child births in a modified midwife managed unit: selection and transfer according to intended place...

Acta Obstet Gynecol Scand 2001; 80: 206–212 Copyright C Acta Obstet Gynecol Scand 2001

Printed in Denmark ¡ All rights reservedActa Obstetricia et

Gynecologica ScandinavicaISSN 0001-6349

ORIGINAL ARTICLE

Child births in a modified midwife managedunit: Selection and transfer according tointended place of deliveryJAN HOLT1, INGAR NIKOLAI VOLD3, BJØRN BACKE3, MAY VOLLNES JOHANSEN2 AND PÅL ØIAN4

From the 1Department of Pediatrics, Nordland Central Hospital, Bodø, and the Department of Obstetrics and Gynecology,2Lofoten Hospital, Gravdal, 3University Hospital of Trondheim, Trondheim and the 4University Hospital of Tromsø, Norway

Acta Obstet Gynecol Scand 2001; 80: 206–212. C Acta Obstet Gynecol Scand 2001

Background. As small obstetrical departments may not be able to give second-level perinatalcare, the delivery unit at Lofoten hospital was for the years 1997–98 reorganized to a modifiedmidwife managed unit. Women at low obstetrical risk were delivered at this unit and womenat high risk were referred to the central hospital. We assessed the effectiveness of the riskselection.Material and methods. The study was a prospective, pragmatic, population-based trial. De-sired outcome was defined as a non-operative delivery at 35–42 weeks gestational age givingan infant not needing resuscitation. Intermediate outcomes: Operative deliveries, infantstransferred to neonatal intensive care unit and infants diverging from normal. The intendedplace of delivery was ultimately decided at admittance to the midwife managed unit.Results. Of the 628 women in study 435 (69.3%) gave birth at the midwife managed unit, 152(24.2%) were selected to be delivered at the central hospital and 41 (6.5%) were transferredto the central hospital after admittance to the midwife managed unit. Desired outcome wasrecorded in 94% of the deliveries at the midwife managed unit as compared to 50.3% at thecentral hospital. Women who intended to be delivered at the midwife managed unit, neededfewer operative deliveries and relatively few infants were transferred to the neonatal intensivecare unit or diverged from normal.Conclusions. As nearly 70% of the births occurred at the midwife managed unit and 94% ofthese deliveries had a desired outcome, this indicates an effective selection process. This modelmight be an alternative to centralization of births in sparsely population areas.

Key words: delivery unit; midwife; perinatal care; selection; transfer

Submitted 11 April, 2000Accepted 10 October, 2000

Over the past 20 years there has been a centraliza-tion of births in Norway. The total number of birthinstitutions has been reduced from 131 in 1974 to58 in 1994. Long distances and unpredictableweather conditions might be in favor of centraliza-tion. However, it can be claimed that Norway still

Abbreviations:NICU: neonatal intensive care unit; SGA: small for gestationalage; LGA: large for gestational age; GA: gestational age; OR:odds ratio; CI: confidence interval.

C Acta Obstet Gynecol Scand 80 (2001)

has a decentralized birth organization for the totalof approximately 60,000 births per year as com-pared to Sweden (68 institutions, 117,000 birthsper year).

Traditionally, the obstetrical care in local hospi-tals (200–400 deliveries per year) has been sharedbetween midwives and hospital doctors. Small hos-pitals most frequently have one obstetrician work-ing as a consultant during day time and sharingcall with the general surgeons. Surgeons no longeraccept responsibility for a delivery unit and most

Child birth in a midwife managed unit: Selection and transfer 207

of them want to bring this involvement to an end.As these hospitals are often located in sparselypopulated areas, the options are either to closedown the delivery unit or to establish new pro-grams for the delivery care.

Following a public debate about the birth organ-ization, The Norwegian Board of Health proposedthree levels of birth institutions: Level 1 is matern-ity homes run by midwives and general prac-titioners with 40–500 deliveries per year; Level 2 issmall obstetrical departments with 500–1500 deliv-eries per year and obstetrician and anesthesiologiston duty; Level 3 is obstetrical departments withmore than 1500 deliveries per year and obstetri-cians, pediatricians and anesthesiologists on dutyall the time.

