Assessing Dutch women’s experiences of labour and birth: adaptations and psychometric evaluations of the measures Mothers on Autonomy in Decision Making Scale, Mothers on Respect Index, and Childbirth Experience Questionnaire 2.0

We mounted the survey on an online platform and recruited through social media and networking sampling through midwifery services and organizations. Women could access the survey through a link on Facebook pages of (1) midwifery care practices, (2) postnatal maternity care organisations, (3) all Dutch Midwifery Academies, or (4) through posts on other Facebook pages related to pregnancy and birth. To minimize recall bias, eligibility was limited to women who gave birth < 1 year prior to filling out the survey. Women were included in the analyses, if a community midwife, hospital-based midwife and/or obstetrician provided care in the intrapartum period, and if the MADM, MORi, and CEQ2.0 were completed. The inclusion of women with any combination of maternal healthcare professionals, resulted in participants completing between one or three measures of the MADM and MORi. To avoid collecting data from women who only shared their (most) positive or negative experience, we only included data of women who had completed both measures for all healthcare professionals attending at birth in our analyses for the psychometric properties reliability, known group validity and convergent validity. To assess the psychometric property acceptability, we did included all data without such an exclusion, since it will provide information if women were willing to complete the MADM and MORi multiple times. Women had to be proficient in the Dutch language and had to consent to anonymous use of their data for this study. Women who were 16 years or younger were excluded. We aimed for a sample size of 500 women or more, which is considered ‘very good’ for psychometric purposes [25, 26]. As previous studies reported difficulties in achieving a representative sample regaring ethnicity and socioeconomic position, increased efforts were made to reach these populations by involving midwfiery care practices in areas with a higher density of underrepresented groups [12, 19, 20].

In a previous study we translated the MADM and MORi according to WHO-guidelines and adapted both measures to the Dutch healthcare system [12]. Afterwards both measures were evaluated on their psychometric properties [12]. The results of this study supported the feasiblity, reliability and knowngroup validity of the MADM and MORi in pregnant women [12]. For the current study, these Dutch versions were adapted to assess experiences in the intrapartum period and instructions were added to ask women to complete the MADM and MORi separately for each health care provider they encountered in the intrapartum period.

The MADM is a measure that consists of 7 items with responses scored on a 6-point Likert scale ranging from 1 (completely disagree) to 6 (completely agree), resulting in a summed scale score ranging from 7 to 42. A higher score indicates higher experienced autonomy. The psychometric properties of the MADM have been evaluated in Canadian settings on pregnancy and birth experiences of women [19]. In our previous study, the MADM was translated and adapted to the Dutch context and consisted –as the original version- the same number of items and scoring method. Atfterwards the Dutch verions of the MADM was psychometrically evaluated regarding women’s experiences during the prenatal period [12]. In both settings the MADM showed good feasibility and a good internal consistency; Cronbach’s alpha ranged from 0.93-0.96 [12, 19]. In addition, the construct validity of the Dutch MADM showed the ability to discriminate between characteristics of women who differed on demographic characteristics (e.g. Dutch Region), and healthcare provision characteristics (e.g. type of maternal healthcare provider, length of consultations) [12].

The MORi consists of 14 items with responses scored on a 6-point Likert scale ranging from 1 (completely disagree) to 6 (completely agree), resulting in a summed scale score ranging from 14 to 84. The psychometric properties of the MORi have been evaluated based on the experiences of respect that women report on for their entire maternity care journey [20] or during pregnancy care alone [12]. In our previous study, the MORi was translated and adapted to the Dutch context and consisted –as the original version- the same number of items and scoring method. Afterwards the Dutch version of the MORi was psychometrically evaluated regarding women’s experiences during the prenatal period [12]. We reported that the internal consistency of the MORi was satisfactory for the Canadian, USA and Dutch versions i, with Cronbach’s alpha ranging from 0.76 to 0.94 [12, 20]. Additionally, the MORi discriminated between several subgroups based on demographic characteristics (e.g. social economic status, history of substance abuse), pregnancy characteristics (e.g. risk factors present, complications during pregnancy) or healthcare characteristics (type of maternity provider) [12, 20].

