Dr S. Anthony, TNO Prevention and Health, Department of Reproduction and Perinatology, P.O. Box 2215, 2301 CE Leiden, The Netherlands.
Objectives In the Netherlands, approximately one-third of births are planned home births, mostly supervised by a midwife. The relationship between maternal demographic factors and home births supervised by midwives was examined.
Design Cross-sectional study.
Setting Dutch national perinatal registries of the year 2000.
Population All women starting their pregnancy care under the supervision of a midwife, because these women have the possibility of having a planned home birth.
Methods The possible groups of birth were as follows: planned home birth or short stay hospital birth, both under the supervision of a midwife, or hospital birth under the supervision of an obstetrician after referral from the midwife during pregnancy or birth. The studied demographic factors were maternal age, parity, ethnicity and degree of urbanisation. Probabilities of having a planned home birth were calculated for women with different demographic profiles.
Main outcome measure Place of birth.
Results In all age groups, the planned home birth percentage in primiparous women was lower than in multiparous women (23.5%vs 42.8%). A low home birth percentage was observed in women younger than 25 years. Dutch and non-Dutch women showed almost similar percentages of obstetrician-supervised hospital births but large differences in percentage of planned home births (36.5%vs 17.3%). Fewer home births were observed in large cities (30.5%) compared with small cities (35.7%) and rural areas (35.8%).
Conclusions This study demonstrates a clear relationship between maternal demographic factors and the place of birth and type of caregiver and therefore the probability of a planned home birth.
In contrast to most other Western countries, planned home birth is relatively more common in the Netherlands. In the year 2000, 30.3% (n= 62,000) of all Dutch births took place at home.1 The majority of planned home births take place under the care of midwives. In rural regions where no midwives practise, these births take place under the care of a general practitioner (GP). Since 1965, the percentage of home births, then 69%,2 decreased, mostly due to the introduction of the short stay hospital birth under the care of a midwife. Over the years 1995–2000, the home birth rate stabilised again around one-third of all Dutch births.1
In the Dutch maternity care system, midwives provide independent care for women with uncomplicated pregnancies. Women with low risk pregnancies can choose to either have a home birth or a hospital birth under the responsibility and care of their midwife. In the latter situation, the women and their babies are generally discharged within a few hours after birth for postpartum home care. Women with high risk pregnancies a priori receive care from an obstetrician in the hospital. As soon as the midwife suspects or diagnoses a complication during pregnancy or birth, a woman is referred to an obstetrician. As a result of this risk selection procedure, three combinations of birth place and caregiver are possible: midwife-supervised planned home birth, midwife-supervised short stay hospital birth and obstetrician-supervised hospital birth (either due to an a priori high risk pregnancy or after referral from the midwife).
So far, most studies into home births focus on outcomes such as mortality and morbidity3–9 or on attitudes towards home births.10–12 Little is known about demographic factors of women such as age, parity, ethnicity and degree of urbanisation and a woman's probability of having a planned home birth. In the present study, the relationship between these maternal demographic factors and home births was examined. The probability of having a home birth was calculated for each demographic profile of the women. Home births may be promoted by using this knowledge of the effect of different maternal factors on the home birth rate.
This study was part of a programme called ‘PROVER’, funded by the National Health Care Insurance Council (CvZ) and was carried out by the Royal Dutch Midwifery Organisation (KNOV). One of the aims of the programme was to monitor the prevalence of home births in the Netherlands during the period 1995–2000.13 Routinely collected data on Dutch home births are not available since 1993, the last year in which Statistics Netherlands (CBS) published these statistics. To monitor the prevalence of home births, a new system had to be developed. Within this new monitoring system, it is also possible to study the relationship between maternal demographic factors and home births.
The monitoring system is based on two professional registers: the National Perinatal Database for Primary Care, a register of midwife-assisted births, and the National Perinatal Database for Secondary Care, a register of obstetrician-assisted births. In these databases, midwives and obstetricians register in anonymous records, detailed information on pregnancy, delivery and puerperium and brief maternal demographic factors. In case of referrals of complicated pregnancies from midwives to obstetricians, both the midwife and the obstetrician separately register information on the same pregnancy. Since 1995, these databases are annually linked and aggregated using a matching procedure based on maternal and child variables to form one Dutch birth cohort per registration year.14 After linkage, an extrapolation is performed to correct for a small number of non-participants in both registers. Because the percentage of non-participants differs between midwives and obstetricians, this extrapolation is based on assigning different weighing factors to the midwife and obstetrician records, depending on their participation rate in that specific year. In the present study, we analysed the linked database for 2000. In that year, 92% of all midwives, all obstetricians in academic and training hospitals and 96% of obstetricians in non-training hospitals participated in the registers.
