Randomised controlled trial of day care for hypertension in pregnancy
Our aim was to assess the effect of the introduction a day-care unit on the care of women with
in pregnancy. A randomised controlled trial was carried out on 54 women who presented at 26 weeks of pregnancy or later with non-proteinuric hypertension (systolic blood pressure 150-170 mm
Hg and/or diastolic pressure 90-105 mm Hg on two occasions at least 15 min apart). 30 women were allocated to care by the day unit and 24 were managed according to the established practice of their clinicians without access to the day unit (control group). Women in the control group spent on average 4·6 times longer as inpatients (difference in mean stay 4·0 days [95% confidence interval 2·1-5·9 days]) than the day-unit group and were 8·8 times (95% Cl 3·0-25·8) more likely to be admitted to hospital. Induction of labour was 4·9 times (95% Cl 1·6-13·8) more likely in the control than in the day-unit group and the development of proteinuria 11·4 times (95% Cl 1·8-71·4) more likely. The control group had a mean of 1·5 fewer hospital outpatient visits (95% Cl 0·36-2·64). The groups did not differ in their use of antihypertensive drugs. Day-unit care for hypertension in pregnancy significantly reduced the need for and the length of antenatal inpatient admissions and the number of medical interventions, at the cost of an increase in outpatient attendances. Our results are further evidence that inpatient care does not improve outcomes or prevent the development of proteinuria in this disorder.
Introduction Antenatal day care is a safe method for the assessment of with hypertension in pregnancy! but the effects of introducing day care on rates of hospital admission, duration of inpatient stay, and induction of labour have not been studied in a randomised trial. The opening of our four-bed day-care unit provided an opportunity for a randomised study to compare the effects of day care and the traditional management of non-proteinuric hypertension in pregnancy. To assess any effect of the trial itself, we also studied the process and outcomes of care before and after the trial
availability of day care could reduce duration of stay in hospital by 50%. At the time of the study the average inpatient stay for women with non-proteinuric hypertension was about 7 days, with a large standard deviation of 5 days. To have at least an 80% chance of detecting a 50% reduction in duration of stay at p < 0-05, with a one-tailed test, we needed 25 subjects in each arm of the study. The eligibility criteria were: at least 26 weeks’ gestation; no previous hospital admission during the pregnancy; and hypertension (blood pressure above 90 mm Hg diastolic and/or 150 mm Hg systolic on two readings at least 15 min apart). The first blood-pressure measurement was made by a midwife; the patient was then seen by a doctor who decided, after a second bloodpressure reading, whether or not to refer the woman to the study. Blood pressure was measured with a standard sphygmomanometer with auscultation of the Korotkov phase IV for the diastolic reading. Patients with blood pressure above 105 mm Hg diastolic or 170 mm Hg systolic or with proteinuria of any degree were excluded.
Subjects were allocated to the study groups by simple randomisation with 60 sealed, opaque, sequentially numbered envelopes. A register was kept in the day unit of all envelopes and the names of the subject and the referring clinician. The envelopes were opened by a midwifery sister, who had no prior knowledge of the patient, when a referral to the study was notified by telephone. All envelopes were accounted for at the end of the trial. Women were informed that they had been randomly assigned one type of care in a study and were given an opportunity to discuss possible alternative management. This procedure was pre-consent randomisation in a form modified from that described by Zelen.1 No patient withdrew from the study at this point. Patients in the control group were managed according to the judgment and established practice of the referring clinician. This management did not necessarily entail admission to hospital. However, one consultant subsequently referred 3 control-group women to the day unit. These "non-compliers" remain in the control group for the analysis. Patients referred to the day unit were seen the morning after referral. Five blood-pressure recordings were made; urine was checked for protein; platelet count and serum urate concentration were measured; and a cardiotocograph was done. Doppler waveform analysis and ultrasound scans for growth were done if they were considered necessary by the clinician responsible for day care. Further visits were arranged according to perceived clinical need by a clinician of at least registrar grade. After review (at each visit), the patient was discharged back to clinic, followed up in the day unit, or admitted to hospital. A record was kept of the patient’s blood pressure, weight, and stage of gestation at referral, and of the action taken after randomisation. Outpatient attendance and admissions and the proportion of time in hospital was recorded prospectively from day of entry to the study to delivery. Clinical complications, the method of management, and labour and delivery details were recorded.
