BlRTH 24:4 December 1997 223

Who Worries That Something Might Be Wrong With the Baby? A Prospective Study of

1072 Pregnant Women

Helen Statham, BSc, josephine M. Green, PhD, and Kostas Kafetsios, PhD

ABSTRACT Background: Little is known about the normal range andpattern ofpregnant women’s worry about something being wrong with the baby, or how this relates to other factors. The objective of this study was to examine the extent to which women are worried about the possibility of something being wrong with the baby relative to other worries they may have, and to determine whether demographic, experiential, attitudinal, and personality characteristics are associated with this worry. Methods: Longitudinal data were collected from 1072 pregnant women who completed postal questionnaires at 16, 22, and 35 weeks of pregnancy. Multivariate statistical techniques were used to determine which variables were independently related to worry about the baby. Results: Worry that something might be wrong with the baby was one of the most prevalent worries at 16 weeks, although not as widespread a source of extreme worry as miscarriage or giving birth. Worry dropped in midpregnancy but rose again at 35 weeks. The most important factors related to this worry were perceived likelihood that something might be wrong with the baby and trait anxiety, although negative mood, previous pregnancy outcomes, and initial reactions to the current pregnancy all demonstrated signif cant, independent effects. Conclusions: Women who answered “Don’t know” to “Have you any reason to think that your baby might be more likely than any other to have some sort of a problem? ’’ had consistently high anxiety throughout pregnancy. We suggest that this question be asked routinely in early pregnancy to identify these potential worries and, it is to be hoped, reassure them at an early stage. (BIRTH 24:4, December 1997)

Screening and diagnostic tests for fetal abnormality are now in widespread routine use, and several studies have examined women’s anxiety levels in relation to the results of the screening tests that they have under- gone (1,2). Most babies do not have abnormalities, and these tests can potentially confirm this fact and

Helen Statham and Kostas Kafetsios are at the Centre for Family Research, University of Cumbridge, Cambridge; and Josephine Green is Senior Lecturer; Midwifery Studies, University of Leeds, Leeds, United Kingdom.

Address correspondence to Helen Statham, Centre for Family Re- search, Free School Lane, Cambridge CB2 3RE United Kingdom.

The Cambridge Prenatal Screening Study was funded by a grant from the Health Promotion Research Trust, Cambridge, United Kingdom.

0 1997 Blackwell Science, Inc.

reassure parents. It has been argued (3), however, that the reassurance is needed only because the tests created doubts in the first place. This raises questions about the extent to which, in the absence of specific consider- ations about testing, pregnant women do worry about the possibility of something being wrong with the baby. Although it is reported to be a universal concern (4), the existing literature on women’s experiences of testing sheds little light, since most studies focus on a particu- lar screening technique, or on a particular subgroup of women, usually “high-risk’’ women, or those given true or false positive results. Early studies were retro- spective, and often used a generalized measure of anxi- ety such as the Spielberger State-Trait Anxiety Inventory score (STAI) ( 5 ) , with the assumption that any anxiety was about fetal abnormality (1,2). More recently, qualitative studies of women found to be high risk after maternal serum screening for Down’s

224 BIRTH 24:4 December 1997

syndrome (6), have highlighted these women’s specific concerns about the health of their baby (7-9), and it is apparent that the nature of some women’s concerns re- lates to factors other than result of the screening test (9).

However, a pregnant woman might worry about many other things. Again, several studies have ad- dressed the issue of generalized anxiety and its implica- tions in the etiology of maternal and fetal complications (10,ll). Others addressed the general concerns of preg- nant women (12-14). Worry about the health of the baby has been identified as a major concern (14), but with relatively small samples it is not possible to explore the contributory factors. For example, person- ality characteristics and past history might be expected to account for some of the variation in women’s feel- ings. In short, little is known about the range of varia- tion or about why particular women may be worried and others not.

This paper examines the extent to which women are worried about the possibility of something being wrong with the baby relative to other worries they may have, and whether demographic, experiential, attitudi- nal, and personality characteristics are associated with this worry. The data were derived from the Cambridge Prenatal Screening Study (1 5-19), a large, prospective study of pregnant women that not only examined the knowledge, attitudes, and experiences of pregnant women booked for antenatal care at hospitals with different screening policies but also monitored their worries throughout pregnancy and after delivery.

