o0.001). Moreover, the psychological, the environment and the Connell, O’Cathain and Brazier (2014) have found that activity and
level of independence domains had significantly predictive effects independence are important aspects in determining the QOL for
on general QOL (p o0.05). The inspection of collinearity indicated this group. Specifically, while some individuals reported the
42 S.E.H. Oliveira et al. / Psychiatry Research 244 (2016) 37–44
benefits of being active, those with severe problems reported
missing the activities they once enjoyed. At the same time, al-
though some individuals mentioned the need for independence
they also recognized that, during periods of illness when they
were not able to deal with daily living activities and work, de-
pendence and requirement for support was particularly necessary.
As acknowledged by the authors, the level of dependence or in-
dependence varied according to current circumstances and dif-
fered over time, and while independence was found to be im-
portant for dignity, pride and privacy, dependency resulted in
feelings of guilt. As suggested by the WHOQOL-Group, con-
ceptually, QOL is a broad ranging concept incorporating in a
complex way the persons’ physical health, psychological state,
level of independence, social relationships, personal beliefs and
their relationships to salient features of the environment (WHO-
QOL Group, 1995). Findings from this study are in line with this
broad concept of QOL, suggesting that theoretically, the degree to
which people depend on others to help them in their daily activ-
ities or to work is also likely to affect their level of independence.
This domain refers to facets of activities of daily living, including
self-care; and working capacity, which is focused on a person’s
ability to perform work, regardless of the type of work. Although
related, this domain seems to be conceptually distinct from the
physical domain in this psychiatric sample, the latter linked to
physical sensations regarding pain and discomfort, energy and
fatigue, sleep and rest.
While the WHOQOL-Bref has been widely field-tested and its
psychometric properties have demonstrated to be adequate to be
used in different cultures and in a variety of population groups,
there are likely to be differences regarding the relative importance
of different domains, depending on the population studied. In fact,
results from studies using the WHOQOL-Bref in different patient
groups (e.g., Chung et al., 2012; Ohaeri et al., 2004; Fang et al.,
2002) have posed major challenges to the WHOQOL-Bref four-
factor model. Simultaneously, while the majority of the existing
studies had not specifically investigated the psychometric prop-
erties of the WHOQOL-Bref, some authors have limited their in-
vestigation to the use of EFA (Usefy et al., 2010; Trompenaars et al.,
2005). Other researchers have developed more detailed psycho-
metric studies using CFA to analyze if collected data replicated the
theoretical 4-domain model of the WHOQOL-Bref (Chung et al.,
2012; Xia et al., 2012; Ohaeri et al., 2007; Ohaeri et al., 2004). In
line with the aforementioned studies this study's results suggest
that regarding the dimensionality of the WHOQOL-Bref, in which
variations of the relative relevance that different domains may
have on the population studied should be taken into account. As
reported by Yao, Wu and Yang (2008) in their study on the content
validity of the WHOQOL-Bref, some standard items may be con-
founded with the meanings of different domains indicating that
they may not be representative of their posited domains. In line
with their findings, the authors suggest that research on the
WHOQOL instrument may necessitate the need to reevaluate the
meanings of items from the respondents’ viewpoints. That is, even
though the WHOQOL-Bref is a valid and reliable tool for academic
research, clinical evaluations, and cross-cultural comparisons
(Szabo, 1996); researchers should consider using facets of the in-
strument without reducing them to the four factors that often do
not replicate across specific groups (von Steinbüchel, et al., 2006).
Hence, this study also extends the applicability of the WHOQOL-
Bref since different valid models may help us to understand the
QOL characteristics of particular cultures or groups (Ohaeri et al.,
2007). Furthermore this allows a more comprehensive assessment
of the dimensionality, construct validity, predictive validity and
reliability of the WHOQOL-Bref in a psychiatric sample. Even
though the WHOQOL-Bref 4-domain model represents a more
parsimonious model, the alternative 5-domain model revealed
higher goodness of fit indices than those reported by the WHO's
4-domain and 6-domain models (Skevington et al., 2004). Con-
struct validity was also supported by the intercorrelations of the
WHOQOL-Bref items and domains since higher correlations were
found for all items within their domains rather than with other
domains.
