Full text of DOI:10.1348/135910700169008

British Journal of Health Psychology
Printed in Great Britain
(2000), 5, 395–411
395
© 2000 The British Psychological Society
Self-rated menopausal status and quality of life in
women aged 40–63 years
Pamela A. Jacobs*, Michael E. Hyland and Ann Ley
Department of Psychology, University of Plymouth, UK
. To determine whether quality of life (QOL) changes during the menopause
Objectives
as a function of menopausal status and other medical and lifestyle variables.
. A postal questionnaire sent to three different samples of women.
Design
. A total of 1188 questionnaires were received from (a) two samples recruited
Method
from two Family Health Service Association (FHSA) lists and (b) one sample recruited
through an advertisement in a women’s magazine. The questionnairesconsisted of (a) a
seven domain, 48-item, condition-speciŽc QOL questionnairewhich was developed for
this study (MQOL), (b) a single item global QOL questionnaire(GQOL), (c) questions
about medical history, (d) questions about work status, (e) questions assessing
menopausal status using two different techniques.
. Both the MQOL and GQOL indicated a relationship with menopausal status.
Results
GQOL and MQOL showed a U-shaped relationship with menopausal status, with
lowest scores associated with the middle of the menopause. However, domain scores of
Sleep and Energy failed to reach the levels reported by women who perceive themselves
to be pre-menopausal, and domain scores of Symptom Impact and Social Interaction
indicate steady decline during the menopausal transition. Women who experiencedthe
menopause long ago reported the highest GQOL, feelings and cognitiondomain scores.
Medical history and work outside the home play an important role in determining
MQOL—women who had undergone hysterectomy, those who had tried but
discontinued HRT while still in the middle of the climacteric, and those with greater
co-morbidity had poorer QOL. Those who worked outside the home reported better
MQOL, and those recruited through the magazine reported poorer QOL.
. QOL is affected by the menopause, but the way it is affected depends on
Conclusions
the measure of QOL used. QOL during the menopause is also affected by medical and
lifestyle variables. QOL during the menopause is a complex interaction of several
different kinds of variable.
The ‘menopause’ has multiple meanings depending on the context in which the term is
used. First, in a medical context, the term is used to refer to the cessation of a women’s
menstrual cycle and is associated with a reduction in oestrogen levels. Typically a
diagnosis of having reached the menopause is applied retrospectively. Second, in popular
*Requests for reprints should be addressed to Pamela A. Jacobs, Department of Psychology, University of Plymouth,
Plymouth PL4 8AA, UK (e-mail. P.Jacobs@plymouth.ac.uk).

Pamela A. Jacobs
396
et al.
literature, e.g., women’s magazines, the term menopause is associated with the whole of
the climacteric, from when a woman begins to notice changes in her menstrual cycle or
she
experiences symptoms that
associates with the menopause. A third use of the term
menopause is found in the social science and feminist literature, (e.g., Hunt, 1994). Here
the menopause is deŽned in much broader terms as a natural transitional phase in
women’s lives that recognizes the biological changes but also gives prominence to the
social and cultural embodiment of women’s lives as they experience the climacteric.
Irrespective of the perspective (which is seldom deŽned), symptoms reported to occur
during the menopause include hot ushes and night sweats, vaginal dryness, sleep
disturbance, anxiety, depression, irritability, fatigue, painful joints, headache, memory
and concentration deŽcits, and sexual problems (Derman, Dawood, & Stone, 1995;
et al
Greene & Cooke, 1980; Hunter, 1992; Pearce, Hawton, & Blake, 1995; Wiklund
.,
1992). However, declining oestrogen levels are unlikely to play a role in the aetiology of all
these symptoms. Depression and poor psychological well-being are often reported during
the climacteric, (Studd, Watson, & Montgomery, 1990) but these symptoms are more
commonly associated with socio-demographic and psychosocial factors, previous experience
of depression and lifestyle changes which coincide with the menopause (Bebbington, 1998;
Dennerstein, Smith, & Morse, 1994; Green, 1992; McKinlay, McKinlay, & Brambilla,
1987). A reporting bias may explain the observed correlation between mental health scores
et al
., 1995) as negative affectivity
and the reporting of menopausal symptoms (O’Connor
leads to greater awareness and reporting of physical symptomatology (Abraham, Llewellyn-
Jones, & Perz, 1995; Watson & Pennebaker, 1989).
The relationship between quality of life (QOL) and menopausal status is controversial.
Studies using the Nottingham Health ProŽle (NHP) have shown a positive relationship
between menopausal status and QOL when comparing pre-menopausal to peri-meno-
pausal and menopausal women in the 45–55 year age range (Ledesert, Ringa, & Breart,
1995; Oldenhave, Jaszmeann, Haspels, & Everaerd, 1993). Women experiencing severe
climacteric symptoms reported lower QOL when compared with those experiencing only
et
mild symptoms as measured by a time trade-off and symptoms impact rating scale (Daly
al
., 1993). QOL differences were also reported among peri-menopausal women by Hunter
(1992) using the Women’s Health Questionnaire. However, Dennerstein (1996) claims that
well-being, a term often used to deŽne QOL measures (Ormel, Lindenberg, Stevennk, &
Vonkorff, 1997), is not associated with menopausal status but is associated with psycho-
social, lifestyle and current health status. Doyal (1994) suggests that women who work have
greater self-esteem, social contacts and support networks which may account for differences
in QOL between women who work and those who do not. Four factors may be contributing
to this diversity of Žndings. First, there is a lack of common understanding of QOL (Hunt,
1997). Second, a variety of instruments are used including simple checklists of symptoms,
measures of distress, measures of well-being, measures of physical and psychological health,
and measures of satisfaction. Few attempts have been made to develop QOL instruments
for research into women’s health that are condition-speciŽc and based upon women’s
experiences. Third, different samples included women attending menopausal clinics,
attending routine ovarian screening, opportunist samples, and studies using small popula-
tion samples. Fourth, there are different ways of determining menopausal status.
Hormonal intervention studies claim to show that hormone replacement therapy
(HRT) can improve QOL in menopausal women. In one study women taking HRT for 4

