Quality of life can both influence and be an outcome of general health perceptions after heart surgery

The Interventional Centre, Faculty Division Rikshospitalet, Faculty of Medicine, University of Oslo, Oslo, Norway.
Health and Quality of Life Outcomes (Impact Factor: 2.12). 02/2007; 5(1):27. DOI: 10.1186/1477-7525-5-27
Source: PubMed
ABSTRACT
Our aim was to investigate the existence of a reciprocal relationship between patients' assessment of quality of life and their appraisal of health. If present, this relationship will interfere with the interpretation of heart surgery's effect on overall quality of life.
Path analysis was used to investigate reciprocal causal relationships between general health perceptions and overall quality of life before and after heart surgery. Longitudinal data from a study of coronary artery bypass surgery were used to model lagged, cross-lagged, and simultaneous paths over four time-points of assessment from before surgery to one year afterwards. The conceptual framework for the analysis was the Wilson and Cleary causal pathway model. General health perceptions were measured with the Short Form 36. Overall quality of life was measured with i) a single question regarding life satisfaction and ii) the multi-item Quality of Life Survey.
Acceptable model fit was obtained for reciprocal causation between general health perceptions and overall quality of life. Regression coefficients changed over different phases of rehabilitation. Serial correlation accounted for much of the variance within variables over time.
The present analysis demonstrates that unidirectional models of causality are inadequate to explain the effect of heart surgery on overall quality of life. Overall quality of life can causally influence as well as be an outcome of health status after coronary artery bypass surgery.

Full-text

Available from: Astrid Klopstad Wahl
BioMed Central
Page 1 of 10
(page number not for citation purposes)
Health and Quality of Life Outcomes
Open Access
Research
Quality of life can both influence and be an outcome of general
health perceptions after heart surgery
Lars Mathisen*
1,2
, Marit H Andersen
1,3
, Marijke Veenstra
4
, Astrid K Wahl
5
,
Berit R Hanestad
5
and Erik Fosse
1
Address:
1
The Interventional Centre, Faculty Division Rikshospitalet, Faculty of Medicine, University of Oslo, N-0027 Oslo, Norway,
2
Dept of
Thoracic and Cardiovascular Surgery, Rikshospitalet-Radiumhospitalet Medical Center, Sognsvannsveien 20, N-0027 Oslo, Norway,
3
Dept of
Surgery, Rikshospitalet-Radiumhospitalet Medical Center, Sognsvannsveien 20, N-0027 Oslo, Norway,
4
Dept. of Biostatistics, Rikshospitalet-
Radiumhospitalet Medical Center, Sognsvannsveien 20, N-0027 Oslo, Norway and
5
The Institute of Public Health/Faculty of Social Sciences,
University of Bergen, N-5020 Bergen, Norway
Email: Lars Mathisen* - lars.mathisen@medisin.uio.no; Marit H Andersen - marit.andersen@rikshospitalet.no;
Marijke Veenstra - marijke.veenstra@nova.no; Astrid K Wahl - a.k.wahl@medisin.uio.no; Berit R Hanestad - berit.hanestad@rektor.uib.no;
Erik Fosse - erik.fosse@medisin.uio.no
* Corresponding author
Abstract
Background: Our aim was to investigate the existence of a reciprocal relationship between
patients' assessment of quality of life and their appraisal of health. If present, this relationship will
interfere with the interpretation of heart surgery's effect on overall quality of life.
Methods: Path analysis was used to investigate reciprocal causal relationships between general
health perceptions and overall quality of life before and after heart surgery. Longitudinal data from
a study of coronary artery bypass surgery were used to model lagged, cross-lagged, and
simultaneous paths over four time-points of assessment from before surgery to one year
afterwards. The conceptual framework for the analysis was the Wilson and Cleary causal pathway
model. General health perceptions were measured with the Short Form 36. Overall quality of life
was measured with i) a single question regarding life satisfaction and ii) the multi-item Quality of
Life Survey.
Results: Acceptable model fit was obtained for reciprocal causation between general health
perceptions and overall quality of life. Regression coefficients changed over different phases of
rehabilitation. Serial correlation accounted for much of the variance within variables over time.
Conclusion: The present analysis demonstrates that unidirectional models of causality are
inadequate to explain the effect of heart surgery on overall quality of life. Overall quality of life can
causally influence as well as be an outcome of health status after coronary artery bypass surgery.
Background
Practice guidelines for chronic stable angina and for the
coronary artery bypass operation target improvement of
survival and symptomatic relief of angina [1], with
improvement of the quality of life (QoL) as an expected
secondary outcome [2]. Although a consensus exists on
the subjective nature of QoL, the issues of what to meas-
ure and how to interpret the results remain areas of con-
Published: 24 May 2007
Health and Quality of Life Outcomes 2007, 5:27 doi:10.1186/1477-7525-5-27
Received: 3 February 2007
Accepted: 24 May 2007
This article is available from: http://www.hqlo.com/content/5/1/27
© 2007 Mathisen et al; licensee BioMed Central Ltd.
This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0
),
which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
Page 1
Health and Quality of Life Outcomes 2007, 5:27 http://www.hqlo.com/content/5/1/27
Page 2 of 10
(page number not for citation purposes)
troversy [3]. Meanwhile, the number of papers returned
from a database query of 'QoL and coronary artery bypass
surgery' average one per week for every year of the last five
years, and clinicians face the task of judging the validity
and significance of this research [4]. The interpretation of
results from correlation research may generate assump-
tions of causality. If disease is a predictor of health, and
health has an effect on QoL, then therapies reducing the
burden of disease are expected to improve QoL. On the
other hand, if the patient's appraisal of QoL is not only an
outcome, but also affects the perception of health, then
the expectation that surgery can improve QoL may be too
simplistic. In other words, reciprocal causality will inter-
fere with the interpretation of heart surgery's effect on
overall quality of life. Valid and reliable measures of QoL
remain at risk of being labeled 'unresponsive' unless this
latent controversy is understood and resolved. In clinical
practice, evidence of reciprocal causality can support pre-
operative screening for patients who rate their QoL as
poor, to guide complementary interventions during reha-
bilitation to ensure that outcomes following surgery are
maximized. In this study, our aim was to investigate the
existence of a reciprocal relationship between patients'
assessment of quality of life and their appraisal of health.
Theoretical framework
In 1995, Wilson and Cleary proposed a causal pathway
model to link clinical variables to QoL (Figure 1), in order
to connect the field of objective measurement to that of
subjective experience [5].