The CEQ2.0 contains 22 items and assesses women’s experiences regarding their childbirth across four domains, i.e. own capacity (8 items), perceived safety (6 items), professional support (5 items), and participation (3 items) [21]. The 22 items of the CEQ2.0 were translated from English to Dutch as follows: (1) a forward translation from English to Dutch, by two independent translators of the Language Centre VUmc resulting in two independent translations. (2) Reconciliation of the forward translations by one native English speaker (PdCsenior lecturer/psychologist) and one native Dutch speaker (LLP, author) to one reconciled forward translation. (3) Independent backward translation from Dutch to English, by two Dutch translators fluent in English who were naïve to the measured construct of childbirth experiences. (4) Consensus with a native English speaker (PdC), native Dutch speaker (LLP) and both backward translators, resulting in one final translated version. Since we aimed to include the Dutch version of the CEQ2.0 in an online survey we changed three items that in the original version used a visual analogue scale into a marking scale ranging from 0 to 10. This scale matches the Dutch school marking system, with 0 reflecting the lowest possible score and 10 the highest. Negatively worded items were reversed in scoing and the marking scale-scores were categorized into 0-2 = 1, 3-5 = 2, 6-8 = 3 and 9-10 = 4. The other nineteen items were scored on a 4-point Likert scale ranging from 1 (completely disagree) to 4 (completely agree) [21]. The scoring of the CEQ2.0 is the following, the item ratings per subscale are aggregated to scale scores by summing the coded values of the items in each scale and dividing by the number of items in that subscale [21] The weighted mean CEQ-score can be calculated by adding all subscale scores (own capacity, perceived safety, professional support, and participation) and dividing by four [21]. The theoretical range is from one to four, a higher score indicates more positive experiences [21]. The psychometric properties of the CEQ2.0 had been evaluated based on the childbirth experiences of women one to three months postpartum [21‐23]. Internal consistency was satisfactory; Cronbach’s alpha ranged from 0.82-0.90 in different settings in Sweden, United Kingdom, and Iran [21‐23]. The CEQ2.0 discriminated between several subgroups based on maternal characteristics (e.g.parity) and birth characteristics (e.g. labour duration) [21]. The Dutch versions of the MADM, MORi and CEQ2.0 are shown in the Additional file 1 including their scoring methods.

The online survey comprised the Dutch versions of the MADM, MORi and CEQ2.0. Women completed all measures at the same time, and could not skip items. Aditionnally, women provided feedback on the clarity of the MADM and MORi regarding its instructions and on individual items by indicating whether or not these items were difficult to complete (yes/no) [12]. In addition, women were asked if they had any general remarks regarding the completion of the MADM and MORi. Aditionally, we collected data on the following individual demographic characteristics: maternal age, (later categorized into ≤25, 26-30, 31-35, and ≥ 36 years); region, ethnicity, marital status, religion, education, and income. Dutch regions was based on Dutch provinces, North (Groningen, Drenthe, Friesland), East (Overijssel, Gelderland, Flevoland), South (Noord-Brabant, Limburg), and West (Noord-Holland, Zuid Holland, Utrecht, Zeeland). Ethnic background was categorised as no migrant background (Dutch) versus migrant background, which was defined as the women or one of the women’s parents being born in the Dutch Antilles, Aruba, Suriname, Morocco, Turkey, Indonesia, or other foreign country [27]. Marital status at moment of birth was registred as single versus partner/spouse. Religion was measured as none, Christianity, Islam or other. Education level was categorized as follows: Low (i.e. primary school, first three years of secondary school or lower level of vocational training), middle (upper secondary school or medium level of vocational training or work-based learning pathways), or high (higher vocational education, and university education). Family monthly net income was assessed by the following categories < €2000, €2000-€2500, €2501-€3500, ≥ €3501 and the option “I do not want to provide this information” was available.

Women also reported on pregnancy characteristics, including parity (nulliparous or mutliparous), and gestational age, which was categorized in to ≤36 + 6, 37 + 0-39 + 6,40 + 0-40 + 6, 41 + 0-41 + 6, ≥42. Women indicated if they had adverse physical health events during pregnancy measured by yes or no on one of the following complications: intra-uterine growth restriction, blood loss first trimester, blood loss second or third trimester, diabetes gravidarum, cholestasis, problems with blood pressure, abruptio placenta, placenta praevia, polyhydramnion, thrombosis, HELLP-syndrome, dysmature, macrosomia, and/or oligohydramnios. Women indicated if they had adverse psychological health status, by reporting if they had experienced depression or anxiety.

Birth characteristics included mode of birth (vaginal birth, instrumental vaginal birth or caesarean section), the number of times women had met the healthcare provider who attended at their births, in their prenatal period (0, 1-2, 3-4 or ≥ 5 times), place of birth (home, maternity hotel/birth centre, hospital), whether birth interventions took place (i.e. induction of labour, CTG monitoring during labour, pain relief treatment, catheterisation, vaginal laceration repair, vaginal laceration repair that required surgery, and manual placenta removal), and neonatal outcomes (mortality, resuscitation).

We pilot tested the online survey with eight women, after which several technical and content changes were made, for example: information on a couple of items was clarified and explanations on what type of birth interventions took place were adjusted to improve understanding.