Methods

The Cambridge Prenatal Screening Study

The purpose of the study was to chart the knowledge, attitudes, anxieties, and experiences of pregnant women from before their first hospital appointment through to the postnatal period, with a particular focus on screening for fetal abnormality. A series of four postal questionnaires, designed for the study, allowed the researchers to look at women’s experiences in the context of other things happening in their lives. Ques- tionnaires were sent at or before 16 weeks (time l), at 22 weeks (time 2), 35 weeks (time 3), and at 6 weeks postnatally (time 4). This paper presents data from the first three questionnaires.

Women were recruited to the study between January and March 1990 in nine district hospitals, all within 60 miles of Cambridge, England. District hospitals are nonteaching hospitals, and are where most pregnant women in the United Kingdom deliver their babies. They were recruited by means of a letter and question- naire forwarded to them with notification of their first hospital appointment, which was generally at around 12 to 14 weeks.

Of 3550 questionnaires mailed, 1876 (53%) were returned. All women booking for antenatal care in the study hospitals during the target period were eligible for the study; none was excluded. Recruitment contin- ued until each hospital had distributed all the question- naires allocated. Valid responses were obtained from 1824 women (51%), and 52 came from women who had miscarried before receiving the questionnaire. Not all questions were answered by all women. Although 1824 women were recruited to the study, the data reported here are restricted to women who completed all three antenatal questionnaires, who were less than 16 weeks pregnant at recruitment, and who were never told that any screening test had revealed (possible) problems with the baby ( n = 1072). Mean age and trait anxiety scores (both measured at the first question- naire) did not differ for this selected group when com- pared with the whole sample, nor did mean baby worry scores between those who had only completed one ( n = 284) or two ( n = 143) or three questionnaires ( n = 1397; F [2,1798] = 0.80, ns).

Measures

Worries: Worries were measured using the Cambridge Worry Scale (CWS) (15,19), which lists pregnancy, health, relationships, and socioeconomic items of pos- sible concern to pregnant women. They were asked to score each item from 0 (not a worry) to 5 (extremely worrying). The items varied slightly at different gesta- tions. Sixteen items were listed in the first question- naire. Only one item will be considered in detail in this paper, that is, “the possibility of something being wrong with the baby,” and this will be referred to as “baby worry.” The 140 women who completed an alternative version of the first questionnaire did not answer any questions specifically about screening and fetal abnormality other than this item within the scale. These women constituted a control group whereby we could check whether the questionnaires were raising awareness or creating anxiety about fetal abnormality. There was no evidence that worry about fetal abnor- mality differed between those receiving the two ques- tionnaires (F [1,1799] = 0.50, ns).

Independent variables: The sample was categorized according to several sociodemographic characteristics to give some of the independent variables; others were derived from responses in the first questionnaire. The variables, together with the questions from which they were derived, are shown in Tables 1 and 2.

Analysis: Data were analyzed by means of the Sta- tistical Package for the Social Sciences (SPSS Inc.), using analysis of variance (ANOVA). One-way ANO- VAs were carried out for each independent variable at each time point, followed by Scheffi contrast tests to establish which pairs of groups differed significantly

BIRTH 24:4 December 1997

Table 1. Definition and Construction of Independent Variables

225

Vuriable Question Response Options Category N O . %

Initial reaction to How did you feel about finding pregnancy that you were pregnant?

Thinking of the baby as a person

Do you think of your baby as being a person yet?

Physical well- How have you been feeling so being far during your pregnancy?“

Likelihood of a Have you any reason to think problem that your baby might be more

likely than any other to have some sort of a problem??

