the relationship between condition-specific morbidity, social support and material deprivation in...

Pergamon Soc. Sci. Med. Vol. 45, No. 9, pp. 1325-1336, 1997 ~ 1997 Elsevier Science Ltd. All rights reserved

PII: S0277-9536(97)00059-2 Printed in Great Britain 0277-9536/97 $17.00 + 0.00

THE RELATIONSHIP BETWEEN CONDITION-SPECIFIC MORBIDITY, SOCIAL SUPPORT AND MATERIAL

DEPRIVATION IN PREGNANCY AND EARLY MOTHERHOOD

DEBORAH BAKER,* H A Z E L T A Y L O R and T H E A L S P A C S U R V E Y T E A M

Institute of Child Health, Royal Hospital for Sick Children, St Michaels Hill, Bristol BS2 8B J, U.K.

Abstract--Poorer health has been consistently associated with both low levels of social support and material deprivation. It has been proposed that social support constitutes a causal link between health and deprivation such that those with lower socio-economic status have poorer health because their lack of social support makes them more vulnerable to disease. This assumption was tested in this study for women moving from pregnancy to early motherhood. The sample of 9208 women was drawn from the Avon Longitudinal Study of Pregnancy and Childhood (ALSPAC). Health status was measured by self-report of morbidity for three contrasting conditions: backache, depression and urinary infection. Data were collected by self-completion questionnaire at eight weeks prepartum and at eight weeks postpartum. The sample was divided into four groups for each condition on the basis of identification of the condition (1) on both occasions, (2) on neither occasion, (3) at eight weeks prepartum only, and (4) at eight weeks postpartum only. Chi-square tests were used to measure the association between presence or absence of a condition as defined above, material deprivation and low social support. Responses on an eight item social support questionnaire tapping emotional, instrumental and communal aspects of perceived social support were compared between these groups for each condition. Results showed significant associations, in late pregnancy and early motherhood, between poorer health and both material deprivation and low social support for two of the three conditions, depression and urinary infection. Changes in levels of perceived social support occurred as a consequence of motherhood; the percentage of women perceiving their partner to be supportive decreased as did the percentage of those who felt that their family or friends would help with financial difficulties; in contrast, the percentage of those perceiving that other mothers/neighbours offered support increased from late pregnancy to early motherhood. The extent and direction of such change was consistently associated with the presence or absence of depression, but not with backache or urinary infection. When mental health "improved", in that depression was present pre- but not postpartum, the percentage feeling supported increased and vice versa. This pattern was in evidence across all items of social support. It was concluded that perceived social support was unlikely in itself to constitute a causal link between poorer health and lower socio-economic status. In this context the relationship between social support and depression requires further consideration, particularly the extent to which they constitute component parts of a more general feeling of emotional well-being, a construct which is, in its turn, in need of further articulation in relation to health outcomes. © 1997 Elsevier Science Ltd

Key words--social support, material deprivation, maternal health, pregnancy

The lack of a s t rong ne twork of social relat ionships

has been located as a major risk factor for heal th

with implicat ions as serious as those for cigarette

smoking and high blood pressure. Such risk factors

are often clustered in popu la t ion groups with lower

socio-economic status and this has led to the hy-

pothesis tha t social suppor t const i tutes a causal link

between pover ty and poore r heal th, part icularly in

cases when a disease is stress related ( M a r m o t

et al., 1984; Wilkinson, 1992; Davey-Smith et al.,

1994). This assumpt ion is buil t on the premise tha t

social suppor t acts as a "buffer" , by decreasing vul-

nerabil i ty to stress and increasing host resistance to

disease (Cassell, 1976). More d isadvantaged groups

*Author for correspondence.

have lower levels of social suppor t to act as a "buf-

fer" against stress and thus they are more vulnerable to disease (House et al., 1988; Berkman, 1984). Support ive relat ionships are likely to be part icularly impor t an t in relat ion to heal th at times when ad- di t ional emot ional and physical demands are made, for example, women coping with the new experience of pregnancy or early mothe rhood , or caring for a new baby in conjunct ion with other siblings (O 'Hara , 1986; Gjerdingen et al., 1991). The study reported in this paper was concerned with testing the assumpt ion tha t social suppor t is a de te rminan t of var ia t ion in menta l and physical illness between different socio-economic groups in the t rans i t ion to mothe rhood .

There are a n u m b e r of reasons why this assump- t ion could be regarded as content ious. The first is

1325

1326 Deborah Baker et al.

that the influence of social relationships on health is somewhat dependent on how this is conceptualised and measured. Social support implies a focus on the quality of social relationships, that is the extent to which they are perceived as supportive, rather than the quantity of social contacts which reflects the extent of a person's social network or its density (O'Reilly, 1988; Orth-Gom6r and Und6n, 1987; Broadhead et al., 1983). This is an important dis- tinction since these two aspects of social relationships have a different relationship to health; whilst it is generally acknowledged that the perception of being supported is universally associated with better health, the quantity of social relationships can be a positive or negative influence on health status. For example, Wilcox (Wilcox, 1981) examined the relationship between mental health and the quality and quantity of social network ties in a group of newly divorced middle class mothers. Those with low levels of depression and high levels of life satisfaction reported receiving higher levels of emotional support than others. However network density, highly reflective of extensive family ties was a positive correlate of depression and low life satisfaction: women with extensive social contacts experienced greater distress than others. Wilcox concluded that kin networks may well provide material support and assistance with childcare, necessities for single parent families, but may exact many costs including interference with the parenting role and personal life.

Secondly, different types of social support have a greater or lesser impact on health and are likely to be invested with varying degrees of importance across the lifespan (Kaplan et al., 1977; Jacobson, 1986). In pregnancy, but particularly in early motherhood, emotional support from the partner and practical help with childcare and household tasks are strongly associated with better mental health. (Power and Parke, 1984; Brown and Harris, 1978; D'Arcy and Siddique, 1984). Mothers who find such support lacking tend to draw heavily on other sources of support. For example, Brown (1996), in a longitudinal study of marital quality and social support in the face of a severe life event, found that as support from the partner fell with worsening quality of marriage, so there was often a steep rise in support from a "very close other" such as mother or friend. Komarovsky and Bott (Komarovsky, 1967; Bott, 1971) also found that mothers lacking support from their partners tended to seek advice relating to childcare and other domestic matters from external support networks.