Several reports have documented the feasibilityand safety of midwife managed care (1–4). Womenat low risk for obstetrical complications may evenbe more satisfied with the maternity care at theseunits (4) and ‘woman centered’ care has increasedthe demand for this care to be offered to womenat or near their place of living. It is not easy, how-ever, to identify those women who will be at riskduring labor and delivery (5). The traditional cri-teria for patients booked for isolated general prac-titioner units and home confinements have beenquestioned (6) and high transfer rates from mid-wife managed units to consultant care been re-ported (2, 4). The effectiveness of the selection pro-cess has not been fully assessed (7).

The selection and transfer process aims at avoid-ing operative deliveries in midwife managed unitsand to ensure that preterm infants are born in aunit located close to a neonatal intensive care unit(NICU). From 1997 the delivery unit of Lofotenhospital has been reorganized to a modified mid-wife managed unit. We report a two year studyfrom this isolated rural hospital with the aim ofestimating the feasibility and the effectiveness ofthe case selection process.

Material and methods

Twenty-three thousand people are living in the Lo-foten islands. Transport from Lofoten to the cen-tral hospital in Bodø (approximately 110 kilomet-ers across the sea) is possible by a coastal liner(four hours), aircraft (30 minutes) or by ferry andcar (five hours). Three fixed wing ambulance air-craft located in Bodø, Tromsø and Brønnøysundor a Sea-king helicopter dedicated for search andrescue might be used.

Before reorganization of the obstetrical depart-ment of Lofoten hospital this unit cared for ap-proximately 300 deliveries and 35 cesarean sectionsannually. High-risk women (approximately 8%)


were delivered at the central hospital. Midwivesand one of two surgeons or an obstetrician wereon duty. For sick newborns a transport team wasavailable from the central hospital (8).

During the study period no cesarean sectionswere planned at the modified midwife managedunit, however, the isolated location of this hospitaland unpredictable weather conditions necessistatedobstetrical back up by surgeons or one obstetri-cian.

Included in the study were all women from thefour communities of Lofoten who gave birth eitherat the midwife managed unit (or outside the hospi-tal in the catchment area) or at the central hospitalbetween 1 January 1997 and 31 December 1998.Births at the other delivery units or from womenliving outside the Lofoten islands were covered butnot included in the analysis.

Antenatal care was provided by a midwife at thelocal health center (from the community or fromthe hospital) and/or a general practitioner. In ad-dition, all women were examined twice at themodified midwife managed unit (at 16–18 weeks-and at 35 weeks of gestational age), by one of thehospital’s midwives. An ultrasound examinationwas performed at 16–18 weeks to estimate the dateof confinement. Prior to study a detailed list ofselection and transfer criteria was agreed upon.This list was used by the midwives to select highrisk cases for referral to the central hospital. Incase of doubt the obstetrician at Lofoten hospitalwas consulted. In recording consecutively the mainreason for selecting a woman to be delivered at thecentral hospital, the cause first pointed out duringthe antenatal period was chosen, for instance: twinpregnancy was given preference to breech presen-tation. Women selected to give birth at the centralhospital were invited to travel to Bodø before onsetof labor or stay home until onset of labor and thenuse a fixed wing ambulance aircraft.

The hospital records for all delivieries wereexamined. Data for infants were gathered fromprotocols at the two delivery units and from caserecords at the NICU. To compare rates for operat-ive deliveries during the study period to rates from1992–96, protocols at Lofoten hospital and thecentral hospital were explored.

Ultrasound-based fetal growth curves (9) wereapplied. Large for gestational age (LGA) was de-fined as mean birth weight π2 s.d. and small forgestational age (SGA) as mean birth weight – 2s.d.

Recordings of flight times from helicopter andfixed-wing ambulance aircraft were combined withthe time of birth to give the time interval betweentransfer and birth.