Menopause and quality of life
397
months reported signiŽcant improvements in (a) sleep, emotions and energy subsections
of the NHP, (b) well-being, anxiety, vitality, self-control, depression and health subscales
of the Psychological General Well-Being Index, and (c) tension and satisfaction with life
et al
subscales of the Feelings Adjective Checklist (Wiklund
., 1992). In another study,
women reporting menopausal symptoms showed an overall improvement in QOL
following 4 months’ treatment with HRT using the Kuppermen Menopausal Index
(which measures severity of symptoms) and the Green index (which measures feelings and
et al
symptoms) (Derman
., 1995).
Despite these reported advantages of HRT, the use of HRT remains low with estimates
of between 7% and 22% of peri-menopausal and menopausal women receiving HRT in
the UK at any one time (Oddens, Baulet, Lehert, & Visser 1994). McCleery and Gebbie
(1994) cite a 1992 National OfŽce of Populations survey of 1000 women in which 25%
of women who started HRT discontinued it within 6 months. Few have compared QOL
in former users with either current or never used, nor with pre- or post-menopausal
women who may be using HRT prophylactically.
This paper provides comparative data for QOL of women in the age range 40–63 years.
The majority of previous studies fail to include women over the age of 55 years, so that
little information exists about the end of the climacteric at a time when women may be
still coming to terms with physical, social, cultural and psychological changes associated
with the menopause and ageing. In this study we have chosen to deŽne QOL as ‘the extent
to which the physical, emotional and social aspects of a person’s life are intact’ (Fletcher &
Bulpitt, 1987). The advantage of using this deŽnition is that the biological, social or
psychological variables that may be responsible for the adverse impact upon women’s
lives become a secondary consideration, and the focus is placed squarely upon a woman’s
experiences, her interpretation of the menopause and the impact she believes it has on her
QOL. In order to effect these comparisons, we developed a menopause-speciŽc QOL scale.
Condition- or disease-speciŽc QOL measures provide a more inclusive characterization of
the QOL of a condition or disease (Hyland, 1992), and therefore can be more sensitive to
cross-sectional and longitudinal differences.
Methodological issues
Levels of circulating oestrogens do not correlate reliably with reported symptoms so that
studies need to Žnd alternative criteria for determining menopausal status. Inferring
menopausal status from menstrual cycle information can be unreliable because the
menstrual cycle is affected by the use of HRT and by surgical interventions. Menstrual
cycle classiŽcation on its own also clumps together women who, because of age differences
alone, may experience differing QOL. An alternative strategy is to ask women to self-rate
their menopausal status directly (e.g., whether they are pre-menopausal, have just reached
the menopause, are in the middle of the menopause, at the end of the menopause, or are
et al
past the menopause long ago, Groeneveld
., 1993).
Issues of health unrelated to menopausal status may be an important consideration in a
QOL study whose population covers a 23-year age span. As people get older chronic
health conditions and co-morbidity increase and levels of medication usage rise (Lewis,
Rook, & Schwarzer, 1994). Accordingly, we asked respondents to list medication usage in
the last 2 months in order to evaluate co-morbidity.

