The comparison of birth outcomes and birth experiences of low-risk women in different sized midwifery practices in the Netherlands

Introduction 

Within the Dutch vision of midwifery care, pregnancy and birth are regarded as normal physiological life events which can take place without intervention.1, 2 The Dutch midwife is the main provider of midwifery care in the Netherlands and works independently. The midwife works at primary care level and carries a caseload that consists of low-risk women for whom she provides care throughout the antenatal, intrapartum and postnatal periods. An average caseload consists of 110 women per midwife per year.3 Low-risk women in the Netherlands enter the maternity system at primary care level and can choose their own midwifery practice in their local area They can also choose whether to have a home birth or to give birth in a local hospital (poly-clinic birth), with both options being facilitated by the independent midwife. In keeping with the responsibilities of the primary caregiver, the midwife screens for deviations in the physiological process of childbirth. Consultation or referral to secondary or tertiary level care occurs when complications arise or threaten to arise at any point during childbirth.4 When referral occurs, care is handed over to an obstetrician and homebirth is no longer an option. The midwife withdraws care and, as the referral takes place, financial remuneration for that client ends, regardless of whether the midwife remains involved with the woman. According to national guidelines5 women are referred to mainstream services in hospital settings. The organization of maternity care in the Netherlands is based on the division between physiology, which is the midwife's domain, and pathology, which is under the auspices of the obstetrician.

Changes in Dutch midwifery care 

Despite the vision and aim of Dutch midwives to achieve a physiological, non-interventionist birth, ideally at home, the intrapartum intervention rate in the Netherlands has been increasing since the late 1990s. Medical interventions such as augmentation and instrumental deliveries as a result of referrals because of dystocia in first and second stage of labour have shown a particular increase.1, 6, 7, 8, 9, 10, 11 Dutch national figures in relation to birth outcomes are shown in Table 1. While medical interventions have increased, Dutch midwifery has simultaneously been faced with a range of employment difficulties caused by a shortage of midwives and a resulting increase in workload.12, 13 Midwives favour part-time jobs in group practices of four or more midwives, which currently form 66.1% of all the midwifery practices in the Netherlands, the remainder being divided between solo practices (15%), and duo practices (18.9%).14 Solo and duo practices have thus become a minority in Dutch midwifery.12, 14, 15 Evidence suggests that the higher the number of practising midwives per practice, the higher the percentages of referrals during birth8, 16, 17 and the higher the number of interventions.18 It is known that a large practice size adversely affects the relationship between the woman and her midwife.9, 20 This is illustrated through anecdotal evidence suggesting a significant decrease in maternal satisfaction among Dutch women as a result of instrumental birth, loss of control during the birth process and the perceived lack of quality of the midwife–woman relationship.21, 22 Rijnders et al.23 reported in their study that in the Netherlands one in five women have a negative childbirth experience as a result of these issues.

Table 1. Registered birth outcomes 2007.

		N	%
	Total birth rate	181.000	100%
	Low-risk women	138.272	76%
	Start labour with primary care midwife	93.365	67.5%
	Referrals during birth	32.841	35.2%
	Birth centre/poly-clinic	5.392	5.8%
	Home birth	27.449	29.4%
	Instrumental births	19.233	20.6%
	Caesarean sections	10.923	11.7%
	Epidural anaesthesia	8.589	9.2%
	Augmented labour	33.798	36.2%