The study was conducted as a pragmatic trial

208 J. Holt et al.

Table I. Baseline characteristics of women (and infants) delivering at the midwife managed unit or at the central hospital. Values are given as medians (interquar-tiles) or numbers (%)

Midwife managed unit Central hospital Total(nΩ435 women) (nΩ193 women) (nΩ628 women)(nΩ436 infants) (nΩ205 infants) (nΩ641 infants)

Maternal age (years) 28.4 (24.7–32.7) 28.8 (24.8–32.9) 28.6 (24.8–32.7)Primipara 137 (62.6) 82 (37.4) 219 (100)Multipara 298 (72.9) 111 (27.1) 409 (100)Twin pregnancies 1 (7.6) 12 (92.3) 13 (100)Birth weight (g) 3570 (3255–3883) 3515 (2830–3955) 3560 (3180–3910)

(10). All cesarean sections performed at Lofotenhospital during the study period were discussed bya panel using the perinatal audit concept (11).

In assessing the effectiveness of the case selec-tion we defined effectiveness according to MuirGray (12): ‘The effectiveness of a healthcare pro-fessional or service is the degree to which the de-sired outcomes are achieved’. Desired outcome wasdefined as a non-operative delivery at 35–42 weeksof gestational age, with the outcome of an infantnot needing resuscitation.

We applied three intermediate outcome meas-ures:– Operative delivery: Deliveries by cesarean sec-

tion, forceps or vacuum extraction.– Transferred to NICU: Infants transferred to the

NICU.– Infants diverging from normal: Infants dis-

playing one or more of the following features:SGA, LGA, born before 35 weeks of gestationalage or in need of positive pressure ventilationimmediately after birth.When comparing intermediate outcomes, we

used the concept of intention to treat: Intermediateoutcomes for women intended to be delivered at

Table II. Main reasons for selection of women to deliver at the central hospital(selection before admission to the midwife managed unit). Values are num-bers (percentages) of those selected

Primipara Multipara Total(nΩ56) (nΩ96) (nΩ152)

Previous obstetric history 0 56 56 (36.8)Breech presentation 12 9 21 (13.8)Current health problems 9 4 13 (8.6)Twin pregnancies 7 4 11 (7.2)Fetal distress 6 4 10 (6.6)Threatening preterm delivery 5 5 10 (6.6)Maternal request 6 2 8 (5.3)Postterm pregnancy (±42.0 weeks) 4 4 8 (5.3)Estimated birth weight ±4500 g 2 3 5 (3.3)Preeclampsia 1 3 4 (2.6)Prolonged rupture of the membranes 3 0 3 (2.0)Growth restriction 1 1 2 (1.3)Placenta previa 0 1 1 (0.7)


the midwife managed unit (women who deliveredat the midwife managed unit added those trans-ferred to the central hospital) were compared tointermediate outcomes for women selected and ac-tually delivered at the central hospital. The ulti-mate intended place of delivery was decided at ad-mittance to the midwife managed unit.

Medians with interquartiles are given for demo-graphic data. Proportions were compared usingthe chi-square test or the Fisher’s exact test. Com-parisons between groups were made using theMann-Whitney test. Results are expressed as oddsratio (OR) with a 95% confidence interval (95%CI). A p-value of 0.05 or less was considered sig-nificant. All statistics were performed using thestatistical package InStat version 3.0, GraphPadSoftware, Inc. San Diego, CA 92121, USA.

Results

A total of 628 women residing in Lofoten gavebirth during the study period, 435 (69.3%) at themidwife managed unit and 193 at the central hos-pital. Baseline variables for the study groups aregiven in Table I.

Selection

Before being admitted to the midwife managedunit 152 women (24.2%) were selected to be de-livered at the central hospital. The main reasonsare shown in Table II. For 56 of 96 multiparouswomen the obstetric history was the main reason(cesarean section: 40; vacuum/forceps: six; retainedplacenta/postpartum hemorrhage: five; perinataldeaths: three; others: two).

Means of transport for 152 women selected tobe delivered at the central hospital were: Fixed-wing ambulance aircraft (107), airliner (26), heli-copter (4), other (13), unknown (two).