This model has influenced the analysis of data from car-
diac [6-8] and other patient populations [9,10]. Wilson
and Cleary structured outcomes along a continuum of
increasing complexity from biological parameters
through symptom status, functional status, general health
perceptions and overall QoL [5]. General health percep-
tions reflect the functional status and symptoms such as
angina pectoris [5], and are important for their predictive
ability on the use of health care services as well as mortal-
ity [11]. Wilson and Cleary used the concepts health status
and health-related QoL interchangeably in the description
of their model. However, both concepts appear clearly
separate from overall QoL, which represents "a stable syn-
thesis of a wide range of experiences and feelings that peo-
ple have" [5]. Interaction effects of individual and
environmental characteristics may occur at each level of
outcomes.
Previous research citing Wilson and Cleary has modeled
unidirectional causal effects from general health percep-
tions towards overall QoL [6-10], under the assumption
that the dominant path of causality is sufficient to guide
data analysis. However, interpretation of results is condi-
tional upon the absence of significant reciprocal effects. It
is debatable whether QoL represents a summary outcome
of different and situational life aspects, or a "top-down"
individual disposition towards the evaluation of life
aspects [12,13]. Integration of these theoretical positions
in a reciprocal causality model of "top-down" personality
factors and "bottom-up" situational variables has been
proposed [14,15]. With repeated measurements of health
status and overall QoL in patients undergoing heart sur-
gery, an opportunity exists to challenge the conventional
direction of causality illustrated in Figure 1, and assess the
strength of causal relationships over time. If reciprocal
causality is possible and the mechanisms can be
explained, the Wilson and Cleary model must be accepted
as more complex than previously recognized in correla-
tion research.
Methods
Patient sample
The data set came from a previously reported [16] rand-
omized clinical study of on-pump versus off-pump coro-
nary artery bypass surgery. The parent study recruited and
included 120 patients between 40 and 80 years of age,
with stable angina pectoris and moderate or good left ven-
tricular function. Exclusion criteria were ejection fraction
< 30 % and/or renal failure (serum creatinine > 200
mmol/L), as well as patients unable to read, write or com-
municate verbally in Norwegian. The study protocol was
approved by the Regional Ethics Committee, and patients
provided written and informed consent. Five patients
were lost to follow-up due to mortality (2 patients) and
withdrawal (3 patients).
A causal pathway model of health-related quality of lifeFigure 1
A causal pathway model of health-related quality of
life. The horizontal arrows indicate the main, but not exclu-
sive, direction of causality. Reproduced with permission from
Wilson IB, Cleary PD. JAMA 1995;273:59–65. Copyright
©
1995 American Medical Association.
Biological and
Physiological
variables
Overall
Quality of
Life
General
Health
Perceptions
Symptom
Status
Functional
Status
Characteristics of
the Environment
Characteristics of
the individual
Nonmedical
factors
Psychological
Supports
Symptom
Amplification
Social and
Economic
Supports
Social and
Psychological
Supports
Personality
Motivation
Values
Preferences
Page 2
Health and Quality of Life Outcomes 2007, 5:27 http://www.hqlo.com/content/5/1/27
Page 3 of 10
(page number not for citation purposes)
For the present path analysis, the patients constituted one
group, as no significant effect of randomization to either
treatment arm was found [16]. Complete sets of data were
required for the analysis, resulting in the exclusion from
analysis of seven more patients where one or more data
points were missing. Thus, the patient sample for the
present study included 108 complete sets out of 120
potential sets of patient data, representing individuals (81
% men) between 47 and 79 years (mean age 64.2 years).
The patients were comparable to the parent study popula-
tion on all subscales of the SF-36 health status survey.
These individuals reported a median angina score of
Canadian Cardiovascular Society class II [17], and 44 %
had previously experienced myocardial infarction.
Procedure
A study database provided demographic data and clinical
parameters. The patients completed a questionnaire 4
times; at hospital admission before surgery (baseline),
and 3, 6 (questionnaire sent by mail) and 12 months after
surgery (follow-up visits at months 3 and 12). While the
questionnaire also included outcome measures such as
symptoms and functional status, the data analyzed in the
present paper only concern overall QoL and general
health perceptions. All in-patient assessments were sched-
uled before any further clinical or research diagnostics.
Self-reported variables
Overall quality of life
Overall QoL can be measured in different ways depending
on the substantive focus of investigation, such as happi-
ness, well-being, life satisfaction [18]. The theoretical
rationale and explicit ambition of the Wilson and Cleary
model, integration of the biomedical and social sciences,
suggested the use of life satisfaction instruments to repre-
sent overall quality of life [5]. Two instruments, Global
Quality of Life (gQoL) and a Norwegian version of the
Quality of Life Survey (QoLS-N), were used in order to
assess the influence of methods' effects.
Global Quality of Life, previously used in epidemiological
research [19], is a single-item overall appraisal of satisfac-
tion with current life, scored on a seven step Likert-type
scale: "Thinking about your life at the moment, would
you say that you by and large are satisfied with life, or are
you mostly dissatisfied?". The labeled response options
ranged from 'very dissatisfied' to 'very satisfied'.
The QoLS-N is a 16-item scale reported as a single sum of
item scores ranging from 16–112 points; higher scores
indicating better QoL. It explores factors such as physical
and material well-being, relationships with other people,
social and civic activities, personal development, recrea-
tion, and independence [20]. Each item is scored on a
seven-step scale with labeled response options that range
from 'very dissatisfied' to 'very satisfied' in the validated
translation [18]. Internal consistency (Cronbach's alpha)
of the QoLS-N is reported at 0.86, with a test-retest relia-
bility of 0.83 [21]. In this study, internal consistency was
0.83 at baseline and 0.90 at all subsequent time-points.
General health perceptions
We used the General Health subscale from the Short Form
36 (SF-36 version 1.2) as a single indicator. The 5 general
health items cover current health, health outlook, and
resistance to illness. Scores range from 0–100 points;
higher scores indicate better health. Internal consistency
(Cronbach s alpha) for this subscale has been reported at
0.84 [22] and varied in our study from 0.73 at baseline
and three months' assessment, to 0.78 (six months) and
0.81 (twelve months). General health perceptions are
associated with physical, mental and social health
domains [23].