CBS41 and TietoEnator.11

Methods 

There are 466 midwifery practices in the Netherlands14 and a total of 273 practices were approached using cluster-stratified sampling. This sampling method was used to affect three proportionate strata (practices with 1–2 midwives, 3–4 midwives, and ≥5 midwives) which covered all levels of urbanization and were divided into urban (at least 1500 households/km2), semi-rural (1000–1499 households/km2) and rural (<1000 households/km2).14 Midwives or practice-assistants gave written information and a consent form to pregnant women who were predominantly in the third trimester of pregnancy. A total of 1900 women were invited to participate. It is unclear how many invitation letters were distributed, but 1020 women (53.6%) agreed to participate. Once consent was given confidentiality and anonymity were guaranteed; and midwives had no access to any women's responses. Coding was undertaken by an independent assistant, who wrote out the participants’ addresses. The researcher was not aware of either the exact geographical location of participating women or their midwifery practices. The researcher was informed if perinatal death took place and these women were then excluded from the study to avoid unnecessary affliction. Women who gave birth between 20 April and 20 May 2007 were included in the study. No reminders were sent to non-responders due to time and financial constraints. The study involved a cohort of women with uncomplicated pregnancies with a gestational age between 37 and 42 weeks of a single child presented by the vertex who had all received midwifery care in various sized practices in the primary care setting; at least until the onset of labour. The onset of labour was defined as regular and painful contractions or a spontaneous rupture of the membranes. The study was approved by a university's ethics committee.

Findings 

Of the 273 practices approached, 141 practices (52.4%) agreed to participate and recruited the participants, which resulted in three equal populations of women in relation to practice size, representing urban, semi-rural and rural settings (Table 2). A natural division of women over three equal groups in relation to practice size occurred spontaneously. The division of level of urbanization shows less equality, as there are slightly more women in the urban group, and there is also more variation within the urban grouping itself in regard to practice size. This variation in urbanization is representative for the Dutch population of childbearing women.14, 41

Table 2. Number and percentage of women represented by practice size (1–2, 3–4 and 5 midwives or more) and level of urbanization (urban, semi-rural and rural).

		1–2	3–4	5 or more	Total
	Urban	89/31.8%	90/32.1%	101/36.1%	280/39%
	Semi-rural	69/32.7%	69/32.7%	73/34.6%	211/29.4%
	Rural	79/34.8%	84/37%	64/28.2%	227/31.6%
	Total	234/32.6%	248/34.5%	236/32.9%	718/100%

Out of the 1020 women who had originally agreed to participate in the study, a total of 870 women responded (85.3%). Of these, 152 (14.9%) could not enter the study, leaving 718 (70.4%) participants who could be included in the study. Of the 152 exclusions, a number of 100 women did not give birth within the period of study and 52 were excluded because of post term gestation after the 42nd week of pregnancy (5), premature birth before the 37th week of pregnancy (2) twin pregnancy (1), referral to obstetrician after 37 weeks gestation but before labour (7), perinatal death (1) and induction of labour between 37 and 42 weeks for medical reasons (4). Thirty women were excluded as their questionnaires were received after the closing date.

The mean age of the participants was between 30 and 34 years of age (range younger than 20 to older than 40). For 36.2% of the participating women the highest educational qualification was on a Dutch medium level (MAVO/MBO/VMBO) and while for 35.9% it was on higher professional level (HBO). Almost all of the women were in a relationship. For 54.2% of the women this was their first birth experience. Table 3 shows the demographic details of the women from each separate group, which shows no significant variation between the groups. The participants were representative of the Dutch population of childbearing women with regard to age, marital status, education and parity41 and all women were equally distributed.

Table 3. Demographic details of women in small practices (1–2 midwives), medium-sized practices (3–4 midwives) and large practices (≥5 midwives).

		1–2 midwives	3–4 midwives	5 or more midwives	Total
	Age
	Younger than 20	1	1	1	3/0.4%
	20–24	12	15	16	43/6%
	25–29	72	70	70	212/29.6%
	30–34	105	108	109	322/44.8%
	35–40	43	46	41	130/18.1%
	Older than 40	3	3	2	8/1.1%
	Education
	LBO	5	6	4	15/2.1%
	MAVO/MBO/VMBO	86	84	90	260/36.2%
	HAVO/VWO	22	17	19	58/8%
	HBO	81	85	92	258/36%
	University	37	50	37	124/17.3%
	None	0	2	1	3/0.4%
	Marital status
	Married	168	178	158	504/70.2%
	Cohabiting	67	60	77	204/28.4%
	Living apart together	0	2	2	4/0.6%
	Single	1	2	3	6/0.8%
	Parity
	Primiparous	131	126	132	389/54.2%
	Multiparous	103	120	106	329/45.8%
	Number of children
	1	131	126	132	389/54.2%
	2	70	76	71	217/30.2%
	3	26	33	28	87/12.1%
	4	6	8	6	20/2.8%
	5 or more	1	3	1	5/0.7%