Descriptive statistics were used to describe the maternal, pregnancy and birth characteristics of the included population. Since the MADM and MORi scores distributions were skewed, median scores with corresponding interquartile ranges (IQR) were reported with no decimal places [12]. Women completed an MADM and MORi measure for each involved healthcare provider (community midwife, hospital-based midwife, and obstetrician). Depending on this answer, women completed one to three measures of the MADM and MORi. To report overall median scores of the included population and subgroups who differed on maternal characteristics, we selected at random a MADM or MORi score if women completed two or three measures. The CEQ2.0 scores were reported by weighted mean with standard deviation (SD), the scores were reported with two decimal places [21‐23].

Reliability was assessed by calculating Cronbach’s alpha for the MADM, MORi and CEQ2.0. Additionally, the Cronbach’s alphas were calculated for each MADM and MORi specified for the involved midwives, hospital-based midwives and obstetricians. A calculated Cronbach’s alpha of ≥0.70 is satisfactory [28].

Construct validity was assessed in terms of known group validity and convergent validity [29]. Regarding known group validity, we hypothesized similar MADM, MORi and CEQ 2.0 scores on women who differed on demographic characteristics [12]. By comparing those measurement scores with women who had a healthy pregnancy and a physiological birth, statistical significant lower scores of the MADM, MORi and CEQ2.0 would have been reported by women who experienced pregnancy complications, birth interventions (e.g. induction of labour, episiotomy or caesarean section) and adverse neonatal outcomes (i.e. mortality and resuscitation) [12, 17, 19, 20, 30‐32]. Statistical differences were calculated on those outcomes if the prevalence was at least 5%. We hypothesized statistical higher scores on the three measures if women received maternal care by a community care midwife, since she can provide continuity of care throughout pregnancy and birth [19‐21, 33‐36]. Mann-Whitney U, Kruskal Wallis tests, Student T-Tests or Anova Tests were used appropriately, to assess the variability of both measures scores among women that differed on these characteristics.

Next, Spearman Rank correlations were performed to examine convergent validity between the MADM, MORi and CEQ2.0. We expected moderate to strong correlations between the CEQ2.0 and MADM and MORi scores, because respectful maternity care and women’s autonomy are key components of high-quality care during childbirth [37]. We used standard interpretations of the correlation coefficients: 0.40-0.59 moderate; 0.60-0.79 strong and 0.80-1.0 as very strong [38].

As this is the first study requiring women to complete the measures MADM and MORi separately for each health care provider they encountered, the acceptability and clarity of these measures were also assessed. Acceptability refers to the question whether or not women were willing to complete the MADM and MORi multiple times [39]. In our study women indicated how many healthcare providers attended at their births and we observed whether for all of these providers the MADM and MORi were completed. With descriptive statistics the clarity on items on the MADM and MORi as well a general remarks were analysed. All statistical analyses were performed with SPSS Statistics 25.0 (SPSS Inc., Chicago, IL). The level of significance was set at p ≤ 0.05.

Regarding the MADM, the Cronbach’s alpha ranged from 0.96-0.97, whereas for the MORi the range was 0.77-0.84. The CEQ2.0 showed a Cronbach’s alpha of 0.92, and the range for its four subscales was 0.78-0.85 (Table 3).

Our analysis (Table 4) showed no statistical differences on MADM, MORi and CEQ2.0 scores between subgroups who differed on demographic characteristics, except for women who differed on monthly income and maternal age (Table 4). Women who differed on maternal age, showed that women in the age group ≥36 years had the highest CEQ2.0-scores compared with the other age groups. Women with a lower income showed statistically significantly lower scores on the MORi.

Compared with women who experienced a healthy pregnancy and had a physiological birth, women who had pregnancy complications, or who experienced birth interventions, had statistically significantly lower scores on the MADM, MORi and CEQ2.0. No statistical significant differences on MADM, MORi and CEQ2.0 scores were observed for women who differed in vaginal laceration repair either with or without surgery (Tables 5, 6, 7). We did not calculate any statistical differences for adverse neonatal outcomes, since none of the women reported newborn/child mortality, and the prevalence of resuscitation was rather low (4.2%).

Women, who had two or three maternal health care providers attending at birth, gave community care midwives statistical significant higher scores on the MADM and MORi compared with a clinical midwife or obstetrician (Supplementary Table). This pattern of decreased MADM and MORi scores was observed for each pregnancy and birth characteristic. For example, major differences in MADM median scores were observed when care was provided by three maternal healthcare providers of which none was familiar with the women, community midwives scored 40 (IQR 33-42), hospital-based midwife 35 (IQR 30-40), and obstetrician 14 (IQR 25-39), respectively (Tables 5, 6 and 7).