The internal consistencies of the reliability of the domains were
satisfactory to good (Cronbach's α ranged between 0.68 and 0.85),
which is consistent with previous findings with psychiatric out-
patients (Trompenaars et al.,, 2005).
Evidence for predictive validity was sustained by regression
analysis showing that the psychological domain was the best
predictor for overall QOL, general health and general QOL, fol-
lowed by the level of independence domain. Meanwhile, though
suggesting that besides psychological health the level of in-
dependence may play an important role and a major concern for
those living with psychiatric conditions, the predictive value of the
level of independence, as an autonomous domain, remains to be
determined in future studies with psychiatric samples.
While the results of the present study add to a growing body
of research findings, providing support for the use of the WHO-
QOL-Bref in psychiatric inpatients and outpatients, there are
limitations of this study to be pointed out. Some participants
were inpatients in short-term and long-term residential pro-
grams with more restricted daily activities might have influenced
the results, in particular, in relation to the environment domain.
Also, due to the cross-sectional nature of this study, findings
should need to be replicated using a longitudinal research
design to deepen understanding the dimensionality of the
WHOQOL-Bref in psychiatric samples and the predictive power of
its domains.
Despite these limitations, implications of our findings and fu-
ture directions for research and clinical practice should be noted.
First, results of this study provide evidence for the need to care-
fully investigate the dimensional structure of the WHOQOL-Bref
across different subgroups, providing further clarity on the
WHOQOL-Bref performance in psychiatric samples. Second, this
study has significance since it was not limited to subjects with a
particular psychiatric diagnosis or attending a particular treatment
context, highlighting that assessing the QOL of people with mental
health problems who attend different treatment contexts, such as
inpatient and outpatient facilities should be at the focus of re-
search and part of the treatment goal. Also, because improving
these individuals’ QOL has become a critical outcome measure
regarding mental health services evaluation, this may contribute
to further understand the QOL of those attending different psy-
chiatric treatment modalities, such as hospital-based inpatient
long-term and short-term care, ambulatory services or commu-
nity-based facilities; as for monitoring potential changes on the
QOL of individuals in transition from psychiatric hospitals to
community facilities. Since deinstitutionalization goes far beyond
discharging patients from long-stay hospitals, it involves providing
adequate services and the transition from a predominantly in-
stitutional perspective into a community-based view (Fakhoury
and Priebe, 2007). Evaluating QOL of inpatients and outpatients is
of great interest in the context of deinstitutionalization since, of
several studies using the WHOQOL-Bref, while some have shown
improvements on the inpatients QOL years after discharge (Gerber
et al., 1994), others have found that psychiatric patients continue
to experience deterioration in QOL (Górna et al., 2005). Moreover,
further studies also need to clarify the influence of some factors,
such as socio-demographic and clinical factors on QOL of people
with psychiatric disorders. Third, measuring the QOL in people
with mental illness should play a major role in helping mental
health professionals to capture more salient aspects of individuals’
lives and to understand their needs. This, in turn, may promote
 S.E.H. Oliveira et al. / Psychiatry Research 244 (2016) 37–44
patients’ involvement in treatment decisions and improve sa-
tisfaction with mental health care.
Future research should address the factor structure of the
WHOQOL-Bref in a larger psychiatric sample, as well as in other
cultures and countries, since QOL is now clearly recognized as a
particularly useful framework to assess treatment outcomes and to
evaluate mental health services.
The extensive body of research systematically developed with
the WHOQOL-Bref, supports the use of this questionnaire as a
valid and reliable tool for addressing QOL, although the in-
vestigation of its dimensionality, validity and reliability in specific
subgroups remains insufficient.