Subjects and methods The hospital ethics committee gave consent for this randomised study, which took place between Aug 1, 1989, and Dec 15, 1989, soon after the unit opened. We aimed to show whether the
ADDRESS: Department of Obstetrics and Gynaecology, St James’s University Hospital, Leeds LS9 7TF, UK (D. J. Tuffnell, MRCOG, Prof R J. Lilford, MRCOG, P C. Buchan, FRCOG, V M Prendiville, RM, A. J Tuffnell, RM, M P. Holgate, RM, M. D Griffith Jones, MRCOG) Correspondence to Mr D J. Tuffnell.
TABLE I-CHARACTERISTICS AT STUDY ENTRY
BP =blood pressure
Patient satisfaction was assessed by a confidential postal questionnaire sent to all women 4-6 weeks after delivery. It consisted of thirty statements that the women scored on a Likert (5-point) scale. There were six other questions about patients’ views on the number of
attendances and admissions in this pregnancy plus the preferred method of management for any future pregnancies. The patients were asked hypothetical questions about preferred future management to evaluate the trade-off between outpatient visits at different frequencies and inpatient admission from Monday to Friday. 3,4 Women were also asked for unstructured comments. A pilot study of the questionnaire was done in 10 patients before the main study, and its suitability was assessed by a psychologist experienced in consumer questionnaires (Dr Frank Ledworth: copies of the questionnaire are available from the authors). To assess any effect that simply being in the trial might have (Hawthorne effect),5 we did a retrospective examination of the management of 36 women, ascertained from the labour ward records, who met the study criteria before the day unit opened. We also studied the management of the 42 patients referred to the day unit for evaluation of non-proteinuric hypertension immediately after the trial. We used Student’st test for parametric data and the MannWhitney U test for non-parametric data. Chi-square tests were used for contingency table analysis except for two x two tables, when Fisher’s exact test was used.6 Confidence intervals (95% CI) are given for all differences and differences between means.’
Results 54 patients were recruited to the study: 30 were randomly allocated to day care and 24 to the control group. No patient declined day care, but 5 eligible patients did not join the study because their clinicians were not prepared to manage them without access to day care. 3 patients allocated to standard care were referred to the day unit by their consultant, who no longer felt able to manage them without
day care. The day-unit and control groups were similar in terms of most relevant characteristics at referral (table I). The systolic blood pressure on referral to the study was higher in the day-unit group, but there was no difference in diastolic blood pressure. After randomisation, all patients in the day-unit group attended the unit. Of the control patients, 16were admitted to hospital without further review, 5 were reviewed by the community midwife 48-96 h later, and 3 were given appointments to the next antenatal clinic. The control group had a mean of 1 ’5 fewer outpatient visits than the day-unit group (table II). The control group
TABLE II-HOSPITAL CONTACTS AND ANTENATAL PROGRESS BY GROUP ALLOCATION
spent more time in hospital, both in terms of number of days and in percentage of time from study entry to delivery. Patients in the control group were significantly more likely (p < 0-0001) to be admitted to hospital than the day-unit patients (odds ratio 8-8 [95% CI 30-258]). Control patients were also more likely to show development of proteinuria (p=0-013; odds ratio 11.4 [1,8-71’4]), but the degree never exceeded + + on dipstick testing. Induction of labour was much more likely in the control than in the day-unit group (15 [65%] vs 8 [27%]; p = 0’004; odds ratio 4-9 [1-6-13-8]). Also there was 1 elective caesarean in the control group. However, there was no difference between the groups in mean birthweight (3340 [SD 476] vs 3321  g; difference 19 [-243 to 281]g), Apgar score at 5 min (9-3 [0.5] vs 9-3 [0-6]), or mode of delivery (control group 17 spontaneous, 4 forceps, 3 caesarean [including 1 elective] vs 20, 8, 2, respectively, in the day-unit group). 2 babies, both of mothers from the day-unit group, were admitted to the neonatal unit. The first was the infant of a diabetic mother whose labour was induced at 36 weeks’ gestation because of difficulty in diabetes control. She had attended the day unit only once, 4 weeks before delivery. The second baby was initially well but became ill 4 h after birth and showed typical signs of hypoxic ischaemic encephalopathy. The mother had attended the day unit only once, 2 weeks before delivery. All her blood-pressure readings and blood tests were normal. There was no clinical evidence of growth retardation and no further investigations were done. The mother’s subsequent care had been in the antenatal clinic. 21 of 24 control patients and 24 of 30 day-unit patients returned their postal questionnaires. There were no TABLE III-COMPARISONS OF CONTROL AND PRETRIAL GROUPS AND DAY-UN IT AND POST-TRIAL GROUPS