Thirdly, the generalisability of the assertion that it is psychosocial factors such as lack of social support that explain the increased vulnerability of lower socio-economic groups to stress related illness is questionable. This hypothesis has been largely developed in cross sectional population studies on mortality from conditions such as coronary heart

disease. Studies that focus on stress related morbidity such as depression have also demonstrated associations between material deprivation and low social support and low social support and depression (Oakley et al., 1994; Brown et al., 1986; Holahan and Moos, 1981; Lin et al., 1986; Daigard et al., 1995), but a causal link between low social support and mental health is not the only possible explanation. Henderson (1984), in his critique of studies focusing on the relationship between social support and stress related conditions such as depression, points towards the possibility that perceived social support is a component of general well-being so that feeling depressed may simply mean that one also sees oneself as less supported. Other studies examining physical as well as mental health have suggested that socio-economic status and social support are independently related to health status. For example, Blaxter (1990), using material from the Health and Lifestyle Survey, found that income level and perceived social support were each related to health status and that this relationship varied across the lifespan. For men and women >40 social support made a significant difference to reported illness only if income was low; for men < 40 income made no difference to reported illness if there was no lack of support. For women <40 social support had a greater effect than income on health. A similar conclusion was drawn by Oakley and her colleagues in their study of the relationship between social support and health for materially disadvantaged mothers in their first year postpartum (Oakley et al., 1994); she found that the strong association between high levels of social support and positive health outcomes could not be explained by the association between social support and social class.

With these three areas of contention in mind, the study below utilised prospective population data to examine the prevalence of three conditions, depression, backache and urinary infection and change in prevalence from pregnancy to early motherhood. Changes in these mental and physical conditions were related to different types of social support and socio-economic status. The following hypotheses were tested:

1. that poorer health will be associated with material deprivation and with lower levels of social support, irrespective of the source of social support or the condition under study; and

2. that improvement or deterioration in health from pregnancy to early motherhood will be associated with increasing or decreasing levels of social support.

METHODS

The sample used was the Avon Longitudinal Study of Pregnancy and Childhood (ALSPAC) for

Maternal morbidity, social support and deprivation 1327

which all women having a baby between April 1991 and December 1992 in the three Health Districts of Bristol were approached to invite their partici- pation. Aspects of their health and the health of their partners and children have been and will continue to be monitored during pregnancy and early motherhood and at yearly intervals from eight months postpartum. Information from self-completion questionnaires administered in pregnancy and at eight weeks postpartum was used for this particular study for a sample of 9208 women.

The measurement o f social support

Social support was measured using a 10 item scale specifically devised for women in the puerperium. The items present statements relating to emotional, instrumental and financial support to which the respondent is offered four response cat- egories "I never feel this way" through to "this is exactly how I feel". Eight of the 10 items were selected as relevant for use in this study. Two related to general feelings of emotional support and were "I have no-one to share my feelings with" and "there is always someone with whom I can share my happiness and excitement about the baby". Two were concerned with emotional and instrumental support from the partner and they were "my partner provides the emotional support I need" and "if I feel tired I can rely on my partner to take over". One concerned emotional support from other mothers--"there are other pregnant women/ mothers with whom I can share my experiences" and a further three tapped instrumental and financial support from family, friends and neighbours: "'I believe in moments of difficulty my neighbours would help me", "if I was in financial difficulty I know my family would help if they could" and "if I was in financial difficulty I know my friends would help if they could". Data from these items of social support were analysed from the questionnaires com- pleted in late pregnancy and at eight weeks postpartum. Those women who reported that they did not have a partner were excluded from the analysis of the items concerning social support from partner.

with childbearing. It is a 10 item psychometric rating scale, each item having four responses from 0 to 3 with a minimum score of 0 and a maximum of 30. While scores on the EPDS do not constitute a continuous scale they can be categorised by well validated cutoff points; those scoring < 10 are classified as "not depressed"; scores between 10 and 12 are used to indicate the presence of mild depressive symptoms and scores over 12 to indicate the likelihood of a major depressive disorder. It has been shown to have a sensitivity of 86% and a specificity of 78% in diagnosing depression according to Research Diagnostic Criteria when the cutoff point of > 12 is used (Spitzer et al., 1978). Even better sensitivities and specificity's (95% and 93%, respectively) have been achieved by using EPDS scores of greater than 12 and comparing the result with clinical interview, using the DSM III criteria (Harris et al., 1989). It has also been validated for identification of depression in pregnancy (Murray and Cox, 1990). In this study the sample was divided into two groups using this measure, those who were not depressed (< 10 on the EPDS) and those who were identified as having mild or severe symptoms of depression (>10 on the EPDS). For backache and urinary infection women were asked at eight weeks prepartum whether they had experienced the condition in the last three months and at eight weeks postpartum they were asked whether the condition had occurred since the birth of the baby. Once again women were divided on the basis of this information into those who had and those who had not experienced a condition on either occasion.

