gms | German Medical Science

Within just ten years (2006-2015), the number of births in hospitals in Berlin increased by 29 percent [1]. In 2016, there were 42,618 births [29]. At the same time, in Berlin, and in Germany as a whole, there is a growing shortage of midwives attending births¹ [5], [29], [35].

In Germany, women are legally entitled to the support of a midwife as well as medical care during labour and delivery (Section 24d, Book Five of the Social Code (SGB V)) [31]. There is, however, no legislative basis stipulating the extent of care, i.e., the duration, content and type of care to which a birthing mother is entitled. According to maternal-foetal guidelines, when giving birth, a woman should be guaranteed 1:1 care by a midwife for at least 95 percent of the time [2]. 1:1 care describes the ratio of labouring mothers to one midwife but does not automatically mean the uninterrupted presence of that midwife. The authors of this paper do assume, however, that 1:1 care generally indicates a high level of midwifery presence and guarantees that the midwife will be immediately available should the birthing mother need her.

The potential benefits of 1:1 midwifery care have been the subject of studies conducted worldwide and there is evidence that it results in fewer interventions and better birth outcomes, e.g., fewer surgical deliveries and greater satisfaction among women with their birth experience [11], [12], [15], [18], [19], [24]. Conversely, a heavier workload and less midwifery presence has a negative impact, e.g., heightened anxiety of the birthing mother, fewer opportunities to monitor the progress of labour and less midwifery support for the mother, even if the midwife is present during labour and delivery [16], [21], [34].

In Germany, 1:1 midwifery care is only the norm for non-hospital births or births supported by private midwives (Begleit-Beleggeburt²). In hospital shift systems, the care ratio is currently between two and more than four birthing mothers per midwife [25]. A hospital birth attended by midwives working shifts for that hospital is the standard form of care³ as the costs for this are met by statutory health insurance.

There are very few studies examining midwifery presence during labour in the German care context. According to a study by Ayerle et al. [3], 60 percent of birthing mothers had a midwife present for approximately 75 percent of the time. Knape et al. [12] studied the effects of midwife presence and workload: On average, a midwife was present for 46 percent of the time and, when questioned eight weeks after the birth of their child, 81 percent of mothers who had been cared for by a midwife reported that they had been satisfied with the level of midwifery presence. Due to the different care structures, international findings are only transferable to a limited extent [13]. As yet, there is no available data on the presence of midwives in labour wards in Berlin hospitals. The study that forms the basis of this article presents current data from a period characterised by a rise in the number of births.

Against the backdrop of an increase in the number of births and a shortage of midwives in Berlin, the aim of this study is first, to determine the amount of time, in a standard care context, for which midwives are present for birthing mothers with spontaneous onset of labour, second, the extent to which mothers perceive specific qualitative care criteria as being fulfilled, and third, how these mothers retrospectively rate the extent of care provided by midwives during labour and delivery. Lastly, the study will particularly examine how this subjective assessment of midwifery presence impacts the mothers’ rating of the extent of care.

All data used to answer the research question are from the Berlin Mothers’ Survey 2016/2017 (Berliner Mütterbefragung 2016/2017) [32], an online questionnaire (cross-sectional survey) on the provision of midwifery care around the time of birth in Berlin. The survey was conducted as part of the author’s Master’s thesis in the field of public health and was funded by the Ludwig Maximilian University Munich.

The data was collected anonymously by means of a structured online questionnaire using Unipark survey software [20]. A semi-quantitative questionnaire⁴ [32] was developed based on expert interviews, a literature analysis and also published articles on surveys of mothers conducted in Germany or internationally [4], [6], [7], [30]. Since the original validated questionnaires were not available for this study, the formulation of the questions followed an explorative approach. The questionnaire was offered in German, English, Arabic and Turkish. A pretest was also conducted among mothers who had given birth in 2016, seven of these were conducted in German and one pretest was performed for each of the other languages offered. The questionnaire and implementation were reviewed by the Berlin Data Protection Office (Datenschutzstelle Berlin) and classified as being in compliance with ethical requirements by the ethics commission at Ludwig-Maximilians-Universität München (Project no.: 17-383 UE).