Women were asked to select all of the words that described how they felt at that moment from a list of 21 adjectives.$

The trait part of the Spielberger State-Trait Anxiety Inventory (3, which measures anxiety as a personality characteristic. 5

Negative mood

Trait anxiety

Overjoyed/pleased Mixed feelings Not very happyhery

Yes No Don’t knowhaven’t

thought about it Sick and tired some of the

time Sick and tired all of the

time Same as usual Very well/blooming Yes No Don’t know

unhappy

0 negative words 1-3 negative words More than 3 negative

Trait score 20-32 Trait score 33-37 Trait score 38-43 Trait score 44-74

words

79 1 Positive Mixed 246 Negative 35

Yes 747 No 176 Don’t know 93

517

428

56 70

Yes 142 No 760 Don’t know 92

357 654

54

Very low 262 Low 280 High 25 1 Very high 242

74 23

3

70 17 9

48

40

5 7

14 77 9

34 61 5

25 27 24 23

-* This question was asked again in later questionnaires, and analyses for times 2 and 3 use the answers given at those times. f This question was not asked in the “ultemative” questionnaire. Percentages are based on the 994 women who answered the qmesfion. $ The number of negative adjectiwr appears to be useful measure ofgeneralized negative mood; we have found it to correbte highly with standardized measures of depression and anxiety ( r = 0.50 with the Edinburgh Postnatal Depression Scale and 0.55 with the state scale o j the STAI; see reference 16). As with physical well-being, this question was asked again in later questionnaires, and analyses for times 2 and 3 use the answers given at those times. $ On the basis of the trait scores, women were grouped in quartiles based on the scores of the original sample of 1824 women. The mean score was 38.4 (SD 8.1).

from each other. Repeated measures ANOVA was used to investigate different patterns over time for different subgroups. Finally, all variables were entered into a factorial ANOVA, which indicated which variables contributed an independent significant effect on baby worry when the others were taken into account.

Results

Baby worry in the context of other worries: Each worry listed was given the full range of scores from 0 to 5. For baby worry, 10 percent scored zero at time 1; this percentage was lower than that scoring zero for any other worry (Fig. 1). More women scored 4 or 5 for miscarriage and giving birth, but the mean score for baby worry was higher than for any of the other worries (mean 2.55, SD 1.54). Sixteen percent scored the maxi- mum of 5 for baby worry, and a further 12 percent scored 4.

Baby worry over time: Baby worry dropped at time 2 and rose again at time 3, although not to the initial levels (Table 3). Repeated measures ANOVA showed

these differences were significant (F [2,1035] = 85.5,

Baby worry by independent variables: The only independent variables that were not significantly re- lated to baby worry at any time point were being married (or living as married) and influence of religion. Thinking of the baby as a person was not significant at times 1 or 2, but at time 3 those who had answered “no” had higher baby worry than those answering “yes” (p < 0.05). Although all one-way ANOVAs were significant, the independent variables were highly interrelated. When all of these variables were taken into account, the factorial analysis of variance carried out for each time point showed that education, age, being in paid employment, and physical well-being were not significant. The variables that showed inde- pendent effects, in order of significance at time 1 , were likelihood of a problem, trait anxiety, negative mood, reproductive history, and initial reaction to pregnancy. F values and significance levels at each time point are shown in Table 4. The total multiple R’ was 0.21 at time 1 and 0.13 at times 2 and 3.

p < 0.0001).

226 BIRTH 24:4 December 1997

Table 2. Definition and Construction of Independent Variables

Variable ~~

Definition Categorj ~ _ _ _ _ _ _ _ _ _ _

No. %

Age

Level of education

Reproductive history

Paid employment

Married or living as

Religious influence married

Age 16 to 20 yr Age 21 to 34 yr Age 35 to 44 yr Left school at earliest possible age, usually

Remained in education up to age 18 yr Education beyond age 18 yr No previous pregnancies Only successful pregnancies (i.e., resulting

in live birth of a healthy baby) Only unsuccessful pregnancies (i.e., not

resulting in live birth of a healthy baby) Mixed pregnancy experiences (i.e., at least

one successful pregnancy and one unsuccessful?)

16 yr

Does your religion or religious upbringing influence the sorts of decisions that you make about your life?

