From the Department of Geriatric Medicine, Ullevaal Hospital, Oslo
(T.B.W., K.L.); National Institute of Public Health, Community Medicine
Research Unit, Verdal (J.H.); and the Section for Medical Statistics,
University of Oslo (P.L.), Oslo, Norway.
Correspondence to Dr Torgeir Bruun Wyller, Nordstrandveien 9, N-1170 Oslo, Norway. E-mail t.b.wyller{at}ioks.uio.no
Methods—Data on all stroke patients (n=1417) and a random
subsample of stroke-free individuals of similar age (n=1439) were
collected from the Nord-Trøndelag Health Survey, a cross-sectional
study of 74 977 persons. Based on a two-sample factor analysis
model, scores of SWB were calculated, and variables explaining SWB
were studied in a regression model.
Results—Four items were a priori believed to measure SWB as
a latent variable ("satisfaction," "strength,"
"calmness," and "cheerfulness"). This was confirmed by factor
analysis. The reliability of these items (the proportion of the
variance of the items that can be explained by the common factor) was
between .42 and .53. Regression analyses showed a significant
effect of having had a stroke, gender (lower SWB in men), age
(increasing SWB with increasing age), perceived general health,
nervousness, loneliness, sleep problems, social support, and use of
analgesics. There was no statistical interaction between these
variables and having had a stroke.
Conclusions—Higher SWB after stroke relates to female gender,
older age, good general and mental health, and a firm social network.
The concept of quality of life (QOL) has gained increasing popularity
among researchers studying the consequences of stroke and other chronic
diseases for the individual. However, important criticism has been
raised against the use of QOL in medical
research.1 The investigators often fail to
specify on what aspect of life they focus and tend not to choose their
instrument on the basis of such
considerations.1 2 3 In its original meaning, QOL
was clearly related to subjectively perceived emotions, eg,
satisfaction and happiness.4 Over the years, the
concept seems to have been expanded by inclusion of aspects of
physical, psychological, functional and social
health.5 In our opinion, the theory supporting
this wider QOL concept is thin.1 3 6 QOL measured
in this way will to a considerable extent overlap with other
measurements, such as activities of daily living (ADL) or motor
function.7 Because the content of the QOL concept
varies widely, several authors have proposed looking for less ambiguous
terms.6 7 In this study the aim was to focus on
emotions such as satisfaction and happiness, and we consider the term
"subjective well-being" (SWB)8 9 more
appropriate for these aspects.10 11
Concepts like SWB and QOL are not easily measurable. Accordingly, they
are treated as latent variables, ie, variables which are not
directly measured but estimated by a common factor model via their
relation to items that can be measured. The proportion of the variance
of the items that can be explained by the common factor is denoted as
the reliability of the items.
Several authors have compared the SWB (or QOL) of stroke patients with
that of nonstroke subjects. With few
exceptions,12 they have found a considerably
lower SWB among stroke patients.13 14 15 16 Our
knowledge is more fragmentary, however, as to which aspects of the
stroke patients' lives are the most important determinants of SWB.
Studies dealing with this issue10 12 14 15 16 17 18 19 20 have
included various sets of possible explanatory variables. The
dependent variables have been operationalized in various ways, and
a majority of the studies have rather small samples and thus have low
statistical power.10 12 14 16 17 18 20
The present work had three objectives. First, to study SWB as a
latent variable; second, to assess the reliability of the items
related to the latent variable; and third, to study variables
explaining SWB in a large, population-based sample of stroke patients
and stroke-free individuals. Four items were a priori believed to
reflect SWB in the stroke as well as the nonstroke populations. These
were "satisfaction," "strength," "calmness," and
"cheerfulness" (Table 1
Screening Procedure
The first questionnaire comprised 31 questions dealing with perceived
health, functional abilities, contact with the health care system,
general well-being, working conditions, and chronic diseases
(hypertension, diabetes, myocardial infarction, angina pectoris, and
stroke). The second questionnaire consisted of 42 questions regarding
lifestyle, housing, educational level, working conditions, medical
symptoms, social support, and well-being. All questions had fixed
answer categories; the number of possible responses varied from two to
six. The exact wording of the questions and the answer categories has
been described elsewhere.22
The validity of self-reported stroke based on the same question as used
in this survey has been investigated
previously.23 This showed 20% false-positive and
3% false-negative self-reports compared with stroke diagnosed
according to the WHO criteria,24 the coefficient
of agreement (
Study Population
Statistical Methods
Using this factor model, we calculated the factor score. We named the
factor score SWB and used this score as the dependent variable in
subsequent analyses. The ordinality of multilevel explanatory
variables was checked, and adjacent categories were collapsed where
appropriate.