The survey period was just under six weeks, from 18.04.2017 to 28.05.2017. Information on and access to the survey could be obtained via a website and social media as well as multilingual flyers regularly distributed mainly in paediatric practices and maternity centres. The software prevented individuals from participating in the survey more than once.

On average, completion of the questionnaire took nine minutes (median). Recruitment was almost exclusively via social media posts of participants. Due to data protection, it was not possible to determine in detail how participants accessed the survey.

The study population of the Berlin Mothers’ Survey 2016/2017 comprised mothers who had given birth to at least one child in Berlin between January 2016 and May 2017. A total of 1,271 mothers took part in the survey. The analyses presented in the current article includes 567 of the mothers surveyed (60 percent of the total number of participants). Exclusion criteria were a planned caesarean section or induced labour in order to, if possible, restrict the analysis to mothers who were already in labour on admission to the labour ward. Mothers surveyed who gave birth in a non-hospital setting and mothers who had planned a non-hospital birth as well as mothers provided with care by a private midwife (Begleit-Beleghebamme)⁵ were also excluded from the analyses. A total of 11 (2 percent) participants receiving standard care who fulfilled all other inclusion criteria did not provide responses to all the questions on midwifery care and were therefore also excluded from the analyses.

Midwifery presence was categorised as follows: Continuously (100 percent), mostly (>50 percent and <100 percent), around half the time (approx. 50 percent), mostly alone or alone with accompanying person (<50 percent), not cared for by a qualified midwife but rather by a trainee midwife, no memory of a midwife being present. The variable “midwifery presence” indicates the amount of time for which a midwife was present during the mother’s labour and contains no information about the content of care, e.g., supportive measures.

The term “rating” corresponds to the assessment of quality on a Likert scale by the mothers surveyed. The mothers rated three statements on the type of midwifery care, i.e., three qualitative aspects of care, on a five-level Likert scale. The scale included the options “true”, “mostly true”, “partly”, “mostly not true”, “not true at all” to rate the following statements: (1) the extent to which the respondent perceived the midwifery support⁶ as competent, (2) the extent to which the midwife involved the respondent in decision-making and (3) the extent to which the midwife addressed the respondent’s needs. In cases where care was provided by several midwives, the participant was asked to refer to the midwife who had spent the most time with her as there are significant differences between midwives in terms of the type of care provided [22], [36]. The term “care” is understood to mean monitoring, directing and supporting (through either emotional assistance or practical measures) the birthing mother and the process of childbirth.

Once they had reported on midwifery presence and the qualitative aspects of care, the mothers questioned were then asked to use a five-level Likert scale to rate the extent of care provided during labour and delivery by midwives and, if relevant, by medical staff. The scale included the options: “far too much”, “too much”, “exactly right”, “too little”, “far too little”. The variable extent of care was used as a measure of the quality of care from the mother’s perspective and was intended to depict their satisfaction with both the amount of care and the content of that care as well as with the qualitative aspects of the care. The terms “rating” and “extent of care” were selected because the question about satisfaction with midwifery care does not take sufficient account of midwifery presence. Moreover, other studies show that women who feel there was a lack of care are not necessarily dissatisfied with midwifery care per se and, because of the perceived heavy workload of midwives there is generally a high level of acceptance for more limited midwife presence [15]. The rating was extended to include the extent of care provided by medical staff due to repeated comments in the pretest that medical care may compensate for a lack of midwifery care.

In a free-text answer, the participants had the opportunity to elaborate on their rating of the extent of care. The following variables were also categorised: age, parity, highest level of educational attainment, insurance status, migration background and the perception of a complicated pregnancy as well as the length of stay in the labour ward from admission until delivery.

Data analysis was conducted using MS Excel and R (3.4.0) [33]. To analyse the effect of midwifery presence (exposition) on the rating of the extent of care (outcome) as well as the influence of other variables, bivariate analyses were first conducted. An χ² test with a significance level of p<0.05 was used to verify significant correlations between all of the captured variables mentioned above and the rating of the extent of care. In categories with case numbers of n≤5, these were combined with the neighbouring smaller categories.