Pamela A. Jacobs
398
et al.
Aims of study
(1) To develop a condition-speciŽc questionnaire that examines menopausal QOL in
women.
(2) To examine the impact of employment, age and medical history on menopausal QOL.
(3) To provide cross-sectional information on differences in QOL in a community based
sample of women as a consequence of self-rated changes in menopausal status.
(4) To examine the relationship between the use of HRT and QOL in the middle of the
menopause.
STUDY 1 DEVELOPMENT OF QUESTIONNAIRE
Interviews and focus groups
Thirty-two women were recruited from an advertisement within the university and an advertisement in a
local newspaper. The group was heterogeneousand included unemployed, housewives, cleaners, clerical staff
and a universitylecturer.The averagelength of interviewwas 25 minutes and all but one were audio-recorded
with permission. An additional 29 women attended one of four focus groups. The focus groups lasted for
around 1 hour 15 minutes and were audio-tapedwith permission.The mean age of the total sample (N 61)
5
was 48 years 9 months (
4 years 7 months, range 40–60 years). Of the total sample, 43% currentlyused
SD
5
HRT, 13% were former users and 44% had never used HRT, and 21% had undergone a hysterectomy.
Questionnaire design
The content of interviews and focus groups were analysed using an adaptationof a contact summary (Miles &
Huberman, 1994). Salient themes, concepts and issues were identiŽed using a coding system and summary
transcripts of each interview and focus group were prepared. This system of analysis allows the frequency of
themes to be determined and highlights exemplar statements from participants that represent each theme,
concept and issue. From this analysis, 63 items arranged in seven domains were constructed for use in the
pilot questionnaire,with a 1–6 response format ranging between ‘I am never like this’ and ‘I am always like
this.’ Thirty-four items were negative (i.e., stating some disadvantage) and 29 were positive to minimize
effects of response bias. QOL domains and number of items per domain were as follows: Energy (14 items),
Sleep (5 items), Appetite (4 items), Cognition (8 items), Feelings (13 items), Interactions (10 items),
Symptoms impact (14 items). The number of items chosen for each domain represented the frequency with
which themes occurred during the interviews and focus groups. Personal details were added to the
questionnaire including self-rated menopausal status, menstrual cycle status, (Groeneveld et al., 1993)
employment status, hysterectomy and ovary removal, current medication, and history of HRT use.
Pilot questionnaire
A total of 191 questionnaireswere distributed via pharmacies, the public library, a local health food outlet,
personal contacts and previous participants in focus groups and interviews. Of these, 99 were returned
(response rate 52%) and analysed. The mean age of the sample was 50 years 10 months (
SD
months). Thirty per cent of women were current users of HRT, 8% were former users and 17% had never
used HRT.
5 years 2
5
Analysis for item selection
The psychometric properties of items were examined in terms of item response frequency, item–total
correlationsand item intercorrelations.Fifteenitems were deleted either becauseof negative skew (more than
75% responding to any one response category) low item–total correlations or very high intercorrelations
with other items leaving a 48-item questionnairewith seven domains which was used for the main study. All

Menopause and quality of life
399
but one of the items associated with the food and drink domain were removed by this procedure. Although
food and drink problemswere indicatedby the focus groupsand interviews, the pilot questionnaireindicated
that this was an infrequent problem.
STUDY 2 MAIN STUDY
Participant recruitment
There were three sources of recruitment for the main study. Two of these were random samples of women
aged between 40 and 63 years from patient lists of two adjacent Health Authorities in the South West of
England (Health Authority 1 and Health Authority 2). A total of 750 questionnaires were distributed to
women in Health Authority 1 and 1000 to Health Authority 2. The third source (magazine) was a very
short editorial piece in a national women’s magazine inviting readers to write to the University if they were
interested in taking part in a survey on the menopause. Those requesting participation in the study and
those from the Health Authority lists were sent anonymous postal questionnaires with freepost envelopes
for return.
Measures
The following assessments were taken: the 48-item menopause-speciŽc QOL questionnaire (MQOL;
developed in Study 1), and a global measures of GQOL (H-scale) in which participants were asked to rate
their overall QOL on a 100-point scale (Hyland & Sodergren, 1996). The global H scale has named end
points (‘perfect quality of life’ and ‘might as well be dead’) as well as seven additional quantiŽers along the
scale, these additional quantiŽers having been found to be associated with greater reliability and responder
preference compared with simple endpoint labelling. Questions on menstrual cycle status, the self-rated
menopausal status classiŽcation (Groeneveld et al. 1993), current and former HRT use and biographical/
medical characteristics were also included. An estimate of co-morbidity was determined by classifying
medication used in the past 2 months according to British National Formulary (BNF) categorization, i.e.
centralnervoussystem, gastro-intestinal,cardiovascular,respiratory,malignantdiseaseetc. then countingthe
number of BNF areas that a respondents medication fell under, i.e., one point for each area.
Results
Sample characteristics
A total of 413 responses were received from Health Authority 1 (response rate 55%) and
403 were received from Health Authority 2 (response rate 40.4%). A cut-off date of 2
months was applied to acceptance of data into the Žnal analysis. Following recruitment
via the magazine, 372 questionnaires were received. The total sample consisted of 1188
responses, mean age 51 years 2 months, SD 6 years 1 month. The relationship
5
5
between self-rated menopausal status, age range and HRT use are shown in Table 1. The
level of current use of HRT is 22.5% of total sample, similar to rates of HRTusage found
in other questionnaire studies, e.g., Quine & Rubin (1997). Of the women who were
N
using or had used HRT (
463) 162 (35%) had used more than one brand. Ninety-
5
two women (8%) aged 54 years or more described themselves as ‘just having reached the
menopause’ or ‘in the middle of the menopause’ which is consistent with general
population Žndings that approximately 90% of women will have experienced the
menopause between the ages of 45 and 54 years (Treloar, 1974). Of the total
sample, 207 (17%) women had undergone hysterectomy, a quarter of whom
N
(
52) had had both ovaries removed. General population rates in the UK for
5

