A randomised controlled trial comparing two schedules of antenatal visits: the antenatal care project

HYPOTHESES

A major clinical concern relating to a reduction in the number of routine antenatal visits is the possibility of increased maternal morbidity associated with late detection of, or failure to detect, pregnancy related hypertensive disorders. We consider that the incidence of caesarean section for such disorders is a reasonable proxy measure for this morbidity. The primary null hypothesis for the trial was that the reduced schedule of visits would not lead to an increase of 1% or more in the rate of caesarean section for pregnancy related hypertensive disorders.

Because antenatal care has a number of other important clinical and non-clinical functions, two further hypotheses were that the reduced and the traditional schedules of visits would not differ in other aspects of clinical effectiveness; and that they would not differ in their psychosocial effectiveness. The final two hypotheses were that the two schedules would not differ in their acceptability to women, nor in their acceptability to the health care professionals who provide antenatal care.

VISIT SCHEDULES

A visit was defined as an encounter between a pregnant woman and a midwife or doctor for the purpose of assessing maternal or fetal wellbeing. Day or inpatient admissions and visits solely for ultrasound or phlebotomy were not counted as visits. The schedules were described as “traditional” (13 visits—the control group) or “new style” (seven visits for nulliparous women and six visits for multiparous women—the study group). They were developed in discussion with obstetricians and midwives within Lewisham and Guy's Hospitals and therefore reflect a local agreement rather than the adoption of external recommendations. (See Appendix for timing of visits.)

All general practitioners within the Lambeth, Southwark and Lewisham Family Health Services Authority who normally referred women to one of the three hospitals were contacted by one of the two research midwives or the project leader. Eighty three out of 105 practices were visited in person. In all, 464 general practitioners (98%) agreed to take part. The study did not attempt to change the balance of professional involvement in women's antenatal care. At the beginning of their pregnancies all but 6% of women (who attended a hospital based midwives' clinic) were receiving shared care—that is, most of their care was given by general practitioners and midwives in the community, with occasional hospital visits.

RANDOMISATION

Randomisation was performed at the booking appointment with random permuted blocks of eight and 16, stratifying by the six offices at which the recruiting midwives were based (Lewisham, Guy's and St Thomas' antenatal clinics, Guy's and St Thomas' community midwives' offices, and the office of a midwifery group practice at Lewisham Hospital).14 Sequentially numbered, non-resealable opaque envelopes containing details of either the traditional or new style visit schedules were used. This number was then assigned to each woman in the trial. The woman's pregnancy health record (maternity notes) was labelled to indicate her number and the schedule to which she had been assigned, and a form giving brief details of her name, address, number, etc was returned to the project office.

MEASURES

Data relating to pregnancy related hypertensive disorders, smallness for gestational age, and other clinical variables, together with measures of health service use, were collected from the pregnancy health records by the two research midwives who were blinded to group allocation by paper covers put over each record. Pre-eclampsia was defined according to the criteria of the International Society for the Study of Hypertension in Pregnancy.15 Assessment of smallness for gestational age (<3rd centile and <10th centile) was made by plotting on centile charts the fetal abdominal circumference as measured by ultrasound scan16 and the baby's birth weight (using birth weight charts from Castlemead Publications, Welwyn Garden City). Twenty one per cent of records were reviewed by both research midwives; the overall interrater agreement was 98.6% (range for different variables 87.5% to 100%).

Women's views and psychosocial effectiveness were assessed by a maternal antenatal questionnaire at 34 weeks of pregnancy and a maternal postnatal questionnaire six weeks after birth. These questionnaires included the Edinburgh postnatal depression scale17 (which has been validated for use in pregnancy as well as postnatally18), the Cambridge worry scale (J Green et al, prenatal screening study, Cambridge), the baby adjective checklist (which contains 16 adjectives and from which the proportion of negative adjectives chosen to describe the baby is calculated),19 and the attitudes to the pregnancy and baby scale (which gives four scores; positive and negative attitudes to the fetus and positive and negative attitudes to the pregnancy),20 together with items from the Office of Population Censuses and Surveys's survey manual on women's experience of maternity care.21 The characteristics and views of women who chose not to participate in the trial were assessed by means of a non-participants' questionnaire devised specifically for the trial. Health professionals' views were ascertained before the trial began and during its last few weeks by using postal questionnaires devised for the trial. These were sent to all midwives, general practitioners, and obstetricians involved in the trial. All questionnaires used were piloted. Two postal reminders were sent at fortnightly intervals to non-responding women and health professionals.