significant differences between the groups. 5 control and 2 day-unit patients felt they spent too long in hospital. When asked about care in any subsequent pregnancy should hypertension recur, 42 said they would have been prepared to attend as outpatients at least twice a week to avoid admission from Monday to Friday. 17 women would have been prepared to attend every day to avoid admission. The appended comments by the patients mostly expressed dissatisfaction about the number of admissions; 15 patients said they would prefer more outpatient care. No patient expressed a preference for more inpatient care. There were no significant differences in management or outcomes between the control group and a similar pretrial group or between the day-unit group and similar women treated in the day unit after the trial (table ill). Discussion This study has shown that day-care assessment can reduce inpatient stay and the rate of induction of labour in patients with non-proteinuric hypertension. In our sample size calculations, we assumed only a 50% reduction in hospital inpatient stay but a reduction of 80% was achieved. The number of inpatient days among the control patients would have been greater but for the 3 women who were referred to the day unit (for a total of 14 visits) by their consultant, despite allocation to the control group. This "non-compliance" occurred despite pretrial agreement by the clinicians for no day-unit care. Induction of labour, which was used frequently in the control group, also curtailed the inpatient days in this group. The use of bed rest in pregnancies complicated by hypertension started after a case-controlled study in 1952.8 The practice was challenged by Matthews and colleagues9 who showed a reduction in perinatal mortality in their unit after a policy of not admitting patients with non-proteinuric hypertension. They reduced hospital bed usage by 72% with urine testing for albumin as the sole screening test. Two controlled trials have investigated the effect of hospital bed and rest on admission non-proteinuric The of frequencies proteinuria, preterm hypertension.lO,l1 delivery, and perinatal death were not reduced by bed- rest. One study showed a reduction in frequency of diastolic blood pressures above 105 mm Hg," but this result contradicted the earlier study." In reviewing this evidence, Crowther and Chalmersl2 stated that "women with nonalbuminuric hypertension should not be coerced into resting in bed at home or in hospital. There is currently no good evidence to support such recommendations". Our finding of proteinuria in 5 women in the control group compared with none in the day-unit group is surprising. This difference is unlikely to be due to our missing substantial proteinuria in the day-unit group, since their urine was also checked on admission in labour. The extent of proteinuria exceeded 05 g in only 1 patient; it is possible that orthostatic proteinuria caused an anomalous result in the other cases. After completion of the randomised trial, proteinuria developed in 6 of 42 women referred to the day unit for non-proteinuric hypertension; day-unit care does not seem to have a protective effect against proteinuria. However, the low level of proteinuria in the outpatientmanaged women in our study strongly suggests that day care is no more likely than inpatient care to be associated with a worsening of the underlying disease process. Obstetricians at the Glasgow Royal Maternity unit’ have used a day unit for 10 years to assess patients with mild to