Transfers

Of 476 women at low obstetrical risk admitted tothe midwife managed unit, 41 (6.5% of the study


Table III. Main reasons for transfer to the central hospital after admission tothe midwife managed unit. Values are numbers (percentages) of transfers

Primipara Multipara Total(nΩ26) (nΩ15) (nΩ41)

Request for epidural analgesia 10 1 11 (26.8)Prolonged rupture of membranes 5 4 9 (22.0)Preeclampsia 4 2 6 (14.6)Infection suspected 0 5 5 (12.2)Fetal heart rate abnormalities 3 1 4 (9.8)Delay in first stage of labor 1 1 2 (4.8)Postterm pregnancy (±42.0 weeks) 2 0 2 (4.8)Breech presentation 1 0 1 (2.4)Preterm hemorrhage 0 1 1 (2.4)

Table IV. Mode of delivery at the midwife managed unit and at the centralhospital

Midwife managed unit Central hospital

P0 P1π tot P0 P1π tot Total

Number of deliveries 137 298 435 82 111 193 628Spontaneous vertex delivery 123 283 406 41 64 105 511Birth before arrival at unit 0 6 6 0 0 0 6Cesarean section 6 7 13 25 36 61 74Forceps or ventouse 8 2 10 11 8 19 29Breech birth 0 0 0 5 3 8 8

P0Ωprimiparous women; P1πΩmultiparous women.

population) were transferred to the central hospi-tal. Table III displays the main reasons for transfer.

Aircraft were used for all transfers from the mid-wife managed unit (fixed-wing ambulance aircraft:36; helicopter: three; airliner: two). For the 29women arriving at the central hospital in labor, themedian time from arrival to birth was six hours

Table V. Complications and adverse clinical indicators occurring at the midwife managed unit and the central hospital

Midwife managed unit Central hospital(nΩ435 women) (nΩ193 women)(nΩ436 infants) (nΩ205 infants) OR (95% CI)

Apgar score 1 minuts ∞8 25 35 0.30 (0.17–0.51)Apgar score 5 minutes ∞8 2 12 0.07 (0.07–0.33)Resuscitation ª (ventilation) 2 18 0.05 (0.01–0.21)Infant transferred to NICU 7 41 007 (0.03–0.15)Preterm birth (∞35.0 weeks) 3 25 0.05 (0.01–0.16)Small for gestational age 14 13 0.49 (0.23–1.06)Large for gestational age 18 26 0.30 (0.16–0.55)Postpartum hemorrhage ±500 ml 27 5 2.64 (1.00–6.96)Postpartum hemorrhage ±1000 ml 3 3 0.46 (0.09–2.33)Third degree tear 6 5 0.56 (0.17–1.85)Retained placenta 1 5 0.09 (0.01–0.79)Stillbirth 11 2 0.23 (0.02–2.59)Neonatal death 0 3 0.07 (0.00–1.29)

NICU: Neonatal intensive care unit.1Born at home.


and 15 minutes. Two women delivered 55 minutesafter arrival. No births occurred during transportfrom Lofoten to the central hospital.

Mode of delivery

During the last five years before the study period,325 operative deliveries (242 cesarean sections)were performed among altogether 1530 womenfrom the study area. The proportion of operativedeliveries was lower during the study period thanduring 1992–96; 16.4% versus 21.2% (OR 0.73 95%CI 0.57–0.93). The cesarean section rate decreasedaccordingly; 11.8% versus 15.8% (OR 0.71 95% CI0.54–0.94). The modes of delivery for the studypopulation are given in Table IV. During the studyperiod 13 cesarean sections were performed at themidwife managed unit (assumed fetal distress: five;prolonged labor: four; partial placental abruption:two; preterm/breech presentation: one; prolapsedcord: one). The audit panel concluded that someof the clinical situations preceding the decision todeliver could have been avoided and that mostprobably all 13 women might have been trans-ferred to the central hospital without any appreci-able risk for fetus or woman. These cesarean sec-tions were recorded as acute cesarean sections (de-cided ,eight hours before delivery). Of 61cesarean sections performed at the central hospital32 were acute and 29 elective operations.

Six infants were accidentally born before arriv-ing in the midwife managed unit (general prac-titioners office: three; unplanned at home: two;ambulance: one). There were no planned home de-liveries.

