Journal of Affective Disorders 67 (2001) 147–158

www.elsevier.com / locate / jad

Research report

Impact of bipolar affective disorder on family and partners

a b,
Glenys Dore , Sarah E. Romans *
a
Wycombe Clinic, 114 Wycombe Road, Neutral Bay 2089, Sydney, NSW, Australia
b
Department of Psychological Medicine, Dunedin School of Medicine, University of Otago, PO Box 913, Dunedin, New Zealand

Received 31 August 1998; accepted 14 January 1999

Abstract

Background: Successful management of major mental illness in the community relies significantly on an informal or
non-professional network of caregivers. The needs and experiences of such caregivers have been little studied with respect to
major chronic mood disorders. Method: A sample of caregivers (n 5 41) of RDC bipolar disorder was systematically
interviewed to determine how this role affected them. Results: Caregivers reported significant difficulties in their
relationships with the patient when s / he was unwell, with considerable impact on their own employment, finances, legal
matters, co-parenting and other social relationships. Violence was a particular worry for partner / parent caregivers of both
male and female patients when the patient was manic. The caregiver’s own mental health appeared unaffected. Despite this,
the caregivers appeared emotionally committed to the patients and showed considerable tolerance of problem behaviours,
which they rank-ordered for difficulty. Among nonfamily partners, knowledge of the illness before cohabitation was poor.
Limitation: The sampling does not capture caregivers who have abandoned their role, such as spouses who have divorced the
bipolar sufferer. Conclusions: Management of this illness requires a partnership between mental health professionals and the
informal caregivers and the authors suggest that each group needs to understand the difficulties encountered by the other.
Although erosion of relationships is a well-known complication of bipolar disorder, findings indicate that treating clinicians
can rely on caregivers committed to the welfare of the patient.  2001 Elsevier Science B.V. All rights reserved.

Keywords: Bipolar disorder; Social adjustment; Occupation; Family; Conjugal relations

1. Background means that family members care for increasing

The move to community-based care for people
with psychiatric illnesses from the 1950s onwards
*Corresponding author. Tel.: 1 64-3-474-7989; fax: 1 64-3-
473-8494.
E-mail address: sarah.romans@stonebow.otago.ac.nz (S.E.
Romans).
numbers of patients in community settings. The
previous assumption that family cause the patient’s
illness through their own behaviours has been re-
placed by a contrary view that the illness in the key
patient causes disturbance in the relatives (Tantum,
1989). In this context, an increasing number of
studies has examined the impact of functional psy-
chiatric illness on family members. Recent work on