moderate non-proteinuric hypertension. In 72% of cases their initial assessment of severity proves correct. However, deterioration in the remaining 28 % shows the importance of continued surveillance of these patients. The standard classification of hypertensive disorders in pregnancy" can be confirmed only in retrospect when after delivery no proteinuria has occurred.’4 A study of more than 3000 women in Glasgow (unpublished) has proved the safety of this form of care. Our study is not large enough to show differences in obstetric outcome, but it showed an 80% reduction in inpatient stay by hypertensive patients with no adverse consequences. Although the pretrial and post-trial groups we studied are unlikely to be identical to the study and control groups, they are similar enough to suggest the absence of a large Hawthorne effect.5 We can conclude that the day unit resulted in a change in process of care but that the change was not due to the trial itself. We are therefore confident that the real differences we have shown will persist under ordinary circumstances. Our postal questionnaire had a response rate of 76%, similar to that found in other studies.15,16 We found no difference in the level of satisfaction between the control and day-unit groups with an objective scoring system. However, 15 patients made favourable subjective comments about the provision of outpatient care. 3 women in the control group said they would have preferred day-unit management in retrospect, and 3 others applied enough pressure on their clinicians to justify non-compliance with the control assignment. The high percentages of women in both groups prepared to attend as outpatients twice a week (93 %) or even every day (37%) show how strongly women prefer outpatient care. The preference for shorter hospital stay has been shown previously for postnatal care.l’ Women with children are especially, and justifiably, averse to long-term
inpatient care. Dawson et ap8 showed that telephonic monitoring of the fetal heart is also associated with fewer hospital admissions and a smaller percentage time in hospital. However, in that study, unlike ours, the standard protocol for the control group at entry to the study was hospital admission. It is appropriate that more attention is being paid to the reasons for antenatal admission; with closer appraisal fewer women will need to be subjected to hospital admission for little or no benefit. REFERENCES case for early recognition and intervention in pregnancy induced hypertension. In: Sharp F, Symonds EM, eds. Hypertension in pregnancy. Proceedings of the Sixteenth Study Group of the Royal College of Obstetricians and Gynaecologists; July, 1986. 2. Zelen MA. A new design for randomised controlled trials. N Engl J Med 1979; 300: 1242-45. 3. Thornton J, Lilford RJ. The caesarean section decision: patients’ choices are not determined by immediate emotional reactions. J Obstet Gynaecol 1989; 9: 283-88. 4. Thornton J, Lilford RJ. Prenatal diagnosis of Down’s syndrome: a method for measuring the consistency of women’s decisions. Med Decis Making 1990; 10: 288-93. 5. Lilford RJ. Clinical experimentation in obstetrics. Br Med J 1987; 295: 1298-300. 6. Johnson N, Lilford RJ. Statistics in obstetrics and gynaecology. In: Philipp E, Setchell M, Ginsburg J, eds. Scientific foundations in obstetrics and gynaecology. London: Butterworth, 1991: 723-38. 7. Gardner MJ, Altman D. Confidence intervals rather than p values: estimation rather than hypothesis testing. Br Med J 1986; 292: 746-50. 8. Hamlin RHJ. The prevention of eclampsia and pre-eclampsia. Lancet 1952; i: 64-68. 9. Matthews DD, Patel IR, Sengupta SM. Outpatient mangement of toxaemia. J Obstet Gynaecol Br Commnwlth 1971; 78: 610-19. 10. Matthews DD. A randomised controlled trial of bed rest and sedation or normal activity and non-sedation in the management of non-
1. Walker JJ. The
albuminuric hypertension in late pregnancy. Br J Obstet
1977; 84: 108-11. 11. Crowther CA, Boumeester A, Ashurst H. Home management versus hospitalisation for women with non-proteinuric hypertension in pregnancy. In: Chalmers I, Enkin M, Keirse MJNC, eds. Effective care in pregnancy and childbirth, vol 1. Oxford: Oxford University Press, 1989: 624-32. 12. Crowther CA, Chalmers I. Bed rest and hospitalisation during pregnancy. In: Chalmers I, Enkin M, Keirse MJNC, eds. Effective care in pregnancy and childbirth, vol 1. Oxford: Oxford University
Press, 1989: 624-32. 13. Davey D, MacGillivray I. The classification and definition of the hypertensive disorders of pregnancy. Am J Obstet Gynecol 1988; 158: 892-98.
14. Lilford RJ. Classification of hypertensive disorders of pregnancy. Lancet 1989; ii: 112-13. 15. Cartwright A. Interviews or postal questionnaires? Comparison of data about women’s experiences with maternity services. Milbank Q 1988; 66: 172-89. 16. Cartwright A. Some experiments with factors that might affect the response of mothers to a postal questionnaire. Stat Med 1986; 5: 607-17. 17. Burnell I, McCarthy M, Chamberlain GVP, Hawkins DF, Elbourne D. Patient preference and postnatal hospital stay. J Obstet Gynaecol 1982; 3: 43-47. 18. Dawson AJ, Middlemiss C, Coles EC, Gough NAJ, Jones ME. A randomised study of a domiciliary antenatal care scheme: the effect on hospital admissions. Br J Obstet Gynaecol 1989; 96: 1319-22.