Out of the 27 breech presentations, 26 were de-livered at the central hospital. Two cases of persist-

210 J. Holt et al.

Table VI. Intention to treat analysis: Intermediate outcomes for women intended to be delivered at the midwife managed unit (women delivering at the midwifemanaged unitπwomen transferred to the central hospital) compared to intermediate outcomes for women selected to be delivered at the central hospital

Midwife managed unit Central hospital(nΩ476 women) (nΩ152 women)

Intermediate outcomes (nΩ477 infants) (nΩ164 infants) OR (95% CI)

Operative delivery 37 66 0.11 (0.07–0.17)Transferred to NICU 18 30 0.18 (0.09–0.32)Infants diverging from normal 33 58 0.14 (0.08–0.22)

ent transverse lie were managed by cesarean sec-tion at the central hospital.

For the total of 23 women having an operativedelivery at the midwife managed unit, one womanwas tranferred to the central hospital four daysafter delivery and one infant with multiple malfor-mations was transferred to a university hospital.

Seven women delivering at the midwife managedunit were, in advance, selected to be delivered atthe central hospital. Three of these women had anoperative delivery and three infants were trans-ferred to the central hospital.

Complications and adverse clinical indicators

Table V shows complications and adverse clinicalindicators recorded at the delivery unit. Of 26women delivering LGA infants at the central hos-pital 24 women were selected before admittance tothe midwife managed unit.

Seven infants (1.6%) were transferred from themidwife managed unit to the NICU (suspected in-fection: three, multiple malformations: two, transi-ent tachypnoe of the newborn: one, preterm/SGA:one).

Two infants were stillborn at the central hospital(placental abruption: one; preterm rupture of themembranes at 23 weeks gestation: one). Three in-fants born at the central hospital died during theneonatal period (major malformations: two; as-phyxia (placental abruption): one).

Outcome measures

The criteria for desired outcome were fulfilled by409 women (94%) at the midwife managed unitand by 97 women (50.3%) at the central hospital.An intention to treat analysis comparing the groupof women intended to be delivered at the midwifemanaged unit to those selected to be delivered atthe central hospital is given in Table VI.

Discussion

During the study period almost 70% of the womenin the catchment area gave birth at the midwife


managed unit. A similar uptake rate of 70% is re-ported from New Zealand (13) where midwivesduring recent years have replaced general prac-titioners as providers of obstetrical care in smalldistrict hospitals. Others (2, 14, 15) have reportedlower rates (46–55%). Of the selected low-riskgroup of women admitted to the midwife managedunit only 41 (8.6%) were transferred to the centralhospital. This transfer rate is lower than reportedfrom other studies (13–16%) (2, 15).

As the obstetrical department of Lofoten hospi-tal in future might be classified as a level 1 deliveryunit, the reorganization aimed at reducing obstet-rical emergencies and avoiding operative deliveriesat this unit. We consider the operative delivery rateof 5.3 as rather low. Also, the cesarean section ratein the study population decreased. The obser-vation that decentralized maternity care for se-lected low-risk women can be associated with a lowrate of intervention, is consistent with previous re-search (13, 16). This effect may be caused by theone-on-one support and care given by the mid-wives (17). Only eight of the 484 low-risk womenoffered to deliver at the midwife managed unit pre-ferred to be delivered at the central hospital. In thepresent study the midwives were responsible for therisk assessment and played a major complement-ary part in the community based antenatal care(18). We believe that the relations and confidencethus created before labor may be of substantial im-portance in explaining a lowered cesarean sectionrate.

Operative deliveries and transfer in labor weremore frequent among primiparas than among mul-tiparas. One in six primiparous women was trans-ferred to the central hospital compared to one in21 multiparous women. For those giving birth atthe midwife managed unit the rates for instrumen-tal delivery differed significantly among primi-paras and multiparas (5.8% versus 0.7%), but theabsolute number was small. Primiparous womenshould be advised that they stand a substantiallyhigher risk than multiparous women for beingtransferred to the central hospital. As maternalheight, age and fetal weight may predict operativedelivery in low risk primiparous women (19), ap-


praisal of these factors might improve the selectionamong primiparas.

Future organization of obstetrical care in thishospital should include the possibility of perform-ing emergency cesarean sections if weather con-ditions exclude transfer. General surgeons must betrained to perform these few cesarean sections (ap-proximately 3–5 per year). The decision to performemergency cesarean sections will be made by themidwife in charge and the surgeon on duty. Con-sultations by phone call or fax (cardiotocograms)might in advance be made to the obstetrical de-partment at the central hospital. A modified mid-wife managed unit might thus be more acceptablefor surgeons.