The computer program EQS29 was used to perform
confirmatory factor analysis, and BMDP30
was used to calculate the factor scores and perform the linear
regression analyses.
The reliability of the four items "satisfaction," "strength,"
"calmness," and "cheerfulness," was .53, .43, .42, and .49,
respectively.
For further analyses we used the SWB scores as the
dependent variable. The explanatory variables applied were age,
gender, and several indicators of general and mental health, functional
capacity, the respondent's social network, and morbidity. We fitted
linear regression models for each of the potential explanatory
variables. Each model contained two covariates: the grouping
variable (stroke/no stroke) and one of the other explanatory
variables (Table 2
To assess whether the role of the explanatory variables differed
between stroke patients and controls, interaction terms were
introduced. We fitted regression models containing the variables
shown in Table 2
The findings confirmed that stroke survivors have significantly lower
SWB than control subjects of similar age.13 14 15 16
However, in another recent study the SWB of stroke patients was
reported to be similar to that of a normative
population.12 That study was, however, carried
out in a highly selected sample.
The variables that were found to be independently associated
with SWB in this study (Table 2
An effect of the respondent's social network on SWB has been
found in most studies in which such variables have been included,
in stroke survivors12 14 16 as well as in other
groups.9 Also, perceived health has been found to
correlate highly with SWB in numerous studies,9
but this relation has, to our knowledge, not been studied specifically
in stroke survivors.
Most stroke studies focus on some measurement of disability, mainly
performance regarding ADL.12 14 16 Our
analysis, including mobility as a typical ADL task, indicated
that this variable, though statistically significant in bivariate
analysis, has no independent effect on SWB when perceived
general health and the respondent's social network are also taken into
account. In a previous study of stroke
patients,10 which included formal assessments of
various impairments and disabilities, we found that arm motor
impairments had a stronger impact than ADL capacity on SWB.
In conclusion, in this population-based sample of stroke patients we
found that SWB is considerably lower than in reference subjects of
similar age. In stroke and nonstroke subjects, high SWB is mainly
explained by female gender, older age, good general and mental health,
and a firm social network. In order to understand, and perhaps improve,
the SWB of individual stroke survivors, the social context as well as
the patient's subjective perception of the situation need to be taken
into consideration.
Received June 2, 1997;
revision received November 6, 1997;
accepted November 7, 1997.
2.
Fletcher A, Gore S, Jones D, Fitzpatrick R,
Spiegelhalter D, Cox D. Quality of life measures in health care, II:
design, analysis, and interpretation. BMJ. 1992;305:1145–1148.
3.
Gill TM, Feinstein AR. A critical appraisal of the
quality of quality-of-life measurements. JAMA. 1994;272:619–626.
4.
Campbell A, Converse PE, Rodgers WL. The Quality
of American Life: Perceptions, Evaluations, and Satisfactions. New
York, NY: Sage; 1976.
5.
de Haan R, Aaronsen N, Limburg M, Hewer RL, van Crewel
H. Measuring quality of life in stroke. Stroke. 1993;24:320–327.
6.
Wade DT. Measuring the quality of life in acute
stroke. In: Amery WK, Bousser M-G, Rose FC, eds. Clinical Trial
Methodology in Stroke. London, UK: Baillière Tindall;
1989:166–176.
7.
Nydegger CN. Measuring morale and life satisfaction.
In: Fry CL, Keith J, eds. New Methods for Old-AgeResearch:
Strategies for Studying Diversity. South Hadley, Mass: Bergin &
Garvey; 1986:213–230.
8.
Diener E. Subjective well-being. Psychol
Bull. 1984;95:542–575.[Medline]
[Order article via Infotrieve]
9.
Fuhrer MJ. Subjective well-being: implications for
medical rehabilitation outcomes and models of disablement.
Am J Phys Med Rehabil. 1994;73:358–364.[Medline]
[Order article via Infotrieve]
10.
Wyller TB, Sveen U, Sødring KM, Pettersen AM,
Bautz-Holter E. Subjective well-being one year after stroke. Clin
Rehabil. 1997;11:139–145.
11.
Wyller TB. Disability models in geriatrics.
Disabil Rehabil. 1997;19:480–483.[Medline]
[Order article via Infotrieve]
12.
King RB. Quality of life after stroke.
Stroke. 1996;27:1467–1472.
13.
Ahlsiö B, Britton M, Murray V, Theorell T.