The measurement of material deprivation

Material deprivation was measured using the household indicators of relative deprivation of tenure and use of car. These were chosen as measures of deprivation in preference to social class on the basis of the well documented biases in occu- pation based classification of women's social and economic status (Moser et al., 1990)

The measurement of health status

Morbidity for three contrasting conditions was also compared at eight weeks prepartum and at eight weeks postpartum. These were depression, backache a common problem associated with pregnancy and childbirth and urinary infection an acute condition frequently experienced by women across the lifespan. The Edinburgh Postnatal Depression Scale (EPDS) (Cox et al., 1987) was used to identify the presence or absence of depression. This was developed specifically for puerperal women exclud- ing as it does somatic symptoms of psychiatric disorder such as sleep disturbance, fatigue, change in appetite and insomnia as these symptoms are often caused by normal physiological changes associated

ANALYSIS

To examine the relationship between health status, social support and socio-economic status for each of the three conditions respondents were placed in one of four groups according to their health status at eight weeks prepartum and eight weeks postpartum. These groups were comprised of:

(a) those who did not have the condition at either eight weeks prepartum or eight weeks postpartum; this was taken as indicative of "best health";

(b) those who reported the condition on both occasions; this was taken as indicative of

1328 D e b o r a h Baker e t al.

Table I. The relationship between the presence or absence of depression, backache or urinary infection and socio-economic circumstances

No symptom on either Symptom on both Symptom prepartum only Symptom postpartum occasion occasions only

No % No % No % No % 4780 (83.0) 781 (67.8) 1022 (72.0) 552 (78.3)

Depression Home ownership Mortgage/ status owned

Council/HA Rented/

other

x~

Use of a car Yes No

Backache Home ownership Mortgage/ status owned

Council/HA Rented/

other

z~

Use of a car Yes No

x~ Urinary infection Home ownership Mortgage/ status owned

Council/HA Rented/

other

x~

Use of a car Yes No

548 (9.5) 245 (21.3) 231 (16.3) 91 (12.9) 433 (7.5) 126 (10.9) 167 (11.8) 62 (8.8)

(100.0) (100.0) (1002) (100.0) 157.0 **** 89.5 **** 10.3 **

5436 (94.3) 984 (85.4) 1258 (88.7) 642 (90.7) 329 (5.7) 168 (14,6) 160 (11.3) 66 (9.3)

(100.0) (100.0) (100.0) (100.0) 113.4 **** 55.8 **** 14.4 ***

677 (82.4) 4001 (78.3) 1599 (76.9) 840 (83.9)

84 (10.2) 646 (12.6) 287 (13.8) 91 (9.1) 61 (7.4) 462 (9.0) 192 (9.2) 70 (7.0)

(100.0) (100.0) (100.0) (100.0) 7.0 * 10.4 ** 0.8 - -

4673 (91.4) 1902 (91.7) 954 442 (8.6) 172 (8,3) 49

(100.0) (100.0) 5 * 3.2 - - 1.8

52 (71.2) 298 (64.6) 437 (76.5)

12 (16.4) 101 (21.9) 71 (12,4) 9 (12.3) 62 (13.4) 63 (11,0)

(1oo.o) (lOO.O) (1oo.o) --- 64.9 **** 6.2 *

(84.9) 394 (85.7) 513 (89.8) (15.1) 66 (14.3) 58 (10.2)

(100.0) (lO0.O) (lO0.O) • 30.8 **** 6.4 *

771 (93.7) 52 (6.3)

(100.0)

6321 (80.1)

918 (11.6) 648 (8.2)

(1oo.o) 3.6

7320 (92.7) 62 574 (7.3) 11

(100.0) 6.5

(95.1) (4.9)

(100.0)

*P < 0.05; **P < 0.01; ***P < 0.001; ****P < 0.0001.

"worst health" although no assumptions have been made about chronicity;

(c) those who reported the condition at eight weeks prepartum only; and

(d) those who reported the condition at eight weeks postpartum only.

Chi-squared tests were used to compare the socio-economic characteristics of women in each of the three groups for which a condition was present [(b)-(d)] with the characteristics of those women for whom there was no report of a condition (a). These comparisons were carried out separately for depression, backache and urinary infection. A similar analysis was used to compare group differences in the perception of social support. For this purpose social support scores for each of the groups [(a)- (d)] were categorised into quartiles; chi-squared tests were then used to compare the percentage of those falling into the lowest quartile (i.e. those who perceived themselves to be least supported) in group (a) with those in groups (b), (c) and (d). These analyses were performed separately for each

condition and for social support data collected in pregnancy and at eight weeks postpartum.

To examine change in health status in relation to change in the perception of social support from pregnancy to early motherhood the percentage of the sample endorsing the response category "this is exactly how I feel" was recorded in late pregnancy and at eight weeks postpartum for each of the eight selected items from the social support scale for groups (a), (b), (c) and (d). The percentage of responses in this category in pregnancy and at eight weeks postpartum were compared for each item by McNemars Test. The difference in the percentage of responses were also calculated with 95% confidence intervals (Gardner and Altman, 1989). These analyses were performed separately for each condition. For the purpose of descriptive comparison the results for those not reporting a condition on either occasion [group (a)] were used as the "gold" stan- dard against which to compare changing patterns of social support for those groups who reported a condition either pre- or postpartum or on both occasions [groups (b), (c) and (d)].

Maternal morbidity, social support and deprivation

Table 2. The relationship between the presence or absence of depression, backache or urinary infection and low social support

1329

Women with lowest social support scores (~<25th centile) Scores in pregnancy Scores at eight weeks postpartum

Depression No % X 2 No % X~

Never 1013 17.5 916 15.8 Prepartum only 497 34.7 205.9**** 437 30.5 163.3"*** Postpartum only 208 29.1 55.7**** 314 43.7 326.6**** Both occasions 540 46.4 466.2**** 678 58.2 992.0**** Backache Never 157 19.2 154 18.6 Prepartum only 531 25.3 12.1"** 509 24.2 10.7"* Postpartum only 198 19.9 0.1- 209 20.9 1.5- Both occasions 1355 26.2 18.3"*** 1460 28.3 33.8**** Urinary infection Never 1904 24 1960 24.6 Prepartum only 139 29.6 7.5** 150 31.6 11.7"** Postpartum only 158 27.6 3.7- 178 31.1 11.7"** Both occasions 29 39.7 9.8** 33 45.8 17.6"***

*P < 0.05; **P < 0.01; ***P < 0,001; ****P < 0.0001.