Only five women were selected to be delivered atthe central hospital because of a suspected macro-somic baby (expected birth weight .4500 g). Sig-nificantly more LGA babies were, however, bornat the central hospital and 24 of these 26 babieswere picked out during the antenatal period. Oneexplanation for this selection might be the generalimpression these women display upon midwivesperforming antenatal care. Considering SGA in-fants the difference between the two birthplaceswas not so marked. Earlier studies (20, 21) haverecorded a low detection rate (7–14%) for intrauterine growth retardation. Thus, a possible expla-nation may be that most of the SGA infants arenot recognized until after delivery.

Postpartum hemorrhage (.500 ml) has been es-timated to occur in 0.4–10% of the deliveries, withthe majority of estimates being between 4% and5% (22). The rates recorded at the midwife man-aged unit (6.2%) and at the central hospital (2.6%)are within the expected range. Various ways of as-sessing blood loss may explain some of the re-corded differences.

As maternal and neonatal death rates are verylow these traditional outcome measures are im-practical to use in a study like ours. For this rea-son we estimated desired outcome and three in-termediate outcomes. The main purpose of selec-tion was to predict the most appropriate place ofdelivery in each pregnancy. As the midwife man-aged unit was supposed to provide care for nor-mal pregnancies and avoid obstetrical- and neo-natal complications, the effectiveness of the selec-tion process might be assessed by the degree towhich this was achieved. No serious adverseevent occurred at the midwife managed unit,and, as 94% of the births had a desired outcomeand fulfilled the criteria for first level care (23),we consider the prediction of birthplace to benear optimal.

Differences in intermediate outcomes (operativedelivery, transfer to NICU and infants diverging


from normal) among the birth institutions arelargely caused by the selection process. To mini-mize selection bias when comparing intermediateoutcomes at the two institutions, we did an inten-tion to treat analysis (Table VI). For women whointended to give birth at the midwife managed unitthe likelihood of having an operative delivery wasreduced approximately nine times and the prob-ability of having a baby transferred to the NICUor an infant diverging from normal decreased fiveand seven times. As the midwife managed unit haslimited resources and the central hospial providesexpert obstetrical- and neonatal care, we believewith confidence that the observed differences inoutcome measures may indicate health benefits formother and infant.

The way the data was collected by the caregiversmay create a potential for bias. However, the datagathered at each center was reviewed several times,information from individual records were con-trolled and the results thus given, might be con-sidered valid.

Pregnant women living in Lofoten were allowedto use aircraft if needed to travel to the centralhospital. As 92% of the women delivering at thecentral hosptial were transferred by air, it has beenfeasible for most of the women to stay home untillabor started. However, being transferred by airmight be an unpleasant experience. Further studiesare required to address maternal satisfaction andevery effort should be made to minimize theemotional impact of transfer in labor.

Conclusions

The reorganization of the delivery unit at Lofotenhospital to a modified midwife managed unit wasassociated with a reduction in operative deliveriesfor the study population. The modified unit playeda complementary part in the community basedantenatal care. The vast majority of preterm andsick infants were born at the central hospital. Asnearly 70% of the births occurred at the midwifemanaged unit and 94% were spontaneous vaginaldeliveries of 35–42 weeks gestation giving an infantnot needing resuscitation, this indicates an effec-tive selection process. This model might be analternative to centralization of births in sparselypopulated areas.

Acknowledgments

The authors would like to thank midwives and doctors at Lo-foten Hospital and Nordland Central Hospital for assistancewith the trial. They also thank Nordland County Council forfinancial support.

212 J. Holt et al.

References

1. Rooks JP, Weatherby NL, Ernst EKM, Stapleton S, RosenD, Rosenfield A. Outcomes of care in birth centres. N EnglJ Med 1989; 321: 1804–11.

2. Hundley VA, Cruickshank FM, Lang GD et al. Midwifemanaged delivery unit: a randomised controlled compari-son with consultant led care. BMJ 1994; 309: 1400–4.