Young Middle Old Minimum

Further Higher No Successful

Unsuccessful

Successful/ unsuccessful

Yes No Yes

Not at all A little Quite a lot Completely Not applicable

97 9 984 83 81 8

599 56

33 1 31 142 13 404 38 390 37

80 8

186 18

663 62 403 38

1013 95

638 60 226 21 89 8 22 2 90 9

* The mean age was 26.9 yr (SD 4.8). f Unsuccessful pregnancies were those that resulted in miscarriage, stillbirfh, the birth of an ill or handicapped child, or a termination for medical reasons. These events are all grouped together for this papec although we have considered them separately elsewhere (15). Tertninations of pregnancy for nonmedical reasons were distinguished from unsuccessful pregnancies; in general, they appear to have little impact on the outcome measures used in this study (15), and therefore they huve been disregarded for these class$cations.

Perceived likelihood of a problem was the most significant predictor of worry at time 1, but was less important at times 2 and 3 (Table 4). The F ratios for all five variables were lower at time 2 than at time 1, but trait anxiety became relatively more important again at time 3. Initial reaction to the pregnancy was the least significant predictor, but unlike the other variables, it maintained the same level of association at each time point.

Baby worry by likelihood of problem: Figure 2 shows the pattern of baby worry for the three groups answering “yes,” “no,” and “don’t know” to this question. Figure 2a shows the basic mean scores for each group; Fig. 2b shows adjusted means that take into account the contributions of the other significant variables. Baby worry was significantly higher throughout pregnancy for those women who answered “yes” or “don’t know” than for those who answered “no.” At time 2 those who answered “don’t know” had significantly higher scores than those who an- swered “yes.” However, it is important to note how the mean score for the “don’t know” group became lower at each time point once other variables were taken into account. This is a reflection of the higher levels of trait anxiety of the women who answered “don’t know.” Repeated measures analysis showed significant differences in the pattern of change over

time (F = 11.47, p<O.O001), with the “yes” group showing a steeper drop from time I to time 2 and a steeper rise from time 2 to time 3 than the other groups.

Baby worry by trait anxiety: Women with higher trait scores had higher baby worry scores (Fig. 3). As before, Fig. 3a shows basic means, and Fig. 3b shows means adjusted for the contribution of the other vari- ables. The two middle groups did not differ signifi- cantly from each other, and by time 3 the group with the lowest trait scores was closer to these two groups than at previous time points. However, the group with the highest trait scores remained significantly more worried throughout. The groups did not differ signifi- cantly in their pattern over time. The effect of taking other significant variables into account (Fig. 3b), was to reduce the differences between the groups slightly, but the basic pattern remained the same.

Baby worry by negative mood: Women’s negative mood, measured by means of the number of negative adjectives selected from a list, appears to be a useful measure of generalized negative mood: it correlates highly with standardized measures of anxiety and de- pression (r = 0.50 with the Edinburgh Postnatal De- pression Scale, and 0.55 with the state scale of the STAI (20)). The relationship at times 1 and 2 was linear: the more negative that adjectives scored, the higher the worry about the baby (Fig. 4). At time 3

BIRTH 24:4 December 1997 227

the most negative group had lower scores than would have been expected from the previous time points, and were not significantly different from the middle group. When the other variables were taken into account, the scores of the most negative group were reduced (Fig.

4b), indicating that this group had a greater contribu- tion from the other variables such as trait anxiety.

Baby worry by reproductive history: The consistent pattern was that women with only successful pregnan- cies were the least worried and those with only unsuc-

Table 3. Baby Worry at Each Time Point

Time 1 Time 2 Time 3

Mean SD Mean SD Mean SD

Baby worry 2.56 1.55 2.00 1.45 2.11 1.44

Table 4. Results of Factorial Analyses of Variance for the Independently Significant Variables at Each Time Point

Time I

Vuriable

Time 2

F fdf)

Time 3

F (df)

Likelihood of problem 36.16f (2,931) Trait anxiety 25.29f (3,931) Negative mood 19.21$ (2,931) Reproductive history 13.943: (3,931) Reaction to pregnancy 7.427 (2,931)

10.563: (2,929) 16.11f (3,929) 14.483: (2,929) 8.933: (3,929) 7.211. (2,929)