Disablement and quality of life after stroke. Stroke. 1984;15:886–890.
14.
Viitanen M, Fugl-Meyer KS, Bernspång B, Fugl-Meyer AR.
Life satisfaction in long-term survivors after stroke. Scand J
Rehabil Med. 1988;20:17–24.[Medline]
[Order article via Infotrieve]
15.
Ebrahim S, Barer D, Nouri F. Use of the
Nottingham Health Profile with patients after a stroke. J
Epidemiol Community Health. 1986;40:166–169.
16.
Åström M, Asplund K, Åström T.
Psychosocial function and life satisfaction after stroke.
Stroke. 1992;23:527–531.
17.
Niemi M-L, Laaksonen R, Kotila M, Waltimo O. Quality of
life 4 years after stroke. Stroke. 1988;19:1101–1107.
18.
Nydevik I, Hulter-Åsberg K. Sickness impact after
stroke: a 3-year follow-up. Scand J Prim Health Care. 1992;10:284–289.[Medline]
[Order article via Infotrieve]
19.
Johansson BB, Jadbäck G, Norrving B, Widner H,
Wiklund I. Evaluation of long-term functional status in first-ever
stroke patients in a defined population. Scand J Rehabil
Med. 1992;(suppl 26):105–114.
20.
Granger CV, Cotter AC, Hamilton BB, Fiedler RC.
Functional assessment scales: a study of persons after stroke.
Arch Phys Med Rehabil. 1993;74:133–138.[Medline]
[Order article via Infotrieve]
21.
Wyller TB, Bautz-Holter E, Holmen J. Prevalence of
stroke and stroke-related disability in North Trøndelag county,
Norway. Cerebrovasc Dis. 1994;4:421–427.
22.
Holmen J, Midthjell K, Bjartveit K, Hjort PF,
Lund-Larsen PG, Moum T, Næss S, Waaler HT. The Nord-Trøndelag
Health Survey 1984–86: Purpose, Background and Methods. Participation,
Non-participation and Frequency Distributions. Verdal, Norway:
Statens Institutt for Folkehelse; 1990.
23.
Wyller TB, Ranhoff AH, Bautz-Holter E. Validity of
questionnaire information from old people on previous cerebral stroke.
Cerebrovasc Dis. 1994;4:57–58.
24.
Hatano S. Experience from a multicentre stroke
register: a preliminary report. Bull WHO. 1976;54:541–553.[Medline]
[Order article via Infotrieve]
25.
Moum T. Yea-saying and mood-of-the-day effects in
self-reported quality of life. Soc Indicat Res. 1988;20:117–139.
26.
Moum T, Næss S, Sørensen T, Tambs K, Holmen J.
Hypertension labelling, life events and psychological well-being.
Psychol Med. 1990;20:635–646.[Medline]
[Order article via Infotrieve]
27.
Næss S, Midthjell K, Moum T, Sørensen T, Tambs K.
Diabetes mellitus and psychological well-being. Results of the
Nord-Trøndelag Health Survey. Scand J Soc Med. 1995;23:179–188.[Medline]
[Order article via Infotrieve]
28.
Bjordal K, Mastekaasa A, Kaasa S. Self-reported
satisfaction with life and physical health in long-term cancer
survivors and a matched control group. Oral Oncol Eur J
Cancer. 1995;31B:340–345.
29.
Dunn G, Everitt B, Pickles A. Modelling
Covariances and Latent Variables Using EQS. London,
UK: Chapman & Hall; 1993.
30.
Dixon WJ. BMDP Statistical Software.
Berkeley, Calif: University of California Press; 1992.
31.
Mastekaasa A, Moum T. The perceived quality of
life in Norway: regional variations and contextual effects. Soc
Indicat Res. 1984;14:385–419.
© 1998 American Heart Association, Inc.
Original Contributions
Correlates of Subjective Well-being in Stroke Patients
Abstract
Top
Abstract
Introduction
Subjects and Methods
Results
Discussion
References
Background and Purpose—Data on
survival and functioning after stroke needs to be supplemented by
measures emphasizing the patients' subjective perception. We studied
(1) subjective well-being (SWB) as a latent variable in a
common-factor model with four items, (2) the reliability of these four
items, and (3) variables related to SWB in stroke
patients.
Key Words: attitude to health • quality of life • stroke • social support
Introduction
Top
Abstract
Introduction
Subjects and Methods
Results
Discussion
References
Until a few years
ago, stroke research was largely focused on survival. A need to improve
the quality of the lives saved is now being increasingly acknowledged.