0165-0327 / 01 / $ – see front matter  2001 Elsevier Science B.V. All rights reserved.
PII: S0165-0327( 01 )00450-5
148 G. Dore, S.E. Romans / Journal of Affective Disorders 67 (2001) 147 – 158
disability associated with psychiatric illness has
emphasised the chronic burdensome nature of most
psychiatric illness to the individual, their immediate
social group, and at a world wide level (Murray and
Lopez, 1996). The burdens of caring for a psychiatri-
cally ill family member have been usefully divided
into two categories, namely ‘objective’ burden and
‘subjective’ burden (Fadden et al., 1987). Objective
burden is any disruption to family or household life,
which is verifiable and observable. It includes high
rates of separation and divorce in marriages with a
psychiatrically ill spouse; restricted social and leisure
activities, often related to fears of stigma; and
financial difficulties (Fadden et al., 1987). Subjective
burden involves the personal feelings of carrying
such a burden, such as feeling distressed, unhappy,
upset (Grad and Sainbury, 1963a,b; Hoenig and
Hamilton, 1966, 1969; Platt, 1985; Fadden et al.,
1987). The types of subjective burden experienced
include guilt if the relative feels responsible for
contributing to the illness; rejection towards the
patient; anger that the illness has spoilt the relative’s
life; grief and a sense of loss contrasting the present
with the person before their illness occurred. Subjec-
tive burden experienced by relatives has been linked
to the development of affective symptoms (Tantum,
1989).
Studies of how family factors influence outcome
in bipolar disorder are scarce. The available data
suggest that family factors may have an important
effect on outcome. In particular, families with high
expressed emotion and / or negative Affective Style
may be associated with an increased tendency to
relapse in bipolar patients (Miklowitz et al., 1988).
Social interventions may be useful in these cases,
particularly for those patients who are only partial
responders to lithium and live with highly emotional
families (Priebe et al., 1989). Another group, in New
York, have been examining the gains to be achieved
when a marital intervention was added to a randomly
selected sample of married inpatients with bipolar
disorder; the control group received only the stan-
dard treatment (Clarkin et al., 1990, 1998). Those
with the additional marital therapy did not differ in
symptoms at follow up but showed higher overall
function and were more adherent to their recom-
mended treatment. The two studies suggest that it is
possible to conduct relevant research in these ques-
tions.
The aim of this study was to examine the impact
of bipolar disorder on the primary caregiver (usually
a family member) in terms of both objective and
subjective burden, and to list those symptoms ex-
perienced as most burdensome.
2. Method
The clinical records were examined of all patients
aged between 16 and 65 years discharged from an
inpatient unit in the Otago region of New Zealand
between 1 January 1985 and 31 December 1988. The
records of all patients with a diagnosis of bipolar
affective disorder (or a related disorder) were further
examined to see if the patient met the Research
Diagnostic Criteria (RDC) for bipolar affective
disorder. This project had approval from the local
ethical committee for health research.
Ninety-one patients were identified, and asked to
undergo a formal diagnostic interview using the
Schedule for Affective Disorders and Schizophrenia
(SADS) (Endicott and Sopitzer, 1978). Two patients
had died, 12 had moved out of the Otago region, and
13 refused to participate in the study. A total of 64
patients, 70% of the original sample attended the
SADS interview of whom 58 patients met the RDC
criteria for bipolar I disorder. This group formed the
cohort for a two year prospective study looking at
socioeconomic characteristics, social networks, and
factors related to relapse (Romans and McPherson,
1992). The study was part of a collaborative New
Zealand / United Kingdom prospective study of bipo-
lar affective disorder (McPherson and Romans,
1993).
Of the 58 bipolar patients in the original cohort,
49 remained in the study by the end of 2 years
(85%). The 49 patients were informed about the
Caregivers Study and asked if they would give their
written consent for the first author to contact the
person they nominated as their closest caregiver. The
nominated caregiver was contacted, given informa-
tion about the study, and gave written informed
consent for participation in the study
The assessment with caregivers involved a face-to-
face interview lasting 60–90 min using a semi-
structured schedule based on that devised by Fadden
et al. (1987). This schedule assessed both objective
and subjective burden of care, difficult behaviours
 G. Dore, S.E. Romans / Journal of Affective Disorders 67 (2001) 147 – 158
and satisfaction with services received. Socioeco-
nomic status was assessed using the usual New
Zealand scales for males and females separately,
based on occupation (Johnston, 1983). The caregiv-
ers were asked to complete the General Health
Questionnaire (Goldberg and Hillier, 1979). Dura-
tion of illness was calculated from the date from the
first contact with psychiatric services till the present.
The median value for duration of illness was 13.6
years. Values above this were taken to indicate long
duration of illness; values below this were taken to
indicate short duration of illness. Severity of illness
was calculated by dividing the number of hospital
admissions by the duration of illness. The severity of
illness values and the median value used as a cut-off
point to define high or low severity of illness. Values
of 0.6 and above indicated high severity of illness;
values of less than 0.6 indicated low severity of
illness.
3. Results
For the 49 patients, 41 caregivers completed the
questionnaire (84%). Of the other eight patients, one
developed leukaemia, two had no nominated care-
giver, two would not agree for their caregiver to be
contacted, in two cases the caregiver did not wish to
participate, and one had episodes of illness only prior
to the caregiver being involved.
3.1. The patient sample
The 41 caregivers were involved with 21 male and
20 female patients (see Table 1). The average age of
these patients with bipolar disorder was 36 years
(20–63 years). All were outpatients at the time of the
study: three (7%) were experiencing a manic or
hypo-manic episode.
Patients were divided into two groups: employed
(full or part-time work, homemaker, student) or not
employed (unemployed, beneficiary, retired). More
patients in paid employment (9 / 22, 41%) were
married compared to those not employed (3 / 19,
16%, x 2 5 3.1, P 5 0.08). More, (14 / 22 64%) of
employed patients were classified as having a low
severity of illness compared with 6 / 19 (32%) pa-
tients who were not working ( x 2 5 4.193; P 5 0.04).
Similarly, most (64%) of employed patients had
149
short duration of illness compared with 32% of
patients not employed ( x 2 5 4.19; P 5 0.04).
Overall, these patients were a group with long-
standing, recurrent illness requiring recurrent hos-
pitalisations. Sixty three percent had had six or more
admissions to hospital since the onset of the illness
and almost one third of the group had been ill for ten
or more years.
3.2. The caregiver sample
Most nominated caregivers were female (61%)
(Table 2). The age distribution of caregivers was
very diverse — mean 46 (range: 14–76 years). A
quarter of the group were 60 years or older, with ten
percent of this group being 70 years or older. All
caregivers over 60 were parents apart from one who
was a friend. The care-giving parent was usually a
mother (73%). Caregivers living with the patient in
the past 6 months (n 5 21, 51%) were significantly
more likely to be a parent or a partner rather than a
friend or other relative ( x 2 5 9.8; P 5 0.002).
Whether a caregiver had lived with the patient in the
past 6 months was unrelated to severity or length of
illness.
3.3. Caregiver’ s relationship with the patient
The illness had a major impact on the caregiver’s
relationship with the patient when they became
unwell. Most caregivers (90%) found the patient
distant and difficult to get close to during episodes of
illness. The patient was experienced as more irritable
when unwell (80%) and this frequently led to
arguments that did not occur at other times. Violence
was a significant concern during episodes of mania
or hypomania. Nearly half the caregivers (44%) had
experienced violence or were frightened that vio-
lence was going to occur when the patient was
unwell. A quarter of the group had in fact ex-
perienced serious forms of violence when the patient
was unwell.
Once he threw me on the floor and tried to butt
my head.
He threatened me with a chain saw once.
150 G. Dore, S.E. Romans / Journal of Affective Disorders 67 (2001) 147 – 158