Statistical analysis
We modeled causal paths with longitudinal data between
single indicator variables for overall quality of life and
general health perceptions. The path analysis used struc-
tural equation modeling [24] where all 4 time-points were
represented in all models tested. This method allows the
inclusion of feedback or reciprocal paths in addition to
unidirectional causal effects [25] and is therefore more
appropriate for our study than standard multiple regres-
sion technique. Figure 2 illustrates the two different sets of
reciprocal relationships that were modeled. Cross-lagged
components model the causal effect as observed at a later
point in time (Figure 2a), while simultaneous compo-
nents are observed at the same time (Figure 2b).
Structural equation modeling does not prove causality,
but it tests whether the data set, with its inherent covari-
ance structure, supports or rejects the postulated effects.
Thus, the data matrix under analysis is the set of covari-
ances between all pairs of variables. The interrelationships
of the observed variables are specified in structural equa-
tions by the researcher, according to hypotheses and the-
ories. Adding, removing or changing the direction of an
effect (arrow in Figure 2) means changing the set of regres-
sion equations. The combination of all equations form a
model, and the fit or appropriateness of this model is
tested by analyzing all equations simultaneously, looking
at the whole landscape rather than the individual parts.
The result of analysis is expressed as a set of fit indices,
indicating how well the specified model fits the observed
reality. The task of interpretation is to accept, reject, or
possibly modify the paths included in the model.
Scoring of the SF-36 was completed according to the man-
ual [23] using the SPSS version 12.0 (SPSS Inc., Chicago
IL). To analyze the extent of selective attrition, χ
2
and t-
Page 3
Health and Quality of Life Outcomes 2007, 5:27 http://www.hqlo.com/content/5/1/27
Page 4 of 10
(page number not for citation purposes)
tests for independent samples were used. Histograms and
descriptive statistics for the individual variables were
screened for deviations from normality. The distribution
of scores indicated reasonable compliance with the
assumptions of linear modeling. The covariance matrix
was analyzed using maximum likelihood estimation in
Lisrel version 8.72 (Scientific Software International, Lin-
colnwood IL). We allowed for each effect component to
vary over time. First, we modeled the time-lagged effects
between general health perceptions and overall QoL over
4 time-points. We included correlations between variables
measured at the same occasion to take into account the
presence of confounding variables [25,26]. The second
model included the cross-lagged effects between the dif-
ferent variables to the following time-point of assessment.
Finally, a third model was introduced to evaluate simulta-
neous effects, where the correlation between variables
measured within one time-point in the previous models
was replaced by reciprocal causal effects. A series of model
fit characteristics were used to evaluate the adequacy of
different causal models: (i) χ
2
analysis, testing that the
model is not significantly different (p > 0.05) from the
underlying population covariance matrix, (ii) compara-
tive fit index (CFI) above 0.90, (iii) root mean square
error approximation (RMSEA) indicating acceptable
(RMSEA < 0.08) or good (< 0.05) fit of the residuals, and
(iv) standardized root mean square residuals (SRMR) less
than 0.10 [24]. The p of Close Fit tested the null hypothe-
sis of RMSEA < 0.05; a non-significant test implies accept-
able model fit. The significance of the cross-lagged effects
and the equivalence of cross-lagged or simultaneous
effects were evaluated with a χ
2
difference test using a crit-
ical value (alpha) of 5 %.
Results
General health and QoL measurements at all time-points
are presented in Table 1, and their bivariate relationships
in Table 2. Relative to preoperative status, the mean
improvement in general health score at all three follow-up
points ranged from 9.5 (SD 18.8) to 12 (SD 18.6) points.
At the group level, the greatest improvement in general
health occurred from baseline to 3 months after surgery,
although at the individual level, one third (32.4 %) of the
individual patients reported no change or decline.
At all assessments, the individual patient scores on the
global Quality of Life item (gQoL) covered the full scale
range from 1 to 7. The group mean gQoL improved from
5.0 before surgery to 5.6 at 12 months' follow-up (Table
1). Overall QoL scores measured with the QoLS-N also
demonstrated individual variation; the group mean at
baseline was 86.5 points increasing to 88.1 after 12
months, while individual scores ranged from 52.6 to 112
points during the same time-span.
Structural equation modeling of global Quality of Life
(gQoL)
Full versions of the cross-lagged and simultaneous path
models, with unstandardized estimates of the paths as
well as correlations of the error variances of the variables,
are available under Additional files. The abbreviated ver-
sions in Figure 3 and 4 serve to outline the main results
with standardized regression coefficients.
We started by estimating a model with longitudinal paths
connecting all assessments within each of the two out-
come measures at all time-points, but no causal paths
between general health and overall QoL. The analysis pro-
vided acceptable model fit and consequently, this model
served as a reference model from which improvements
through the addition of hypothesized causal paths could
be evaluated (Table 3). We also tested an alternative and
more parsimonious lagged effects model, where the only
paths specified were the connections from one time-point
to the immediate next within each variable, omitting the
bridging connections between baseline and 6 months,
baseline and 12 months, and 3 months and 12 months'
assessments. This model fitted the data poorly and was
discarded.
The cross-lagged model added reciprocal effects from the
previous to the next assessment between QoL and general
health perceptions, and the standardized regression coef-
ficients of these causal paths are presented in Figure 3 (see
also Additional file 1: Additional file
1_xlagged_gQOL.pdf). A statistically significant cross-
Reciprocal causal paths, illustrating a) cross-lagged effects and b) simultaneous effectsFigure 2
Reciprocal causal paths, illustrating a) cross-lagged
effects and b) simultaneous effects. Single arrows indi-
cate causal paths. Only two time-points are illustrated, while
four time-points were analyzed in all models reported in this
paper.
a) b)
General
health
Quality
of Life
Quality
of Life
General
health
Time YTime X
General
health
General
health
Quality
of Life
Quality
of Life
Time X Time Y
Page 4
Health and Quality of Life Outcomes 2007, 5:27 http://www.hqlo.com/content/5/1/27
Page 5 of 10
(page number not for citation purposes)
lagged effect from overall QoL at baseline to the three
months assessment of general health was present, indicat-
ing independent predictor properties of the baseline QoL
appraisal during the greatest change in general health sta-
tus. This model fitted the data well (Table 3) although the
change in χ
2
compared to the lagged model was not statis-
tically significant. The simultaneous reciprocal model
(Figure 4) demonstrated best fit and, by chi-square test, a
significant model improvement, with significant path
coefficients observed at three and six months after surgery
from overall QoL toward general health perceptions (see
also Additional file 2: Additional file
2_simultaneous_gQOL.pdf). To contrast this analysis
with an assumption of no causal effects, we set the bidirec-
tional paths within each time-point to equal size. This
resulted in an unidentifiable model.