10.093 (2,921) 20.513: (3,921) 14.503 (2,921) 3.86" (3,921) 7.537 (2,921)

~~~

* p < 0.01. f p < 0.001. $ p < 0.0001.

Baby Worry

Miscarriage

Giving Birth

Money

Coping With New Baby

Internal Examination

Going to Hospital

Own Health

Someone Else's Health

Housing

Giving Up Work

Relationships With Family and Friends

Relationship with HusbandlPartner

Employment Problems

Whether Partner Present at Birth

Problems With The Law 1

0% 20% 40% 60% 80% 100%

Fig. 1. Worries at time 1.

228 BIRTH 24:4 December 1997

3.75

3.5

3 25

3

2.75

2 P 1 2.5

2.2:

1.7E

1: 1

Figure 2(a): Unadjusted

2

Time

3

Fig. 2. Baby worry by perceived likelihood of problem.

Figure 3(a): Unadjusted

1 2 Time

3

Figure 2(b): Adjusted For Other Variables

3 75

325 ;.h Fl -&-Don't Know

t 1

2 3

Time

Figure 3(b): Adjusted For Other Variables

3 25

1 2 Time

3

Fig. 3. Baby worry by trait anxiety groups.

BIRTH 24:4 Decemher 1997 229

Figure 4(a): Unadjusted Figure 4(b): Adjusted For Other Variables

35

3 25

1 3

f 275

P 2 5

2 25

2

1 7 5

cessful pregnancies before this study were consistently the most worried (Fig. 5) . Women with mixed preg- nancy experiences were slightly less worried than those who had never been pregnant before. The group with only unsuccessful pregnancies comprised women whose mean score changed most after adjusting for other variables. This is a reflection of their significantly higher trait anxiety scores. The groups did not differ significantly in their pattern over time.

Baby worry by initial reaction to pregnancy: The unadjusted means showed women with mixed feelings about the pregnancy as consistently the most worried about the possibiIity of something being wrong with the baby (Fig. 6). However, adjustment for the other significant variables had the effect of substantially de- creasing the means for both the negative and the mixed groups (Fig. 6b). This indicated that nonpositive reac- tions to the pregnancy were highly related to other significant factors such as previous pregnancy experi- ences. The result of this finding made the positive and mixed groups more similar; the significant effect in the factorial analysis can be seen to be primarily be- cause the negative group had the lowest mean scores. Although the negative group did not show the charac- teristic rise in scores between times 2 and 3, there was no significant difference in the way the groups changed over time.

Fig. 4. Baby worry by negative mood (number of negative adjectives selected).

Discussion

The sociodemographic characteristics of the study group (51 % of those invited to join the study) reported here defined a population broadly representative of the childbearing population of the areas served by the study hospitals. The location is largely rural or with light industry, and it has areas of social disadvantage but no inner-city hospitals. Only one hospital had a large ethnic minority group, and we were unsuccessful in recruiting from that sector. Use of the term ‘‘married or living as married” was left to the women to interpret for themselves, and the apparently high rate of babies conceived within stable relationships may reflect dif- ferent interpretations of the term. Within the whole sample, no differences were found in mean age, trait anxiety, or baby worry between the women reported in this paper who completed all three antenatal ques- tionnaires and those who only completed one or two questionnaires. In addition, women were excluded if they were informed during the pregnancy of a possible or definite fetal abnormality.

This study showed that worry that something might be wrong with the baby is highly prevalent among these pregnant women. Most rated their worry in the midrange, but for some it was a major worry. In most cases, worry dropped by midpregnancy and rose again

230 BIRTH 24:4 December 1997

-

Figure .5(a): UnadJUSted

325 I

Fig. 5. Baby worry by previous pregnancies.

Figure 5(b): Adjusted For Other Variables

3 75

35

3.25

3

.. 1 375

I 35

3 25

3

225

2

1 7 5

__ -

Initial Reaction to Pregnancy (a)

I - , 1 5 1 2 3

Tlme

~ -~ ~ .. ~

Fig. 6. Baby worry by initial reaction to pregnancy.

-B- Successful -A- Mixed

, 1 2 3

Time

~

375

35

325

3

g 2 , 3 2 2 25

2 25

2

1 7 5

15 1

Initial Reaction to Pregnancy (b)

.

2

Tlme

3

BIRTH 24:4 December 1997 23 1

at the end, corresponding to the characteristic U- shaped curve for mood during pregnancy first de- scribed by Lubin et al (21).