Next we used a binary logistic regression with all bivariate significant variables allowing for possible interactions. The use of two subjectively assessed variables (self-reported dependent and independent variables) did not violate the model assumptions of the logit regression because only the independence of the women questioned is required (woman A must be independent from woman B). Further, we did not perform an adjustment of the a level due to the multiple testing because the information provided in the questionnaire does not consist of random variables and correcting on the basis of a universal null hypothesis is therefore not recommended [23]. In order to be able to integrate it into the logistic regression model, the outcome was dichotomised. The cases of “(far) too much” care (n=5, <1 percent were excluded from the analysis. The categories “too little” and “far too little” were incorporated into a single category entitled “too little” and contrasted with the answer “exactly right”. To select the model with the best goodness-of-fit, i.e., the model in which all contained variables had a significant effect on the outcome, a backward selection of the variables was performed.

The free-text answers elaborating on the rating of the extent of care were analysed using qualitative content analysis following Mayring’s approach [17]. Categorisation was initially deductive based on the fourth part of the research question and was then extended by adding inductive subcategories.

The features of the sample are described in Table 1 [Tab. 1]. Exclusion criteria were: planned Caesarean section, induction of labour, non-hospital birth, planned non-hospital birth in the case of hospital births, birth supported by a private midwife and lack of information about midwife care.

One-third (n=187, 33 percent) of mothers surveyed reported that their midwife was present for less than 50 percent of the time from labour ward admission to delivery. A further 124 (22 percent) of mothers surveyed stated that the midwife was present for approximately half the time (Figure 1 [Fig. 1]).

The answers provided by the mothers surveyed on the qualitative aspects of care are summarised in Figure 2 [Fig. 2]. A total of 77 percent (n=438) of the mothers surveyed considered the midwife care they received to be (mostly) competent. 69 percent (n=395) reported that the midwife (mostly) involved them in decision-making and 74 percent (n=421) said that the midwife addressed their needs.

58 percent (n=330) or the vast majority of mothers surveyed rated the extent of care provided by the midwife as exactly right. 41 percent (n=232) rated the extent of care as too little.

202 (36 percent) of the mothers surveyed elaborated on their rating of the extent of care in their own words using the spaces for free text. Women felt the care they received was “too much” when there were too many people in the room, it was not calm enough and also when the individual caring for them behaved in an authoritarian manner resulting in them losing their autonomy to make their own decisions. Care classified as “exactly right” comprised the midwife being available when needed (n=36). 31 of the mothers surveyed also stated that good care and/or a low number of births happening at the same time as their own was fortunate and they frequently (n=16) expressed a link between the two. Participants receiving 1:1 care (n=13) described uninterrupted availability as creating the ideal condition for the process of childbirth and saw the midwife as having an important impact on the course of childbirth and on their personal experience.

The majority of explanatory comments were provided by participants who had rated their care as “too little” (53 percent of all comments). Here, descriptions of capacity overload, i.e., a shortage of midwives and delivery rooms and a high number of births were pivotal (n=123). Some of the mothers surveyed (n=15) felt that the time pressure those caring for them were under was transferred onto them and they perceived this as having a negative impact on the progress of their birth. Some participants who had experienced complications during and after the birth reported that they had a sense of having been left alone (n=37) and that the lack of care and their concerns about this hindered the process of childbirth (n=47). Reference was made to a lack of medical care from doctors in connection with the lack of availability of an epidural (n=2). One woman reported that medical personnel compensated for the lack of midwife presence.

Assessments of the role of non-professional birthing companions, such as the woman’s partner, varied. While two of the women surveyed felt that the support offered by a birth partner compensated for the lack of midwife care, four of the participants emphasised that this was no substitute for the professional care provided by a midwife. More specifically, mothers surveyed also felt that there was a lack of practical and emotional support, a lack of individualised care and a lack of communication, particularly when it came to providing information on the process of childbirth and obstetric measures. This criticism was also voiced by women who rated the availability of the midwife as adequate (n=4).

Correlations between the rating of the extent of care and midwife presence are illustrated in a cross table in Table 2 [Tab. 2] containing the dependent and other significant independent variables in the logistic regression. The less the midwife was present, the less she addressed the needs of the mothers in labour, the less competent the midwifery care was felt to be, or the more likely the birth was to be perceived as complicated, the greater the chance that the extent of care would receive a low rating. The less present the midwife was, the less likely the qualitative criteria of care would be rated positively. For example, 24 percent (n=45) of the mothers surveyed who reported that the midwife was present for less than 50 percent of the time stated that the midwife did not address their needs compared to just 2 percent (n=4) of the mothers who reported that the midwife had been continuously present or mostly present.