Structural equation modeling of the Quality of Life Survey
(QoLS-N)
We re-ran the previous analyses with the QoLS-N results
(Table 3). Full versions of the cross-lagged and simultane-
ous path models are available as Additional files 3 and 4
(Additional file 3_xlagged_QOLSN.pdf, and Additional
file 4_simultaneous_QOLSN.pdf). By fitting identical
models with the two instruments, it was possible to assess
the extent of instrument-specific results, which constitutes
a step towards cross-validation (Figure 5).
In the cross-lagged model, the path from general health at
six months to QoL at 12 months was statistically signifi-
cant. Each point increase in general health at six months,
resulting from all directional paths assigned, was associ-
ated with 0.11 point increase in QoLS-N at one year after
Table 2: Bivariate relationships of general health perceptions and overall quality of life at four time-points
Baseline 3 months 6 months 12 months
General
health
gQoL QoLS-N General
health
gQoL QoLS-N General
health
gQoL QoLS-N General
health
gQoL QoLS-N
Baseline
General health 446 .32* .38* .57* .27* .18 .58* .24 .20 .54* .22 .24
gQoL 1.25 .42* .37* .32* .23 .36* .43* .34* .35* .46* .35*
QoLS-N 104 .28* .22 .58* .35* .30* .59* .21 .24 .59*
3 months
General health 354 .55* .42* .71* .39* .30* .75* .46* .42*
gQoL .76 .52* .44* .53* .30* .47* .54* .46*
QoLS-N 107 .36* .37* .61* .35* .37* .66*
6 months
General health 387 .53* .54* .66* .41* .51*
gQoL .81 .57* .37* .62* .42*
QoLS-N 120 .24 .43* .64*
12 months
General health 453 .43* .49*
gQoL .80 .52*
QoLS-N 106
The diagonal (italic) represents variance; the upper diagonal contains Pearson correlation coefficients, asterisks indicate p = 0.01. Common to all
three scales, higher scores indicate better health or quality of life.
Table 1: General health and overall quality of life at 4 time-points (n = 108)
Baseline 3 months 6 months 12 months
General
health
57.7 (21.1) 69.7 (18.8) 67.2 (19.7) 68.7 (21.3)
gQoL 5.0 (1.1) 5.6 (0.9) 5.5 (0.9) 5.6 (0.9)
QoLS-N 86.5 (10.2) 88.0 (10.3) 87.4 (11.0) 88.1 (10.3)
Each observation is presented as mean (standard deviation).
For all three scales, higher scores indicate better health or quality of life.
General health scale range 0–100.
gQoL scale range 1–7.
QoLS-N scale range 16–112.
Page 5
Health and Quality of Life Outcomes 2007, 5:27 http://www.hqlo.com/content/5/1/27
Page 6 of 10
(page number not for citation purposes)
surgery. In the simultaneous reciprocal model, significant
standardized regression coefficients from QoL to general
health dominated at 3 and 6 months after surgery (0.26
and 0.28, respectively), while the reverse effect from gen-
eral health towards QoL (0.27) was present at one year
after surgery. However, the model fit indices suggested
that the cross-lagged model was superior to the model
allowing simultaneous and bidirectional causal paths
(Table 3). Cross-lagged effects also demonstrated better
model fit than two unidirectional and simultaneous
effects models, either according to the conventional paths
of the Wilson and Cleary model or with reversed causa-
tion only – from QoL towards general health. An
extended version of Table 3 details the latter results and is
available on request.
Discussion
In this study, structural equation modeling supported the
existence of reciprocal causal paths between general
health perceptions and overall QoL. Our longitudinal
analysis indicated that changes in general health percep-
tions may be conditional upon as well as contributing to
the appraisal of overall QoL. The clinical implication is
that the evidence used for preoperative counseling on
expected changes in symptoms and functioning, should
not be extrapolated to health in general or life satisfaction
following coronary artery bypass surgery. Restraining the
level of abstraction to outcomes that are conceptually
closer to clinical parameters and the surgical intervention,
such as symptoms and functional status domains [5], may
prevent misunderstandings and facilitate joint decision-
making between patient and provider. Bearing on
research, the present study describes a potential source of
error when interpreting cross-sectional associations
between overall QoL, general health, and heart surgery.
We based our analysis on research demonstrating that
general health perceptions and overall QoL represent con-
ceptually and empirically distinct dimensions [27]. Previ-
ous studies of heart patients where the Wilson and Cleary
model has been evaluated as a conceptual framework,
present different interpretations of overall QoL: either as
life satisfaction similar to the present study [6], or by refer-
ring to health-related QoL (HQoL) [28], or disease-spe-
cific health-related QoL [7]. If the main causal direction
from general health perceptions towards overall QoL were
dominant, blurring of the conceptual distinction between
health status or HQoL and overall QoL would be less con-
sequential, although possibly not desirable [3]. However,
our results indicate a more complex causal network,
which precludes conceptualizing overall QoL as health or
subsumed within health. Examples from neighbor
research fields expand on the causal networks involved,
where explanatory variables associated with overall QoL
include a genetic component [29,30], personality trait
characteristics [31], and – in cardiac patients – life orien-
tation or Sense of Coherence [32].
In our study, the regression coefficients between general
health perceptions and overall QoL did not demonstrate
a stable pattern. This variation may represent true varia-
tion of structurally stable constructs, or there may be uni-
dentified structural variation due to response shifts from
changing values or beliefs of respondents [33], possibly
mediated by alterations in cognitive processing after heart
surgery. Sample size was in this study insufficient for
Table 3: Model fit indices
χ
2
df p RMSEA CFI SRMR ∆χ
2
df
gQoL
1. Lagged effects only 18.55 12 .100 .072
a
.990 .125 reference
2. Cross-lagged effects 7.71 6 .260 .052
a
.997 .023 10.84 6
(Figure 3)
3. Simultaneous effects 9.76 9 .371 .028
a
.998 .026 8.79
b
3
(Figure 4)
QoLS-N
4. Lagged effects only 17.50 12 .132 .066
a
.990 .076 reference
5. Cross-lagged effects 7.21 6 .302 .044
a
.998 .022 10.29 6
6. Simultaneous effects 17.49 9 .042 .095
a, c
.986 .066 .01 3
a
p of Close Fit >.05, the null hypothesis is RMSEA <.05.
b
P < .05
c
The 90% confidence interval is .018 to .161
CFI Comparative Fit Index, values > 0.90 indicates a good fit of the model.