Age df=2, p=0.99; education df=2, p=0.99; marital status df=2, p=1.00; parity df=2, p=0.99.

Maternal birth outcomes 

Table 4 shows the results in relation to maternal birth outcomes: place of birth and interventions. Home births occurred more often in small practices and referral during the birth process was less frequent in practices with one or two midwives compared to practices with more than two midwives. Births in small practices were less likely to require administration of pethidine, CTG monitoring or an unplanned caesarean section.

Table 4. Place of birth and interventions in small (1–2 midwives), medium-sized (3–4 midwives) and large (≥5 midwives) practices.

	Maternal birth outcomes	Small N=234 N (%)	Medium N=248 N (%)	Large N=236 N (%)	p-Value small practices compared to medium-sized and large practices together
	Homebirth (midwife-led)	126 (53,8)	86 (34,7)	56 (23,7)	0.016*
	Poly-clinic birth (midwife-led)	31 (13,2)	35 (14,1)	30 (12,7)	0.06
	Referral to obstetrician during birth (consultant-led)	77 (32.9)	127 (51,2)	150 (63,6)	0.0006*
	Pethidine administration	22 (9,4)	29 (11,7)	43 (18,2)	0.044*
	Epidural	10 (4,3)	23 (9,3)	43 (18,2)	0.106
	Acceleration of birth by means of IV syntocinon	36 (15,4)	69 (27,8)	72 (30,5)	0.51
	CTG	61 (26,1)	108 (43,5)	129 (54,7)	0.015*
	Instrumental delivery	21 (9)	31 (12,5)	44 (18,6)	0.62
	Unplanned caesarean section	8 (3,4)	14 (5,6)	25 (10,6)	0.046*
	Episiotomy	47 (20,1)	81 (32,7)	96 (40,7)	0.067

* Significant p-value <0.05.

Birth experience 

Women were asked to assign a grade to how they had experienced the birth process, ranging from very negative (0) to very positive (10). The score was analyzed in relation to occurrence of interventions (Chart 1) and practice size (Chart 2). Of the women who had an interventionist birth 34.5% reported this experience as negative (score ≤2) and 29.6% as positive (score ≥8). Of the non-interventionist group, 3.8% of the women had a negative experience and 73.1% of the women viewed the birth as a positive experience (p=0.001). Of the women in small practices 60.3% described the birth as a positive experience (score ≥8) and 13.2% as a negative one (score ≤2). In medium-sized practices 47.2% of the women reported their experience of the birth as positive and 22.9% as negative. In large practices these percentages were respectively 36% and 27.1% (p=0.003).

View Large Image Download to PowerPoint

Chart 1. Birth experiences of women reported from very negative (0) to very positive (10) in relation to interventions.

View Large Image Download to PowerPoint

Chart 2. Birth experiences of women reported from very negative (0) to very positive (10) in relation to practice size.

Woman–midwife relationship 

The participants were asked if they knew the midwives in the practice and whether they had met the midwife from the practice who attended the birth beforehand (Chart 3). Women in practices with 1–2 midwives knew the midwife who attended the birth more frequently in comparison to women in practices with more than 2 midwives (p=0.003). Women were also asked if they experienced a relationship with the midwives in their practices by allocating a score ranging from very negative (0) to very positive (10) (Chart 4).

View Large Image Download to PowerPoint

Chart 3. The percentage of women knowing a midwife during pregnancy and birth in relation to practice size.