Pamela A. Jacobs
400
et al.
hysterectomy are usually given at around one in Žve of all women who reach the age of
60 years (Dickinson & Henriques, 1994) although regional variations exist. Three
hundred and ninety-Žve women (33%) who responded to the survey reported taking
prescribed drugs other than HRT in the last 2 months, and a further 60 women (5%)
reported taking non-prescribed medication such as vitamin supplements and primrose
oils. Of the total sample, 713 women (60%) were employed or doing unpaid voluntary
work outside the home.
Table 1. Self-rated menopausal status and age, and HRT status
Age range (years)
HRT use
Self-rated menopausal status
Never Former Current
N
1
(
1188)
40–46 47–53 54–60 60
used
user
user
5
Not reached menopause
Just reached menopause
In middle of menopause
Reached end of menopause
Reached menopause long ago
193
81
46
2
71
120
167
34
0
9
75
61
58
0
0
8
30
78
254
144
101
66
13
24
75
40
42
7
48
144
29
39
3
28
93
Totals
325
420
213
116
654
196
267
(27%) (35%) (18%) (10%) (55%) (16.5%) (22.5%)
Sample differences
F
p
9.36, < .001), indicates a signiŽcant
A one-way between-group ANOVA (
5
difference in the age of women by source of recruitment. Scheffe’s follow-up indicates
that women who responded to the Magazine advert were younger than respondents from
both of the Health Authority lists. Inspection of frequencies by source of recruitment
indicates that the magazine sample as compared with the two health authority samples
was more likely to be on HRT, more likely to have undergone hysterectomy, and were at
an earlier stage of the menopause.
Inspection of mean scores on the MQOL and GQOL scales revealed that women
recruited through the Magazine advertisement were reporting lower levels of QOL than
the Health Authority groups on all of the menopause-related domains, total MQOL and
GQOL. No signiŽcant differences were found between the two health authority groups.
Factor analysis
A principal components analysis was conducted on the full data and separately for each of
the subsamples. For all these analyses there was a strong one-factor solution which for the
full data set accounted for 35% of the data variance with an eigen value of 16.01. This
one-factor solution appears to be a general severity factor. Because QOL items can form
highly intercorrelated hierarchical structures (Hyland, Bellesis, Thompson, & Kenyon,
1997) we also conducted exploratory factor analyses using oblimin rotation, for the total
sample and for each of the subsamples. In the case of the full data set a meaningful seven-
factor, hierarchical structure was obtained which accounted for 57% of the data variance

Menopause and quality of life
401
with eigen values of 16.01, 3.04, 2.21, 1.72, 1.65, 1.51, 1.39, but this structure was not
stable across subsamples. There was a close but not identical relationship for the full data
set between the factor structure and the domain structure. As we wished to combine the
three subsamples for analyses, we adopted the simpler solution (which was common to all
subsamples) of a single general factor of menopausal QOL. Therefore we aggregated
overall (equally weighted) items to produce a total menopausal QOL score (the MQOL).
In addition we calculated domain subscores based on our original classiŽcation of items to
domains. Factor loadings and their association between items and domains for the Žrst
factor are shown in Appendix 1.
As a measure of internal consistency of the scale, Cronbach’s alpha for the total MQOL
score and for each of the domains was calculated. Results are shown in Table 2.
Table 2. Correlation matrix and Cronbach’s alpha: Domains: Sleep, Energy, Cognition, Symptoms
Impact, Feelings, Social Interaction, Appetite, Total Menopausal Quality of Life (MQOL) and
N
Global Quality of Life (GQOL) (
1188)
5
Symptoms
Social
Sleep Energy Cognition impact Feelings interaction Appetite MQOL
Energy
.33*
Cognition
.35* .44*
Symptoms impact .36* .68**
Feelings .41* .54*
Social interaction .24* .68**
.32*
.65**
.37*
.27*
.67**
.43*
.44*
.67**
.26*
.78**
.40*
.56*
.35*
.84**
.63**
Appetite
MQOL
GQOL
.25* .29*
.51* .83**
.33* .42*
.26*
.82**
.43*
.40*
.29*
.61**
.92
Cronbach’s alpha .75 .69
.76
.75
.91
.90
–
p
p
< .001.
*
5
< .05; **
5
Relationship between menopausal QOL domains and total scores and global QOL scores
The menopausal questionnaire items can be aggregated to produce a measure of overall
QOL deŽcit (MQOL), but in addition, respondents rated their global QOL using a global
H-scale (GQOL). Table 2 also shows the correlations between domain subscale, the total
menopausal score and the global score. The total MQOL and GQOL scores have slightly
less than 40% variance in common, indicating that these are highly related, but never-
theless distinct. All domains were statistically related with GQOL. Intercorrelations
between menopausal domains scores range from .25 to .68. The highest reported
correlations between the domains scores were: energy and social interaction (.68); energy
and symptoms impact (.68); symptoms impact and social interaction (.67); and cognition
and feelings (.65). As a further exploration of the determinants of global QOL we carried
out a multiple regression with GQOL as the dependent variable and all seven domain
scores, age, self-rated menopausal status and HRT use as independent variables (Table 3).
Scores on the feelings domain were the best predictors of global QOL, but impact of
symptoms, appetite, sleep and age also contributed independently to global QOL, with
all of these variables accounting for some 43% of the variance in global QOL score.