This study contributes to bridging this gap, providing con-
firmatory evidence to the suitability of the WHOQOL-Bref with
psychiatric inpatients and outpatients. Our proposed 5-domain
model may also represent a valid alternative solution, particularly
of interest in studies with psychiatric samples. Structural differ-
ences of the WHOQOL-Bref may provide clarity regarding the QOL
concept and its variations, particularly important among specific
subgroups, suggesting a broader concept of QOL, reflected in the
WHOQOL original instrument (WHOQOL-100). Simultaneously,
due to the recognizable and well-defined influence of QOL-do-
mains, the intervention on particular aspects of QOL may have
important benefits for the rehabilitation and recovery process,
helping professionals to assist people with mental illness to afford
a better QOL.
Conflict of interests
The authors have no competing interests to report.
Acknowledgment
This study was supported by a grant from the Portuguese Sci-
ence Funding Agency–Fundação para a Ciência e Tecnologia to the
first author (grant number: SFRH/BD/75379/2010).
References
American Psychiatric Association, 1994. Diagnostic and Statistical Manual of Mental
Disorders, 4th ed., American Psychiatric Association, Washington DC.
Berlim, M.T., Pavanello, D.P., Caldieraro, M.A., Fleck, M.P., 2005. Reliability and va-
lidity of the WHOQOL-Bref in a sample of Brazilian outpatients with major
depression. Qual. Life Res. 14, 561–564.
Cabrera-Nguyen, P., 2010. Author guidelines for reporting scale development and
validation results in the journal of the society for social work and research. J.
Soc. Soc. Work Res 1, 99–103.
Canavarro, M.C., Simões, M.R., Vaz Serra, A., Pereira, M., Rijo, D., Quartilho, M.J.,
et al., 2007. Instrumento de avaliação da qualidade de vida da Organização
Mundial de Saúde: WHOQOL-Bref. In: Simões, M., Machado, C., Gonçalves, M.,
Almeida, L. (Eds.), Avaliação psicológica: Instrumentos validados Para a popu-
lação portuguesa Vol. III. Quarteto Editora, Coimbra, pp. 77–100.
Canavarro, M.C., Vaz Serra, A., Simões, M.R., Rijo, D., Pereira, M., Gameiro, J., et al.,
2009. Development and psychometric properties of the World Health Organi-
zation Quality of Life Assessment Instrument (WHOQOL-100) in Portugal. Int.
Soc. Behav. Med. 16, 116–124.
Chand, P.K., Matoo, S.K., Sharan, P., 2004. Quality of life and its correlates in patients
with bipolar disorder stabilized on lithium prophylaxis. Psychiatry Clin. Neu-
rosci. 58, 311–318.
Chen, W.C., Wang, J.D., Hwang, J.S., Chen, C.C., Wu, C.H., Yao, G., 2009. Can the web-
form WHOQOL- Bref be an alternative to the paper-form? Soc. Indic. Res. 94,
97–114.
Chung, W.S., Lan, Y.L., Yang, M.C., 2012. Psychometric testing of the short version of
the world health organization quality of life (WHOQOL-BREF) questionnaire
among pulmonary tuberculosis patients in Taiwan. BMC Public Health 12, 630.
Connell, J., O’Cathain, A., Brazier, J., 2014. Measuring quality of life in mental health:
are we asking the right questions? Soc. Sci. Med. 120, 12–20.
Costello, A.B., Osborne, J.W., 2005. Best Practices in Exploratory Factor Analysis:
four recommendations for getting the most from your analysis. Pract. Assess.
Res. Eval. 10, 1–9.
43
Da Silva Lima, A.F., Fleck, M., Pechansky, F., De Boni, R., Sukop, P., 2005. Psycho-
metric properties of the world health organization quality of life instrument
(WHOQOL-Bref) in alcoholic males: a pilot study. Qual. Life Res. 14, 473–478.
Fakhoury, W., Priebe, S., 2007. Deinstitutionalization and reinstitutionalization:
major changes in the provision of mental healthcare. Psychiatry 6, 313–316.