Since the study was large, we could examine the relationship between baby worry and a range of inde- pendent variables to try and explain why some women are more worried than others, and why not all women follow the same pattern over time. A range of experien- tial, attitudinal, personality, and mood factors were found to be related to baby worry. Perhaps not surpris- ingly, women with previous pregnancies that had not resulted in the birth of a healthy baby demonstrated the highest baby worry scores, especially if they also had an unsuccessful pregnancy. In terms of reproduc- tive history, however, the group who were statistically most different from the rest were women who had only previous successful pregnancies; these women were significantly less worried than the other groups. Thus not only do previous unsuccessful pregnancies increase baby worry, but also experiencing at least one successful pregnancy decreases baby worry.

Trait anxiety score was also a strong predictor of the extent of women’s baby worry. Groups with lower trait anxiety scores did not necessarily have signifi- cantly lower baby worry scores, but the top quartile group were consistently more worried throughout pregnancy, indicating that high trait anxiety was a suffi- cient but not a necessary condition for baby worry. In other words, if a woman was of a highly anxious disposition, she would probably worry about some- thing being wrong with her baby, but she might still worry for other reasons even if she was not an anxious tY Pe.

Initial reaction to the pregnancy was also predictive of baby worry: in the basic analysis, women with mixed feelings were the most worried. The most frequently cited reasons for mixed feelings were previous preg- nancy experiences, unplanned pregnancy, or social and housing difficulties. Thus, accounting for other signifi- cant variables, particularly previous pregnancy experi- ence and trait anxiety, reduced the mean score of this group. However, the multivariate analysis also lowered the mean score of the group who were most negative about their pregnancy and who already had low scores, because this group had higher trait anxiety scores, which therefore masked their relative lack of worry about the baby.

One of the strongest predictors of baby worry was the answer given in early pregnancy to the question, “Have you any reason to think that your baby might be more likely than any other to have some sort of a problem?” As might be expected, women answering “yes” to this question were significantly more worried about the baby than women answering “no.” The pattern of worry of the former group differed from the

latter group, dropping more in midpregnancy and rising more steeply again at the end. This was characteristic of women who had undergone an amniocentesis (15), which many of these women will have experienced. The particular interest of this variable is in the women who answered “don’t know.” They were just as wor- ried as those answering “yes,” and, indeed, had signif- icantly higher scores at time 2. What leads a woman to answer “don’t know” to this question? It may reflect general nervousness and a desire not to tempt fate by answering ‘‘no”; they showed significantly higher trait anxiety scores than other women, but were still signifi- cantly more worried than the “no” group when this was accounted for. Alternatively, it may reflect general ignorance; this group contained a disproportionate number of women with the minimum level of educa- tion (x’ = 20.6, 4 d j p<O.OOl). Of course, in some cases “don’t know” may also be the right answer, for example, if a woman does not know whether some previous event or current lifestyle puts her baby at risk. Depending on the reason, some obvious strategies might be adopted for turning some “don’t know” answers into a “no,” and thus reducing women’s wor- ries about the baby. We are currently undertaking fur- ther analysis of women who answered “don’t know” to this and a number of other questions in the study, since they might be a vulnerable group.