Pamela A. Jacobs
402
et al.
Table 3. Multiple regression analysis: GQOL
Partial
correlations
R
R2
F
Predictor
Multiple
.653
Beta
.427
87.73**
Age
.069
.012
.015
.060
.015
.017
.109
.379
.014
.069
.055*
.011
.011
.052*
.018
Self rated menopausal status
HRT usage
Sleep
Energy
Cognition
Symptoms impact
Feelings
Social interaction
Appetite
.018
.130**
.495**
.017
.057*
p
p
* < .05; ** < .0001
Measuring menopausal status
Table 4 shows the relationship between self-rated menopausal status and menstrual cycle
status by age range. Respondents who are currently taking HRT or who have undergone
hysterectomy have been excluded due to the effect of these factors on the menstrual cycle.
Cross-tabulations show that for the age range 40–53 years, 70% of the women who
report having regular menstrual periods in the last 12 months describe themselves as not
having reached the menopause. Of the women who report having no menstrual periods in
the last 12 months, irrespective of age, 97% describe themselves as being in the middle or
at the end of the menopause. Most variability in self-rated menopausal status is associated
with women who have experienced changes in their menstrual periods during the last 12
months.
Medical history, work status and menopausal QOL
In order to examine the impact of employment and health status on menopausal QOL we
calculated a multiple regression analysis. Employment status, co-morbidity, HRT use,
hysterectomy, ovary removal, self-rated menopausal status age and source of recruitment
were used as independent variables. Overall, 39% of the variance was accounted for by
these variables.
Partial correlations (Table 5) indicate that the best unique predictor of MQOL was
employment status, with women working outside the home reporting signiŽcantly
higher levels of MQOL than those not working outside the home. SigniŽcant unique
contributions to MQOL were also found for: source of recruitment, age, co-morbidity,
self-rated menopausal status and HRT use (in order of importance).
Sensitivity of questionnaire scores to self-rated menopausal status
et al
If the severity of menopausal symptoms affects QOL as reported previously (Daly
.,
1993), then responses to the questionnaires should covary with menopausal status.
In order to determine relationships between stages of the menopause and QOL,

Menopause and quality of life
403

Pamela A. Jacobs
404
et al.
Table 5. Multiple regression analysis: MQOL
Partial
correlations
R
R2
F
Predictor
multiple
.624
Beta
.396
70.31*
Source of recruitment
Age
.37
.21
.32*
.28*
Hysterectomy
Oophrectomy
HRT use
Co-morbidity
Employment
.03
.04
.14
.20
.42
.16
.03
.04
.13*
.17*
.41*
.21*
2
2
2
2
2
2
Self-rated menopausal status
2
2
p
* < .00001.
between-group analyses of covariance were calculated to examine effects of self-rated
menopausal status by GQOL, MQOL and each of the seven domain subscale scores, with
source of recruitment, age, hysterectomy, ovary removal, HRT use, co-morbidity and
employment status as covariates. Results indicate that when health differences and
employment status are accounted for, menopausal status interacts with: GQOL
F
p
5
F
p
.00002), MOOL ( (4, 810) 7.09,
5
( (4, 1009) 8.05,
0.00011), symptom
5
5
5
F
F
p
p
.000001), sleep ( (4, 986) 8.32,
5
impact ( (4, 876)
energy ( (4, 966) 3.80,
11.10,
.000001),
.002), feelings
5
5
F
p
F
p
.004), cognition ( (4, 1010) 4.16,
5
5
p
5
5
F
F
p
.00004), social interactions ( (4, 1019) 7.44,
5
( (4, 986) 6.38,
.000007),
5
F
5
5
5
appetite ( (4, 1026) 0.66, n.s.). Scheff’s follow-up analysis indicates that signiŽcant
differences in QOL scores occur between the middle two stages of stages of the menopause
(just reached and in the middle of the menopause) and the pre- and post-menopausal
stages (have not reached, have reached the end, and reached menopause long ago) on
GQOL, MQOL and the subscales of sleep, cognition and feelings.
Table 6 shows the adjusted mean values for each of the self-rated menopausal status
categories on the seven domain subscales as well as MQOL and GQOL. Domain scores
show different patterns of relationship with menopausal status. Overall mean scores
follow an inverted U-shaped pattern across the stages of the menopause with the lowest
scores achieved in the middle of the menopause for each of the domains, MQOL and
GQOL scores. Two patterns can also be distinguished: (a) decline and recovery above pre-
menopausal level (GQOL, cognition and feelings); (b) some initial recovery but does not
reach pre-menopausal levels (MQOL, sleep, social interaction, symptom impact and
energy).
Relationship between QOL, co-morbidity and HRT use
HRT has been found to improve QOL for some women who undergo a natural menopause
et al
et al
(Daly
. 1993; Derman
. 1995; Karlberg, Mattsson, & Wiklund, 1995) but it
seems plausible that HRTuse is more likely to be used by women who suffer QOL deŽcits
during the menopause. Differences in QOL may also provide information on the efŽcacy
of HRT use and provide insight into women who stop taking during the middle of the
menopause. A between-group ANOVA of HRT use by MQOL, GQOL and co-morbidity