Fang, C.T., Hsiung, P.C., Yu, C.F., Chen, M.Y., Wang, J.D., 2002. Validation of the World
Health Organization quality of life instrument in patients with HIV infection.
Qual. Life Res. 11, 753–762.
Fleck, M.P., Louzada, S., Xavier, M., Chachamovich, E., Vieira, G., Santos, L., et al.
2000. Application of the Portuguese version of the abbreviated instrument of
quality of life WHOQOL-Bref. Revista Saúde Pública. 34, 178–283.
Gerber, G.J., Coleman, G.E., Johnston, L., Lafave, H.G., 1994. Quality of life of people
with psychiatric disabilities 1 and 3 years after discharge from hospital. Qual.
Life Res. 3, 379–383.
Górna, K., Jaracz, K., Rybakowski, F., 2005. Objective and subjective quality of life in
schizophrenic patients after a first hospitalization. Rocz. Akad. Med Bialymst.
50, 225–227.
Hair, J., Black, W., Babin, B., Anderson, R., 2010. Multivariate Data Analysis, 7th ed.,
Prentice-Hall, Inc,, Upper Saddle River, NJ, USA.
Hawthorne, G., Herrman, H., Murphy, B., 2006. Interpreting the WHOQOL-Bref:
preliminary population norms and effect sizes. Soc. Indic. Res. 77, 37–59.
Herrman, H., Hawthorne, G., Thomas, R., 2002. Quality of life assessment in people
living with psychosis. Soc. Psychiatry Psychiatr. Epidemiol. 37, 510–519.
Hooper, D., Coughlan, J., Mullen, M., 2008. Structural Equation Modelling: Guide-
lines for Determining Model Fit. EJBRM 6, 53–60.
Hu, L., Bentler, P.M., 1998. Fit indices in covariance structure modeling: sensitivity
to underparameterized model misspecification. Psychol. Methods 3, 424–453.
Hu, L., Bentler, P.M., 1999. Cutoff criteria for fit indexes in covariance structure
analysis: conventional criteria versus new alternatives. Struct. Equ. Model. 6,
1–55.
Katschnig, H., 2006. How useful is the concept of quality of life in psychiatry?. In:
Katschnig, H., Freeman, H., Sartorius, N. (Eds.), Quality of Life in Mental Dis-
orders, 2nd ed., John Wiley & Sons Ltd, West Sussex, pp. 3–17.
Jaracz, K., Kalfoss, M., Gorna, K., Baczyk, G., 2006. Quality of life in Polish re-
spondents: psychometric properties of the Polish WHOQOL-Bref. Scand. J.
Caring Sci. 20, 251–260.
Li, K., Kay, N.S., Nokkaew, N., 2009. The performance of the World Health Organi-
zations WHOQOL-Bref in assessing the quality of life of Thai college students.
Soc. Indic. Res. 90, 489–501.
Liou, H.H., Chen, R.C., Chen, C.C., Chiu, M.J., Chang, Y.Y., Wang, J.D., 2005. Health
related quality of life in adult patients with epilepsy compared with a general
reference population in Taiwan. Epilepsy Res. 64, 151–159.
Mas-Expósito, L., Amador-Campos, J.A., Gómez-Benito, J., Lalucat-Jo, L., 2011. The
World Health Organization Quality of Life Scale Brief Version: a validation study
in patients with schizophrenia. Qual. Life Res. 20, 1079–1089.
Michelone, A.,P., Santos, V.L., 2004. Quality of life of cancer patients with and
without an ostomy. Rev. Lat. Am. De. Enferm. 12, 875–883.
O’Connell, K., Skevington, S., 2012. An international quality of life instrument to
assess wellbeing in adults who are HIV-positive: a short form of the WHOQOL-
HIV (31 items). AIDS Behav. 16, 452–460.
Ohaeri, J.U., Awadalla, A.W., Ghoulum, A., El-Abassi, A.M., Jacob, A., 2007. Con-
firmatory factor analytical study of the WHOQOL-BREF: experience with Su-
danese general population and psychiatric samples. BMC Med. Res. Methodol.
