Trends in the medicalisation of childbirth in Flanders and the Netherlands
Report of a Short Term Scientific Mission (STSM) for COST action ISO907
Wendy Christiaens1, Raymond Devries2,3 and Marianne Nieuwenhuijze2
1 Department of Sociology, Ghent University
2 Center for Bioethics and Social Sciences in Medicine, University of Michigan
3 Department Midwifery Science, Academie Verloskunde Maastricht, Zuyd University
Summary
Although they are neighbouring countries sharing the same language, political system and geography, Belgium and the Netherlands are characterised by a different organisation of health care, particularly in maternity care. In Belgium the medical risks of childbirth are emphasised but neutralised by a strong belief in the merits of the medical model. In line with the midwifery model of care, in the Netherlands childbirth is defined as a normal physiological process and family event. These different approaches to childbirth are reflected in the organisation and utilisation of maternity care facilities. Not surprisingly, Belgium has been characterised by higher obstetric intervention rates compared to the Netherlands. However, over time the obstetric intervention rates of both countries have been converging and even reversing.
In this paper we give an overview of the trends in obstetric intervention rates as indicators of medicalisation of childbirth in both Flanders and the Netherlands. In addition contextual factors, such as health care finance, the medical-legal system and professional guidelines, are taken into account.
Background
Although they are neighbouring countries sharing the same language, political system and geography, Flanders and the Netherlands are characterised by a different organisation of health care, particularly in maternity care. In Flanders the medical risks of childbirth are emphasised but neutralised by a strong belief in the merits of the medical model. In line with the midwifery model of care, in the Netherlands childbirth is defined as a normal physiological process and family event [46]. These different approaches to childbirth are reflected in the organisation and utilisation of maternity care facilities. In the Netherlands, for example, home births are encouraged by directing women expecting a normal birth into primary care (DeVries, 2001), resulting in a 21.5% home birth rate (Stichting Perinatale Registratie Nederland (SPRN), 2007). The option of a policlinical birth, or a ‘home birth away from home’, (11.3%) provides women with the possibility of having a midwife-led hospital birth with a short stay after the baby is born. In case of difficulties during pregnancy and labour, women are referred to specialist care. The relatively high proportion of home births and the emphasis on normality result in low rates of obstetric interventions. In contrast, in Flanders 97.9% of childbearing women prefer to have their babies in hospital, finding reassurance in the proximity of obstetric technology (Studiecentrum voor Perinatale Epidemiologie (SPE), 2007). Not surprisingly, Flanders has been characterised by higher obstetric intervention rates compared to the Netherlands. However, over time some obstetric intervention rates of both countries have been converging and even reversing (see below).
Although the Netherlands stand out for their unique maternity care system characterized by a relatively high proportion of home births, the emphasis on normality of childbirth and low intervention rates, there are some tendencies towards increasing medicalisation. For example, a trend analysis in referrals during pregnancy and labour in Dutch midwifery care from 1988 to 2004 showed an increase of 14.5%. The main reasons for referral were medical or obstetrical history, particularly that of caesarean section, and the growing demand for pain relief (Amelink-verburg et al., 2009). Also the 2007 year report of the “Stichting Perinatale Registratie Nederland” (SPRN), signals a yearly increase in the use of epidural analgesia during labour with one to two percent. Other trends are an increasing number of inductions (28.8% in 2003 versus 33% in 2007), a decreasing number of women starting prenatal care in primary care (83.1% in 2003, versus 77.7% in 2007) and a decreasing number of births in primary care (36.2% in 2000 versus 32.9% in 2007) (SPRN, 2003; SPRN, 2005; SPRN 2007). In addition, home births are under increasing pressure of a scientific debate closely followed by popular media.
The medicalisation of the birth process is not a new phenomenon, but until recently the Netherlands were known as the country which seemed less influenced by that trend. Now, the global medicalisation trend seems to affect both the attitude of Dutch women in their demands (Pavlova et al., 2009; Van der Hulst et al., 2007), and of Dutch caregivers (e.g. midwives in their assessment of normality, obstetricians in their readiness to intervene) (Amelink-verburg et al., 2009).
In this paper we look at childbirth from a medicalisation perspective, because that theoretical frame allows us to distinguish between several motivations to seek medical support. Medical care is not merely the response to health problems or pathology, as is suggested by the definition of Conrad (1992): “a process by which nonmedical problems become defined and treated as medical problems, usually in terms of illnesses or disorder” (Conrad, 1992, p. 209). Other motivations to health care use lay on top of the primary aim to cure, for example the search for quick and easy solutions for inconveniences both on the part of the patient and the physician (e.g. induction of labour to fit birth into personal schedules or in function of the hospital organisation), or the optimalisation or enhancement of normal characteristics (e.g. optimalisation of concentration, energy, or fitness, suppressing bodily cycles such as menstruation etc.).