SAMPLE SIZE

The intended sample size was based on the primary hypothesis. Review of the obstetric database at one of the hospital centres (Euroking, Guy's Hospital, 1991) indicated an incidence of 1% for caesarean section for pregnancy related hypertensive disorders. If this rate had increased by 1% or more (from 1% in the traditional group to 2% or more in the new style group) a minimum of 1415 women would have been needed in each group for the difference to be detected (95% significance level and power of 80%).22 The sample size needed to detect this one tailed effect was therefore 2830. Assuming a loss rate of not more than 10%, we estimated that 3144 women would need to be enrolled into the project.

ANALYSIS

Analysis was by intention to treat, using the Statistical Package for the Social Sciences (SPSS-PC),23 Confidence Interval Analysis,24 and Epi Info.25 To test the primary hypothesis a one tailed Fisher's exact test was used. Two tailed tests were carried out for all other variables, since the hypotheses to which they relate were not unidirectional. For continuous variables, Student's t test (for parametric variables) and the Mann-Whitney U test (for non-parametric variables) were used. For the continuous parametric variables, 95% confidence intervals were also calculated for differences between means but are not reported here. For categorical variables, odds ratios and 95% confidence intervals were calculated with Confidence Interval Analysis, except where there were small numbers, in which cases exact limits were calculated with Epi Info. Confidential interim analysis was performed.

CLINICAL VARIABLES

The two groups did not differ significantly in the primary variable of caesarean section rate for pregnancy related hypertensive disorders (traditional 14/1396 (1.0%), new style 11/1359 (0.8%); P=0.630).

Table 2 shows that there were no significant differences between the two groups for variables relating to pregnancy related hypertensive disorders, labour, maternal morbidity, or perinatal morbidity. Analysis of variables relating to smallness for gestational age showed a significantly lower incidence of antenatal suspicion of smallness for gestational age in women receiving new style care than in those receiving traditional care (odds ratio 0.73; 0.54 to 0.99). However, there were no significant differences in correct antenatal diagnosis of smallness for gestational age for birth weights either below the 3rd centile or below the 10th centile. One multiparous woman in the new style group died postnatally from a cerebrovascular accident, having attended 20 antenatal visits.

ORGANISATIONAL VARIABLES

Women in the traditional group received significantly more ultrasound scans than did women in the new style group (1.7 v 1.6; P=0.003); they also had more day admissions (1.0 v 0.8; P=0.002). The number of antenatal inpatient admissions and the number of nights spent in hospital antenatally did not differ significantly (table 3).

Regarding continuity of caregiver, women in the new style group saw significantly fewer caregivers (5.2 v 5.7; P<0.001) and received less conflicting advice (odds ratio 0.77; 0.62 to 0.95). They also, however, saw their usual caregiver fewer times (3.1 v 4.1; P<0.001) and were more likely to say that they did not feel remembered from one visit to the next (odds ratio 1.52; 1.25 to 1.89). There was no difference in the proportions who had met their intrapartum midwife during their antenatal care (0.93; 0.70 to 1.23).

PSYCHOSOCIAL VARIABLES

Information—The groups were similar in feeling encouraged to ask questions at their antenatal visit (1.12; 0.90 to 1.39), feeling that things had been explained well (0.91; 0.66 to 1.26), and in wanting more medical (1.16; 0.96 to 1.39) or general (1.04; 0.87 to 1.25) advice. They were also similar in whether they discussed their birth preferences at visits (1.09; 0.87 to 1.36), made a birth plan (1.11; 0.91, 1.34), attended antenatal classes (0.94; 0.78, 1.14), or read more than three books about pregnancy (1.14; 0.93 to 1.38). The maternal postnatal questionnaire asked if women would have liked to have known more about various aspects of childbirth and life with a new baby. The only difference found was that women in the new style group were more likely to wish they had known more about feeding their baby (285/846 (33.7%) v 232/843 (27.5%); odds ratio 1.34; 1.09 to 1.65).

Social support—Women in the new style group were less likely to feel listened to at visits (0.75; 0.60 to 0.95) and were more likely to say that they wanted more time talking at visits (1.26; 1.05 to 1.52). There were no differences in antenatal or postnatal depression scores (table 4).

Worry—Women receiving new style care were significantly more worried during pregnancy about whether their baby was thriving (1.7 v 1.5 (maximum score 5); P=0.010). Postnatally they were also more worried about coping with their baby (0.6 v 0.5 (maximum score 5); P=0.014; table 4).

