By Caroline Flint

A randomized trial of continuity of care by a team of midwives

A team of four midwives provided the care during pregnancy, labour and the puerperium to 503 women at low obstetric risk, over a two-year period. Compared with standard hospital care randomly allocated to 498 women this `Know Your Midwife' scheme was associated with greater continuity in all phases of maternity care. The scheme appeared very acceptable to women: they spent less time in the antenatal clinic, and overall, felt more satisfied, better prepared and better able to discuss problems. The scheme was characterized by less obstetric intervention particularly in respect of augmentation of labour and intrapartum analgesia; labours tended to be longer. Neonatal outcome was generally similar in the two groups but the size of the trial did not allow a precise assessment of differential effects in these terms. The `Know Your Midwife' scheme is feasible. It should now be introduced more widely but in a way which allows continuing evaluation.

The care received by many women during pregnancy, labour and puerperium is fragmented and lacking in continuity (Royal College of Obstetricians and Gynaecologists, 1982). The distress which this can cause is well recognized and there have been calls for changes in the organization of perinatal care (Social Services Committee, 1980; Association of Radical Midwives, 1986; Royal College of Midwives, 1987) so that a woman `can build up a relationship of trust with the staff she meets' (Maternity Services Advisory Committee, 1982). The Maternity Services Advisory Committee, for example, has suggested that `efforts should be made to involve the same group of staff at each visit' and ideally', `women should have the same midwife to attend them in labour as in antenatal period' (Maternity Services Advisory Committee, 1982).

The `Know Your Midwife' scheme was established at St George's Hospital, London, in 1983 to try and meet these ideals. The in tention was that a team of four midwives would give continuity of care during pregnancy, labour and the puerperium to 250 women each year, with the back up of the hospital obstetricians. Suitability for inclusion in this scheme, of women who wished to have all their care at the hospital, would be assessed after review of the past medical history and first physical examination. Thereafter most care would be given by the team of four midwives but an obstetrician would be consulted routinely at 36 and 41 weeks of pregnancy (if applicable) and at any other time when the midwife requested it.

To evaluate the scheme it was introduced in the context of a randomized controlled trial. The findings of this study are summarized here. A more detailed report (Flint and Poulengeris, 1986) is available on request from one of us (CF).

There were two main purposes of this study. The first was to assess whether the provision of total perinatal care by a team of four midwives is feasible and does, indeed, result in greater continuity of care. The latter was measured by the number of caregivers seen during pregnancy and labour and whether women had met their intrapartum caregiver during pregnancy. The second was to assess the scheme's acceptability to women at low obstetric risk. The differential effects of the two policies on the clinical management and outcome of perinatal care were also examined, but it was recognized in advance that a trial of the size envisaged would provide only imprecise estimates of these, particularly in respect of relatively uncommon adverse outcomes.

The trial was conducted at St George's Hospital, London. Women `booking' for delivery at the hospital between April 1983 and March 1985 were considered for entry if they were likely to receive their full antenatal care at the hospital. The reasons for attending the hospital for all care were either the general practitioner did not offer `shared care' or the women preferred to attend the hospital. The entry criteria were:

• the woman was over five feet tall (132 cm) o she had no serious medical problems
• there had been no previous uterine surgery o there was no past obstetric history of more than two miscarriages or terminations of pregnancy, or of stillbirth or neonatal death, or of a baby with intrauterine growth retardation or preterm delivery
• there were no Rhesus antibodies in the cur rent pregnancy.

After the first visit to the hospital, women who met these criteria of low risk were then randomly allocated, using sealed opaque envelopes, to one of two forms of hospital- based care. Four hundred and ninety eight women received the hospital's standard antenatal, intrapartum and postpartum care for low risk women. The other group (503 women) were sent a letter offering antenatal, intrapartum and postpartum care by a team of four midwives as part of the `Know Your Midwife' scheme (Flint, 1986). Forty three (nine per cent) declined; in practice these women received the standard form of hospital care but for the purpose of analysis they have been retained in the `Know Your Midwife' group to avoid selection bias. The final sample size was dictated by the maximum length of time (two years) that recruitment could continue.

All data were collected in the same ways for both groups. Clinical data on all the women in the trial were transcribed from the casenotes after delivery. The casenotes of all 101 women booked to deliver in a 2-month period during the trial were studied in greater detail to describe the caregivers who were actually seen by the women.