Table 1
Sociodemographic and clinical characteristics of patients
Total sample Percent
(n 5 41)
Time since 1st contact psychiatric services
3–4 years 4 10
5–9 years 9 22
10–20 years 21 51
More than 20 years 7 17
Sex
Male 21 51
Female 20 49
Marital status
Single 17 42
Separated / divorced 12 29
Married / defacto 12 29
Number of children
None 14 34
One or more 27 66
Current living situation
With partner (spouse / defacto) 12 29
With other family member / s 8 20
Living alone 12 29
With friends / flatmates 7 17
Other (residential programme, halfway house) 2 5
Ethnicity
European 37 90
Maori 2 5
Other 2 5
Socioeconomic status
High: Skilled, professional 12 29
Low: Unskilled 9 22
Other: housewife, retired 15 37
Student 5 12
No. of hospital admissions since
first contact psychiatric services
1–5 admissions 15 37
6–10 admissions 13 32
11–20 admissions 12 29
More than 20 admissions 1 2

She was about to stab me with a kitchen knife
so I knocked her out.
Female patients (43%) were as likely as males
(45%) to be violent towards their caregiver. Partners
were more likely to be exposed to violence than
parents (77 vs. 33%: x 2 5 5.3; P 5 0.02) or other
non-parental caregivers (77 vs. 23%: x 2 5 7.5; P 5
0.006). Levels of violent behaviour were similar for
those with high and low severity of illness (43 vs.
45%: x 2 5 0.02; P 5 0.9).
Most caregivers (81%) were distressed by the way
the patient related to them when unwell; 64%
described the level of personal distress as ‘severe’.
While most caregivers were able to discuss at least
some of their concerns about the illness with the
patient, a significant number (24%) were reluctant to
discuss it at all. Willingness to discuss the illness was
 G. Dore, S.E. Romans / Journal of Affective Disorders 67 (2001) 147 – 158 151

Table 2
Characteristics of care-giver sample
Total sample Percent
(n 5 41)
Sex
Male 16 39
Female 25 61
Ethnicity
European 39 95
Maori 1 2
Other 1 2
Relationship to patient
Parent 15 37
Sibling 1 2
Other relative 9 22
Partner: spouse / defacto 13 32
Living together (11) (27)
Separated (2) (5)
Friend 3 7
Marital status
Single 6 15
Separated / divorced 6 15
Married / defacto 24 59
To patient (11) (27)
To other (13) (32)
Widow / widower 5 12
Employment status
Working (full / part-time, homemaker, student) 30 73
Not working (unemployed, beneficiary) 3 7
Retired 8 20
Socioeconomic status
High: skilled, professional 17 41
Low: unskilled 7 17
Other: housewife, retired 14 34
Student 3 7
Carer’s experience of illness
Both mania and depression observed 37 90
Mania / hypomania only observed 4 0

not significantly related to the sex of the patient, the
nature of the relationship, severity or length of
illness or the occurrence of violence towards the
caregiver.
The caregiver’s relationship with the patient usual-
ly improved significantly when they became well
again, with 80% of the group feeling that the
relationship remained close during times of remis-
sion. Almost half the group (49%) felt the illness had
brought them closer. This increased sense of close-
ness was not significantly related to the severity or
length of the illness, or to the nature of the caregiver.
It was, however, significantly related to the sex of
the patient and caregiver. A closer relationship was
more common if the patient was male ( x 2 5 5.5;
P 5 0.02) and the caregiver female ( x 2 5 3.2; P 5
0.07).
Our relationship is stronger now because of the
illness.
He’s more open with me now that he used to
152 G. Dore, S.E. Romans / Journal of Affective Disorders 67 (2001) 147 – 158