Menopause and quality of life
405

Pamela A. Jacobs
406
et al.
was calculated with scores of women in the middle of the menopause only. In order to
identify the impact upon women undergoing a natural menopause, respondents who had
undergone hysterectomy were excluded from the analysis. Results suggest a signiŽcant
F
relationship between QOL and HRT use: MQOL by HRT use ( (2, 232) 10.02,
5
p
F
p
.0001), GQOL by HRTuse ( (2, 198) 4.54,
5
.001), co-morbidity by HRTuse
.0001). Scheffe’s follow-up analysis revealed that signiŽcant
5
5
F
p
5
( (2, 217) 5.97,
5
differences lay between former HRTusers and both current users and those who had never
used HRT on MQOL, GQOL and co-morbidity. Inspection of mean scores by HRT use
indicates that former users of HRT score signiŽcantly lower on both QOL measures and
have signiŽcantly higher co-morbidity scores when compared with current users and
never-used HRT groups.
Discussion
Development of the MQOL scale
We developed a condition-speciŽc QOL questionnaire (MQOL) for menopausal women
on the basis of interviews and focus groups with 61 women. The Žnal 48-item
questionnaire (developed from an earlier 63-item version) covered seven domains;
sleep, energy, cognition, symptom impact, feelings, social interactions and appetite.
All items were highly intercorrelated, and psychometric analysis showed a large Žrst
factor accounting for 35% of the variance, upon which all items loaded. A large Žrst
factor is a common Žnding in QOL measures and is a general severity factor. We also
conducted an exploratory factor analysis using the full data set, with an oblimin rotation.
This produced a seven factor hierarchical structure which accounted for 57% of the data.
The seven factors corresponded closely with the domain structure but the factor structure
was not consistent across samples. Consequently, we scored the MQOL questionnaire (a)
by an overall score and (b) subscale scores for each of the seven domains.
Menopausal quality of life both for total scale and for domains was highly correlated
with global quality of life. While we would expect good correlations between a
r
condition-speciŽc and global measures of QOL, the size of the correlation (
.61)
5
supports the use of condition-speciŽc QOL measures to provide information to that of
self-rated global QOL. Cronbach’s alpha ranged from .92 for the total MQOL scale, and
from .91 to .69 for each of the menopausal QOL domains, suggesting that the scale has a
reasonable level of internal consistency.
Measuring menopausal status
Determining menopausal status in surveys remains problematic. The term ‘menopause’
has many possible interpretations and we would expect some differences in interpretation
among respondents. We found that the self-rated menopausal status classiŽcation has
advantages over a broad menstrual cycle classiŽcation. Firstly, it allows for the classiŽca-
tion of women who for medical reasons no longer have a menstrual cycle but who perceive
themselves to be at a different stage of the menopause than their menstrual cycle status
would suggest. Second, if the menopause is not just a biological event linked to the
menstrual cycle, then other indicators, or a woman’s own perceptions become important

Menopause and quality of life
407
in determining their position in that process. Third, the menstrual cycle classiŽcation
assumes that women who no longer have a menstrual cycle are similar and form a
heterogeneous group. However, we found that QOL is different for women who perceive
that they experienced the menopause long ago when compared with women who perceive
that they have just reached the end of the menopause, irrespective of age. Thus self-rated
menopausal status has discriminability beyond that of menstrual cycle status, as both
women at the end of the menopause and those who reached it long ago are equivalent in
terms of menstrual cycle status.
Sample effects and self-selection
We obtained data from three different samples. Although women recruited through two
different Health Authorities had similar levels of QOL, those recruited through the
Magazine sample had signiŽcantly poorer QOL. Nevertheless, all of our samples are
biased by self-selection, and the perception that the study concerns the menopause may
possibly have biased the samples towards greater severity of problems compared with the
total population.
Medical history and work history and MQOL
Doyal’s (1994) research highlighted the role of women’s employment status in their
physical and mental health. Our results suggest that working outside the home also
makes a positive contribution to menopausal quality of life: women who work outside the
home reported signiŽcantly higher MQOL.
In addition, current health and medical history also affect MQOL: women with high
levels of co-morbidity and those who undergo hysterectomy or oophrectomy report lower
MQOL compared with women with good health. Poorer QOL was also reported by those
who were former users of HRT compared to those who were present users or had never
used HRT. However, we also found that women who were former users of HRT reported
signiŽcantly higher levels of co-morbidity. Thus, employment, gynaecological history
and co-morbidity are all related to MQOL.
QOL changes over the menopause
Previous research has been equivocal about the relationships between menopausal status
and QOL changes. Dennerstein (1996), reported that menopausal status did not affect
et al
et al
. (1995) show a positive
well-being. However, Oldenhave
. (1993) and Ledesert
relationship between menopausal status and QOL in pre-menopausal and peri-
menopausal women. We found that when lifestyle, health differences and HRT
usage were accounted for, QOL was related to the menopause, but the relationship
differed as a function of the measure used. In the global quality of life, cognition and
the feelings domains, there was an inverted U-shaped relationship in which women at
the end of the menopause were more positive than women who were at the beginning.
Menopausal quality of life, and the domains of sleep and energy also show an inverted
U-shaped relationship across the stages of the menopause, but these scores did not
reach levels reported by women at the beginning of the menopause. However, scores
for the domains of social interaction and symptom impact indicate a steady decline in