RESULTS

Health status and deprivation

Table 1 shows that the presence of both depression and urinary infection was consistently associated with relative deprivation in the sample selected for analysis. Those reporting this symptom at eight weeks prepartum or at eight weeks postpartum or on both occasions were significantly more likely to live in rented accomodation and significantly less likely to have use of a car when compared to women who did not report the condition on either occasion. For backache women reporting the symptom at eight weeks prepartum or on both occasions were similarly significantly more likely to live in rented accomodation and significantly less likely to have use of a car when compared with women who did not report the condition on either occasion. However there were no significant differences in levels of deprivation between the latter group and those reporting backache at eight weeks postpartum only.

Health status and social support

Table 2 shows the results when chi-squared tests were used to compare scores on the social support scale for those women who reported the presence of a symptom with those who did not. For all conditions women who reported a symptom either at eight weeks prepartum or on both occasions were significantly more likely to record low social support scores than women who did not report symptoms either pre- or postpartum. This was the case for social support scores collected at both eight weeks prepartum and at eight weeks postpartum. There was some variation in the results for each condition when comparisons were made between those reporting symptoms at eight weeks postpartum only and those who did not report the symptom either pre- or postpartum. For depression and

urinary infection those women who had the condition at eight weeks postpartum only were significantly more likely to have low social support scores than those women who did not have the condition on either occasion (X2 = 326.6, df = 1, P < 0.0001, X2= l l.7, d f = 1, P = < 0.001, respectively). In contrast there were no significant differences in social support scores for women reporting backache at eight weeks postpartum only when compared with those who did not report backache on either occasion.

Changes in health status and changes in perceived social support from pregnancy to early motherhood

Table 3 presents changes in the eight selected items of social support from pregnancy to early motherhood for women who did not report a condition on either occasion. These women were thus neither suffering from poor health, or as the results above demonstrate, experiencing relative deprivation. The table shows that the perception of social support changed from pregnancy to early motherhood and that the extent and direction of change varied considerably according to the item of support being considered. For example, the percentage of women feeling that their partner gave them the support they required fell significantly pre- to postpartum for the item "if I feel tired I can rely on my partner to take over" so that for women who did not report backache 52% rated this item as "this is exactly how I feel" in late pregnancy compared with 45.1% at eight weeks postpartum (change in % = - 6 . 9 1 , 95% CIs = - 1 0 . 6 4 , -3.19). Other items showing significant decreases pre- to postpartum were "if I was in financial difficulty I know my family would help if they could" and "if I was in financial difficulty I know my friends would help if they could". In contrast for other aspects of social support, particularly emotional and instrumental support from other


Table 3. Changes in perceived social support from pregnancy to early motherhood for those women not reporting symptoms on either occasion

Rating of "exactly how I feel" on social support item No in Eight weeks Eight weeks Difference % Lower Upper McNemars

Items of social support category prepartum %~' postpartum %b pre- postpartum 95% CI 95"/0 CI Test P-value

I have no one to share my feelings with a No depression pre- or postpartum 5738 78.4 78.9 0.44 -0.75 No backache pre- or postpartum 809 77.9 76.9 -0.99 -3.99 No urinary infection pre- or postpartum 7860 70.3 69.1 -1.23 -2.33

There is always someone with whom I can share my happiness and excitement about the baby No depression pre- or postpartum No backache pre- or postpartum No urinary infection pre- or postpartum

No depression pre- or postpartum No backache pre- or postpartum No urinary infection pre- or postpartum

No depression pre- or postpartum No backache pre- or postpartum No urinary infection pre- or postpartum

1.62 2.02

-0.14

5677 65.8 72.3 6.52 5.16 7.87 795 64.0 68.9 4.91 1.24 8.57

7783 58.7 63,6 4.84 3.66 6.03 Myparmer provides the emotionalsupport I need 5572 49.7 51.0 1.24 -0.16 2.64 777 49.7 47,5 -2.19 -5.81 1.44

7583 43.7 42.8 -0.95 -2.12 0.22 l f l f e e l t ~ e d 1 can rely on my partner to take over

5566 50.1 45.4 -4.73 -6.12 -3.33 781 52.0 45.1 -6.91 -10.64 -3.19

7576 45.4 39.9 -5.45 -6.64 --4.26 There are other pregnant women~mothers with whom I can share my experiences

9.13 11.97 5.78 13.39 7.07 9.42

7.93 10.57 3.90 10.98 6.56 8.74

-6.50 --4.15 -10.14 -3.79 -8.50 -6.38

-9.88 -7.24 -14.91 -8.15 -9.69 -7.50

No depression pre- or postpartum 5650 26.6 37.1 10.55 No backache pre- or postpartum 793 25.2 34.8 9.58 No urinary infection pre- or postpartum 7752 23.6 31.8 8.24

I believe in moments o f difficulty my neighbours wouM help me No depression pre- or postpartum 5673 26.8 36.0 9.25 No backache pre- or postpartum 793 28.6 36.1 7.44 No urinary infection pre- or postpartum 7778 24.1 31.7 7.65

I f I was in financial difficulty 1 know my family would help i f they could No depression pre- or postpartum 5686 80.0 74.6 -5.33 No backache pre- or postpartum 804 78.5 71.5 -6.97 No urinary infection pre- or postpartum 7786 76.0 68.5 -7.44

l f l was in financial difficult), I know my friends would help i f they could No depression pre- or postpartum 5644 40.8 32.2 --8.56 No backache pre- or postpartum 798 41.6 30.1 -11.53 No urinary infection pre- or postpartum 7750 36.6 28.0 -8.59

NS NS

NS NS NS

aFor this item the values recorded were for the % reporting "I b% reporting "This is exactly how 1 feel". *P < 0.05; **P < 0.01; ***P < 0.001: ****P < 0.0001.

never feel this way";

mothers, friends or neighbours, the percentage of

women feeling supported increased from late preg-

nancy to early motherhood. For example, for

women who did not report depression on either oc-

casion the percentage of those who rated the cat-

egory "this is exactly how I feel" in response to the

item "there are other pregnant women/mothers

with whom I can share my experiences" increased

from 26.6% in late pregnancy to 37.1% at eight

weeks pos tpar tum and this increase was significant

(change in % = 10.6, 95% CIs = 9.1, 12.0). Similar

pat terns were identified for the items "I believe in

moments of difficulty my neighbours would help

me" and "there is always someone with whom I

can share my happiness and excitement about the

baby".