In the sample, reported birth complications were associated with low midwife presence. The share of the mothers surveyed who reported that they mostly had no midwife care was 40 percent (n=78) in the group of women with birth complications and 29 percent (n=107) for those who did not report any complications. In contrast, 35 percent (n=67) of the mothers with complications reported that they had continuous midwife care and 45 percent (n=164) of those who did not experience complications.

Neither parity nor the duration of stay in the labour ward up until delivery were significant factors influencing the rating of the extent of care, although the longer a woman spent in the labour ward, the more likely she was to rate the extent of care as low. At the same time, however, midwifery presence actually fell. With shorter stays of up to three hours, 19 percent (n=45 of 242) of the women were mostly alone or only with their birth partner. With labour ward stays of 5-10 hours, however, 80 percent (n=56 of 70) stated that there had mostly been no midwife present.

All data used in this study were gathered using the Mothers’ Survey.

55 percent of the mothers surveyed in this study reported that, in the context of standard care in Berlin hospitals, a midwife was present for 50 percent of the duration of their stay in the labour ward during birth. This is considerably lower than the percentage reported in the study by Ayerle et al. [3] who identified a midwife presence of around 60 percent for more than 75 percent of the duration of labour. Also in the study by Knape et al. [12], at 46 percent, midwifery presence was, on average, higher.

One possible explanation for the differences could be that the data used in the study by Ayerle et al. in 2005 were collected in maternity hospitals in Lower Saxony and the data in the study by Knape et al. between 2007 and 2009 were gathered as part of the introduction of midwife-led maternity support in four hospitals in Hamburg, Bremerhaven and Bad Cannstatt. Regional differences might be the result of the different workloads of obstetrics departments, e.g., because of different birth rates or the distance from the next closest obstetrics department. A heavier workload tends to go hand in hand with a higher number of births per midwife and thus presumably also a lower midwife presence. The setting might also have influenced the presence of midwives. In a midwife-led labour ward, for example, midwifery presence might, due to the particularly high rating of midwifery care by the mother or the midwife, be higher than in the context of a doctor-led labour ward. An important factor influencing the substantially lower presence of midwives reported in the Berlin Mothers’ Survey might also be the rise in the number of births observed in recent years and the growing shortage of midwives attending births (e.g., due to a deterioration in working conditions) [25]. Both of these factors also lead to a higher number of births per midwife.

However, the comparability of the findings is limited due to the different methods used to capture information on midwifery presence and the different samples. In the two studies used for comparison purposes, the midwives documented their presence during births and the birthing mothers included in the studies showed no or only marginal perinatal risks. Known perinatal risks (e.g. hypertension) can necessitate increased supervision during childbirth, which might increase midwife presence. In the Berlin Mothers’ Survey, midwifery presence was captured retrospectively through an assessment by the mothers questioned. The presence of perinatal risks was not an exclusion criterion.

The responses given on the qualitative aspects of care were largely positive. 77 percent of the mothers surveyed considered their midwife to be (mostly) competent, 69 percent reported that the midwife (mostly) involved them in decision-making and 74 percent said that the midwife addressed their needs.

Another finding of this analysis is that the mothers surveyed mainly rated the extent of care provided by midwives during the birth as adequate but 41 percent rated the care as too little. An online survey on midwifery care in Baden-Württemberg conducted in 2016 produced similar findings [6]. In the survey, slightly more than half of all the mothers questioned stated that they would have liked more care during birth and only very few mothers would have been satisfied with less care. In 2016, Baden-Württemberg recorded its highest birth rate since 1999. Between 2010 and 2015, the number of births increased by 11 percent [14]. These figures are comparable with the corresponding statistics for Berlin referred to at the start of this paper. This can be seen as an indication that a large proportion of mothers giving birth during a period with a high birth rate receive insufficient midwifery care during birth.

The results of the regression analysis show that with low midwife presence of 50 percent of the duration of stay in the labour ward or less, the rating of the extent of care was far poorer, whereas it made no significant difference whether the midwife was continuously or mostly present during the birth. This seems to indicate that the perceived midwife presence is decisive for the perceived quality of care, and more midwifery presence goes hand in hand with higher quality of care.

