The Dutch Birth Centre Study: study design of a programmatic evaluation of the effect of birth centre care in the Netherlands

The Dutch Birth Centre Study consists of five sub studies which are linked to one another:

Different qualitative and quantitative methods will be used in the sub studies. Data collection includes observations, interviews (individual and group interviews), questionnaires (clients, caregivers, managers of birth centres), standard registered data and additional registrations. Data collected in one sub study will be shared with other sub studies as much as possible to make sure that the birth centres and other professionals involved in birth and postpartum care are minimally burdened by participating in the various evaluations. This study will be conducted in the period 2013–2015.

Serious adverse outcomes are expected to be very low as the study population consists of women with an uncomplicated pregnancy who will start labour under midwifery-led care. Therefore the two main outcome measures will be composite measures: the Optimality Index (OI) and a composite measure of adverse neonatal and maternal outcomes [16].

The Optimality Index is a composite score combining background and outcome data based on a simple scoring system: optimal or not optimal. The optimal score is maximal perinatal outcome with minimal intervention placed against the woman’s health status. The OI is very suitable to compare groups with comparable risk profile or to correct group comparisons for differences in risk profile [17, 18]. Background data include age, parity, obstetric history, postal codes to characterize neighbourhood effects and social economic status, origin (Dutch or non-Dutch), together indicating the risk profile [19]. Elements included in the outcome part of the Optimality Index are for example: colour of amniotic fluid, induction/augmentation of labour, episiotomy, instrumental (vaginal) birth, Caesarean section, placental retention (>30 min) and Apgar score at 5 min.

The composite adverse outcome score will include maternal and neonatal outcome indicators. Adverse maternal outcome indicators are maternal death (within 42 days of giving birth), third or fourth degree of perineal trauma, placental retention, postpartum haemorrhage (>1000 ml), and admission to an intensive care unit or obstetric high care unit. Adverse neonatal outcome indicators are stillbirth after presentation in labour, early neonatal death (<7 days), Apgar score <7 after 5 min, neonatal encephalopathy, meconium aspiration, admission to neonatal unit within 48 h of birth and birth weight below 5th percentile.

Birth centres in their current presentation are relatively new in the Netherlands. Therefore no clear definition and no list of birth locations that can be considered a birth centre is currently available. To examine the large variety of possible birth centres criteria for inclusion are selected: birth settings where out-of-home community midwifery led care is provided in a home-like environment to women at low risk of medical complications at the onset of labour. Every birth location that can be included will be invited to participate in this part of the study. Based on the characteristics birth centres are examined by three independent researchers and a selection of all potential Dutch birth centres is made.

A comprehensive set of structure and process quality indicators will be developed to evaluate birth centre care using a multi-staged approach. The development process consists of three phases: 1) identification of existing structure and process quality indicators in birth care (literature study); 2) translating indicators for maternity care in general into determinants for measuring structure and process quality of birth centre care; 3) determinant selection of relevant structure and process quality indicators (two-step web-based Delphi consultation) [20]. The web-based, anonymous nature of the Delphi technique ensures that a single individual cannot dominate the consensus formation. Professionals from different disciplines who are working with or in a birth-centre-like setting with several years of experience, representatives of health insurance companies, policymakers and advisors will be invited to participate in the Delphi consultation. The experts are instructed to rate the determinants both on relevance to a birth centre setting and on feasibility of use and, if necessary, to comment on them or add new topics. Each determinant will be rated by each expert on a seven-point Likert scale (1 = not at all relevant/feasible; 4 = neutral; 7 = very much relevant/feasible). Agreement among experts is defined as 80 % or more of the ratings within a range of three (i.e. 5-6-7 of 4-5-6). In the first round determinants with a median score of ≥ 6 with agreement on both ratings are considered to be relevant and feasible to collect and are accepted instantly. Determinants scored with a median score of ≤3 are rejected. Median scores of >3 and <6 with agreement or ≥6 without agreement are scored again in the second Delphi round. In the second round, the experts are informed about the median scores of relevance and feasibility of the total expert group, their own scores and the comments of the respondents regarding determinants for which no consensus is reached in the first round. They are instructed to re-consider their own rating of the determinants presented in the first round as well as to rate and comment possible new elements the same way as in the first round.