View Large Image Download to PowerPoint

Chart 4. Experiencing the relationship with midwives reported from very negative (0) to very positive (10) in relation to practice size.

Women in small practices experienced a relationship with their midwives more often when compared to women in practices with more than two midwives (p=0.001).

Discussion 

The women participating in the study are representative of ethnic Dutch women who use midwifery services in the primary care setting according to the background variables of age, parity, marital status and education, homebirth, referral, pain relief, CTG and unplanned caesarean sections.11, 41 As speaking Dutch was one of the inclusion criteria, the women in the study are not representative of ethnic minority groups in the Netherlands. In the study the number of referrals, interventions and administrations of pain relief during birth were consistent with national figures. It was noted, however, that women in the study had lower rates of acceleration of the birth process and instrumental deliveries but higher rates of episiotomies in comparison with national data.11, 41 The percentages of primiparous and multiparous women in the study were representative for the Dutch low-risk population and distribution of women in the various practices was representative for the level of urbanization and division of practices per region.11, 14, 41 There was no significant variation between the demographic details of women among the different practice sizes which would have been a possible explanation for the differences in outcomes between women in the study and national data. The difference between the national data and the study data can possibly be explained in the disproportionate distribution of women in the differently sized practices within the study as compared to the true division of various sized practices within the Netherlands. As a result of a variable sampling fraction small practices were over-represented in the study when compared to the actual number of small practices in existence.14

As women were approached regarding participation in the study by their own midwives selection bias could have occurred. It is unknown exactly how many women were approached and whether midwives consciously and categorically asked certain women to participate rather than others which could have resulted in socially desired answers being given. As 53.6 percent of the consent forms were returned, it raises the question of whether selection by midwives had occurred, and forms had not been handed out at all, or whether only women personally interested in the topic of study had decided to take part. Of the women who initially agreed to participate 85.3% returned the questionnaires, which raises the question of whether more articulate women who wanted to voice either their satisfaction or dissatisfaction had been addressed. In order to decrease selection bias women were approached regarding participation during pregnancy, as at that time women had only experienced a part of their care during pregnancy and the study was intended to examine the total period of the childbirth experience, from the antenatal to postnatal period onward. Although midwives were aware of the nature of the study, they were however not familiar with the exact content of the questionnaire. Gratitude bias was reduced as questionnaires were returned directly to the researcher. The process of coding and recoding made it impossible to locate the women's practices. It was possible that over-representation of practices with definitive characteristics influenced the findings. As the data covers only a period of one month, over-representation is very unlikely, however the sampling technique employed could be open to criticism as it may seem to give rise to an element of bias.

The questionnaire was sent out six weeks after the women's estimated due dates which can still be a time in which women are recovering from birth and readjusting to their life.35 Retrospective questionnaires can create potential for selectivity and inaccuracy in recall and could have influenced the reliability of the findings,42 but as Hodnett43 concluded there may not be an optimum time to evaluate the process of childbirth.

The study showed that women in small practices had a home birth more often and fewer medical interventions, as well as more positive overall birth experiences in comparison to women in medium-sized and large practices. Referral and interventions had a negative effect on how women experienced the birth process. These findings correlate with those of Rijnders et al.,23 who concluded that women with unplanned interventionist hospital births are less satisfied with their birth experience in comparison to women with an uncomplicated homebirth. Olde44 also concluded in his study that obstetrical interventions such as instrumental and operative births contribute to a negative experience of birth.

The study showed that women in small practices had a non-interventionist (home) birth more often and experienced higher levels of satisfaction with their birth experience in comparison to women in practices with more than two midwives. Women in small practices experienced a relationship with their midwife more commonly, knew their midwife more often and were more frequently supported by their own familiar midwife during birth in contrast to women in larger sized practices. These findings confirm results from earlier international studies in relation to continuity of care and carer19, 20, 31, 43, 45, 46 and support the idea that a small team of midwives, who are familiar to the woman, contribute positively to an uncomplicated birth process.