Pregnancy and childbirth are textbook cases of medicalisation, because they are, as such, healthy and normal physiological phenomena, which are defined as hazardous and a potential pathology under influence of the biomedical ideology (Oakley & Houd, 1990). Consequently childbirth can only be called “normal” in retrospect (Lumley, 1993). Several authors (Oakley, 1983; Oakley, 1992; Martin, 2001; Johanson et al., 2002) from diverse academic disciplines have criticised the extent of, and the manner in which, medical professionals control the birth process. Research into birth-experiences suggests that the use of medical technology may result in alienation through the erosion of women’s control over the birth process (Davis-Floyd, 1994; Martin, 2001). The medicalisation critique stands in contrast to the observation that many women feel comfortable about medical control and are satisfied with medical interventions and hospital deliveries (Lazarus, 1994; Sargent & Stark, 1989). Many women desire and actively seek medical control over the unpredictable process of birth. In the childbirth literature, a growing number of studies emphasises the active role of mothers and fathers in the doctor-patient interaction and in decision-making regarding pregnancy and birth (VandeVusse, 1999; Van der Hulst et al., 2007; Shorten et al., 2005). Zadoroznyj (2001) found considerable evidence to support the notion of childbearing women as reflexive consumers. She argued that the medical encounter is a highly complex form of interaction, with women taking up the roles of both consumers and ‘patients’.
To understand why the experiences of childbearing women diverge from traumatic in one extreme to empowering in the other extreme, we need to identify the determinants of the meanings attached to medical solutions. By contextualizing childbirth we can gain insight in the diversity of meanings driving the medicalisation of childbirth. Applied to the Flemish and Dutch case, this means that the comparison of trends over time, embedded in their social and political context, should allow us to map the medicalisation of childbirth and gain insight in the diversity of driving and counteracting forces. The aim of this paper is to compare the medicalisation of childbirth in Flanders and the Netherlands and search for determinants.
Methods
The perinatal statistics in this paper are collected by the National Perinatal Database of the Netherlands (PRN, Perinatale Registratie Nederland) and Flanders (SPE, Studiecentrum voor Perinatale Epidemiologie). What concerns the Netherlands the registration of data takes place in the practices of the participating midwives (LVR1) and obstetricians (LVR2). A form or an automated system is used for registration. The data to be recorded are to a large extent similar for primary and secondary care. The remaining part contains data that are more specific for the performance of either midwives or obstetricians. For Flanders the registration takes place in Flemish maternity units and independent midwifery practices. In 2009, all 67 maternity units in Flanders participated, and almost all Flemish home births (N=665) were registered by participating independent midwifes.
In addition, contextual information regarding health care finance, the medical-legal system and professional guidelines, are taken into account in order to contextualise the medicalisation trends in Flanders and the Netherlands. More specifically in the second part of this report, the national context is described in function of the argument that childbirth can be considered as a social problem in the Netherlands, while it is not in Flanders.
Results
1. Trends in obstetrical interventions (1992-2009)
We started the comparison of official perinatal statistics for Flanders and the Netherlands, but encountered a lot of incompatibilities which hamper cross-national comparison. In the following we will discuss these difficulties and address two obstetric interventions (induction and epidural analgesia), for which comparison was methodologically legitimate.
The first cause of lack of comparability originates in the clear distinction between primary and secondary care in the Netherlands, which is reflected in the national perinatal registration system. As specified in the method section, the Dutch registration of data takes place in the practices of the participating midwives (LVR1) and obstetricians (LVR2). Rates of obstetric interventions specific for secondary care (e.g. Caesarean section) are calculated on the total number of hospital births, not on the total number of hospital and home births. In Belgium however, percentages are calculated on the total including primary and secondary care. Since the number of home births is low in Flanders, only around 1%, the cross-national comparison is mainly based on hospital births, which seems not that problematic. However, if Dutch home births would be taken into account in the denominator, the Dutch obstetric intervention rates would be lower and hence provide a more optimistic image.
There is a second problem with the denominators in both countries: in the PRN data the denominator refers to all children born dead or alive after a pregnancy of at least 22 weeks. In the SPE data, however, they are calculated on the total number of deliveries. A delivery is not specified in terms of length of pregnancy, but as the birth of one or more children with a birth weight of minimum 500 g. Since one delivery can result in more than one child, the Dutch denominator is more inclusive. In 2007 PRN reports 3182 (1.9%) multiples in the Netherlands, in 2009 SPE counted 2466 (3.6%) multiples in Flanders.
1.1 Induction rates
Induction rates were quite stable during the nineties: in the Netherlands between 27 and 28%, in Flanders around 30%. From 2003 a divergent pattern develops with an increase for the Netherlands and a decrease for Flanders. In 2007 the induction rates of both countries lay 7.4% apart, with 33% for the Netherlands and 25.6% for Flanders. In 2004, Prof. dr. M. Temmerman, an influential Flemish gynecologist, launched a debate about elective induction in the Flemish media. She argued for the introduction of an informed consent to be signed by mothers who want an induction for other than medical reasons, a strategy which reduced the induction rate in the University hospital of Ghent, from 33 to 26%. An informed consent was integrated in the guidelines for good clinical practice in low risk deliveries, published in 2010. In addition the guidelines state that an elective induction is not recommended between 39 and 41 weeks of pregnancy. From 37 weeks of pregnancy an induction is recommended in case of ruptured membranes and lack of spontaneous labour during the subsequent 24 hours. Also after 41 weeks of pregnancy it is acceptable to induce labour (Mambourg, Gailly and Wei-Hong, 2010).