RMSEA Root Mean Square Error of Approximation, values < 0.08 indicating acceptable, and < 0.05 good fit of the residuals.
SRMR Standardized Root Mean Residual, values < 0.10 indicating good fit.
df Difference in degrees of freedom from reference model
∆χ
2
Difference in chi-square from reference model chi square
Page 6
Health and Quality of Life Outcomes 2007, 5:27 http://www.hqlo.com/content/5/1/27
Page 7 of 10
(page number not for citation purposes)
modeling of latent variables to control for time invariance
of item factor loadings on each construct. However, we
used instruments where the variance was fully modeled –
as in the case of the single-item gQoL – or where the factor
structure and reliability has been extensively reported
across patient groups. The adequacy of the lagged effects
models suggested that longitudinal construct validity was
satisfactory. A homeostatic concept of subjective well-
being offers a possible explanation to the changing "bot-
tom-up" and "top-down" causal paths over situations and
time. It has been suggested that mechanisms to restore
homeostasis are triggered when a challenge to an individ-
ual set point occurs from contributing domains [13].
Severe health impairments may trigger through negative
feedback, whereas positive health transitions following
surgery may alter the relative strength of bidirectional
paths connecting health perceptions and QoL. In the
present study, the patients were in a transition from a pre-
operative state through rehabilitation. During the period
of greatest change in general health perceptions, from pre-
operative status to three months after surgery, paths from
overall QoL towards general health appeared dominant in
the simultaneous effects models (gQoL and QoLS-N) as
well as in the cross-lagged model using gQoL observa-
tions. In contrast, the cross-lagged QoLS-N model was
indifferent from baseline to three months after surgery.
Furthermore, from six months to one year after surgery
and during less magnitude of health transition, the regres-
sion coefficients in the conventional direction from gen-
eral health perceptions towards overall QoL were
significant in the cross-lagged and simultaneous effects
QOLS-N models (Figure 5, Additional file 3: Additional
Simultaneous effects modelFigure 4
Simultaneous effects model. This model demonstrates
bidirectional causal paths at each time-point observed after
baseline. The path from general health towards overall qual-
ity of life occurs at the same time as the path from overall
quality of life towards general health perceptions is observed.
Two of the six simultaneous paths are statistically significant
in the direction from 3 and 6 months' gQOL towards the
corresponding scores on the SF-36 General health subscale.
As with the cross-lagged paths, this model includes strong
serial associations over time within each variable except
between 6 and 12 months' general health perceptions.
Straight and curved single arrows indicate the causal paths
modeled. The corresponding decimals are standardized
regression coefficients. Bold face coefficients indicate p <
0.05 while broken lines are used for paths with a corre-
sponding p 0.05. The curved line between baseline varia-
bles represents a correlation; the number is the
corresponding correlation coefficient. Model fit indices are
summarized in Table 3. For correct model specification of
simultaneous effects, the present baseline assessment pro-
vides exogenous and correlated variables, without simultane-
ous effects at this time-point.
General
health
General
health
gQOL
General
health
gQOLgQOL
gQOL
R
2
= .55
General
health
R
2
= .61
.33
.41
.23
.40
.11 .07
.28
.28 .07
.49 .19.46
.22
.24
.32
.25 .51
.34
baseline 3 months 6 months 12 months
.13
Cross-lagged modelFigure 3
Cross-lagged model. This model demonstrates bidirec-
tional causal paths from one time-point to the following,
from general health towards overall quality of life as well as
from overall quality of life towards general health percep-
tions. One of these cross-lagged paths is statistically signifi-
cant; the path from baseline gQOL towards 3 months' scores
on the SF-36 General health subscale. Together with the
cross-lagged paths, the model also includes strong serial
associations over time within each variable except between 6
and 12 months general health perceptions. Straight and
curved single arrows indicate the causal paths modeled. The
corresponding decimals are standardized regression coeffi-
cients. Bold face coefficients indicate p < 0.05 while broken
lines are used for paths with a corresponding p 0.05. The
curved line between baseline variables represents a correla-
tion; the number is the corresponding correlation coefficient.
Model fit indices are summarized in Table 3. Figures 3 and 4
are available in more detailed versions as Additional files 1
(Additional file 1_xlagged_gQOL.pdf) and 2 (Additional file
2_simultaneous_gQOL.pdf) with this paper, including the
unstandardized parameter estimates.
General
Health
gQoL
General
Health
gQoLgQoL
General
Health
gQoL
R
2
=.55
General
Health
R
2
=.60
-.01.02.16
.17 .08 .07
baseline 3 months 6 months 12 months
.32
.31 .46
.26
.23
.34
.52.26
.42
.51 .52 .17
.14
Page 7
Health and Quality of Life Outcomes 2007, 5:27 http://www.hqlo.com/content/5/1/27
Page 8 of 10
(page number not for citation purposes)
file 3_xlagged_QOLSN, and Additional file 4: Additional
file 4_simultaneous_QOLSN).
Variation between regression coefficients were in the
present study associated with the choice of overall QoL
instrument. While a consensus exists as to how QoL
should be assessed, i.e. as a subjective appraisal obtained
by asking the patient [34], there is no gold standard or ref-
erence criterion for evaluation of content validity of over-
all QoL instruments [35]. In our observations, we selected
measures that emphasize life satisfaction as a critical com-
ponent of overall QoL. It should be noted that Wilson and
Cleary [5] also cite subjective well-being and happiness
along with life satisfaction as representative for overall
QoL, although their paper does not enter the discussion
on structural relationships of indicators of overall QoL.
The correlation matrix of Table 2 indicates only a moder-
ate overlap of content between the gQoL and the QoLS-N.
Their intercorrelation coefficients remain below 0.60, and
the extent of common methods variance is unknown. Two
complementary explanations may be offered for the mod-
est strength of association and the different results
obtained when modeling with different instruments:
First, compared to the QoLS-N, the single-item gQoL
emphasizes "top-down" effects towards general health
perceptions during changing health conditions. It is pos-
sible that the single question favors a life orientation
response, as the response options neither are anchored to
specific life domains nor impose any assumptions of
weighting due to the number or order of items. Con-
versely, the sum score of the QoLS-N may represent a bot-
tom-up perspective of overall QoL as a sum of experiences
and appraisals. However, although the selection of items
is empirically grounded [20], each item is given equal
weight in the summary score and this may not adequately
reflect the preferences and priorities held by respondents.