The maternal demographic factors that are registered in the linked perinatal database are age, parity and ethnicity of the mother and the degree of urbanisation of the maternal place of residence. In the present study, these variables were defined as follows: parity was categorised as primiparous (no prior birth) and multiparous (one or more prior birth). The age of the mother was calculated using the birth date of mother and child and categorised as under 25 years, 25–29 years, 30–34 years and above 34 years. In the National Perinatal Database, ethnicity is categorised as ‘Dutch’, ‘Mediterranean’, ‘other European’, ‘Black’, ‘Hindu’, ‘Asian’ and ‘Other’. For this study, the ethnicity of the mother was divided into the dichotomous variable ‘Dutch’versus‘non-Dutch’. The degree of urbanisation was defined using the definition of Statistics Netherlands.15 This definition uses five categories based on the number of households per km2. For this study, we reduced these categories to three categories: large city (at least 1500 households/km2), small city (1000–1499 households/km2) and rural area (less than 1000 households/km2).
GPs supervise around 5–6% of all births in the Netherlands.13 Due to the fact that GPs do not yet register in the perinatal database, these births were not included in this study. This study focuses on all births under the care of a midwife or an obstetrician. First, an overview is given of births in the different categories: midwife-supervised home births, midwife-supervised short stay hospital births, obstetrician-supervised hospital births after referral from the midwife during pregnancy or birth and obstetrician-supervised hospital births of women with an a priori high risk pregnancy and therefore starting pregnancy care directly with the obstetrician. For all further analysis, the latter group of women starting their pregnancy care directly in hospital with an obstetrician due to an a priori high risk pregnancy was excluded because these women did not have any possibility of having a planned home birth. Next, the relationship between the maternal demographic factors and the combination of place of birth and caregiver, with a special focus to planned home births, was explored in univariate analyses. Finally, within the same subgroup of women the probability of a planned home birth was determined for the different combinations of demographic factors. By this stratification into the different demographic profiles possible confounding between the different demographic factors was taken into account.
All statistical analyses were performed in SPSS, version 11.
The number of births under the care of midwives and obstetricians was 191,471 in the year 2000.Table 1 shows the distribution of these births by place of birth and type of caregiver. More than a quarter of these births (n= 50,314) were home births under the supervision of a midwife and 10% (n= 19,164) were hospital births supervised by a midwife. Another 47.5% (n= 90,851) of the births took place under the supervision of the obstetrician after referral from the midwife.
Table 1. Distribution of births under the care of midwives and obstetricians by place of birth and type of caregiver in the year 2000.
Of 80 midwife-supervised births, the place of birth (home or hospital) was unknown. These births were therefore not included in this table and further analyses.
Midwife-supervised home birth
Midwife-supervised short stay hospital birth
Obstetrician-supervised hospital birth—after referral from midwife
Obstetrician-supervised hospital birth—a priori high risk pregnancy
Total births under care of midwives or obstetricians*
The distribution of the maternal demographic factors for the subgroup of women starting their pregnancy care with the midwife and therefore initially having the option of a planned home birth is shown in Table 2. Almost 50% of the women starting their pregnancy care with a midwife are primiparous. The age group of 25–29 years includes the largest group of primiparous women (39.6%), whereas the age group of 30–34 years is the largest age group for the multiparous women (46.7%). Almost 85% of the women starting pregnancy care with a midwife are of Dutch origin. Most of the women starting pregnancy care with a midwife live in a large city (44.1%) followed by the group of women living in rural areas (34.6%).
Table 2. Distribution of maternal demographic factors for the subgroup of women starting their pregnancy care with the midwife in the year 2000.
The total number of women per factor differs from the overall number of women starting pregnancy care with midwife due to a varying number of missing values depending on the factor under study.