For the purposes of clarity four social support

items have been selected from those used in this

analysis to illustrate changes in health status as they

related to these patterns of changing social support.

The four chosen were those that reflected the con-

trast in perception of support from partner ("my

par tner provides the emotional support I need",

"when I am tired I can rely on my partner to take over") when compared with support from other

mothers, friends and neighbours ("there are other pregnant women/ mothers with whom I can share

my experiences", "I believe in moments of difficulty my neighbours would help me").

D e p r e s s i o n ( T a b l e 4 )

For depression there was a similar pat tern across all four items between the presence or absence of symptoms and the percentage of those reporting high levels o f social support . If depression was present, then the percentage of those feeling supported was lower; if depression was absent, then the percentage of those feeling supported was higher. This pat tern is well illustrated by change from pre- to pos tpar tum in the percentage of those responding "this is exactly how I feel" to the item "there are other pregnant women/mothers with whom I can share my experiences" (Fig. 1). As reported above for women who were not depressed on either occasion the percentage responding in this way was highest of all the four groups (26.6) and significantly increased pre- to postpar tum (change in % = 10.6, 95% CIs = 9.1, 12.0). Fo r women report ing depression at eight weeks prepar tum only the percentage responding "this is exactly how I feel" was lower by comparison (18.7%) but significantly increased pre- to pos tpar tum (change in % = 7.9, 95% CIs = 5.1, 10.6). For those with depression at eight weeks pos tpar tum only the percentage responding "this is exactly how I feel" was

Maternal morbidity , social support and deprivat ion 1331

Table 4. Changes in four selected items of perceived social suport from pregnancy to early motherhood in relation to the presence or absence of depression

Rating of "exactly how 1 feel" on social support item Difference %

No in Eight weeks Eight weeks pre- Lower Upper McNemars Item of support category prepartum %a postpartum%a postpartum 95% CI 95% CI Test P-value

My partner provides the emotional support I need No depression either pre- 5572 49.7 51.0 1,24 -0.16 2.64 NS or postpartum Depression at eight weeks 1335 34.0 35.7 1.72 -1.05 4.50 NS prepartum only Depression at eight weeks 691 36.6 24.7 -11.87 - 15.70 -8 .04 **** postpartum only Depression at both eight 1071 25.3 16.2 -9.06 -11.77 -6 .34 **** weeks pre- and eight weeks postpartum

l f l feel tired I can rely o n my partner to take over No depression either pre- 5566 50. l 45.4 -4.73 -6.12 -3.33 **** or postpartum Depression at eight weeks 1339 36.8 35.0 -1.79 --4.61 1.02 NS prepartum only Depression at eight weeks 689 41.7 27.6 -14.08 -18.05 -10.10 **** postpartum only Depression at both eight 1068 31.1 21.9 -9.18 -12.22 -6.13 **** weeks pre- and eight weeks postpartum

There are other pregnant women/mothers with whom I can share my experiences No depression either pre- 5650 26.6 37.1 10.55 9.13 11,97 **** or postpartum Depression at eight weeks 1398 18.7 26.6 7.87 5.12 10,61 **** prepartum only Depression at eight weeks 701 23.5 22.7 -0.86 -4.63 2.92 NS postpartum only Depression at both eight 1130 12.9 12.7 -0.27 -2 .76 2.23 NS weeks pre- and eight weeks postpartum

I believe in moments of difficulty my neighbours would help me No depression either pre- 5673 26.8 36.0 9.25 7.93 10.57 **** or postpartum Depression at eight weeks 1395 18.0 27.8 9.82 7.38 12.26 **** prepartum only Depression at eight weeks 704 21.2 22.9 1.70 -1.79 5.20 NS postpartum only Depression at both eight 1143 16.4 17.1 0.61 -1.82 3.04 NS weeks pre- and eight weeks postpartum

~% reporting "'This is exactly how I feel"; *P < 0.05; **P < 0.01; ***P < 0.001; ****P < 0.0001.

relatively high in late pregnancy (23.5%) but did not significantly increase pre- to postpartum, in fact there was a slight decrease (change in % = - 0 . 9 , 95% CIs = - 4 . 6 , 2.9). For those women who reported the condition on both occasions the percentage responding "this is exactly how I feel" was the lowest of all four groups in late pregnancy (12.9%) and there was no significant increase in this percentage by eight weeks postpartum (change in % = - 0 . 3 , 9 5 % CIs = - 2 . 8 , 2 . 2 ) .

Backache (Table 5)

In contrast for backache whilst the percentage feeling supported was associated with the presence or the absence of the condition, the direction or the extent of change in perceived social support from late pregnancy to early motherhood was not associated with changes in health status. And so, as Fig. 2 illustrates, the percentage of those responding "this

is exactly how I feel" to the item "there are other pregnant women/mothers with whom I can share my experiences" was highest for those not reporting the condition on either occasion (25.2%) and those reporting backache at eight weeks postpartum only (25.2%) and lowest for those reporting backache at eight weeks prepartum only (24.4%) and on both occasions (22.3%).

There was a significant increase in the percentage responding "this is exactly how I feel" on this item for all four groups from late pregnancy to early motherhood and the relative levels of perceived social support between these groups remained the same.

Urinary infection (Table 6)

For urinary infection there was little consistency across items in either the relationship between the presence or absence of the condition and the percentage feeling supported or the extent and direc-


40 O

35

30

o 25

"There are other pregnant women/mothers with whom I can share my experiences"

.~ 20 ~--

15 z

10 • ~. 8 weeks 8 weeks

prepartum postpartum

No depression pre- or postpartum • Depression at 8 weeks postpartum only • Depression at 8 weeks prepartum only • Depression both pre- and postpartum

Fig. 1. Changes in perceived social support in relation to the presence or absence of depression.

tion of change in perceived social support pre- to

postpartum. For example, Fig. 3 shows that the

percentage of women who responded "this is

exactly how I feel" to the item "there are other

pregnant women/mothers with whom I can share

my experiences" was highest for those women who

reported the condition on both occasions (24.6%).