Pamela A. Jacobs
408
et al.
QOL during the menopause. Overall the lowest QOL scores irrespective of measure
were reported by women who self-rated themselves as being in the middle of the
menopause.
Thus, our data support the apparently contradictory assertion of Dennerstein (1996)
et al
et al
that lifestyle is important to QOL and Oldenhave
. (1993) and Ledesert . (1995)
that menopausal status is important to QOL. Our data also show that both are important.
In addition, our data show several other factors that need to be taken into account when
considering QOL during the menopause.
First, women who were not working outside the home also reported lower QOL when
compared with women who worked outside the home. As women who declare themselves
to be at the end of the menopause and beyond are mostly close to, or past retirement age,
these data suggest that some of the decline in QOL at the end of the menopause may be
reecting other social changes in women’s lives associated with ageing, rather than
hormonal changes.
Second, our data show the effect of sampling method. We found that women who were
invited to take part in an menopausal study through a magazine advert were much more
likely to be suffering from QOL deŽcits, and had tried and given up on HRT despite
these deŽcits. Other researchers have used a variety of sampling procedures. Dennerstein
(1994) used a population sample recruited through random digital telephone dialling,
et al
Ledesert
. (1995) used volunteers from the French national gas and electricity
company, Hunter (1992) took a sample of women attending routine ovarian screening
service at a London hospital.
Third, our data show the advantage of self-rated menopausal status compared with
et al
information about the menstrual cycle, used by some other researchers, e.g. Ledesert
.
(1995), Hunter (1992) and Dennerstein (1996). We found that self-rated menopausal
status provided additional discriminability.
In conclusion, our results show a relationship between self-rated menopausal status and
QOL, and between lifestyle and QOL. They also indicate the importance of medical
history and the opportunity to work outside the home in determining QOL. For many
women any decline in QOL during the menopause is followed by improvement in QOL,
and the end of the climacteric coincides with high levels of QOL evaluations on some
dimensions.
References
Abraham, S., Llewellyn-Jones, D., & Perz, J. (1995). Changes in Australian women’s perception of
menopausal symptoms before and after the climacteric. Maturitas, 20 (2–3), 121–128.
Bebbington, P. E., Dunn, G., Jenkins, R., Lewis, G., Brugha, T., Farell, M., & Meltzer, H. (1998). The
inuence of age and sex on the prevalence of depressive conditions: Report from the national survey of
psychiatric morbidity. Psychological Medicine, 28, 9–19.
Daly, E., Gray, A., Barlow, D., McPherson, K., Roche, M., & Vessey, M. (1993). Measuring the impact of
menopausal symptoms on quality of life. British Medical Journal, 307, 836–840.
Dennerstein, L. (1996). Well-being, symptoms and the menopausal transition. Maturitas, 23, 147–157.
Dennerstein, L., Smith, A. M. A., & Morse, C. A. (1994). Psychological well-being mid-life and the
menopause. Maturitas, 20, 1–11.
Derman, R. J., Dawood, M. Y., & Stone, S. (1995). Quality of life during sequential hormone replacement
therapy—a placebo controlled study. International Journal of Fertility and Menopausal Studies, 40, 73–78.
Dickinson, A., & Henriques, N. (1994). Hysterectomy, the women’s view. London: Quartet Books.

Menopause and quality of life
409
Doyal, L. (1994). Waged work and well-being. In S. Wilkinson & C. Kitzinger (Eds.), Women and health:
Feminist perspectives (pp. 65–84). London: Taylor & Francis.
Fletcher,A. E. & Bulpitt, C. J. (1987). Measurementof quality of life in clinical trials therapy. Cardiology,75,
41–52.
Greene, J. G. (1992). The cross-sectional legacy: an introduction to longitudinal studies of the climacteric.
Maturitas, 14, 95–101.
Greene, J. G., & Cooke, D. J. (1980). Life stress and symptoms at the climacterum. British Journal of
Psychiatry, 136, 486–491.
Groeneveld, F. P. M. J., Bareman, F. P., Barentsen, R., Dokter, H. J., Drogendijk, A. C., & Hoes, A. W.
(1993). Relationship between attitude toward menopause, well-being and medical attention among
women aged 45–60 years. Maturitas, 17, 77–88.
Hunt, K. (1994). A cure for all ills? Constructionsof the menopause and the chequeredfortunes of Hormone
Replacement Therapy. In S. Wilkinson & C. Kitzinger (Eds.), Women and health: Feminist perspectives (pp.
141–151). London: Taylor & Francis.
Hunt, S. M. (1997). The problem of quality of life. Quality of Life Research, 6, 205–212.
Hunter, M. (1992). The women’s health questionnaire:a measure of mid-aged women’s perceptions of their
emotional and physical health. Psychology and Health, 7, 45–54.
Hyland, M. E. (1992). Selection of items and avoidanceof bias in Quality of Life scales. PharmacoEconomics,1,
182–190.
Hyland, M. E., Bellesis, M., Thompson, P. J., & Kenyon, C. A. P. (1997). The constructsof asthma quality of
life: Psychometric, experimental and correctional evidence. Psychology and Health, 12, 101–121.
Hyland, M. E., & Sodergren, S. C. (1996). Development of a new type of global quality of life scale, and
comparison of performance and preference for 12 global scales. Quality of Life Research, 5, 469–480.
Karlberg, J., Mattsson, L.-A., & Wiklund, I. (1995). A quality of life perspective on who beneŽts from
estradiol replacement therapy. Acta Obstetricia et Gynecologica Scandinavica, 74, 367–372.
Ledesert,B., Ringa, V., & Breart, G. (1995). Menopauseand perceivedhealth status among the women of the
French GAZEL cohort. Maturitas, 20, 113–120.
Leiblum, S. R., & Swartzamn, L. C. (1986). Women’s attitudes toward the menopause:an update. Maturitas,
8, 47–56.
Lewis, A. M., Rook, K. S., & Schwarzer, R. (1994). Social support, social control and health among the
elderly. In G. N. Penny, P. Bennett, & M. Herbert (Eds.), Health Psychology: a lifespan perspective (pp. 191–
211). Chur, Switzerland: Harwood Academic.
McCleery, J. M., & Gebbie, A. E. (1994). Compliance with hormone replacement therapy at a menopause
clinic in a community setting. The British Journal of Family Planning, 20, 73–75.
McKinlay, J. B., McKinlay, S. M., & Brambilla, D. (1987). The relative contributionsof endocrine changes
and social circumstancesto depressionin mid-agedwomen. Journalof Health and Social Behaviour, 28, 345–
363.
Miles, M. B., & Huberman, A. M. (1994). Qualitative data analysis. New York: Sage.
O’Connor, N. M., Del Mar, C. B., Sheehan, M., Siskind, V., Fox-Young, S., & Cragg, C. (1995). Do psycho-
social factors contribute more to symptom reporting by middle-aged women than hormonal status?
Maturitas, 20, 63–69.
Oddens, B. J., Boulet, M. J., Lehert, P., & Visser, A. P. (1994). A study on the use of medication for
climacteric complaints in Western Europe–II. Maturitas, 19, 1–12.
Oldenhave, A., Jaszmeann, L. J. B., Haspels, A. A., & Everaerd, W. T. A. M. (1993). The impact of
climacteric on well-being: A survey based on 5213 women 39–60 years old. American Journal of Obstetrics
and Gynecology, 168, 772–780.
Ormel, J., Lindenberg, S., Steverink, N., & Vonkorff, M. (1997). Quality of life and social production
functions: a framework for understanding health effects. Social Science and Medicine, 45, 1051–1063.
Pearce, J., Hawton, K., & Blake, F. (1995). Psychological and sexual symptoms associated with the
menopause and the effects of hormone replacement therapy. British Journal of Psychiatry, 167, 163–173.
Quine, L., & Rubin, R. (1997). Attitude, subjective norm and perceived behavioural control as predictors
of women’s intentions to take hormone replacement therapy. British Journal of Health Psychology, 2,
199–216.
Studd, J., Watson, N. R., & Montgomery, J. (1990). Depression and the menopause. British Medical Journal,
300, 1653.