Figure 1 shows the distribution of place of birth and type of caregiver for primiparous and multiparous women by age distribution. In all age groups, the home birth rate for primiparous women is lower than that for multiparous women. A similar trend is observed in all age groups for the proportion of short stay hospital births with a midwife. Overall, 23.5% of the primiparous women delivered at home, 10.3% delivered in hospital under the supervision of the midwife and 66.1% delivered in hospital with an obstetrician after referral. For the multiparous women, these figures are 42.8%, 15.0% and 42.1%, respectively.
A low home birth rate is observed in the youngest age group both in the primiparous and multiparous women. This is not caused by more referrals for obstetrician-supervised hospital births but by the high percentage of midwife-supervised short stay hospital births in this age group. In primiparous women, the 25–29 year age group shows the highest home birth rate (25.9%); whereas in multiparous women, the highest home birth rate is seen in the 30–34 year age group (45.7%).
A small difference exists between Dutch and non-Dutch women in proportion of obstetrician-supervised hospital births after referral (54.2% and 50.4%, respectively). A large difference exists in percentage of home and hospital births with the midwife. In the Dutch group, 36.5% are home births and 9.3% short stay hospital births; whereas in the non-Dutch group, 17.3% are home births and 32.3% short stay hospital births (Fig. 2).
In rural areas, referral from the midwife to the obstetrician occurs slightly more often than in large and small cities (56.0% instead of 52.8% and 52.3%). Fewer home births are observed in large cities (30.5%) compared with small cities (35.7%) and rural areas (35.8%). In large cities, more short stay hospital births, supervised by a midwife, take place.
Within the subgroup of women starting their pregnancy care with a midwife, the probability of having a planned home birth was determined for women with different demographic profiles (Table 3). Overall, multiparous women and Dutch women are more likely to have a home birth than primiparous women and non-Dutch women. The probability that a Dutch primiparous woman within the age group of 25–29 years living in a rural area has a planned home birth is 31%, whereas the probability that a multiparous woman with the same ethnicity, age and urbanisation characteristics has a home birth is 55%. A Dutch multiparous woman aged between 30 and 34 years and living in a large city is at least two times more likely to have a home birth than a woman with the same characteristics who is of non-Dutch origin (52%vs 23%).
Table 3. Probability (in %) of having a home birth for women with different profiles of demographic factors, calculated within the subgroup of women starting their pregnancy care with a midwife in the year 2000.
The Dutch maternity care system is based on a risk selection procedure under the responsibility of midwives, who make a distinction between women at elevated risks of obstetric complications and requiring referral to the obstetrician, and women with a low risk and remaining under their care. Whether a woman with a low obstetric risk chooses a home birth or a short stay hospital birth, both under the supervision of a midwife, is a matter of personal preference.
The aim of this study was to examine the relationship between place of birth and type of caregiver and demographic characteristics of women starting pregnancy care with a midwife, with a special focus on home births.
Multiparous women under the care of a midwife deliver at home almost twice as frequently as primiparous women (42.8%vs 23.5%). This difference is largely explained by the a priori risk selection of multiparous women with obstetric problems in previous pregnancies. Multiparous women with obstetric problems in a previous pregnancy will be brought directly under the care of the obstetrician in a next pregnancy. Therefore, multiparous women starting pregnancy care with a midwife are selected women with low risks of complications during pregnancy or birth. A similar selection obviously cannot be made in the group of primiparous women. In this study, referral from the midwife to the obstetrician was observed in 66.1% of the primiparous women and in 42.1% of the multiparous women. The observed difference in home birth percentage is largely explained by this difference.
Low home births percentages are seen in the youngest age group under 25 years. This is not explained by a higher referral rate to the obstetrician but by more short stay hospital births under the supervision of the midwife. It is not clear why younger women more often choose a short stay hospital birth. Their living conditions might be less conducive to a home birth, for example, if they are living in an apartment building without an elevator. Social economic characteristics such as level of education may also explain this difference. Perhaps additional patient information and better instructions by the midwife might increase the proportions of home births in this group of young pregnant women.
Dutch and non-Dutch women starting their pregnancy care with the midwife show similar percentages of referrals to the obstetrician. Dutch and non-Dutch women giving birth with a midwife, however, make different choices concerning their place of birth. Twice as many Dutch than non-Dutch women choose a home birth (36.5%vs 17.3%). This difference in choice may possibly be explained by the existing attitude towards birth in most other countries than the Netherlands, where childbirth is more medicalised and hospital centred. Additional patient information and better instruction by the midwife may possibly increase the number of non-Dutch women choosing a home birth.