3. Turnbull D, McGinley M, Fyvie H et al. Implementationand evaluation of a midwife development unit. Br J Mid-wifery 1995; 3: 465–8.

4. Turnbull D, Holmes A, Shields N et al. Randomised, con-trolled trial of efficacy of midwife-managed care. Lancet1996; 348: 213–18.

5. Wall EM. Assessing obstetric risk. A review of obstetricrisk-scoring systems. J Fam Pract 1998; 27: 153–63.

6. Reynolds JL, Yudkin PL, Bull MJV. General practitionerobstetrics: does risk prediction work? J R Coll Gen Pract1988; 38: 307–10.

7. Informed choice for professionals 10. Place of birth. Mid-wives information and resource service (MIDIRS). Clifton,Bristol UK 1997: 8.

8. Holt J, Fagerli I. Air transport of the sick newborn infant:audit from a sparsely populated county in Norway. ActaPædiatr 1999; 88: 66–71.

9. Marsal K, Persson P-H, Larsen T, Lilja H, Selbing A, Sul-tan B. Intrauterine growth curves based on ultrasonicallyestimated foetal weights. Acta Pædiatr 1996; 85: 843–8.

10. Roland M, Torgerson DJ. Understanding controlled trials:What are pragmatic trials? BMJ 1998; 316: 285.

11. Dunn FH, McIlwaine GM. Perinatal audit. A report pro-duced for the European Association of Perinatal Medicine.Prenat Neonat Med 1996; 1: 160–94.

12. Muir Gray JA. Evidence-based healthcare. New York, Ed-inburgh, London, Madrid, Melbourne, San Fancisco,Tokyo,: Churchill Livingstone, 1997.

13. Rosenblatt RA, Reinken J, Shoemack P. Is obstetrics safein small hospitals? Lancet 1985; 2: 429–32.

14. MacVicar J, Dobbie G, Owen-Johnstone L, Jagger C, Hop-kins M, Kennedy J. Simulated home delivery in hospital: arandomised controlled trial. Br J Obstet Gynaecol 1993;100: 316–23.


15. Baird AG, Jewell D, Walker JJ. Management of labor in anisolated rural maternity hospital. BMJ 1996; 312: 223–6.

16. Butler J, Abrams B, Parker J, Roberts JM, Laros RK. Sup-portive nurse-midwife care is associated with a reduced in-cidence of cesarean section. Am J Obstet Gynecol 1993;168: 1407–13.

17. Kennell J, Klaus M, McGrath S, Robertson S, Hindley C.Continuous emotional support during labor in a US hospi-tal. A randomized controlled trial JAMA 1991; 265: 2197–201.

18. Parbosingh J, Kerr M. Innovations in the role of obstetrichospital in prenatal care. In: Enkin M, Chalmers I (eds):Effectiveness and Satisfaction in Antenatal care. Clinics inDevelopmental Medicine Nos. 81/82. Spastics InternationalMedical Publications. London: William Heinemann Medi-cal Books Ltd, Philadelphia; J. B. Lippincott Co., 1982:254–65.

19. Read AW, Prendiville WJ, Dawes VP, Stanley FJ. Cesareansection and operative vaginal delivery in low-risk primipar-ous women, Western Australia. Am J Public Health 1994;84: 37–42.

20. Langhoff-Roos J, Lindmark G. Obstetric interventions andperinatal asphyxia in growth retarded term infants. ActaObstet Gynecol Scand 1997; 76 (Suppl 165): 39–43.

21. Backe B, Nakling J. Effectiveness of antenatal care: a popu-lation based study. Br J Obstet Gynaecol 1993; 100: 727–32.

22. Varner M. Postpartum hemorrhage. Crit Care Clin 1991;7: 883–7.

23. Agostino R, Antsaklis A, Breart G, Carrapato MRG,Chabernaud JL, Debauche C et al. Recommendations fromthe European Association of Perinatal Medicine 1. Newdecision rules: regionalization in perinatal care and indi-cations for perinatal transfer. Prenat Neonat Med 1999; 4(Suppl 1): 104–7.

Address for correspondence:

Jan Holt, M.D.Department of PediatricsNordland Central Hospital8092 BodoeNorway

child births in a modified midwife managed unit: selection and transfer according to intended place...

Documents