Identification of variables related to life satisfaction in stroke
patients is a prerequisite for such efforts. 
).
View this table:
[in a new window]
Table 1. Scoring on Each Item According To Group
Subjects and Methods
Top
Abstract
Introduction
Subjects and Methods
Results
Discussion
References
Screened Population
Nord-Trøndelag county is situated in the middle of
Norway. About 127 000 of Norway's 4.2 million inhabitants live there,
mostly in rural areas. All inhabitants aged 
20 years (85 100
persons) were invited to participate in the Nord-Trøndelag Health
Survey21 22 during the period from January 1984
to February 1986, and 74 977 persons (88.1%) attended. The attendance
was 76.7% in the age group 20 to 29 years, increased steadily to a
maximum of 93.8% in the age group 50 to 69 years, and then fell to
89.1% in the age group 70 to 79 years, and to 71.6% among persons
aged 80 years. Causes for nonattendance have been studied in
detail.21 22 No differences in health status or
mortality were found between the younger nonattendants and the
attendants of similar age. The nonattendants in these age groups were
mainly men, stating "lack of time" as the reason for not attending.
In the age group 70 to 79 years, however, 25% of the nonattendants
stated "health problems" as the reason for not participating, and
among nonattendants aged 80 years, the corresponding percentage was
35. Among these older nonattendants, the mortality during the screening
period was approximately fourfold that of attendants of similar age.
After the screening period, the mortality among the older nonattendants
fell to approximately twice that of attendants of the same age. This is
thought to reflect the fact that some did not attend the screening
because of terminal illness.21 22
A postal questionnaire was distributed in advance and collected
at attendance. The screening nurses verified that all the questions had
been answered and assisted the respondents in filling out the
questionnaire if help was needed. The participants were asked to return
a second questionnaire by mail, and 64 543 (86.1% did so).
Furthermore, the screening procedure included a chest x-ray and
measurements of height, weight, and blood pressure. Participants who
were >40 years old or suffered from diabetes also had their blood
glucose (nonfasting) measured. 
) being .79. The items reflecting SWB have been used,
and to some extent cross-validated, in studies of the general
population25 and of patients with
hypertension,26 diabetes,27
and cancer.28
Of those screened, 1417 persons reported having suffered a
stroke previously. Their mean age was 71.6 years (SD=12.6), and 48.9%
were men. Details about these subjects have been published
elsewhere.21 From the remaining records, we
collected a stratified random reference sample of 1439 individuals
(46.3% men), with the same age distribution as the stroke
patients.
A two-sample confirmatory factor analysis was performed,
with the stroke and the nonstroke subjects as the two groups. Four
items (Table 1
) were used as indicators of SWB as a latent
variable. The variables were assumed to have multinormal
distributions, and maximum likelihood estimates of the factor model
parameters were calculated. Since the variables are
ordinal rather than continuous, factor analysis methodology
that takes the ordinality of the data into account was also applied.
Similar results were obtained, and methods based on multinormality were
used for further analyses.
Results
Top
Abstract
Introduction
Subjects and Methods
Results
Discussion
References
The mean scores of the variables are given in Table 1. A
two-sample factor model was fitted with use of the four variables.
A factor model that constrained the factor loadings to be equal between
the samples, but to vary over the items, explained the data adequately
(Figure). A higher SWB score reflects a better well-being. The
estimated difference in mean SWB score between stroke and nonstroke
persons was -0.44 (95% confidence interval, -0.54 to -0.34;
P=.0004). We also fitted factor models that in addition to
the four SWB items included items mainly tapping mental health. This
resulted, however, in a significant loss of fit of the one-factor
model.
View larger version (16K):
[in a new window]
Figure 1. Two-sample factor model for subjective well-being (SWB).
e indicates residual error. Model fit: 
2=15.6,
df=9; P=.07.). Variables
with P<.10 were then considered for inclusion in a
subsequent multivariate analysis. Gender and
age were included in this model regardless of their statistical
significance. A model with 12 explanatory variables (Table 2)
explained 50.3% of the variance in the SWB score, and none of the
remaining variables contributed significantly to the model. We
found statistically significant effects of having had a stroke, gender
(higher SWB in women), age (increasing SWB with increasing age),
perceived general health, nervousness, loneliness, sleep problems,
social support, and use of analgesics. The fit of the model was
established by inspection of standardized residuals and Cook distances,
ie, different plots illustrating to which degree the observed data
deviate from the estimated.