The remaining variable that made a significant inde- pendent contribution to baby worry was negative mood. At time 3 a difference in baby worry was no longer present between women selecting one to three negative words and those selecting more than three; the significant difference was between these two groups and women selecting no negative adjectives. This may reflect the general rise in the number of negative words selected between times 2 and 3; thus, the women who selected no such words may have been exceptional. At times 1 and 2 the relationship was a straightforward linear one: the more negative a woman felt, the higher the worry score. This was true even when other variables were accounted for. In other words, the relationship with baby worry is not just because women who choose more negative adjectives have higher trait anxiety. Furthermore, we noted that negative feelings about the pregnancy were associated with lower worry. Thus, the direction of the relation- ship between baby worry and negative mood is not necessarily what one would have predicted: we might have hypothesized that negative mood would be asso- ciated with apathy and worrying less about the baby. Since this is not so may help shed a small amount of light on a question raised in an earlier study (22), where it was reported that women with low emotional well-being postnatally were more negative about their babies. The question was whether these women were

232 BIRTH 24:4 December 1997

more negative about their babies because of their low mood, or whether their low mood was because of having a less likeable baby. The present finding would argue against the first interpretation if we take baby worry to be an indicator of positive concern about the baby.

Demographic variables, notably age and education, initially showed a significant relationship with baby worry, but then were found not to be significant once other factors were taken into account. This is an inter- esting finding. Seldom can we expect a direct relation- ship between demographic variables and outcomes; more usually they are proxies for something else (23). In this particular case, however, a more direct link might have been postulated; for example, older women might be expected to be more worried about something being wrong with the baby because they are indeed at somewhat higher risk, and recent developments in diagnostic testing have emphasised this fact. Despite a strong statistical relationship from a simple analysis of variance in this study and similar findings elsewhere (14,24), which appeared to support that hypothesis, the factorial analysis showed that the effect was mediated through other variables. Age was, in fact, significantly related to every one of the other significant indepen- dent variables: older women were much more likely to have mixed feelings on finding that they were preg- nant; were much more likely to have been pregnant before and thus more likely to have had at least one unsuccessful pregnancy; and were least likely to circle no negative adjectives. They were also most likely to say that they thought their baby was more likely than others to have a problem, whereas the youngest group, who also had the highest trait anxiety scores, were more likely to say “don’t know.” Such a situation of highly interrelated independent variables is very com- mon, and only through the use of multivariate statistical techniques can we disentangle all the separate effects. The choice of 20 years as the upper age of the women assigned to the youngest age group might appear high. We wished, however, for this group to contain similar numbers to the oldest age group (over 3 3 , and rates of teenage pregnancy in the United Kingdom are lower compared with the Americas (25). Although age 18 years is now the legal age of majority in the United Kingdom, this group would be widely thought of as “young” within the local population of childbearing women.

We report here the findings of an empirical study of aspects of worry about fetal abnormality which have not been investigated before, to our knowledge, in such a large sample, and at this stage we have not related the findings to theoretical concepts regarding worry. For the pregnant women in this sample some degree of worry about whether or not the baby will be all

right is normal: only 10 percent indicated that they were not at all worried. We have shown elsewhere (15) that baby worry is, in general, unrelated to the tests that a woman undergoes, except when results are abnormal, and that it is unrelated to knowledge about prenatal tests (18). The present study showed that the extent of this worry is largely determined by aspects of the woman’s life that she brings with her to the pregnancy: previous experiences and current attitudes are important. We cannot say how women in different cultural or socioeconomic situations may feel. These women are pregnant in a health care system that allows them free access to antenatal, perinatal, and postnatal care. Ultrasound scanning was offered to all women, and the belief was widespread that antenatal care was reassuring (unpublished data). For most of these women the pattern of worry was consistently U- shaped, except for those women who answered “no” to the question, “Have you any reason to think that your baby might be more likely than any other to have some sort of a problem?” If raised anxiety has implications for fetal health (1 l), we would suggest that the main message for practitioners to be derived from these data must therefore be with regard to this question. If, for example, every woman’s caregiver were to ask it of her at the beginning of pregnancy and to follow up on her answer if it is “yes” or “don’t know,” it would probably be possible to reassure many woman who would otherwise continue to worry throughout pregnancy.