be. And I’m more nurturing than I used to be — I the illness. Most (67%) found work a strain as a
take more care of him and listen more. consequence of the illness and half of this group
estimated the strain as being severe.
We’re really close when she’s all right. It’s
really hurtful when she’s off: she blames me for
all her problems. 3.4.2. Finances
Over one quarter (27%) of caregivers had ex-
perienced a reduction in their income since the onset
3.3.1. Caregivers experience of stress
of the patient’s illness. Reduced income was not
Caregivers were asked a general question about
significantly related to severity of illness or length of
how much they experienced stress as a consequence
illness ( x 2 5 0.2; P 5 0.6), nor to the gender of the
of the patient’s illness. Most (71%) of caregivers
patient ( x 2 5 0.9; P 5 0.3). However, partners were
described the stress as major: for 44% this was
significantly more likely to experience a reduction in
primarily when the patient was ill, while another
income related to the illness when compared with
27% felt this stress even when the patient was well.
other caregivers ( x 2 5 7.0; P 5 0.007). Patients fre-
Similar numbers of partners and parents experienced
quently had difficulty managing their finances during
the stress as major (85 vs. 87%). Partners and
an episode of illness and almost half the caregivers
parents were far more likely to see the stress as
(49%) took on significant responsibility for the
major compared with other caregivers (86 vs. 38%:
patient’s finances at those times. This financial
x 2 5 9.6; P 5 0.002). For those parents experiencing
support often needed to continue once the patient
the illness as a major stress, 54% had been living
was well (37%). Such financial responsibility was
with the patient over the past 6 months and 46% had
not significantly related to illness severity, length of
been living elsewhere. All but one partner had been
illness, patient gender, nor the nature of the caregiver
living with the patient over the previous 6 months.
relationship.
3.3.2. Psychiatric morbidity
The GHQ-28 was used to screen caregivers for the I’ve had to watch his bank statements and
presence of minor psychiatric morbidity, with a 4 / 5 cheque books when he’s high — he has mini
cutoff used to identify caseness (Romans-Clarkson et spend-ups.
al., 1989). Only six caregivers (17%) had a score of
5 or higher. The GHQ score was not significantly I got his main savings in an account which
related to the nature of the caregiver, nor to overall requires a signature in addition to his own, to
levels of stress experienced as a consequence of the protect him from over-spending when he’s manic.
illness.
Caring for a patient with bipolar disorder was often
3.4. The practical consequences of the patient’ s
an expensive exercise. Almost a third (29%) of
illness for the caregiver
caregivers had incurred major costs as a result of the
illness. Parents and partners bore the brunt of these
3.4.1. Occupation
costs more frequently than other caregivers ( x 2 5
Relatively few caregivers were required to make
7.8; P 5 0.005). The costs involved were often
significant changes to the nature of their work
sizeable; for 15% of the group this involved short-
because of the illness; only 12% experienced limita-
term payments of $5000 or more.
tions on the type of work they were able to do. In
only one case had the breadwinner in the household
changed as a result of the illness. However, three- She spent three and a half months in a private
quarters (76%) of the caregivers working outside the hospital which cost us $7000.
home had to reduce their hours of work or take time
off work during episodes of illness. One caregiver She became unwell during an overseas trip and
reduced permanently their hours of work because of it cost about $8000 to get her home.
 G. Dore, S.E. Romans / Journal of Affective Disorders 67 (2001) 147 – 158
He spent up about $5000 in a couple of weeks,
but most of it was recuperated.
He wrote off one car and badly damaged the
replacement car when he drove while high and
alcohol-effected. It cost about $10,000.
Coping with the patient’s financial affairs was seen
as a significant stress by many caregivers (46%).
3.4.3. Legal difficulties
Police involvement with patients during at least
one episode of illness was usual (66%), although this
seldom resulted in legal repercussions (17% of the
total group). There were no gender differences for
legal difficulties (female 2 / 20, males 5 / 21) or police
involvement (female 13 / 20, male 14 / 21). However,
both legal difficulties and police involvement were
significantly related to the severity of illness. The
majority (6 / 7) of those with legal difficulties had
high severity of illness, while those without legal
difficulties tended to have low severity of illness
(19 / 34, x 2 5 4.02; P 5 0.05j. For those with police
involvement, high severity of illness was more
common (17 / 27), while those with no police in-
volvement tended to have low severity of illness
(10 / 14) h x 2 5 4.4; P 5 0.04j. Police involvement
caused significant stress for many caregivers (42%),
with one quarter of the group felt the stress had been
severe.
The police brought her home when she was
found wandering the streets and couldn’t find her
way home.
She ran along the street taking her clothes off
and the police were called.
She complained to the police that her parents
had abducted her, and they had to check it out.
3.4.4. Marital and sexual relationships
Thirteen of the caregivers (32%) described them-
selves as the patient’s partner (spouse or defacto
partner). Eleven were living with the patient, while