Second, one may question whether the gQoL and the
QOLS-N represent the same latent variable, life satisfac-
tion. Exploring their relationships, we correlated the gQoL
at baseline and at one year after surgery with a three-factor
solution of the QOLS derived from analysis of healthy
subjects' responses [20]. Amongst these factors, Health
and Functioning demonstrated the greatest strength of
association to the gQoL, followed by Relationships and
Material Well-Being and finally Personal, Social and Com-
munity Commitment. Of note, the QoLS-N scale contains
one item specifying physical health. To control for unto-
ward loss of variance and inflated regression coefficients
between the observed variables in our analyses, we ran
separate models with a 15-item modification of the QoLS-
N score in which the health item was deleted. No substan-
tial change in model fit was observed (data available on
request).
Some limitations of this study should be acknowledged.
We assumed that the baseline values of our observed var-
iables carried adequate adjustment for numerous candi-
date background variables such as gender, age,
socioeconomic status, level of education and co-morbid-
ity. A larger sample size would allow for more parameters
and variables – observed or latent – to be included. How-
ever, our analysis used the p of Close Fit indicator (see leg-
end, Table 3) to provide an estimate of sufficient power to
detect poor model fit due to misspecification. As this
study is an early investigation of reciprocal effects, we
could not locate publications that could validate the tim-
Comparison of significant a) cross-lagged and b) simultaneous paths from two sets of modeling with two different quality of life instruments: gQoL and QoLS-NFigure 5
Comparison of significant a) cross-lagged and b)
simultaneous paths from two sets of modeling with
two different quality of life instruments: gQoL and
QoLS-N. Figure 5 summarizes only the statistically signifi-
cant paths observed between General Health and overall
Quality of Life, indicated as arrows in the direction of causal-
ity. The causal paths are labeled with their corresponding
QoL instrument and standardized regression coefficients
derived from structural equations. Paths within each concept
from one time-point to another, for example from General
Health at three months to General Health at 12 months, are
not drawn. See Table 3 for model fit indices, and Additional
files 1 through 4 for separate model parameters.
a) Cross-lagged models
QoL QoL QoL QoL
gQoL .17
QoLS-N .21
SF-36 GH SF-36 GH SF-36 GH SF-36 GH
Baseline 3 months 6 months 12 months
b) Simultaneous (reciprocal) effects models
QoL QoL QoL QoL
gQoL .33 gQoL .28
QoLS-N .26 QoLS-N .28 QoLS-N .27
SF-36 GH SF-36 GH SF-36 GH SF-36 GH
Baseline 3 months 6 months 12 months
Page 8
Health and Quality of Life Outcomes 2007, 5:27 http://www.hqlo.com/content/5/1/27
Page 9 of 10
(page number not for citation purposes)
ing of assessments as more or less sensitive to the causal
paths investigated.
Conclusion
Unidirectional models of causality are inadequate to
explain the effect of cardiac surgery on overall QoL. Over-
all quality of life can causally influence as well as be an
outcome of health status after coronary artery bypass sur-
gery. Our analysis substantiates the potential for recipro-
cal effects within the Wilson and Cleary model. This study
offers a pilot design for confirmatory modeling with more
frequent sampling of a larger patient population.
Abbreviations
gQoL Single-item Global Quality of Life question
QoL Quality of Life
QoLS-N Quality of Life Survey-Norwegian
SF-36 GH Short Form 36 General Health subscale
Competing interests
The author(s) declare that they have no competing inter-
ests.
Authors' contributions
LM initiated this paper as part of a larger study of patient
reported outcomes and drafted the manuscript together
with MV, who provided statistical advice. MA participated
in data collection. All authors critiqued revisions of the
paper and approved the final manuscript. EF, BRH and
AKW supervised LM and MA; EF was principal investigator
for the research program on off-pump versus on-pump
coronary artery bypass surgery.
Additional material
Acknowledgements
We thank Arne Kolstad for statistical advice, Per Kristian Hol for coordi-
nating the clinical study, the staff at the Interventional Centre and the
Department of Thoracic and Cardiovascular Surgery for skilful and compas-
sionate patient care, the respondents for sharing their experiences, and Jan
L. Svennevig for reviewing a draft of this paper. Kari Skinningsrud reviewed
a late draft as a medical writer funded by the Interventional Centre. Anon-
ymous referees provided constructive critique during the review process.
LM was financially supported by unrestricted grants from the Research
Council of Norway, the Norwegian Nurses Association, the Norwegian
Association of Heart and Lung Patients, and the Center for Patient Partici-
pation and Nursing Research at Rikshospitalet.
References
1. Eagle KA, Guyton RA, Davidoff R, Edwards FH, Ewy GA, Gardner TJ,
Hart JC, Herrmann HC, Hillis LD, Hutter AM, Lytle BW, Marlow RA,
Nugent WC, Orszulak TA: ACC/AHA 2004 guideline update for
coronary artery bypass graft surgery: a report of the Ameri-
can College of Cardiology/American Heart Association Task
Force on Practice Guidelines (Committee to Update the
1999 Guidelines for Coronary Artery Bypass Graft Surgery).
Circulation 2004, 110:e340-e437 [http://circ.ahajournals.org/cgi/
reprint/110/14/e340?].
2. Gibbons RJ, Abrams J, Chatterjee K, Daley J, Deedwania PC, Douglas
JS, Ferguson TB Jr, Fihn SD, Fraker TD Jr, Gardin JM, O'Rourke RA,
Pasternak RC, Williams SV, Gibbons RJ, Alpert JS, Antman EM,
Hiratzka LF, Fuster V, Faxon DP, Gregoratos G, Jacobs AK, Smith SC
Jr: ACC/AHA 2002 Guideline update for the management of
patients with chronic stable angina: a report of the American
College of Cardiology/American Heart Association Task
Force on Practice Guidelines (Committee to Update the
1999 Guidelines for the Management of Patients with
Chronic Stable Angina). [http://www.acc.org/qualityandscience/
clinical/guidelines/stable/stable_clean.pdf].
3. Tennant A: Quality of life–a measure too far? Ann Rheum Dis
1995, 54:439-440.
4. Holmes S: Assessing the quality of life-reality or impossible
dream? A discussion paper. Int J Nurs Stud 2005, 42:493-501.
5. Wilson IB, Cleary PD: Linking clinical variables with health-
related quality of life. A conceptual model of patient out-
comes. JAMA 1995, 273:59-65.