There was no significant increase in the percentage

of those responding "this is exactly how I feel"

from late pregnancy to early motherhood for those

reporting the condition at eight weeks prepartum

(change in % = 4.4, 95% CIs = - 0 . 5 , 9.2) and yet

there was a significant increase in this percentage

for those reporting this condit ion at eight weeks

postpartum (change in % = 6.3, 95% CIs = 2.1, 10.6).

40

-~ 35

30

25

.~ 20

15

10

"There are other pregnant women/mothers with whom I can share my experiences"

= 8 weeks 8 weeks o prepartum postpartum

.o ~ No backache pre- or postpartum

• Backache at 8 weeks postpartum only • Backache at 8 weeks prepartum only • Backache both pre- and postpartum

Fig. 2. Changes in perceived social support in relation to the presence or absence of backache.

DISCUSSION

For two of the three conditions s tudied--depression and urinary infec t ion--a positive association between poorer health and both relative deprivation and low social support was confirmed. When compared with women who did not report the symptom at either eight weeks prepartum or eight weeks postpartum (those with the "best health") those who reported symptoms at eight weeks prepartum only, eight weeks postpartum only or on both occasions were significantly more likely to live in rented accomodation, not to have use of a car and to be less likely to see themselves as being supported.

For backache the pattern was the same, with the exception of women reporting the condition at eight weeks postpartum. For these women there was no significant difference in either levels of deprivation or perception of social support when compared with those women who did not report the condition on either occasion. Evidence suggests that this could be due to the stronger influence of other factors in predicting backache at this time. MacArthur (MacArthur et al., 1991) in her study of health after childbirth found that primiparity and epidural anaesthesia were the primary predictors o f backache postpartum and that socio-economic characterstics were secondary. Similarly our own study of maternal morbidity at eight months postpartum (Baker et al., 1997) identified primiparity as an important factor associated with the higher self-report of backache at this time, which was not associated with socio-economic characteristics.

The results for backache illustrate the fact that the strength of the relationship between health status, relative deprivation and social support can change from one time point to the next and that other factors may have a more important influence on health in particular contexts.

Whilst the findings of this study in general confirmed the relationship between poorer health and both material deprivation and lower social support, it appeared unlikely that levels of perceived social support were simply the consequence of either socio-economic conditions or health status. In this regard results suggested that changes in patterns o f perceived social support were triggered by the event of motherhood itself and that the nature of these changes had little to do with either poorer health or relative deprivation. For those with the "best health", who were also relatively affluent, the percentage feeling supported by their partners never- theless decreased from late pregnancy to early motherhood, as did the percentage of those feeling that they could rely on family and friends for financial support; in contrast the percentage of those deriving support from other mothers and from neighbours increased. There are two possible explanations for this, both of which relate to the event of


Table 5. Changes in four selected items of perceived social support from pregnancy to early motherhood in relationship to the presence or absence of backache

Rating of "exactly how I feel" on social support item Eight weeks Difference %

No in Eight weeks postpartum pre- Lower Upper McNemars Item of support category prepartum %a %a postpartum 95% CI 95% CI Test P-value

My partner provides the emotional support I need No backache either pre- or 777 49.7 47.5 -2.19 -5.81 1.44 NS postpartum Backache at eight weeks 1980 44.6 45.4 0.76 - 1.55 3.07 NS prepartum only Backache at eight weeks 942 48.6 45.1 -3.50 -6.97 -0.03 * postpartum only Backache at both eight weeks 4943 40.9 39.8 -1.09 -2.54 0.35 NS pre- and eight weeks postpartum

I l l feel tired I can rely on my partner to take over No backache either pre- or 781 52.0 45.1 -6.91 -10.64 -3.19 *** postpartum Backache at eight weeks 1977 44.4 42.7 -1.62 -3.93 0.69 NS prepartum only Backache at eight weeks 944 47.7 40.6 -7.10 -10.70 -3.50 *** postpartum only Backache at both eight weeks 4933 43.8 37.2 -6.63 -8.09 -5.17 **** pre- and eight weeks postpartum

There are other pregnant women~mothers with whom I can share my experiences No backache either pre- or 793 25.2 34,8 9.58 5.78 13,39 **** postpartum Backache at eight weeks 2048 24.4 32A 7.67 5.31 10.02 **** prepartum only Backache at eight weeks 960 25.2 34.0 8.75 5.37 12.13 ** ** postpartum only Backache at both eight weeks 5048 22.3 29.8 7.55 6.13 8.97 **** pre- and eight weeks postpartum

I believe in moments ofdi~iculty my neighbours wouM help me No backache either pre- or 793 28.6 36,1 7.44 3,90 10.98 *** postpartum Backache at eight weeks 2052 23.7 32.8 9.16 7.04 11.28 **** prepartum only Backache at eight weeks 971 25.5 35.3 9.78 6.61 12.96 **** postpartum only Backache at both eight weeks 5069 22.6 29.2 6.53 5,20 7.86 **** pre- and eight weeks postpartum

~'% reporting "'This is exactly how I feel"; *P < 0,05; **P < 0.01; ***P < 0,001; ****P < 0.0001.

motherhood itself: one is that as a reaction to lack of support from their partner with the task of caring for a new baby, women turn to other forms of support outside the home; or these changes may reflect structural changes in social network that occur as a consequence of motherhood, whereby women make a new circle of friends, from whom they derive support that is directly relevant to the experience of caring for a baby (Komarovsky, 1967; Bott, 1971; Baker, 1989).