Pamela A. Jacobs
410
et al.
Treloar, A. E. (1974). Menarche, menopause and intervening fecundability. Human Biology, 46, 89–107.
Watson, D., & Pennebaker,J. W. (1989). Health complaints,stress, and distress. Exploringthe centralrole of
negative affectivity. Psychological Review, 96, 234–254.
Wiklund, I., Holst, J., Karlberg, J., Mattsson, L-A., Samsioe, G., Sandin, K., Uvebrant, M., & Van Schoultz,
B. (1992). A new methodologicalapproach to the evaluation of quality of life in postmenopausalwomen.
Maturitas, 14, 211–224.
Received 13 January 1999; revised version received 29 July 1999
Appendix
Menopausal Quality of Life Scale
N
Factor Loadings: Principal components analysis, unrotated (
1188)
5
Scale items
Domain
Energy
Feelings
Feelings
Feelings
Feelings
Energy
Symptoms impact .69
Energy
Feelings
Interactions
Feelings
Energy
Energy
Interactions
Interactions
Interactions
Symptoms impact .64
Feelings
Loading
Things I used to enjoy have become a bit of a chore
I have a general sense of well-being
I feel enthusiastic about things
.73
.72
.71
.71
.71
.70
I feel cheerful
I am depressed about things that didn’t bother before
I Žnd I have the energy to do the things I want
Because of my symptoms, I miss out on leisure activities
I am Žnding it increasingly difŽcult to do my work
I feel conŽdent
I’m afraid to tell anyone at work how I feel
I suffer from unpredictable mood swings
I can keep going all day without any difŽculty
I am too tired to do everyday tasks
I am more reclusive than I would like
I worry that I might snap at friends or at people at work
I get very irritable with people at home
I worry about missing work because of my symptoms
I feel stable
.68
.67
.66
.66
.66
.65
.65
.64
.64
.63
.62
.62
.61
.60
I get tearful easily
At times I want to lock myself away at work
I lose my temper over small things
I can concentrate easily
Feelings
Interactions
Feelings
Cognition
Because of my symptoms I sometimes have to get out of places
I have a problem remembering everyday things
I think my memory is quite good
I feel isolated
I do less than I would like
I start a conversation and can’t remember what I was saying
I feel inadequate in comparison to other people of my age
I scream and shout at people at home
I enjoy chatting as much as I ever did
Symptoms impact .59
Cognition
Cognition
Feelings
.58
.58
.58
.57
.56
.56
.56
.55
Energy
Cognition
Feelings
Interactions
Interactions

Menopause and quality of life
411
Scale items
Domain
Energy
Loading
.53
I am too tired for sex
I enjoy sex as much as ever
Symptoms impact .52
I feel good about my appearance
I can concentrate on hobbies for as long as I used to
I Žnd housework easy
Feelings
Cognition
Energy
.51
.50
.50
My ushes/night sweats keep me awake at night
I Žnd hot ushes embarrassing
At night I throw off all the bedclothes and then feel cold
Symptoms impact .47
Symptoms impact .47
Symptoms impact .46
I can work hard if I want to
I usually sleep well
I sleep through the night
Energy
Sleep
Sleep
.46
.45
.44
I Žnd intercourse uncomfortable because of dryness
I have a good appetite
Symptoms impact .41
Appetite
.40
I Žnd I can’t get back to sleep if I wake at night
My symptoms do not interfere with my work
I am more interested in sex
Sleep
.37
Symptoms impact .35
Symptoms impact .31
I take short naps during the day
Energy
.25