Attitudes to fetus and baby—Women receiving new style care were found to have a more negative attitude to their fetus: they rated themselves as feeling more detached, concerned, and uncertain about their unborn baby (6.6 v 6.0 (maximum score 21); P=0.002). This effect continued into the postnatal period, when women in the new style group chose a higher proportion of negative words to describe their baby than women in the traditional group (0.207 v 0.188; P=0.028).

Health related behaviour—Women in the two groups did not differ significantly in whether they smoked (odds ratio 1.18; 0.95 to 1.47) or in their reported alcohol consumption in late pregnancy (P=0.854). There were also no differences in the proportions of women who gave up breast feeding (odds ratio 1.14; 0.90 to 1.44) or never started (0.99; 0.79 to 1.25).

Birth experience—The two groups did not differ in any variables relating to the birth, such as antenatal confidence about giving birth (P=0.374), satisfaction with the birth (P=0.777), or self ratings of coping with giving birth (P=0.815).

ACCEPTABILITY OF VISIT SCHEDULE

Women—For the purposes of this study we classified women as dissatisfied if, in the new style group, they would have preferred more visits or if, in the traditional group, they would have preferred fewer visits. Dissatisfaction is therefore a preference in the direction of the alternative type of care. Women in the new style group were much more likely to be dissatisfied with the number of visits they received than women in the traditional group (odds ratio 2.50; 2.00 to 3.11; table 5). This disparity in satisfaction was most evident in early pregnancy but persisted into the postnatal period (2.41; 1.90 to 3.07). Women's reporting of their partners' views (where an opinion was expressed) showed greater dissatisfaction in the new style group (10.75; 7.23 to 15.99). Women in the new style group were also more likely to feel that some of the gaps between visits were too long (2.20; 1.83 to 2.66). Although a greater proportion of women receiving new style care said they would choose the same schedule in a future pregnancy (1.41; 1.16 to 1.71), women in the traditional group rated their antenatal care more highly (3.8 v 3.6 (maximum score 5); P<0.001).

Professionals—The response rate for the questionnaire sent to all health professionals at the end of the trial was 62%. The majority (187/229; 82%) of professionals who replied reported that they would like to see a reduction (from the traditional 13 visits) in the overall number of routine antenatal visits. The remainder would not like to see a reduction (17; 7%) or were undecided (25; 11%). However, professionals were of the opinion that traditional care met the non-clinical needs of nulliparous women better than new style care (53% v 22%).

LIMITATIONS OF THE STUDY

We accept the problems associated with choosing a proxy measure for the morbidity associated with pregnancy related hypertensive disorders. The many potential influences on caesarean section rates are well documented.26 The final sample size was marginally lower than intended, which meant that we were able to test the primary hypothesis with a power of only 79.6% rather than 80%. The primary hypothesis on which the sample size was based was unidirectional, reflecting the concern of clinicians that a new style of care should not lead to an increased caesarean section rate for pregnancy related hypertensive disorders. The interesting possibility that new style care might lead to a decrease in the primary variable was not tested; this would require a substantially larger sample size. Furthermore, our sample size precludes examination of rare but serious events such as perinatal loss, eclampsia, and maternal mortality.

We made no attempt to examine the economic aspects of care, nor was it possible in a randomised controlled trial to examine fully the impact that widespread introduction of our intervention might have. Should reduced antenatal visit schedules be introduced for the majority of women there may, for example, be benefits for women and health professionals resulting from the extra time available. As the women following the new style schedule formed a minority of the total pregnant population in south east London at the time, any such effects could not be measured.

Our study also illustrates the difficulties of maintaining changes in delivery of services in a pragmatic trial. That the difference in the number of visits actually received was smaller than intended also needs to be borne in mind when interpreting the findings. We were unable to identify the reasons for the convergence seen between the two groups in the mean number of visits received, although we expected that medical problems arising during pregnancy and women changing their booking for maternity care to other hospitals would lead to some non-adherence to the schedules. Women's responses to open ended questions on the maternal antenatal questionnaire suggested that the convergence in the mean number of visits received was the result of non-adherence to the schedules by both women and professionals.

CLINICAL EFFECTS

Our findings confirm that, for the clinical variables examined, new style care is as effective as traditional care. This supports the conclusions reached from observational data, both internationally9 and in the United Kingdom.2

It is interesting that women in the traditional group had more antenatal ultrasound scans and day admissions. The first of these findings replicates that of Binstock and Wolde-Tsadik.11 These extra scans and day admissions may have been associated with the higher index of suspicion of smallness for gestational age among clinicians caring for women in the traditional group. This increased use of resources was not associated with any other measurable clinical effect, beneficial or otherwise.