6. Janz NK, Janevic MR, Dodge JA, Fingerlin TE, Schork MA, Mosca LJ,
Clark NM: Factors influencing quality of life in older women
with heart disease. Med Care 2001, 39:588-598.
7. Hofer S, Benzer W, Alber H, Ruttmann E, Kopp M, Schussler G,
Doering S: Determinants of health-related quality of life in
coronary artery disease patients: a prospective study gener-
ating a structural equation model. Psychosomatics 2005,
46:212-223.
Additional file 1
Cross-lagged model with Global Quality of Life (gQOL) displaying
unstandardized and standardized estimates, together with correlation
coefficients between error variances.
Click here for file
[http://www.biomedcentral.com/content/supplementary/1477-
7525-5-27-S1.pdf]
Additional file 2
Simultaneous reciprocal effects model with Global Quality of Life (gQOL)
displaying unstandardized and standardized estimates.
Click here for file
[http://www.biomedcentral.com/content/supplementary/1477-
7525-5-27-S2.pdf]
Additional file 3
Cross-lagged model with Quality of Life Scale – Norwegian version
(QOLS-N) displaying unstandardized and standardized estimates,
together with correlation coefficients between error variances.
Click here for file
[http://www.biomedcentral.com/content/supplementary/1477-
7525-5-27-S3.pdf]
Additional file 4
Simultaneous reciprocal effects model with Quality of Life Scale – Norwe-
gian version (QOLS-N) displaying unstandardized and standardized esti-
mates.
Click here for file
[http://www.biomedcentral.com/content/supplementary/1477-
7525-5-27-S4.pdf]
Page 9
Publish with BioMed Central and every
scientist can read your work free of charge
"BioMed Central will be the most significant development for
disseminating the results of biomedical research in our lifetime."
Sir Paul Nurse, Cancer Research UK
Your research papers will be:
available free of charge to the entire biomedical community
peer reviewed and published immediately upon acceptance
cited in PubMed and archived on PubMed Central
yours — you keep the copyright
Submit your manuscript here:
http://www.biomedcentral.com/info/publishing_adv.asp
BioMedcentral
Health and Quality of Life Outcomes 2007, 5:27 http://www.hqlo.com/content/5/1/27
Page 10 of 10
(page number not for citation purposes)
8. Arnold R, Ranchor AV, Sanderman R, Kempen GIJM, Ormel J, Suur-
meijer TPBM: The relative contribution of domains of quality
of life to overall quality of life for different chronic diseases.
Qual Life Res 2004, 13:883-896.
9. Wettergren L, Bjorkholm M, Axdorph U, Langius-Eklof A: Determi-
nants of health-related quality of life in long-term survivors
of Hodgkin's lymphoma. Qual Life Res 2004, 13:1369-1379.
10. Sousa KH, Holzemer WL, Henry SB, Slaughter R: Dimensions of
health-related quality of life in persons living with HIV dis-
ease. J Adv Nurs 1999, 29:178-187.
11. Kelli L, Heller DA, Ahern FM, Gold CH: Relationship of health-
related quality of life to health care utilization and mortality
among older adults. Aging Clin Exp Res 2002, 14:499-508.
12. Diener E: Subjective well-being. Psychol Bull 1984, 95:542-575.
13. Cummins RA: Normative Life Satisfaction: Measurement
Issues and a Homeostatic Model. Soc Ind Res 2003, 64:225-256.
14. Brief AP, Butcher AH, George JM, Link KE: Integrating bottom-up
and top-down theories of subjective well-being: The case of
health. J Pers Soc Psychol 1993, 64:646-653.
15. Lance CE, Mallard AG, Michalos AC: Tests of the causal direc-
tions of global-life facet satisfaction relationships. Soc Ind Res
1995, 34:69-92.
16. Mathisen L, Hol PK, Lingaas PS, Lundblad R, Rein KA, Tonnessen TI,
Mork BE, Svennevig JL, Wahl AK, Hanestad BR, Fosse E: Patient
reported outcome after randomization to on-pump versus
off-pump coronary artery surgery. Ann Thorac Surg 2005,
79:1584-1589.
17. Campeau L: Grading of angina pectoris. Circulation 1976,
54:522-523.
18. Wahl AK, Rustoen T, Hanestad BR, Lerdal A, Moum T: Quality of
life in the general Norwegian population, measured by the
Quality of Life Scale (QOLS-N). Qual Life Res 2004,
13:1001-1009.
19. Holmen J, Midthjell K: The North Trøndelag Survey 1984–86 Norwegian
Institute for Public Health (SIFF): Report no.4; 1990.
20. Burckhardt C, Anderson K: The Quality of Life Scale (QOLS):
Reliability, validity, and utilization. Health Qual Life Outcomes
2003, 1:60.
21. Wahl A, Burckhardt C, Wiklund I, Hanestad BR: The Norwegian
version of the Quality of Life Scale (QOLS-N). Scand J Caring
Sci 1998, 12:215-222.
22. Loge JH, Kaasa S: Short form 36 (SF-36) health survey: norma-
tive data from the general Norwegian population. Scand J Soc
Med 1998, 26:250-258.
23. Ware JE, Snow KK, Kosinski M, Gandek B: SF-36 Health Survey: Man-
ual and interpretation guide Boston: The Health Institute; 1993.
24. Kline RB: Principles and practice of structural equation modeling 2nd edi-
tion. New York:Guilford; 2005.
25. Finkel SE: Causal analysis with panel data Thousand Oaks, CA: Sage;
1995.
26. Zapf D, Dormann C, Frese M: Longitudinal studies in organiza-
tional stress research: a review of the literature withrefer-
ence to methodological issues. J Occup Health Psychol 1996,
1:145-169.
27. Smith KW, Avis NE, Assmann SF: Distinguishing between quality
of life and health status in quality of life research: a meta-
analysis. Qual Life Res 1999, 8:447-459.
28. Heo S, Moser DK, Riegel B, Hall LA, Christman N: Testing a pub-
lished model of health-related quality of life in heart failure.
J Card Fail 2005, 11:372-379.
29. Tellegen A, Lykken DT, Bouchard TJJ, Wilcox KJ, Segal NL, Rich S:
Personality similarity in twins reared apart and together. J
Pers Soc Psychol 1988, 54:1031-1039.