This procedure will result in a list of potential structure and process quality indicators for birth centres in the Netherlands. In order to test whether these quality indicators actually can measure the quality of birth centres, they will be validated within the presumed selection of birth centres.

The Dutch Birth Centre questionnaire will be sent to a (management) representative of each birth location as identified in the first step of the study. A definition for different types of birth centres in the Netherlands will be developed based on internationally used definitions and the information obtained through our questionnaire. The characteristics of all Dutch birth centres will be described.

To construct a typology of birth centres we will use the concept of integrated care. This concept was developed first for the increasing number of people with a chronic disease. Different (health-related) disciplines are involved in the continuous care for persons with a chronic disease. For instance, care for a person with diabetes mellitus type II may involve a general practitioner, a dietician, and a physiotherapist, but also an endocrinologist. The essence of integrated care is a continuum of care for service users which crosses the boundaries of primary, secondary, tertiary and public health care [20‐22]. The definition of the WHO illustrates the extensive conceptualization of integrated care: “a concept bringing together inputs, delivery, management and organization of services related to diagnosis, treatment, care, rehabilitation and health promotion. Integration is a means to improve the services in relation to access, quality, user satisfaction and efficiency” [23]. Domains of integration are 1) clinical, 2) professional, 3) organisational, 4) systemic, 5) functional, and 6) normative integration [24]. Based on the scores on the different domains an overall score of integration will be calculated to define the level of integration for each birth centre as low, medium or high.

Midwifery practices in the area of all birth locations in this study will record the data for each birth under their care during 3 months: data routinely recorded in the Netherlands Perinatal Registry and additional process indicators not available from the Netherlands Perinatal Registry.

To assess client experiences the postnatal part of the Repro-Q will be used. Especially for this study, questions about facilities and transfer are added for women who received care in a birth centre. The same midwifery practices as in sub study 2.1 will be asked to distribute information of this part of the study and a acceptance paper form to each woman that receives care in their postpartum period regardless who gave natal care to them. These women will be approached 6 to 8 weeks after they give birth by the way they preferred to answer the questions on client experiences i.e. by email, by post or by telephone. A reminder will be sent after 4 weeks.

To assess the experiences of professionals working within and with a birth centre a questionnaire will be developed based on earlier questionnaires used in workforce planning [25, 26]. The development will be a joint effort with other Dutch researchers to create a multipurpose questionnaire. The questionnaire will contain questions about personal background, current job situation, cooperation with other care providers, current job evaluation and future job situation and will be sent to all care givers working in or with a birth centre like community midwives, clinical midwives, obstetricians, paediatricians and maternity care assistants.

In May 2011 a baseline assessment was performed in areas where a birth centre was intended to start before June 2013. Ten regions collected data for more than 3 months. Midwifery practices in the area of an intended birth centre recorded the following data for each birth under their care: data routinely recorded in the Netherlands Perinatal Registry and additional process indicators not available from the Netherlands Perinatal Registry (see sub study 2). The follow-up measurement has been conducted in the second half of 2013. Data collection has resulted in around 3000 births for the pre-test period and will result in 3000 births for the post-test period. These numbers are sufficient to describe changes in the region between the period before the birth centre started and afterwards. Logistic regression analysis will be performed to study the difference in planned place of birth between the period before and after the start of the birth centre. Linear regression analysis will be performed to test the mean differences in the Optimality Index between the period before and after the birth centre started. All analyses will be adjusted for potential confounders such as maternal age, parity and gestational age.

The outcome measure for the effect study will be the Optimality Index. At least three midwifery practices in the area of each birth centre in this study will record data for each birth that started under their care during 3 months: data routinely recorded in the Netherlands Perinatal Registry and additional process indicators not available from the Netherlands Perinatal Registry (see sub study 2). A sample size of nine birth centres per level of integration (low, medium, high, see sub study 1.4) with 66 women per centre achieves 80 % power to detect an effect size of 0.2 (ICC = 0.005, alpha = 0.05).