7, 37.
Ohaeri, J.U., Olusina, A.K., Al-Abassi, A.-H.M., 2004. Factor analytical study of the
short version of the World Health Organization Quality of Life Instrument.
Psychopathology 37, 242–248.
Sartorius, N., 2006. Quality of life and mental disorders: a global perspective. In:
Katschnig, H., Freeman, H., Sartorius, N. (Eds.), Quality of life in mental dis-
orders, 2nd ed., John Wiley & Sons Ltd, West Sussex, pp. 321–327.
Schreiber, J.B., Nora, A., Stage, F.K., Barlow, E.A., King, J., 2006. Reporting Structural
Equation Modeling and Confirmatory Factor Analysis Results: A Review. J. Educ.
Res. 99, 323–338.
Skevington, S.M., Lotfy, M., O’Connell, K.A., 2004. The World Health Organization's
WHOQOL-BREF quality of life assessment: Psychometric properties and results
of the international field trial A Report from the WHOQOL Group. Qual. Life Res.
13, 299–310.
Szabo, S., 1996. On behalf of the WHOQOL Group. The world health organization
quality of life (WHOQOL) assessment instrument. In: Spiker, B. (Ed.), Quality of
Life and Pharmacoeconomics in Clinical Trials, 2ed., Lippincott-Raven Publish-
ers, Philadelphia.
Trompenaars, F.J., Masthoff, E.D., Van Heck, G.L., Hodiamont, P.P., De Vries, J., 2005.
Content validity, construct validity, and reliability of the WHOQOL-Bref in a
population of Dutch adult psychiatric outpatients. Qual. Life Res. 14, 151–160.
Ullman, J.B., 2001. Structural equation modeling. In: Tabachnick, B.G., Fidell, L.S.
(Eds.), Using Multivariate Statistics, 4th ed.., Allyn & Bacon, Needham Heights,
MA, pp. 653–771.
Usefy, A., Ghassemi, G.R., Sarrafzadegan, N., Mallik, S., Baghaei, A., Rabiei, K., 2010.
Psychometric Properties of the WHOQOL-Bref in an Iranian Adult Sample.
Community Ment. Health J. 46, 139–147.
Vaz Serra, A., Canavarro, M.C., Simões, M.R., Pereira, M., Gameiro, S., Quartilho, M.J.,
et al., 2006. Estudos psicométricos do instrumento de avaliação da Qualidade
de Vida da Organização Mundial de Saúde (WHOQOL-Bref) para Português de
Portugal. Psiquiatr. Clín. 27, 41–49.
44 S.E.H. Oliveira et al. / Psychiatry Research 244 (2016) 37–44
Von Steinbuchel, N., Lischetzke, T., Gurny, M., Eid, M., 2006. Assessing quality of life
in older people: Psychometric properties of the WHOQOL-Bref. Eur. J. Ageing 3,
116–122.
WHOQOL Group, 1995. The World Health Organization Quality of Life Assessment
(WHOQOL): Position paper from the world health organization. Soc. Sci. Med.
10, 1403–1409.
World Health Organization’s. 1996. Quality of Life Group: WHOQOL-Bref: In-
troduction. Administration and Scoring. Field (Trial) (version).
WHOQOL Group, 1998. The World Health Organization Quality of Life Assessment
(WHOQOL): development and general psychometric properties. Soc. Sci. Med.
46, 1569–1585.
Xia, P., Li, N., Hau, K.T., Liu, C., Lu, Y., 2012. Quality of life of Chinese urban com-
munity residents: a psychometric study of the mainland Chinese version of the
WHOQOL-Bref. BMC Med. Res. Methodol. 12, 37.
Yao, G., Wu, C.H., Yang, C.T., 2008. Examining the content validity of the WHOQOL-
BREF from respondents' perspective by quantitative methods. Soc. Indic. Res.
85, 483–498.