View this table:
[in a new window]
Table 2. Linear Regression Models of Subjective Well-being
Factor Score plus an interaction term consisting of the grouping
variable (stroke/no stroke) and one of the other explanatory
variables. In neither of these models did the interaction term
reach statistical significance (results not shown). This indicates that
the effect of the explanatory variables on SWB is similar for
stroke patients and nonstroke subjects.
Discussion
Top
Abstract
Introduction
Subjects and Methods
Results
Discussion
References
The concurrent validity of a measure of SWB is difficult to
establish, because no accepted gold standard exists. One has to rely on
the construct validity, which in this study was indeed supported by the
confirmatory factor analysis (Figure). Some authors include in
their measurement of SWB also items mainly tapping mental health. Our
results indicate, however, that SWB and mental health, though
interrelated, are separate concepts. ) can be roughly summed up as gender,
age, physical and mental health, and the social network. The finding of
a higher SWB among women is surprising, because most authors report
that SWB is either gender independent31 or lower
in females.8 10 The gender difference on SWB was
slight but statistically significant in the full
multivariate analysis in our sample. In this
model we also found a positive effect of increasing age on SWB. With
our cross-sectional study design, we cannot assess whether this is an
effect of age or cohort. Other authors have found that the effect of
age is dependent on whether the dependent variable (eg, QOL or SWB)
is weighted toward satisfaction or happiness; self-reported happiness
seems to decline whereas satisfaction may increase with
age.31
Acknowledgments
The Nord-Trøndelag Health Survey was carried out by the
National Health Screening Service of Norway, and we are grateful for
being allowed to have access to the data.
References
Top
Abstract
Introduction
Subjects and Methods
Results
Discussion
References
1.
Mor V, Guadagnoli E. Quality of life measurement:
a psychometric tower of Babel. J Clin Epidemiol. 1988;41:1055–1058.[Medline]
[Order article via Infotrieve]
This article has been cited by other articles:
 |
|
 |
|
  V. de Albuquerque Citero, P. B. de Araujo Andreoli, L. A. Nogueira-Martins, and S. B. Andreoli New Potential Clinical Indicators of Consultation-Liaison Psychiatry's Effectiveness in Brazilian General Hospitals Psychosomatics, February 1, 2008; 49(1): 29 - 38. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J. W. Gargano, M. J. Reeves, and for the Paul Coverdell National Acute Stroke Regis Sex Differences in Stroke Recovery and Stroke-Specific Quality of Life: Results From a Statewide Stroke Registry Stroke, September 1, 2007; 38(9): 2541 - 2548. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 T. Kikusui, J. T Winslow, and Y. Mori Social buffering: relief from stress and anxiety Phil Trans R Soc B, December 29, 2006; 361(1476): 2215 - 2228. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 C Donnellan, D Hevey, A Hickey, and D O'Neill Defining and quantifying coping strategies after stroke: a review. J. Neurol. Neurosurg. Psychiatry, November 1, 2006; 77(11): 1208 - 1218. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 T Hillen, S Davies, A G Rudd, T Kieselbach, and C D Wolfe Self ratings of health predict functional outcome and recurrence free survival after stroke J Epidemiol Community Health, December 1, 2003; 57(12): 960 - 966. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
W. M. Hopman and J. Verner Quality of Life During and After Inpatient Stroke Rehabilitation Stroke, March 1, 2003; 34(3): 801 - 805. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
Y Ben-Shlomo, L Camfield, and T Warner What are the determinants of quality of life in people with cervical dystonia? J. Neurol. Neurosurg. Psychiatry, May 1, 2002; 72(5): 608 - 614. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
P. Clarke, V. Marshall, S. E. Black, and A. Colantonio Well-Being After Stroke in Canadian Seniors: Findings From the Canadian Study of Health and Aging Stroke, April 1, 2002; 33(4): 1016 - 1021. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J. Carod-Artal, J. A. Egido, J. L. Gonzalez, and E. Varela de Seijas Quality of Life Among Stroke Survivors Evaluated 1 Year After Stroke : Experience of a Stroke Unit Stroke, December 1, 2000; 31(12): 2995 - 3000. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 H C Hanger, B Fogarty, T J Wilkinson, and R Sainsbury Stroke patients' views on stroke outcomes: death versus disability Clinical Rehabilitation, April 1, 2000; 14(4): 417 - 424. [Abstract] [PDF]
|
|
|
|
|
|
E. K. Pisa, M. Wiesmann, and U. Missler Serum S-100 Protein in Stroke and Cardiac Surgery • Response Stroke, May 1, 1999; 30 (5): 1153 - 1154. [Full Text] [PDF]
|
|
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||