DISSATISFACTION

Women were more likely to be dissatisfied with the number of visits in new style care than the number in traditional care. Over half of the women in the group receiving new style care felt that some of the gaps between antenatal visits were too long. Although more women in this group said that they would choose the same schedule in future, this indirect measure of satisfaction involves speculation about future feelings and is less relevant than the dissatisfaction women reported in their current pregnancy. This dissatisfaction may be due to their feeling that they were less likely to feel listened to and that they were less happy with the time they had to talk at visits, than women receiving traditional care. Since non-participants' dissatisfaction with the idea of new style care also needs to be taken into consideration, it is important to remember that 26% of apparently eligible women chose not to participate, mainly because they did not wish to have fewer antenatal visits. The major problem with the new style schedule seems to be its lack of acceptability to women. Dissatisfaction is important because of the increasing importance accorded to patients' views in health care evaluation, together with the high priority it has been given in Changing Childbirth.10

Although it is clear that, in the context of this trial, women in the new style group were more dissatisfied, it is not known whether this dissatisfaction would persist if new style care became the norm. There is evidence from surveys of preference and satisfaction that individuals tend to be happier with existing rather than new forms of care.27 Widespread implementation of new style care may change expectations, and with them, satisfaction with care, but there is no evidence that this would be the case.

PSYCHOSOCIAL EFFECTIVENESS

The findings relating to the psychosocial effectiveness of the two schedules need careful interpretation. Although the reported differences in psychosocial effectiveness between the groups are significant, the magnitude of the differences is small. We recognise that, when multiple comparisons are made, some findings will occur by chance. However, since all the differences in psychosocial effectiveness are in the same direction—that is, in favour of traditional care—and since they tend to cluster together around similar concepts such as feelings about the fetus or baby and talking or listening, we feel that the differences in psychosocial effectiveness are meaningful. It should also be borne in mind that health professionals in the trial felt that traditional care was better than new style care for meeting the non-clinical needs of nulliparous women.

That significantly more women receiving new style care wanted more time talking at visits is indirect evidence supporting Hall et al's finding that reducing the number of antenatal visits does not lead to any increase in consultation length.2 We have reported elsewhere that, among maternity caregivers, longer antenatal consultations were the least supported of a number of possible changes in antenatal care.28

Our findings on continuity confirm those of Hall et al that reducing the number of visits leads to reductions in the number of different caregivers seen.2 We were able to show that, although this led to a decrease in conflicting advice, it also led to women feeling less well remembered from one visit to the next. This may be because they saw their usual caregiver fewer times.

CHANGES IN CARE SCHEDULES

Any policy changes in antenatal care should aim to maintain current levels of safety and psychosocial support for women. Since the two schedules in this trial were equally effective on all the clinical variables measured, providers may decide to offer women fewer visits than traditionally scheduled. However, we have shown that a reduced schedule of care is less effective psychosocially and less acceptable to women, even though women in the trial were encouraged to book extra appointments if they felt them to be necessary. Women may find it difficult to justify, both to their caregivers and to themselves, the use of health professionals' time for non-clinical, but important, reasons such as the need for support, information, or reassurance.

If fewer antenatal visits are to be offered, it is important to ensure that women feel able to have extra visits for psychosocial reasons, not solely when visits are clinically necessary. Flexible patterns of visits that are tailored to individual needs might be developed. It may also be useful to develop forms of care, other than the routine antenatal visit, for meeting women's psychosocial needs in pregnancy. Whatever system of antenatal care is adopted, or retained, it is important that it is evaluated to ensure that a high quality of service is provided to women.

Further research on antenatal visit schedules would permit the use of meta-analysis to clarify the effect of reduced schedules on rare clinical problems such as eclampsia and perinatal mortality; it should also address the health economic implications of different schedules. In addition, future studies might usefully focus on the acceptability and psychosocial effects of reduced schedules in other service contexts, such as in midwifery group practices, or in conjunction with alternative sources of support, information, and reassurance. Finally, research will be required to evaluate any development of flexible, individualised approaches to antenatal care.

We thank the women who took part in our study and gave us their views on the care they received. We are also grateful to our secretarial and clerical staff, Maureen Statham and Becky Green, and all the midwives, general practitioners, obstetricians, and administrative staff whose cooperation ensured the completion of the trial. We thank the many individuals who gave valuable advice and help throughout the trial, in particular our steering committee. JS thanks his partners, Drs H Graver, B Essex, A Platman, and S Thomas, for supporting his research interest.