However, home birth, knowing the midwife and continuity of care after referral are not only associated with the number of midwives in a practice but also with practice organization and local policies. At the time of the study there were certain areas in the Netherlands where home birth was not an option due to the closure of local hospitals resulting in women's homes being too far away from the nearest hospital to be able to get there in time should an emergency arise.47 This could have been a possible confounding variable out with the midwives control but with a profound effect on midwifery care and subsequently on women's choices. However as this was more an exception than a rule and involved only a small number of women, it can be assumed that this had no effect on the study's findings. It is known that women in rural areas give birth at home more frequently than women in urban areas. It is, however, known that the rate of home births is influenced by the higher numbers of women from ethnic minorities group within the urban areas.48, 49 The study did not look at the correlation between place of birth and level of urbanization or at local policies or the organization of care in individual practices. The under-representation of women from ethnic minorities within the short time-span of the study, makes it is unlikely to have had any significant influence on the study's findings.

Variables such as practices’ annual caseload, time-management, age, workload and the experience of individual midwives are known to influence midwifery care15, 50 and could have influenced the validity of the findings. In a large practice with five midwives or more, care can be organized in such a way that women only see one or two midwives antenatally. In a small practice with only one or two midwives it does not automatically imply that these midwives spend more time and attention to women than midwives in a practice with, for example, four midwives. Midwives in a duo practice with a shared caseload of 250 women per year might have less time to spend per antenatal visit or birth in comparison to a practice of four midwives with a shared caseload of 350 women per annum. In small practices it is also quite possible that women are confronted with unfamiliar locum midwives more often than in large practices, simply because the large practice has more midwives to share the roster. Initially the study attempted to examine the relation between practice size, caseload size and the organization of care within the respective practices. A questionnaire was developed in order to do this. The responses from participating practices were, however, too small to use for reliable analysis.

The identification of causal relationships between maternal birth outcomes and experiences and caseload size per practice might give more insight into how time is managed within practices and how this is associated with referrals and interventions. Further study is required in order to gain more insight into this issue and to provide a deeper understanding of women's individual thoughts and feelings in relation to birth experiences. It would then be interesting to examine midwives’ attitudes, motivation and views on continuity of carer further, in order to assess whether this is a feasible, desirable and sustainable option for midwives as a method of care or as a vision to fit in with Dutch maternity services.

Despite the limitations of the study, it can be cautiously suggested that the number of midwives in a practice plays an important role in the care of childbearing women. Reflection on and discussion of the study's results seem to be relevant for Dutch midwifery care and maternity services. The development of larger sized practices has been the result of an attempt to deal with increasing workloads. There is currently no shortage of midwives in the Netherlands, but a decrease in the birth rate has resulted in a reduced average annual caseload from 120 to 110 women per midwife per year.3 In the interests of women and women's health a system should be created based on a higher level of continuity of carer. To address this issue there is a need for support from the Dutch government and health insurance companies in particular, in order to support the development of small midwifery practices and to acknowledge continuity of care financially after referral. This can play a profound role in a reduction of referrals and interventions during birth in the Netherlands as well as increasing women's satisfaction with their experiences of the birth of their children.

The findings of the study are relevant to Dutch midwifery care and the culture of the setting where it was conducted; however it may not necessarily reflect generically either to secondary or tertiary care or to other cultures.

Conclusion 

In this study women in the Netherlands in midwifery practices consisting of one or two midwives were referred less frequently and underwent fewer medical interventions during birth compared to women in practices with more than two midwives. Women in solo and duo practices had higher levels of satisfaction with the birth experience than women in larger practices. Knowing the midwife and the presence of a known midwife after referral during birth is important to women and occurred more often in practices with a maximum of two midwives than in practices with more than two midwives. Caution is in relation to the transferability of these findings is advised. Discussion of the study's findings seems relevant in order to reorganize care, however, further study is required.