The calculated Spearman Rank correlations between the separate measures of the MADM, MORi and CEQ2.0 ranged, as hypothesized, from 0.49-0.63 (Table 8).

Regarding the acceptability of completing the MADM and MORi for each healthcare provider that attended during the intrapartum period 639 (97.6%) women out of the eligible population of 655, who completed the online survey, had filled in the measures multiple times (Fig. 1). Regarding the clarity of the seven items of the MADM, 7 women (10.8%) indicated problems for completing item 1 “My midwife/obstetrician asked me how involved in decision making I wanted to be” and none had problems understanding item 4 “My midwife/obstetrician helped me understand all the information”. Regarding the clarity of the fourteen items of the MORi, women indicated the most problems with completing item 5 “I chose the care options that I received” (2.7%), item 2 “Comfortable declining care that was offered”(2.4%), and item 4 “Coerced into accepting the options my (midwife, doctor) suggested” (2.1%, Table 9).

Additionally, women provided some general remarks on the overall completion of the MADM and MORi, they questioned if all items were applicable to complete since they had experienced a precipitous labour or an elective caesarean section.

The first strength is that in our sample, women completed the measures MADM and MORi for each healthcare professional that attended at their births. Another major strength of our study was the detailed information that we collected regarding the course of their childbirth, including several birth interventions. These data were of great value for evaluating the known group validity of the MADM, MORi and CEQ2.0-. The second strength of the study was that of all eligible women who completed the online survey. Only 2% gave inconsistent answers in the questionnaire and where therefore excluded in the analyses. The third strength of our study was the inclusion of a substantial part (41.2%) of women living in the North of the Netherlands. Compared to other Dutch regions, those women are significantly more likely to be classified in the lowest socio-economic status quartile [41].This is reflected in our included population since 7.6% had obtained a low education level and 17.6% had a monthly income less than €2000. Our study showed that women who differ in socio-economic status are able to complete the three measures. To minimize recall bias, only women who gave birth < 1 year prior were included in the study. It is difficult to determine what time frame between childbirth and conducting the study is best, as there no clear evidence supporting these decisions [42]. Literature does show that studies carried out soon after childbirth has a higher level of positive experiences compared to studies conducted later, as women need time to reflect [43]. However, studies conducted later have a higher chance on recall bias. The current study takes these aspects into account by the < 1 year time range, as it includes both women who recently gave birth and women who had given birth 11-12 months ago.

We are aware that we did not reach women who had a low literacy, and those who did not read Dutch. In total 6.4% of the women had a minority ethnic background, which is an underrepresentation compared to the complete Dutch population. Our study showed that women with a migrant background had lower scores on the MADM, MORi and CEQ, however due to the small sample size of this subgroup, no statistical differences could be captured. Previous studies reported that women who lived in poverty or had a migration background had statistically significantly lower scores on the MADM and MORi [19, 20].

Of our sample, 54.9% started their births in community midwifery care and 82.0% delivered vaginally. Compared with the Dutch population of pregnant women of singletons the corresponding population numbers in 2019 were 50.7, and 74.5%, respectively [44]. Women who give birth vaginally often rate a better birth experience than women who received care had an instrumental birth or caesarean section [45]. Therefore, it is possible there is an overrepresentation of higher scores in this study.

It is likely that the included women were self-selected by choosing to participate in the online voluntary survey. This is a known limitation and can cause selection bias [35]. Women with a negative birth experience may be more prone to take part in a study covering their birth experience than others, which could lead to an overrepresentation of lower scores in the questionnaire.

Our results regarding the acceptability and clarity of the MADM and MORi indicated a promising uptake and completion of those measures. However, women provided feedback to enhance clarity even more by adding examples to indicate what their involvement could be in e.g. decision making regarding childbirth interventions (e.g. pain relief treatment). Therefore, our first recommendation is to explore with women by using qualitative research if future (small) adaptions are preferred to the original items. A second recommendation would be to evaluate the clarity of the separate items of the CEQ2.0, by either qualitative or quantitative methods. A recommendation for clinical practice would be to use the MADM, MORi and CEQ2.0 as an exit survey or in facility quality improvement programs. These measurement scores could optimize future maternal healthcare since it will help healthcare professionals to be more reflective about their care provision including their communication skills. For maternal health care providers, the results can help to acknowledge women’s autonomy during labour and birth, especially during situations with referrals or complications and when continuity of care is under threat. Women’s right to decision making is an essential aspect of respectful maternity care provision, and should be secured as best as possible in any circumstance in order to achieve a positive birth experience.