Referral from the midwife to the obstetrician occurs slightly more often in rural areas than in cities. Midwives may anticipate earlier to problems in rural regions where the distance to the nearest hospital is larger than in cities. Women in large cities more often choose a short stay, midwife-supervised hospital birth than women in small cities and rural areas. This may be related to the fact that the more urbanised an area is, the more hospital facilities are available within a short distance.
One of the limitations of this study is that births under the supervision of the GP, representing around 5% of all Dutch births, could not be included as these births are not yet registered in the National Perinatal Databases. These births take place especially in rural areas where no midwife is practising. The results from this study are not directly applicable to births under the care of GPs because the distribution of the maternal demographic factors, such as degree of urbanisation, of women giving birth with a GP will be different from both births with a midwife and births with an obstetrician.
This study is based on routinely collected data. Misclassification within the used categories of demographic factors is possible. For example, no clear instructions for the registration of ethnicity exist. One caregiver may classify a second generation immigrant as ‘Dutch’, whereas another caregiver may register the same woman as ‘non-Dutch’. It is, however, assumed that this misclassification is random and not dependent on the place of birth or type of caregiver. Therefore, it cannot explain the observed relationships between demographic factors and place of birth and type of caregiver. It remains possible that these relationships are weakened due to the potential misclassification.
When comparing the results of this study with international studies, one should bear in mind that maternity care systems are totally different in different countries and therefore difficult to compare. Australian data confirm the observed higher referral rate for primiparous women.16 A study in the United States into home births concluded that mothers who gave birth at home were more likely to be of higher parity.17 In the same study, a maternal age of 30 years and over was associated with more home births. This was also confirmed in another Dutch study where women choosing a home birth were older on average than women choosing a short stay hospital birth.18 As in our study, Statistics Netherlands showed in the year 1990 that within the group of primiparous women, the age group of 25–29 was most likely to have a home birth; whereas for all women, the age group of 30–34 had the highest home birth rate.2 In that year also, more than twice as many Dutch women chose a home birth than non-Dutch women. This relationship was also confirmed in a prospective Dutch cohort of 1836 women, showing more short stay hospital births than home births for women belonging to an ethnic minority.19 In the United States also, it has been described that white women more often choose a home birth.17 Differences in home birth rates depending on the degree of urbanisation have already been described for the Dutch situation. Statistics Netherlands showed that the home birth rate decreased the more inhabitants lived in a municipality.2 Hingstman and Boon20 concluded that ‘the supply of hospital beds and population density in a region (which are intercorrelated to a certain extent) have a negative effect on the proportion of home confinements’. In another Dutch study examining the determinants of the choice for home or hospital birth, urbanisation was not found to be a predictor of choice.12
This study demonstrates clear relationships between maternal demographic factors and the place of birth and type of caregiver and the probability of a home birth. The place of birth is partly determined by the risk of referral from the midwife to the obstetrician during pregnancy or birth. This risk is, for example, higher for primiparous than for multiparous women. Apart from this medical risk, women giving birth with a midwife can choose between a home birth and a short stay hospital birth. This choice is dependent on the preference of the woman. Non-Dutch women, for example, prefer a short stay hospital birth above a home birth. If home births are to be promoted, special attention should be focussed on non-Dutch women, a growing number in the Netherlands, young pregnant women and women in large cities. The choice of place of birth is often made early in pregnancy or even before pregnancy21 and is often influenced by ‘significant others’ such as family and friends.12 Apart from information via the midwife, alternative ways of informing young and non-Dutch women and their ‘significant others’ about the Dutch maternity system with its possible places of birth should be explored.
The authors would like to thank the Royal Dutch Midwifery Organisation (KNOV), the Dutch Association for Obstetrics and Gynaecology (NVOG) and the Paediatric Association of the Netherlands (NVK) and all their members. Without their help and their continuing data collection, this work would not have been possible. The authors thank the PROVER committee for their scientific contribution.
The authors also thank the Netherlands Perinatal Registry (Stichting Perinatale Registratie Nederland) and Prismant who gave permission to analyse the perinatal registries for this study and who made the registries available for use.
This study was subsidised by the National Health Care Insurance Council (CvZ).