Also, in this study, two of the qualitative aspects of care covered by the survey had a significant effect on the rating of the extent of care. If the midwife (mostly) did not address the needs of the mothers surveyed, these mothers rated the extent of care as too low almost four times more frequently than the mothers who reported that the midwife (mostly) addressed her needs. The effect on the rating of the extent of care was almost equally strong (OR 3.6) when the mothers surveyed considered the midwifery care they received as (mostly) not competent as compared to the mothers with (mostly) competent midwife care.

Moreover, the present study shows that the qualitative aspects of care covered by the survey were more likely to be fulfilled, the higher the midwifery presence. In their study, Ross-Davie et al. [21], [22] also found that midwives support birthing mothers less, the more they labour outside the delivery room and, in addition, the midwives themselves stated that continuous midwifery presence was a core component of the mother-to-be feeling supported.

In the study by Knape et al. [12], 87 percent of women benefited from supportive activities carried out by a midwife (e.g. massage). The type of support a midwife offers depends crucially on their self-concept as a midwife and also the prevailing care culture [13]. Presumably, this type of support was partly responsible for the high level of satisfaction with midwife presence of 80 percent.

The duration of stay in the labour ward and the respondent’s parity did not have a significant impact on the rating of the extent of care. This is presumably due to the higher level of midwife presence during a short stay, or faster delivery [8]. The study by Ayerle et al. [3] also made the assumption that a high level of midwifery presence was correlated with rapid births or births which were already at an advanced stage on admission. The quantitative and qualitative results of the present study suggest that women with longer labours, including first-time mothers versus multiparous mothers, were more likely to perceive a care deficit. This suggests that the presence of a midwife is not only important at the end of the birth.

This study produced differing qualitative results with regards to the role of non-professional birth partners. Internationally, the influence of continuous support from laypeople is rated positively [9] but there is no study comparing this form of support with professional care provided by midwives in settings that correspond with conditions in Germany.

It is assumed that midwifery presence and support depend on the course of labour and that midwives respond to high-risk or pathological labours by increasing their presence and support. However, this assumption is not proven [12]. In the study by Ayerle et al. [3], the presence of the midwife did not change as a result of medical intervention (as an indicator of deviation from the standard physiological process) as compared to births where no medical interventions were necessary.

The current study even showed that midwives supervising mothers who experienced birth complications were less present than for those who experienced no complications. The mothers surveyed who experienced complications during labour and delivery were significantly more likely to rate the extent of care as too little (OR 1.65) compared to respondents who experienced no complications. A causal link between more limited midwife presence and the occurrence of complications can, however, not be proven due to the study design. It is also possible that the occurrence of complications influenced the mothers’ perception of the midwife’s presence and the rating (reverse causation). What we are able to derive from this finding, however, is that, particularly the potentially increased care needs in the event of birth complications were not met. It is conceivable that, due to capacity shortages, despite birth complications, there could be no significant increase in the presence of a midwife or the midwife on duty did not notice the deviations from the normal physiological process in good time due to her limited presence.

The online survey proved to be a suitable method for surveying mothers with adequate German language skills [4]. The use of an unvalidated questionnaire should generally be considered a limitation. Thanks to the high number of plausible responses, the questionnaire did prove to be suitable but there is still a need to develop the tool further and to validate it, particularly when it comes to operationalising midwifery presence and capturing the qualitative aspects of care.

Moreover, the sample was only representative to a limited extent particularly because it was characterised by a level of education that was considerably higher than average [26]. Mothers with a migration background and multiparous mothers were underrepresented for Berlin [10], [27]. The share of mothers with an instrumental vaginal delivery was slightly higher than the share with c-section deliveries and the share of c-sections was slightly lower than the average rate for Berlin [27]. In addition, a self-selection bias is probable due to respondents having strong emotions about the subject and, because of the study design, there is a risk of recall bias. In this study, the perception that there was a lack of care could have led to mothers underestimating midwife presence retrospectively.