Assessment of maternal satisfaction was limited to the final 559 women recruited in the trial. A standardized questionnaire was given out at 37 weeks of pregnancy to 285 women in the `Know Your Midwife' group and to 274 women in the control group; this was returned by 277 (97 per cent) and 268 (98 per cent) respectively. Further questionnaires were distributed at two days and six weeks after delivery to 279 women and 267 women (for compassionate reasons women whose baby was in the Special Care Nursery were not included). The first of these perinatal questionnaires was returned by 275 (99 per cent) and 261 (98 per cent) women, and the second by 249 (89 per cent) and 227 (85 per cent) women. Where appropriate, the source of data is indicated on the Tables. Clinical outcome data are not available for 15 women allocated to the `Know Your Midwife' scheme and for 19 women in the control group because they moved away during pregnancy. Individual items are missing for a few women in both the casenotes and questionnaire datasets and this is why the denominators vary slightly.

Chi square and Student's `t' statistical tests were used as appropriate. Confidence intervals of the relative rates were calculated using the method recommended by Morris and Gardner (1988), 99 per cent rather than 95 per cent intervals are presented because of the many comparisons made between the groups. The confidence interval gives a range in which the true effect is likely to lie. If the 99 per cent confidence interval does not include unity the difference between the two groups is statistically significant at the one per cent level (p<0.01).

The two groups derived by random allocation were similar at entry in most respects (Table 1); there were, however, more Asian women in the control group and more smokers in the `Know Your Midwife' group.

The `Know Your Midwife' scheme was associated with greater continuity of care (Table 2). Fewer caregivers were seen during pregnancy. During labour, the `Know Your Midwife' group saw both fewer midwives and fewer doctors (Table 2). The caregiver during labour already known to the women in the standard care group was usually a junior doctor whom they had met in the antenatal clinic; the caregiver known to the women in the `Know Your Midwife' group was invariably one of the four `Know Your Midwife' midwives who stayed with the women throughout labour.

The `Know Your Midwife' scheme appeared to be more acceptable to the women in almost every aspect investigated. Some of the parameters which differed between the groups are described in Table 3. In the antenatal clinic there were fewer delays, easier discussion of anxieties and greater satisfaction generally. Two days after the delivery, women in the `Know Your Midwife' group were more likely to feel that they had been well prepared for labour and to rate those who had been looked after them during labour as `very caring'. Six weeks after delivery there were striking differences in the women's recollection of labour, about their preparedness for child care and the ease with which they could discuss problems during their puerperium.

Clinically, the `Know Your Midwife' scheme was characterized by less obstetric intervention (Table 4). The difference in the rates of augmentation, epidural and intramuscular analgesia and episiotomy were statistically significant. Overall, rates of perineal trauma were, however, almost identical because of a reciprocal increase in vaginal tears in the `Know Your Midwife' group. The tendency for labours to be longer in the ` Know Your Midwife' group may be related to the difference in augmentation of labour, but could also be due to a different definition of the onset of labour.

Neonatal outcome was generally similar in the two groups (Table 5). There were more low five minute Apgar scores in the `Know Your Midwife' group but neonatal resuscitation was used significantly less often in this group and the final status on the delivery suite was similar as judged by the numbers of babies who were transferred to the Special Care Nursery. There were four stillbirths and four neonatal deaths in the `Know Your Midwife' group compared with two stillbirths and two neonatal deaths in the standard care group. Review of these 12 cases by an experienced obstetrician suggests that it is unlikely that any could have been prevented by a change in care. With such a small number of events the confidence intervals are wide (Table 5) and include both a clinically important increase and a clinically important decrease in the risk of perinatal death associated with the `Know Your Midwife' scheme.

Secondary analyses showed that adjustment for ethnicity and smoking makes no statistically significant difference to these findings. Women of European origin were more likely to have an epidural and less likely to have pethidine analgesia. These adjustments tended to accentuate the contrast between the groups in epidural use and slightly reduce the difference in pethidine use.

The experiment has shown that a team of four midwives successfully cared for 500 women at low obstetric risk over a two-year period.

The difference in the five-minute Apgar scores is likely to reflect the difference in the use of neonatal resuscitation; the rates of admission to the Special Care Nursery do not suggest a major adverse effect of the `Know Your Midwife' policy on pregnancy outcome. Although there were more deaths in the `Know Your Midwife' group the total number is small and the confidence interval is very wide; as recognized at the outset, the trial was too small to assess with any precision the safety of the scheme in terms of perinatal mortality.

The scheme does demand extra responsibility from the midwives and this change in practice was sometimes stressful, particularly at first, for both those implementing the scheme and those administering it. The provision of continuity of care over a 24-hourperiod also meant that occasionally a midwife worked for long hours without formal back-up at nights or at the weekends. Expansion of the team to five or six midwives would allow a second midwife `on call' and greater flexibility in the rota, although this would probably reduce continuity.