the remaining two had separated. The impact of the
illness on the partner was often marked. Most
partners (92%) said that they had found it difficult to
153
keep the relationship going because of the illness; for
39% this was the case even when the patient was
well again.
She’s suspicious of me when she’s hypomanic:
inspecting hankies for lipstick, accusing me of
having affairs.
Our marriage is a lot different than it would
have been without the illness: even when she’s
well, I’m vigilant for signs of illness all the time.
I keep hoping that S might return to the person
she was before the marriage.
Separation because of illness-related difficulties was
common (62%). While most had got back together
again, two of the group (15%) were separated from
their partner at the time of the interview.
Problems were commonly experienced within the
sexual relationship because of the illness (77%) and
for half (54%) of the group, these difficulties had
continued when the patient was well. Two of the
partners were aware of the patient being involved in
an extramarital relationship as a result of a manic or
hypomanic episode.
3.4.5. Knowledge of illness
The partner’s awareness of the illness before
marriage or cohabiting was variable. For 15% of
partners, the illness developed after living together.
Another 39% knew about the illness beforehand;
while 45% either had the illness kept from them, or
were aware there was a problem but didn’t ap-
preciate its significance.
He told me he had manic depression but I
didn’t know what it was.
The illness was carefully kept from me until he
suddenly went mad five years ago.
The majority of the partners (62%) felt that they
would probably not have entered into the relationship
if they had had more knowledge about the illness and
its effects. This belief was more common for those
who didn’t know about the illness or had limited
knowledge before committing to the relationship
154 G. Dore, S.E. Romans / Journal of Affective Disorders 67 (2001) 147 – 158
than for those who were aware of the illness (88 vs.
20%; x 2 5 5.9; P 5 0.01).
3.4.6. Parenting
For the 13 couples, ten were co-parenting (of nine,
the children were from the relationship; for one
couple, the children were from a previous relation-
ship). In 60% of cases, the caregiver stated that the
parenting within the relationship had changed as a
consequence of the illness. This change was not
significantly related to severity of length of illness.
Ninety percent of caregivers had found it more
difficult to function because of the illness in their
parent role.
For those four couples without their own children,
none said that the illness was a reason why they had
not had children together. Most (89%) partners were
aware that bipolar disorder could be inherited, but
only half knew this prior to the birth of their
children. The majority of partners (67%) were
concerned that their children might have a chance of
developing bipolar disorder. Some (44%) of partners
felt that knowledge about the inheritance of bipolar
disorder might have influenced their decision to have
children.
3.4.7. Social and leisure activities
Social activities and leisure pursuits were fre-
quently interrupted when the patient became unwell
(61%). This disruption was not significantly related
to the gender of the caregiver or patient, nor to the
severity or length of the illness. However, partners
did experience significantly more disruption to their
social and leisure activities than parents or other
caregivers ( x 2 5 7.9; P 5 0.005).
Over half the group (56%) reported that the illness
had a significant negative impact on their relation-
ships with friends, family and acquaintances. This
impact often continued when the patient was well
(44%). Some caregivers (15%) stated that they had
lost one or more friends because of the illness.
Partners were more likely to have lost friends than
other caregivers ( x 2 5 4.0; P 5 0.05). Loss of
friends was not significantly related to the severity or
length of the illness. Over a third (39%) had
experienced problems or tensions with neighbours
because of the patient’s behaviour when unwell.
Many caregivers (42%) were embarrassed by the
illness and almost a quarter (24%) had experienced a
sense of stigma because of it. Experiencing stigma
was not significantly related to severity or length of
illness, sex of patient or caregiver, nature of care-
giver, nor to strain of legal difficulties or police
involvement.
The family blamed me. They didn’t see him as
ill — they saw me as putting him into hospital.
One of them wouldn’t speak to me for years.
We’ve lost several friendships as we’ve had no
time to nourish them.
Half the group (49%) found it distressing to cope
with the impact of the illness on their relationships
with other people. The others were remarkably
resilient, and were able to cope well, or had de-
veloped a sense of resignation to the changes in their
relationships. Distress from relationship difficulties
was not significantly related to the sex of the
caregiver or the length or severity of the illness.
3.4.8. Problem behaviours
Most commonly, caregivers found aggressive and
violent behaviours to be most disturbing (17%), with
suicidal ideas and acts next. Other manic behaviours
which distressed included odd behaviours (10%),
overactivity, overtalkativeness and impulsive spend-
ing (each 4%). For depression, suicidal ideas and
attempts the next most common. For depressive
episodes, the depressed mood itself (with its accom-
panying misery and hopelessness) disturbed the
caregivers most. Ninety percent of caregivers
believed the disturbing manic behaviours were
caused by the illness, with only two caregivers