It was also clear that if social support does act to reduce vulnerability to disease, then in this study it had a much more potent effect on mental as oppose to physical health. Changes in perceived social support were most consistently associated with changes in depression. The relationship between social support and depression suggested that when mental health "improved" in that depression was present pre- but not postpartum, the percentage of those feeling supported increased and vice versa. This pattern was in evidence across all items of social sup-

port. In contrast there was little evidence that women became any more or less vulnerable to a chronic physical condition such as backache or an acute condition such as urinary infection as a consequence of increasing or decreasing levels of social support, It would thus be plausible to suggest that if lack of social support is associated with the onset of depression (Brown e t al. , 1986), it could be that low social support gives rise to depression which in its turn underlies the relationship between relative deprivation and vulnerability to stress related ill- nesses such as coronary heart disease. But an alternative explanation is that social support and depression are component parts of a more general feeling of emotional well-being and cannot be regarded as distinct from one another. This would raise serious questions about the meaning of the extensive literature using social support and depression as independent variables in "risk factor" analysis.

1334 D e b o r a h B a k e r et al.

Table 6. Changes in four selected items of perceived social support from late pregnancy to early motherhood in relation to the presence or absence of urinary infection

Rating of "exactly how I feel" on social support item Eight weeks Eight weeks Difference %

No in prepartum postpartum pre- Lower Upper McNemars Item of support category %~ %a postpartum 95% CI 95% CI Test P-value

My partner provides the emotional support I need No urinary infection either pre- or 7583 43.7 42.8 -0.95 -2.12 postpartum Urinary infection at eight weeks 435 38.6 37.0 -1.61 -6.68 prepartum only Urinary infection at eight weeks 545 42.6 41.7 -0.92 -5 .54 postpartum only Urinary infection at both eight weeks 66 40.9 31.8 -9.09 -22.20 pre- and eight weeks postpartum

I f 1 feel tired I can rely on my partner to take over No urinary infection either pre- or 7576 45.4 39.9 -5.45 -6.64 postpartum Urinary infection at eight weeks 436 43.8 37.6 -6.19 -11.14 prepartum only Urinary infection at eight weeks 544 45.4 37.3 -8.09 -12.74 postpartum only Urinary infection at both eight weeks 66 28.8 28.8 0.00 -11.40 pre- and eight weeks postpartum

There are other pregnant women~mothers with whom I can share my experiences No urinary infection either pre- or 7752 23.6 31.8 8.24 7.07 postpartum Urinary infection at eight weeks 460 20.9 25.2 4.35 -0 .46 prepartum only Urinary infection at eight weeks 554 22.9 29.2 6.32 2.06 postpartum only Urinary infection at both eight weeks 69 24.6 24.6 0.00 -12.70 pre- and eight weeks postpartum

I believe in moments ofdi~culty In), neighbours would help me No urinary infection either pre-or 7778 24.1 31.7 7.65 6.56 postpartum Urinary infection at eight weeks 461 19.3 26.5 7.16 2.57 prepartum only Urinary infection at eight weeks 562 22.8 30.1 7.30 3.57 postpartum only Urinary infection at both eight weeks 71 18.3 21.1 2.82 -7 .55 pre- and eight weeks postpartum

0.22 NS

3.47 NS

3.70 NS

4.01 NS

-4 .26 ****

-1 .24 *

-3 .44 **e,

- 11.40 NS

9,42 ****

9,15 NS

10.57 **

12,70 NS

8,74 ****

11.75 **

11.02 ***

11.80 N S

"% reporting "This is exactly how I feel"; *P < 0.05; **P < 0.01; ***P < 0.001; ****P < 0.0001.

It could be argued that measures of "perceived social support" should be validated by more objective measurement to add weight to the kind of evidence presented in this paper. But, as was pointed out in the introduction, the construct of social support is concerned with feelings about being supported and it is this that is so closely associated with good or poor health. Validity would thus not necessarily be established by objective measurement such as the number and intensity of social contacts, since people may feel supported by only one family member or friend, or the same person may be supportive at one time point, but not at another. Objective measurement would have to take account of the main source of social support, the quality of the supportive relationship and its constancy over time. Brown (1996) has used such techniques to good effect in his illuminating small scale longitudinal study of social support and its relationship to the onset of depression; one conclusion he drew from this study was that perceived social support was a good approximation for his detailed measures of actual social support. Whilst such measures are impractical for large scale population studies, they

are likely to provide additional insight into the mechanisms linking the quality of social support to mental and physical health outcomes.

C O N C L U S I O N S

The results presented in this paper are not condu- cive to the generalised assumption that social support is a protective factor that ensures the better health of more affluent groups by decreasing vulnerability to disease. "Social support" sits uneasily amongst the configuration of social, biological and environmental risk factors that are conventionally used to explain heightened morbidity and mortality for groups with lower socio-economic status. This study found that poorer health was associated with both material deprivation and lower social support for two out of the three conditions studied. Changes in perceived social support were triggered by the event of motherhood and occurred for affluent as well as poor mothers and those who had the best health as well as those whose health was poor. The extent and direction of such change was signifi-


"There are other pregnant women/mothers 2. with whom I can share my experiences"

- - ~ 40 f 35 "=3o ~ , ; ~ 25

.-~ 2o

10 e~ '~ 8 weeks 8 weeks

prepartum postpartum

No U1 pre- or postpartum • UI at 8 weeks postpartum only • UI at 8 weeks prepartum only • UI both pre- and postpartum

Fig. 3. Changes in perceived social support in relation to the presence or absence of urinary infection.

cantly influenced by mental , but not by physical health.

These findings indicate tha t a firmer unders tand- ing of the funct ion of social suppor t in relat ion to heal th could be developed by establishing the conti- nuities and discontinuit ies in this relat ionship across condi t ions and in different social contexts.

UNLINKED REFERENCES

Broadhead et al., 1983not cited in text

Acknowledgements--We are extremely grateful to all the mothers who took part in the ALSPAC study and to the midwives for their cooperation and help in recruitment. The whole ALSPAC study team comprises interviewers, computer technicians, laboratory technicians, clerical workers, research scientists, volunteers and managers who continue to make the study possible. This study could not have been undertaken without the support of the Wellcome Trust, the Department of Health, the Department of the Environment and British Gas. The ALSPAC study is part of the WHO initiated European Longitudinal Study of Pregnancy and Childhood. The authors also wish to acknowledge the constructive and helpful comments of one of the anonymous referees for this paper.