For the births included in the effectiveness part the costs will be assessed. Costs of birth in this study include the health care costs from the start of labour until 7 days after birth. These costs consist of a) medical interventions during birth such as: referral, augmentation, pharmaceutical and non-pharmaceutical pain relief, continuous foetal monitoring, intra partum antibiotics prophylaxis, continuous support of labour, birth by caesarean section, instrumental vaginal birth, manual placenta removal and blood transfusion, b) use of hospital facilities such as: hospital admission and length of stay, and c) staffing such as: attending midwife or obstetrician or both, maternity care assistance during childbirth and in the days thereafter.

Volume of health care resource use will be registered prospectively on the case record form used by the attending midwife. Costs of birth and postpartum care are estimated by a detailed cost price analysis. Other resource use (e.g. hospital days) will be translated into costs using standard prices [25].

Total costs per woman according to planned place of birth will be calculated. Mean differences between the groups and their 95 % confidence intervals will be estimated using non parametric bootstrapping due to the skewed nature of cost data.

The aim of the economic evaluation is to study the cost-effectiveness of the care provided by different types of birth centres compared to home birth and hospital birth under midwifery led care. The economic evaluation will be performed from a health care perspective. The time horizon of the economic evaluation is from the start of labour until 7 days after birth. Due to this short time frame no discounting will take place.

Costs and effects (as measured by the Optimality Index) will be transformed in a net-monetary benefit (NMB) estimate. Using the net benefit regression approach cost-effectiveness acceptability curves will be generated which show the probability of being cost-effective for the different planned places of birth: at home, the different types of birth centres and hospital birth for all acceptable levels of the willingness to pay [27].

This longitudinal qualitative research focuses on arriving at a deeper level of understanding of the process of care and cooperation and its development over time. The research design for this study is that of a process study using the grounded theory methodology [28]. Seven birth centres will be selected after an initial first exploratory round of visits by theory-driven case selection [29].

Data will be collected through investigating from a so-called engaged scholarship/quasi-ethnographic perspective, in which from a variety of data sources over a substantial period of time conclusions will be drawn. This means that observations will take place in each of the birth centres for a number of days at a time, during daytime and during night-time, to observe ongoing activities and to interview care providers as well as clients.

Data will be analysed using the constant comparative method. The purpose is to attain new insights by breaking through standard ways of thinking about phenomena reflected in the data [28]. In this way concepts emerge as theory is formed. Analysis will start as soon as the first data are collected and continue with each additional data collection. The first step in the analysis will be coding the transcripts of the observations and interviews. The analysis and findings will be based on a triangulation of different types of data [30]. First, the researcher will make comprehensive detailed field notes of the observations and informal conversations. Second, surveys will be used. Third, qualitative dimensions such as distances between birth centre and obstetrical ward and time needed for transfer in case of referral will be measured. Fourth, a member-check will be conducted to verify the collected information. Fifth, the researcher will keep a diary in which she reports her own behaviour and feelings, as distinct from her observations in the field notes. Sixth, peer-reviewing will be conducted by evaluation of the work by one or more colleagues.

The insights of all sub studies will be put together, whereby the various elements of the research will be integrated. The national quantitative results will be combined with the insights from the interviews, the cost effectiveness results, client and professional experiences and the information and mirror sessions of the in-depth study to provide insight in the quality of birth centre care in the Netherlands. The regional quantitative results will provide insight into the development over time in a changing health care setting. Based on the confrontation of the various kinds of information more insight can be gained about birth centre care in general and about the strengths and weaknesses of different ways to organise birth centres in particular.

This form of triangulation of information that results from various scientific paradigms is an exciting process that will be carried out by the principal investigators of the participating organizations. It will lead to recommendations for further development of birth centres in the Netherlands.

This study will be carried out by an unique collaboration of several organizations, each with their own proven expertise in the field of the organization of health care and perinatal care in particular. Prior to as well as during the study period all organizations will be involved in both the planning and execution of all related sub studies. A broad advisory committee will be formed by representatives of all different kind of maternity care providers, research consortia, professional organizations, health insurance companies, national health department and clients to discuss the process and preliminary outcomes of this study. Design and planning of the study were presented to the Medical Ethics Committee of the UMCU (University Medical Centre Utrecht). They confirmed that this study agrees with Dutch legal regulations for the methods used for this study and because of that official ethical approval of this study is not required [31].