believing that the behaviour was under the patient’s
control. Similarly, 97% believed the disturbing be-
haviours during depression were part of the illness.
Ten percent of caregivers had witnessed episodes of
mania only. Of those caregivers who had ex-
perienced both episodes of mania and depression,
30% found mania to be more distressing, 19% found
depressive episodes more distressing, while 46%
found both equally distressing. Comments from
caregivers included:
 G. Dore, S.E. Romans / Journal of Affective Disorders 67 (2001) 147 – 158
At least with manic behaviour I can get some
help around the house.
When manic she’s very bubbly. When de-
pressed she wouldn’t do anything; I couldn’t get
her up to have meals. She lay in bed and wouldn’t
talk.
His hostility and threatening attitude to me
personally is upsetting and at times frightening
(during mania).
4. Discussion
This is the largest study of the impact of bipolar
disorder on caregivers to be published to date and it
generates a number of interesting findings. Whilst it
describes a clinical sample, these patients were
recruited from a public hospital psychiatric service
and the results typify families who have to manage a
serious bipolar illness in their midst. In this sense,
the findings are likely to generalise to families with
members who have this serious mental disorder, with
its fluctuating, relapsing and remitting course (Sil-
verstone and Romans-Clarkson, 1989; Goodwin and
Jamison, 1990). The modest size of the care-giver
sample means that we can have confidence in the
statistically significant results.
Of interest first are the demographic characteristics
of the people who the patients with bipolar affective
disorder identified as their primary care-givers. The
majority were female. A number of people have
written about the care-giving role with respect to
gender, suggesting that women are socialised to care
for others and that often women do this at consider-
able personal cost to themselves with inadequate
recognition from the community at large. Collings
and Seminiuk (1998) have described how women are
often assumed to have the required skills, to see it as
an extension of their role as mothers and often to be
the only potential people prepared to consider taking
on a caregiving role. These authors note that a
coherent sociological theory of caregiving work has
not yet emerged. There is no conceptual framework
readily applicable to the analysis of tasks and
experiences which cross the ‘traditional’ divisions
between public and private, and home and work. In
155
this study, the majority of caregivers were family
members: parents (37%); a partner (32%) or another
relative (24%). The care-givers showed a wide range
of age (average 46 years; range: 14–76). These data
suggest that a diverse group of people become
involved as primary care givers for patients with
bipolar disorder and should caution against quick
presumptions about the impact of the care-giving
role.
Most caregivers were significantly distressed by
the way the patient related to them when unwell.
Partners and parents particularly found this stress
major. When ill, the patient became more distant and
difficult to get close to, irritable, with increased
frequency of arguments. Violence was more common
towards partners than other caregivers and was not
determined to the gender of the patient. Most part-
ners (92%) found it difficult sustaining the relation-
ship because of the illness. Despite the negative
impact during episodes of illness, most caregivers
felt the relationship remained close during times of
remission and only a small proportion of caregivers
(17%) had a GHQ acute score of 5 or higher when
studied.
There is little information about how spouses of
bipolar patients cope with the illness. Clearly many
of them do not do so and separation / divorce is the
consequence. The literature indicates that the mar-
riages of bipolar patients often lead to separation and
divorce (Brodie and Leff, 1971; Carlson et al., 1974;
Weeke et al., 1975; Dinicola, 1989; McPherson et
al., 1992; Kessler et al., 1998). Some authors have
described bipolar marriages as being ‘intermittently
incompatible’, i.e. they can be stable much of the
time but are threatened by enormous upheaval
whenever the patient becomes unwell (Greene et al.,
1976; Frank et al., 1981). Our data support this idea.
Fadden et al. (1987) examined a group of persis-
tently depressed patients where one third of the
group had bipolar disorder. Generally, spouses saw
the patient and themselves as having little or no
control over the difficulties. They were critical about
the services they had received. Over half reported
they had not been provided with adequate infor-
mation outlining the nature of the patient’s difficul-
ties, the cause and the prognosis. Being married
improves patient outcome, presumably by improving
lithium compliance (Demers and Davis, 1971; Con-
156 G. Dore, S.E. Romans / Journal of Affective Disorders 67 (2001) 147 – 158
nelly et al., 1982; Lesser, 1983). Close collaboration
between professional services and caregivers is an
essential part of good community care. It is im-
portant to understand the precise nature of stresses
experienced by caregivers as they contribute much of
the overall management of the patient’s illness.
For patient–caregiver couples with children in this
study, almost all reported finding difficulties func-
tioning as a parent because of the illness and in 60%
of cases the parenting roles had altered significantly.
There is a need for further understanding about the
ways in which this illness disables parents from
providing the necessary practical and emotional