Acknowledgments

We would like to thank Claire Snowdon, our original co-worker on the project, Sally Roberts for statistical assistance, and Owen Green for graphics.

References

1.

2.

3.

4.

5.

6.

Green JM. Calming or Harming?A Criticul Review of Psycho- logical Effects of Fefal Uiugnosis on Pregnanf Women. Occa- sional papers, Second Series, No. 2. London: Galton Institute, 1990. Green JM. Women’s experiences of prenatal screening and diagnosis. In: Abramsky L, Chapple J, eds. PvenataE Diugno- sis: The Humun Side. London: Chapman and Hall, 1994: 37-53. Kolker A. Advances in prenatal diagnosis: Social-psychologi- cal and policy issues. hf J Technoi Assess Health Care 1989;

Reid M. Consumer-oriented studies in relation to prenatal screening tests. Eur I Obstet Gynecol Reprod Biol 1988;

Spielberger CD, Gorsuch RL, Lushene RE. The State-Trait Anxiety Inventory. Palo Alto, CA: Consulting Psychologists Press, 1970. Green J, Statham H. Testing for fetal abnormality in routine antenatal care. Midwifery 1993;9:124-135.

5 :60 1-6 17.

28(suppl):79-92,

BIRTH 24:4 December 1997 233

7.

8.

9.

10.

I I .

12.

13.

14.

15.

16.

Statham H, Green J. Serum screening for Down’s syndrome: Some women’s experiences. BMJ 1993;307:174-176, Roelofsen EEC, Kamerbeek KI, Tymstra T. Chances and choices. Psycho-social consequences of maternal serum screening. A report from The Netherlands. J Reprod lnfant Psychol 1993;11:4147. Santalahti P, Latikka A-M, Ryynanen M, Hemminki E. Wom- en’s experiences of prenatal serum screening. Birth 1996;23:

MacDonald R, Christakos A. Relationship of emotional ad- justment during pregnancy to obstetric complications. Am J Obstet Gynecol I963;86:341-348. Reading AE. The influence of maternal anxiety on the course and outcome of pregnancy outcome: A review. Health Psychol

Pleshette N, Asch SS, Chase J. A study of anxieties during pregnancy, labor, the early and late puerperium. Bull N Y Acad Med 1956;32:436455. Light HK, Fenster C. Maternal concerns during pregnancy. Am J Obstet Gynecol 1973;118:46-50. Glazer G. Anxiety levels and concerns among pregnant women. Res Nurs Health 1980;3:107-113. Statham H, Green J, Snowdon C, Psychological and social aspects of screening for fetal abnormality during routine ante- natal care. In: Robinson S, Thomson A, Tickner V, eds. Pro- ceedings of the 1992 Research and the Midwife Conference. Manchester, UK: School of Nursing, University of Manches- ter, 1993. Green J, Statham H, Snowdon C . Screening for fetal abnor-

101- 107.

1983;2: 187-202.

17.

18.

19.

20.

21.

22.

23.

24.

25.

mality: Attitudes and experiences. In Chard T, Richards MPM, eds. Obstetrics in the 1990s: Current Controversies. London: MacKeith, 1992. Green JM, Snowdon C, Statham H. Pregnant women’s atti- tudes to abortion and prenatal screening. J Reprod lnfant

Green JM, Statham H, Snowdon C. Women’s knowledge of prenatal screening tests. 1 : Relationships with hospital screening policy and demographic factors. J Reprod Infant Psychol 1993;ll:ll-20. Statham H, Green JM. The effects of miscarriage and other “unsuccessful” pregnancies on feelings early in a subsequent pregnancy. J Reprod Infant Psychol 1994; 12:45-54. Green JM. Postnatal depression or periiiatal dysphoria? Find- ings from a longitudinal, community-based study using the EPDS. J Reprod Infant Psychol. In press. Lubin B, Gardener SH, Roth A. Mood and somatic symptoms during pregnancy. Psychosom Med 1975;37: 136-146. Green JM, Coupland VA, Kitzinger JV. Great Expectations: A Prospective Study qf Women’s Expectations and Experiences of Childbirth. Cambridge, England: Child Care and Develop- ment Group, 1988. Oakley A. Social Support and Motherhood. Oxford: Black- well, 1992. Berryman J, Thorpe K, Windridge K. Older Mothers: Concep- tion, Pregnancy and Birth after 35. London: Pandora, 1995. World Health Organization. Challenges in Reproductive Health Research. Biennial report 1992-1993. Geneva: Author, 1994: 120.

Ps)jchol 1993;11:31-39.