30. Roysamb E, Tambs K, Reichborn-Kjennerud T, Neale MC, Harris JR:
Happiness and health: environmental and genetic contribu-
tions to the relationship between subjective well-being, per-
ceived health, and somatic illness. J Pers Soc Psychol 2003,
85:1136-1146.
31. DeNeve KM, Cooper H: The happy personality: A meta-analy-
sis of 137 personality traits and subjective well-being. Psychol
Bull 1998, 124:197-229.
32. Motzer SU, Stewart BJ: Sense of coherence as a predictor of
quality of life in persons with coronary heart disease surviv-
ing cardiac arrest. Res Nurs Health 2005, 19:287-298.
33. Rapkin BD, Schwartz CE: Toward a theoretical model of qual-
ity-of-life appraisal: Implications of findings from studies of
response shift. Health Qual Life Outcomes 2004, 2:14.
34. Fayers PM, Machin D: Quality of life: Assessment, analysis and interpreta-
tion Chichester Wiley; 2000.
35. Anderson KL, Burckhardt CS: Conceptualization and measure-
ment of quality of life as an outcome variable for health care
intervention and research. J Adv Nurs 1999, 29:298-306.
Page 10
  • Source
    • "By the term 'social support', we refer mainly to the resources provided by other persons (Cohen and Syme, 1985). It has also been defined as the cognitive appraisal of being 'reliably connected to significant others in a given social environment (Mathisen et al, 2007). Interestingly, it also fits with another identified contributor to overall QOL, social function. "
    [Show abstract] [Hide abstract] ABSTRACT: Background: The definition of health for people with cancer is not focused solely on the physiology of illness and the length of life remaining, but is also concerned with improving the well-being and the quality of the life (QOL) remaining to be lived. This study aimed to identify the constructs most associated with QOL in people with advanced cancer. Methods: Two hundred three persons with recent diagnoses of different advanced cancers were evaluated with 65 variables representing individual and environmental factors, biological factors, symptoms, function, general health perceptions and overall QOL at diagnosis. Three independent stepwise multiple linear regressions identified the most important contributors to overall QOL. R(2) ranking and effect sizes were estimated and averaged by construct. Results: The most important contributor of overall QOL for people recently diagnosed with advanced cancer was social support. It was followed by general health perceptions, energy, social function, psychological function and physical function. Conclusions: We used effect sizes to summarise multiple multivariate linear regressions for a more manageable and clinically interpretable picture. The findings emphasise the importance of incorporating the assessment and treatment of relevant symptoms, functions and social support in people recently diagnosed with advanced cancer as part of their clinical care.
    Full-text · Article · Apr 2013 · British Journal of Cancer
  • Source
    • "Other evaluations of the Wilson–Cleary model have examined such reciprocal effects. For example, Mathisen and colleagues [27] note reciprocal causal effects over time between general health perceptions and overall quality of life. Wilson and Cleary [5] note the possibility of such effects, and other researchers have called for an examination of these bidirectional relationships [7, 9]. "
    [Show abstract] [Hide abstract] ABSTRACT: Purpose: Using the Wilson-Cleary model of patient outcomes as a conceptual framework, the impact of functional status on health-related quality of life (HRQoL) among older adults was examined, including tests of the mediation provided by life-space mobility. Methods: Participants were enrollees in a population-based, longitudinal study of mobility among community-dwelling older adults. Data from four waves of the study equally spaced approximately 18 months apart (baseline, 18, 36, and 54 months) were used for participants who survived at least 1 year beyond the 54-month assessment (n = 677). Autoregressive mediation models using longitudinal data and cross-sectional mediation models using baseline data were evaluated and compared using structural equation modeling. Results: The longitudinal autoregressive models supported the mediating role of life-space mobility and suggested that this effect is larger for the mental component summary score than the physical component summary score of the SF-12. Evidence for a reciprocal relationship over time between functional status, measured by ADL difficulty, and life-space mobility was suggested by modification indices; these model elaborations did not alter the substantive meaning of the mediation effects. Mediated effect estimates from longitudinal autoregressive models were generally larger than those from cross-sectional models, suggesting that mediating relationships would have been missed or were potentially underestimated in cross-sectional models. Conclusions: These results support a mediating role for life-space mobility in the relationship between functional status and HRQoL. Functional status limitations might cause diminished HRQoL in part by limiting mobility. Mobility limitations may precede functional status limitations in addition to being a consequence thereof.
    Full-text · Article · Nov 2012 · Quality of Life Research
  • Source
    • "In contrast, the patients' experience of their health (i.e., HRQOL) is substantially influenced by the fracture. To be a patient with a hip fracture was a strong predictor of worsened physical health in an elderly population , even when known correlates of decreased physical health such as co-morbidity, age, and marital status404142 were adjusted for. This indicates a strong association between a hip fracture and worsened health. "
    [Show abstract] [Hide abstract] ABSTRACT: The long-term effect of hip fracture on health-related quality of life (HRQOL) and global quality of life (GQOL) has not been thoroughly studied in prospective case-control studies. a) to explore whether patients with low-energy hip fracture regain their pre-fracture levels in HRQOL and GQOL compared with changes in age- and sex-matched controls over a two year period; b) to identify predictors of changes in HRQOL and GQOL after two years. We examined 61 patients (mean age = 74 years, SD = 10) and 61 matched controls (mean age = 73 years, SD = 8). The Short Form 36 assessed HRQOL and the Quality of Life Scale assessed GQOL. Paired samples t tests and multiple linear regression analyses were applied. HRQOL decreased significantly between baseline and one-year follow-up in patients with hip fractures, within all the SF-36 domains (p < 0.04), except for social functioning (p = 0.091). There were no significant decreases within the SF-36 domains in the controls. Significantly decreased GQOL scores (p < 0.001) were observed both within patients and within controls between baseline and one-year follow-up. The same pattern persisted between baseline and two-year follow-up, except for the HRQOL domain mental health (p = 0.193). The patients with hip fractures did not regain their HRQOL and GQOL. Worsened physical health after two years was predicted by being a patient with hip fracture (B = -5.8, p < 0.001) and old age (B = -1.0, p = 0.015), while worsened mental health was predicted by co-morbidity (B = -2.2, p = 0.029). No significant predictors of differential changes in GQOL were identified. A hip fracture has a long-term impact on HRQOL and is a strong predictor of worsened physical health. Our data emphasize the importance of preventing hip fracture in the elderly to maintain physical health. This knowledge should be included in decision-making and health care plans.
    Full-text · Article · Sep 2010 · BMC Musculoskeletal Disorders
Show more