resources to their children. The children’s perspec-
tives also need documenting. The caregiving spouse
needs accurate information about the likely conse-
quences of bipolar affective disorder on family
activities. Only 56% of partners were aware that
bipolar disorder could be inherited prior to the birth
of their children and many (44%) felt that knowl-
edge about the heritability of bipolar disorder might
have influenced their decision to have children.
Future research needs to focus on age specific needs
of these children, so that where ill parents cannot
provide appropriate resources be they physical, so-
cial or educational, the system knows about the
vacuum and can fit it flexibly.
A study by Targum et al. (1981) found that,
compared with their spouses, patients appear to
minimize the impact of bipolar disorder on their
family, and the potential for offspring to develop the
illness. In contrast, spouses often indicated they
would have reconsidered marrying and having chil-
dren if they had had more knowledge about the
illness. A recent study however, showed almost all
patients and spouses want to know about genetic test
information on their potential children; however few
felt that knowledge that their partnership had a
genetic loading for them of their children would have
deterred them from marriage or childbearing (Tri-
ppitelli et al., 1998). The more optimistic tone of this
article may arise from the questions being hypotheti-
cal, rather than documenting past experiences, as
with the Targum study.
It is very clear that most caregivers experienced
significant disruption in social activities and leisure
pursuits common when the patient was unwell
(61%). This pertained particularly for partner
caregivers. Relationships with friends, family and
acquaintances were often negatively affected, re-
sulting in the loss of friendships, tensions with
neighbours, and the experience of being stigmatised,
especially with respect to police involvement with
the patient. The majority of caregivers reported that
they needed to take time off work and a third of
caregivers experienced a reduction in income since
the onset of the illness. These findings show the
various ways in which the caregiving role com-
promises other social roles occupied by the caregiver
and begin to define the heavy social cost of bipolar
affective disorder. Surprisingly, few of these vari-
ables related to the patients’ gender or the length of
illness. Seventy of illness (defined as percentage of
time unwell) was associated with greater caregiver
burden.
Aggressive and violent behaviour were most dis-
turbing during mania, and depressed mood during
depressive episodes. These behaviours were seen as
symptomatic of illness by most caregivers and not
under the patient’s control.
Previous work has listed a range of symptoms
which trouble (Targum et al., 1981; Fadden et al.,
1987). It seems clear the caregivers cope better when
they clearly understand that the symptoms are caused
by the illness and not by the patients waywardness or
personality. This finding echoed in the data reported
here, highlights the need for good psychiatric educa-
tion, not only of the patient, but also of his / her
immediate social group.
Janowsky et al. (1970) outlined the types of
behaviours which alienate the manic patient from
others: these have their greatest impact on the marital
relationship. These behaviours included manipulation
of others, projection of responsibility and progressive
limit testing.
The data summarised above show that many
caregivers (spouses and relatives) wish to continue in
their committed role to the patient with bipolar
disorder. We need to understand how best profession-
al services can support their essential contributions.
The role of external stressors in the precipitation of
relapse in bipolar affective disorder remains unclear.
Some like Post (1992) have argued that psychosocial
stress may precipitate an affective episode, whilst
others have essentially failed to uncover a role for
life event stress in the relapse pattern of bipolar
 G. Dore, S.E. Romans / Journal of Affective Disorders 67 (2001) 147 – 158
disorder (McPherson and Romans, 1993). It is
probable that the discriminating factor between these
disparate research results lies in the strength and
availability of the social support experiences by the
bipolar sufferer. The buffering effect of social sup-
port against a number of stressors has been outlined
(Anashensel and Stone, 1982; Alloway and Bebbin-
gton, 1987; Brugha, 1995).
It is important to understand the precise nature of
stresses experienced by caregivers as they contribute
much of the overall management of the patient’s
illness. The ways in which professionals can inter-
vene also needs further study and will certainly
involve psychotherapy for many (McAlpin and
Goodnick, 1998). Then may the professional sector
and non-professional care-givers work together effi-
ciently in their support of people with bipolar
affective disorder.
References
Alloway, R., Bebbington, P., 1987. The buffer theory of social
support: a review of the literature. Psychol. Med. 17, 91–108.
Anashensel, C.S., Stone, J.D., 1982. Stress and depression; a test
of the buffering model of social support. Arch. Gen. Psychiatry
39, 1392–1396.
Brodie, H., Leff, M., 1971. Bipolar depression — a comparative
study of patient characteristics. Am. J. Psychiatry 127, 126–
130.
Brugha, T.S., 1995. Social support and psychiatric disorder:
overview of the evidence. In: Brugha, T.S. (Ed.). Social
Support and Psychiatric Disorder: Research Findings and
Guidelines For Clinical Practice, Vol. 2. Cambridge University
Press, Cambridge, UK.
Carlson, G., Kotin, J., Davenport, Y., Adland, M., 1974. Follow-up