It is unclear how the increasing induction rate in the Netherlands can be explained. Could it be the impact of the publication in 2002 of “Preventive support of Labour”, a manual for midwives and gynaecologists, suggesting more active support during labour (Reuwer & Bruinse, 2002)?
1.2 Epidural rates
For the Netherlands epidural rates are available for 2003 until 2007 only. It is clear from the Dutch-Flemish comparison that the country statistics are 57 percentage points apart in 2007, with 66.6% in Flanders and 9.6% in the Netherlands. It seems that both countries have a completely different strategy to manage labour pain, but this divergence is not reflected in the clinical guidelines of both countries.
In Flanders the clinical guidelines (Mambourg, Gailly and Wei-Hong, 2010) recommend that the choices of the pregnant woman regarding pain relief, should be respected. In addition, it is recommended to inform women during labour about the consequences of loco-regional analgesia (e.g. more frequent monitoring, limited movement). Finally, it is recommended to start loco-regional analgesia when labour started and the woman requests it, regardless of the stage of cervical dilatation.
In the Netherlands a guideline regarding pharmaceutical pain relief during delivery was formulated in 2008 (NVA and NVOG, 2008). Overall the Dutch guidelines converge with the Flemish, but goes more into detail. Every women in labour should on her request have access to an adequate form of pain relief. Also she should be informed about consequences of analgesia, but the list of consequences is more inclusive: a longer hospital stay, a higher risk of oxytocin use, an instrumental delivery, hypotension, locomotive block and urine retention. It is recommended to start epidural analgesia regardless of the stage of cervical dilatation.
Table 1: Percentage of epidural analgesia in Flanders and the Netherlands (1991-2009)
% of epidural analgesia(sectio's included) / % of epidural analgesia
in vaginal births 7
Year / Flanders 1 / the Netherlands / Flanders 2 / the Netherlands 3,4
Total / LVR1 / LVR2 / Primiparae / Multiparae / Total / Primiparae / Multiparae / Total
1991 / 32.00 / 30.8
1992 / 34.80 / 33.7
1993 / 39.00 / 39.0
1994 / 43.70 / 42.4
1995 / 48.20 / 46.8
1996 / 51.60 / 50.1
1997 / 55.00 / 62.0 / 41.4 / 53.3
1998 / 58.10 / 65.4 / 44.1 / 56.5
1999 / 61.30 / 68.3 / 47.4 / 59.7
2000 / 61.70 / 68.3 / 47.9 / 60.0
2001 / 62.70 / 68.7 / 49.3 / 58.0
2002 / 63.20 / 69.2 / 49.7 / 58.5
2003 / 64.40 / 5.4 / 69.7 / 50.7 / 59.3 / 3.2 / 1.2 / 4.4
2004 / 61.60 / 6.2 / 66.9 / 48.4 / 57.1 / 3.7 / 1.4 / 5.1
2005 / 64.70 / 7.0 / 68.5 / 50.3 / 58.8 / 4.2 / 1.6 / 5.8
2006 / 66.50 / 8.2 / 70.7 / 51.6 / 5.0 / 1.9 / 6.9
2007 / 66.60 / 1.8 5 / 9.6 4 / 70.9 / 51.6 / 6.0 / 2.2 / 8.2
2008 / 67.40 / 71.5 / 52.6
2009 / 66.50 / 71.0 / 51.7
1 Source: SPE, Perinatale Activiteiten in Vlaanderen (Year reports 1991 - 2009)
2 Source: SPE, Perinatale Activiteiten in Vlaanderen (Year reports 2000, 2005 - 2009)
3 Source: PRN, Perinatale zorg in Nederland (Year reports 2003-2007)
4 This figure is restricted to LVR2, hence the denominator counts only deliveries in secondary care.
5 This figure contains only spinal, epidural and anesthesia in primary care deliveries.
7 Instumental deliveries included.
Recently I published an article on the use of labour pain medication in Belgium and the Netherlands (Christiaens et al., 2010). In Belgium labour pain is perceived as a needless inconvenience easily resolved by means of pain medication, while in the Netherlands it is perceived as an ally in the birth process. We found that this country difference cannot be explained by labour pain acceptance, since Dutch and Belgian women giving birth in a hospital setting are characterised by a similar labour pain acceptance. However personal control in pain relief can partially explain the country differences in coping with labour pain. For Dutch women we find that the use of pain medication is lowest if women experience control over the reception of pain medication and have a positive attitude towards labour pain. In Belgium however, not personal control over the use of pain relief predicts the use of pain medication, but negative attitudes towards labour. Apart from individual level determinants, such as length of labour or pain acceptance, our findings suggest that the maternity care context is of major importance in the study of the management of labour pain. The pain medication use in Belgian hospital maternity care is high and is very sensitive to negative attitudes towards labour pain. In the Netherlands, on the contrary, pain medication use is already low. This can partially be explained by a low degree of personal control in pain relief, especially when co-occurring with positive pain attitudes.