Contraceptive efficacy of lactational amenorrhoea KATHY I. KENNEDY
Pregnancy is rare among breastfeeding women with lactational amenorrhoea. The lactational amenorrhoea method (LAM) is the informed use of breastfeeding as a contraceptive method by a woman who is still amenorrhoeic, and who is not feeding her baby with supplements, for up to 6 months after delivery. Under these three conditions, LAM users are thought to have 98% protection from pregnancy. It can be difficult, however, to determine when supplementation of the baby’s diet begins. We have analysed data from nine studies of the recovery of fertility in breastfeeding women to assess the effectiveness of lactational amenorrhoea alone, irrespective of whether supplements have been introduced, as a fertility regulation method post partum. Cumulative probabilities of ovulation during lactational amenorrhoea were 30·9 and 67·3 per 100 women at 6 and 12 months, respectively, compared with 27·2 at 6 months when all three criteria of the LAM were met. Cumulative pregnancy rates during lactational amenorrhoea were 2·9 and 5·9 per 100 women at 6 and 12 months, compared with 0·7 at 6 months for the LAM. The probability of pregnancy during lactational amenorrhoea calculated from these studies is similar to that of other modern contraceptive methods, and it seems reasonable for a woman to rely on lactational amenorrhoea without regard to whether she is fully or partly breastfeeding. So that amenorrhoea and fertility suppression can be maintained, counselling about good breastfeeding and weaning practices
CYNTHIA M. VISNESS
fully breastfeeding are 98% protected from pregnancy for 6 months after delivery.2,3 The lactational amenorrhoea method (LAM) has been developed as a way for individual women to apply this advice for family planning.4 The Bellagio guidelines suggest that natural contraceptive protection is conferred only before the introduction of supplements to the baby’s diet. Supplementation is thought to reduce the baby’s need to suckle, which reduces the neuroendocrine stimulus at the breast and hastens the return to fertility. Thus, supplementation is often associated with the return of menses and the recovery of fertility.5,6 However, the introduction of supplements does not automatically mean a return to fertility, since supplements do not, at least initially, always replace or substitute for breastfeeding. Few women solely breastfeed their babies for a long time, so most women may think that they cannot use the LAM.7 Since post-partum fertility is influenced more strongly by lactational amenorrhoea than by supplementation, interest has emerged in quantifying the degree of contraceptive protection associated with lactational amenorrhoea irrespective of whether have been introduced. supplements Only 3-10% of women are likely to conceive during lactational amenorrhoea8-lO and time post partum is positively associated with the risk of pregnancy .11,12 We have calculated the risk of ovulation and pregnancy among breastfeeding women from four continents irrespective of whether there was amenorrhoea or supplemention of the infant’s diet; when there was amenorrhoea but irrespective of supplementation of the infant’s diet; and when women were fully breastfeeding and amenorrhoeic. The sample size is larger than in other studies11,12 and women from both developed and emerging countries are included, so there is a wide variety of infant-feeding patterns.
Methods were collected from nine prospective studies of lactating by Family Health International in Mexico,13 Thailand,14 Egypt,15 Pakistan,16 rural17 and urbanlg Philippines, and Canada, Australia, and England (unpublished; copies available from the authors). Six studies included ovarian hormone assays of serum or
Introduction Extended periods of breastfeeding can have a substantial effect on population fertility rates,1 but breastfeeding is widely regarded as being unreliable for personal contraception. These two seemingly contradictory truths have been reconciled in the Bellagio consensus statement on the use of breastfeeding as a family planning method, which concluded that amenorrhoeic women who are fully or nearly
ADDRESS: Family Health International, Research Triangle Park, North Carolina 27709, USA (K. I. Kennedy, MA, C M Visness, MPH). Correspondence to Ms Kathy I Kennedy.