of 53 bipolar manic-depressive patients. Br. J. Psychiatry 124,
1.
Clarkin, J.F., Carpenter, D., Hull, J., Wilner, P., Glick, I., 1998.
Effects of psychoeducational intervention for married patients
with bipolar disorder and their spouses. Psychiatr. Serv. 49,
531–533.
Clarkin, J.F., Glick, I.D., Haas, G.L., Spencer, J.H., Lewis, A.B.,
Peyser, J., DeMane, N., Good-Ellis, M., Harris, E., Lestelle, V.,
1990. A randomized clinical trial of inpatient family interven-
tion. V. Results for affective disorders. J. Affect. Disord. 18,
17–28.
Collings, S., Seminiuk, D.H., 1998. The caring role. In: Sarah, R.
(Ed.), Folding Back the Shadows: A Perspective On Women’s
Mental Health. University of Otago Press, Dunedin.
Connelly, C., Davenport, Y., JI, N., 1982. Adherence to treatment
regimen in a lithium carbonate clinic. Arch. Gen. Psychiatry
39, 585–588.
157
Demers, R., Davis, L., 1971. The influence of prophylactic lithium
on the marital adjustment of manic-depressives and their
spouses. Comp. Psychiatry 12, 348–353.
Dinicola, V., 1989. The child’s predicament in families with a
mood disorder: research findings and family interventions.
Psychiat. Clin. N. Am. 12, 933–949.
Endicott, J., Sopitzer, R., 1978. A diagnostic interview: the
schedule for affective disorders and schizophrenia. Arch. Gen.
Psychiatry 35, 837–844.
Fadden, G., Bebbington, P., Kuipers, L., 1987. The burden of care:
the impact of functional psychiatric illness on the patient’s
family. Br. J. Psychiatry 150, 285–292.
Frank, E., Targum, S., Gershon, E., Anderson, C., Stewart, B.,
Davenport, Y., Ketchum, K., Kupfer, D., 1981. A comparison
of nonpatient with bipolar patient–well spouse couples. Am. J.
Psychiatry 138, 764–768.
Goldberg, D.P., Hillier, V.F., 1979. A scaled version of the General
Health Questionnaire. Psychol. Med. 9, 139–145.
Goodwin, F.K., Jamison, K.R., 1990. Manic-depressive Illness.
Oxford University Press, New York.
Grad, J., Sainbury, M., 1963a. Evaluating a community care
service. In: Freeman, H., Farndale, J. (Eds.), Trends in Mental
Health Services. Pergamon Press, Oxford.
Grad, J., Sainbury, M., 1963b. Mental illness and the family.
Lancet 1, 1.
Greene, B., Lustig, N., Lee, R., 1976. Marital therapy when one
spouse has a primary affective disorder. Am. J. Psychiatry 133,
827–830.
Hoenig, J., Hamilton, M., 1966. The schizophrenic patient in the
community and his effect on the household. Int. J. Soc.
Psychiatry 12, 165–176.
Hoenig, J., Hamilton, M., 1969. The Desegregation of the
Mentally Ill. Routledge and Keegan-Paul, London.
Janowsky, D., Leff, M., Epstein, R., 1970. Playing the manic
game; interpersonal maneuvers of the acutely manic patient.
Arch. Gen. Psychiatry 22, 252–261.
Johnston, R., 1983. A Revision of Socio-Economic Indices For
New Zealand. New Zealand Council for Educational Research,
Wellington.
Kessler, R.C., Walters, E.E., Forthofer, M.S., 1998. The social
consequences of psychiatric disorders. III. Probability of
marital stability. Am. J. Psychiatry 155, 1092–1096.
Lesser, A., 1983. Hypomania and marital conflict. Can. J. Psychi-
atry 28, 362–366.
McAlpin, C.A., Goodnick, P.J., 1998. Psychotherapy. In: Good-
nick, P.J. (Ed.), Mania; Clinical and Research Perspectives.
American Psychiatric Press, Washington, DC, pp. 363–381,
Chapter 17.
McPherson, H.M., Dore, G.M., Loan, P.A., Romans, S.E., 1992.
Socioeconomic status of a Dunedin sample of bipolar patients.
NZ Med. J. 105, 161–162.
McPherson, H.M., Romans, S.E., 1993. Life events and relapse in
established bipolar affective disorder. Br. J. Psychiatry 163,
381–385.
Miklowitz, D., Goldstein, M., Nuechterlein, K., Synder, K., Mintz,
J., 1988. Family factors and the course of bipolar affective
disorder. Arch. Gen. Psychiatry 45, 225–231.
158 G. Dore, S.E. Romans / Journal of Affective Disorders 67 (2001) 147 – 158
Murray, C.J.L., Lopez, A.D. (Eds.), 1996. The Global Burden of
Disease. A Comprehensive Assessment of the Mortality and
Disability From Diseases, Injuries and Risk Factors in 1990
and Projected To 2020. Harvard School of Public Health on
behalf of the World Health Organisation, and the World Bank,
Boston, MA.
Platt, S., 1985. Measuring the burden of psychiatric illness on the
family: an evaluation of some rating scales. Psychol. Med. 15,
383–393.
Post, R.M., 1992. Transduction of psychosocial stress into the
neurobiology of recurrent affective disorder. Am. J. Psychiatry
149, 1.
Priebe, S., Wildgrube, C., Muller-Oerlinghausen, B., 1989.
Lithium prophylaxis and expressed emotion. Br. J. Psychiatry
154, 396–399.
Romans, S.E., McPherson, H.M., 1992. The social networks of
bipolar affective disorder patients. J. Affect. Disord. 25, 221–
228.
Romans-Clarkson, S.E., Walton, V.A., Herbison, G.P., Mullen, P.E.,
1989. The validity of the GHQ-28 and its subscales in New
Zealand. Aust. NZ J. Psychiatry 23, 187–196.
Silverstone, T., Romans-Clarkson, S.E., 1989. Bipolar affective
disorder: the causes and prevention of relapse. Br. J. Psychiatry
154, 321–335.
Tantum, D., 1989. Familial factors in psychiatric disorder. Curr.
Opin. Psychiatry 2, 296–301.
Targum, S., Dibble, E., Davenport, Y., Gershon, E., 1981. The
family attitudes questionnaire: patients’ and spouses views of
bipolar illness. Arch. Gen. Psychiatry 38, 1.
Trippitelli, C.L., Jamison, K.R., Folstein, M.F., Bartko, J.J., De
Paulo, J.R., 1998. Pilot study on patients’ and spouses’
attitudes toward potential genetic testing for bipolar disorder.
Am. J. Psychiatry 155, 899–904.
Weeke, A., Billie, M., Videbech, T., Dupont, A., Juel-Nielson, N.,
1975. Incidence of depressive syndromes in Danish County.
The Aarhus County investigation. Acta Psychiat. Scand. 57,
28–41.