Criticism could be levelled at the fact that the retrospective and subjective statements made by the participating mothers do not permit objective analyses. Care needs and the rating of care also depend on the personality and attitude of the mother. However, in the assessment instrument test in the study by Ross-Davie et al. [22] on quantitative and qualitative support provided by midwives, the information supplied by the mothers and the observing researchers were highly correlated. Ultimately, the quality of care cannot be assessed without the rating provided by the recipients of that care [22]. However, objective criteria are also essential when trying to capture care data in order to examine the relationship and differences between actual care and perceived care in a methodologically sound way. The main disadvantage of a retrospective cross-sectional study design is that it is not possible to derive causality from the results.

Another drawback of the survey method is that findings relevant for the course of labour, such as perinatal risks, cannot be measured in a valid manner, as they would be in a professional study for instance. These findings could have an impact on both the mother’s perspective as well as the midwife’s and may also be connected to the course of labour.

When classifying the results of the logistic regression, it is important to bear in mind that the subjective assessment of both the independent and the dependent variables by the same person makes correlations between the different variables likely. This may influence the size of the effect measure and the significance of the results.

The findings of this survey suggest that midwife presence in standard care in Berlin is currently too low. From the mothers’ perspective, there is a substantial quantitative and qualitative care deficit, which is correlated with this low level of midwife presence. First-time mothers and mothers with long or complicated births seem to be at a particular disadvantage. Our data also shows the perceived presence of the midwife to be a significant factor in the mothers’ satisfaction with the care received during labour and delivery.

In light of the lack of studies in the German care context, there is a need for more research to be conducted on the actual presence of midwives in standard care and their impact on both the care itself (e.g., the number and type of interventions) and on birth outcomes for both mothers and their children. A comprehensive study requires both an objective measurement of midwifery presence and for perinatal risks to be recorded using proper, standardized documentation as well as by surveying mothers. When conducting a survey of mothers on their satisfaction with midwifery care, the duration, content and type of care provided must be taken into account.

In Berlin, a Midwifery Roundtable (Runder Tisch Geburtshilfe) was established because of the shortage of midwives. The roundtable brings together representatives from politics, the health insurance funds, the medical profession and maternity hospitals as well as the midwives in Berlin who attend births and the Berlin Midwifery Association (Berliner Hebammenverband e. V.). On 1st February 2018, the Roundtable adopted a package of measures. One of the objectives of the package was to improve midwives’ employment conditions, by increasing the workforce in hospitals, for instance. In addition, training capacity was also to be increased and the academisation of the midwife profession should be actively developed [28]. Increasing training capacity without having an adequate number of certified midwives working on labour wards must be seen in a very critical light in terms of the impact on the quality of training and the quality of birth assistance provided, not least because of the risk that trainee or student midwives will be used to cover shifts instead of certified midwives. An evaluation report on the implementation and effects of the package of measures is currently being compiled⁷.

¹ In this paper, the term midwife refers to both male and female practitioners in this field.

² A Begleit-Beleggeburt is defined as a birth supported by a private midwife who is not part of a hospital’s shift system or its on-call service but rather accompanies her client, with whom she has developed a rapport during pregnancy, to her planned hospital birth.

³ The concept of standard care was introduced to statutory health insurance (Section 75 SGB V) as part of dental care. There is no generally accepted definition of standard care in other service areas. Generally, this term indicates care provided on the basis of the collective contract with the health insurance fund and thus includes all the services that all individuals with statutory health insurance are entitled to, irrespective of the specific health insurance fund they pay into. According to this definition, benefits that are not part of standard care are, for instance, optional benefits, such as subsidies for the midwife’s on-call flat fee.

⁴ The questionnaire used in the Berlin Mothers’ Survey 2016/2017 can be found in the appendix of the report on the findings of the survey.

⁵ For a full definition, see Begleit-Beleggeburt in footnote 2.

⁶ Based on suggestions from the pretest regarding the formulation of the items, the term “care” was replaced by “support” to emphasise the supportive nature of midwifery care, which is geared towards the needs of the birthing mother. In this statement, support is used as a synonym for care.

⁷ Response from the Press Office of the Senate Department for Health, Care and Equality on 15.02.2019 to an inquiry on the status of the results of the package of measures.

The authors declare that they have no competing interests.

The present study was financed by Ludwig-Maximilians-Universität München for strengthening the research focus and for the implementation of student research projects.