REFERENCES

Baker, D. (1989) Social identity in the transition to motherhood. In The Social Identity of Women, eds S. Skevington and D. Baker. Sage, London.

Baker, D., Taylor, H. and the ALSPAC Survery Team (1997) Inequality in health and health service use for mothers of young children in South West England. Journal of Epidemiology and Community Health 51(1), 79-79.

Berkman, C. F. (1984) Assessing the physical health effects of social networks and social support. Annual Review of Public Health 5, 413-432.

Blaxter M. (1990) Health and Lifestyles. Tavistock/ Routledge, London.

Bott E. (1971) Family and Social Network. Tavistock, London.

Broadhead, W. E., Kaplan, B. H., Sherman, A. J., Wagner, E. H., Schoenback, V. J., Grimson, R., Heyden, S., Tibblin, G. and Gehlbach, S. H. (1983) The epidemiologic evidence for a relationship between social support and health. American Journal of Epidemiology 117, 521 537.

Brown, G. W. (1996) Onset and course of depressive disorders: summary of a research programme. In The Origin and Course of Affective Disorders, eds C. Mundt et al. Royal College of Psychiatrists, Gaskell.

Brown, G. W. and Harris, T. O. (1978) Social Origins of Depression: A Stucl)' of Psychiatric Disorder in Women. Tavistock, London.

Brown, G. W., Andrews, B., Harris, T., Alder, Z. and Bridge, L. (1986) Social support, self-esteem and depression. Psychological Medicine 16, 813-831.

Cassell, J. C, (1976) The contribution of the social environment to host resistance. American Journal Of Epidemiology 104, 107 123.

Cox, J. L., Holden, J. M. and Sagovsky, R. (1987) Detection of postnatal depression: development of the 10 item Edinburgh Postnatal Depression Scale. British Journal of Psychiatry 150, 782-786.

D'Arcy, C. and Siddique, C. M. (1984) Social support and mental health among mothers of preschool and school age children. Social Psychiatry 19, 155- 162.

Dalgard, O. S., Bjork, S. and Tambs, K. (1995) Social support, negative life events and mental health. British Journal <~[" Psychiatry 166, 29 34.

Davey-Smith, G.. Blanc, D. and Bartley, M. (1994) Explanations for socio-economic differentials in mortality: evidence from Britain and elsewhere. European Journal o]" Public Health 4(2), 131 144.

Gardner, M. J. and Altman, D. G. (1989) Statistics with Confidence." Confidence Intervals and Statistical Guidelines. British Medical Journal, London.

Gjerdingen, D., Froberg, D. G. and Fontaine, P. (1991) The effects of social support on women's health during pregnancy, labour and delivery and the postpartum period. Family Medicine 23(5), 371 375.

Harris, B., Huckle, P., Thomas, R., Johns, S. and Fung, H. (1989) The use of rating scales to identify postnatal depression. British Journal of Psychiatry 154, 813--817.

Henderson, A. S. (1984) Interpreting the evidence on social support. Social Psychiato' 19, 49-52.

Holahan, J. C. and Moos, R. H. (1981) Social support and psychological distress: a longitudinal analysis. Journal of Abnormal Psychology 49, 365-370.

House, J. S., Landis, K. R. and Umberson, D. (1988) Social relationships and health. Science 241, 540-545.

Jacobson, D. E. (1986) Types and timing of social support. Journal ~1" Health and Social Behaviour 27, 250 264.

Kaplan, B. H., Cassel, J. C. and Gore, S. (1977) Social support and health . Medical Care 15(5 supp), 47 88.

Komarovsky, M. (1967) Blue Collar Marriages. Random House, New York.

Lin, N., Dean, A. and Essel, W. (eds) (1986) Social Support, Lifb Events and Depression. Academic Press, New York.

MacArthur, C., Lewis, M. and Knox, E. G. (1991) Health ,4tier Childbirth. HMSO, London.

Marmot, M., Shipley, M. J. and Rose, G. (1984) Inequalities in death: specific explanations of a general pattern. Lancet 5, 1003 1006.

Maser, K., Pugh, H. and Goldblatt, P. (1990) Mortality and the social classification of women. In Longitudinal Stud)'." Mortality and Social Organisation 1971-1981, ed. P. Goldblatt. OPCS series LS no 6, HMSO, London.

Murray, D. and Cox, J. (1990) Screening for depression during pregnancy with the Edinburgh Postnatal Depression Scale (EPDS). Journal of Reproductive and brant Psychology 8, 63 65.

1336 Deborah Baker et aL

Oakley, A., Hickey, D. and Ryan, L. (1994) Love or money? Social support, class inequality and the health of women and children. European Journal o f Public Health 4, 265-273.

O'Hara, M. W. (1986) Social support, life events and depression during pregnancy and the puerperium. Archives of General Psychiatry 43, 569-573.

O'Reilly, P. (1988) Methodological issues in social support and social network research. Social Science & Medicine 26(8), 863-873.

Orth Gom6r, K. and Und+n, A.oL. (1987) The measurement of social support in population surveys. Social Science & Medicine 24(1), 83-94.

Power, T. G. and Parke, R. D. (1984) Social network factors and the transition to parenthood. Sex Roles 10, 949-972.

Spitzer, R. L., Endicott, J. and Robins, E. (1978) Research diagnostic criteria. Rationale and Reliability. Archives of General Psychiatry 35, 73-82.

Wilcox, B. (1981) Social support in adjusting to marital disruption: a network analysis. In Social Networks and Social Support, ed. B, H. Gottlieb. Sage, London.

Wilkinson, R. (1992) Income distribution and life expect- ancy. British Medical Journal 304, 165-168.

the relationship between condition-specific morbidity, social support and material deprivation in...

Documents