Background

Measuring childbirth experiences is an important tool to assist clinicians and researchers in quantifying women’s experiences relating to labour and birth to evaluate practice. With the current pandemic and government restrictions on non-accompanied labour [1], it is important to measure the impact of these factors on outcomes of women’s labour experiences. The lack of a robust, validated tool for evaluating labour experiences in Hong Kong during the pandemic is a topical issue. Even though various tools measuring women’s perinatal experiences exist [2], in the literature, most scales lack of complete testing of the psychometric properties [2].

In 2010, a multidimensional Childbirth Experience Questionnaire (CEQ) [3] was developed in Sweden, measuring four domains of the childbirth experiences. The four domains include: Own capacity, Perceived safety, Professional support, and Participation. The instrument was robust and reliable and has been translated and validated in several languages [4,5,6,7] and used in studies of culturally diverse samples [8,9,10]. However, CEQ was recently revised in Sweden by the same team of developers, as it was found that two domains, Participation and Professional support, showed weaker performance. A total of 14 new items were revised. In Professional support, reversed items were developed to avoid high ceiling effects. In Participation, more relevant items relating to information and decision-making were added [11]. The Childbirth Experience Questionnaire (CEQ2.0) was then developed. It was validated and showed good criterion validity in relation to the nationally used Maternity Survey, test-retest reliability, and differences between known groups, in both primiparous and multiparous women [11].

Measuring the impact of an intervention or a policy on women’s childbirth experiences is as important as measuring its impact on outcomes such as mode of birth, perinatal outcomes, maternal postpartum health, which is scarce in the literature. As aforementioned, there was a lack of a reliable instrument that measures the birthing experiences in Chinese. A recent study has validated CEQ in the Chinese population [6]. As CEQ was modified and CEQ2.0 was developed, the objectives of this study are to validate CEQ2.0 in the Chinese population and to assess its effectiveness in evaluating labour experiences.

Methods

Design, setting, and participants

A cross-sectional survey was conducted in Hong Kong from July 2020 to February 2021. The details of the study were published elsewhere [12]. In brief, participants were recruited by (1) a research assistant who distributed leaflets to women in the obstetrics clinic, (2) an independent researcher in five maternal and child health centers in three regions in Hong Kong, and (3) online promotions on social media and mothers’ groups. The participants completed the survey online or were interviewed by the independent researcher in the maternal and child health centers. The survey, available in English and Traditional Chinese, includes sociodemographic characteristics, obstetrics, maternal health histories, infant feeding practices, the fear level, depressive symptoms, and the childbirth experiences. Eligible participants included women (1) aged 18 or above who (2) were pregnant or gave birth since 2020. As the objective of this study was to validate the Chinese version of CEQ2.0, only women who gave birth and completed the traditional Chinese CEQ2.0 were included in the analysis.

Childbirth Experience Questionnaire (CEQ2.0)

The childbirth experiences of the women were measured using CEQ2.0. We obtained permission to use the questionnaire from Dr. A. Dencker, who was the developer of the original Swedish version of CEQ2.0 [11]. In this study CEQ2.0 was first translated from English to traditional Chinese using a forward-backward translation procedure. The English CEQ2.0 was forward translated into Chinese by two bilingual translators who are native Chinese speakers and health professionals. Then, both translators compared their translations in a consensus meeting with a senior team member and discussed any discrepancies. Two other native Chinese-speaking bilingual translators who were completely blinded to CEQ2.0 translated the Chinese version back into English and this translation was compared with the English version by a native English speaker (to identify discrepancies in the Chinese translation). The final Chinese version of CEQ2.0 was then prepared.

CEQ2.0 comprises 22 statements assessing four domains of the childbirth experiences. These include (i) Own capacity, (ii) Perceived safety, (iii) Professional support, and (iv) Participation. Responses are scored using a 4-point Likert scale ranging from 4 (totally agree), 3 (mostly agree), 2 (mostly disagree), to 1 (totally disagree). Three items referring to labour pain, sense of security, and control are assessed with visual analogue scales (VAS). The VAS-scale scores are transformed to categorical values, 0-40=1, 41-60=2, 61-80=3, and 81-100=4. Item rating was summed and a higher score indicates a better childbirth experience.

Sample size

The planned sample size was 220 women. This was based on the recommendation of a sample size of ten times the number of observed variables in the health measurement tool being evaluated [13]. CEQ2.0 comprises 22 items; therefore 220 completed questionnaires would be required. Assuming a 70% response rate from the original Swedish study and a 20% attrition rate, approximately 400 women are needed to achieve a final sample size. However, as this study is not the primary study outcome of the survey, we calculated the sample size of the survey based on the main study outcome. Therefore, we obtained a larger sample size.

Statistics and data analysis

Baseline descriptive statistics were reported in all participants who completed a baseline CEQ2.0, using frequencies with proportions or mean values with standard deviation (SD). In addition, the mean and SD of each item of CEQ2.0 were presented. To validate CEQ2.0 in the Chinese population, the factor structure, internal validity, known-group validity, and reliability of the scale were evaluated.

Factor structure

First of all, data were fitted into the model proposed by the original authors using the confirmatory factor analysis (CFA) [11]. However, the model did not fit well with the data. Therefore, the data were randomly split into a training set (n=487) and a validation set (n=488).

The training set was used in the exploratory factor analysis (EFA) using the maximum likelihood factoring with promax rotation of factors. The number of factors was identified by using scree plot. The Bartlett’s test of sphericity was used to test the assumption of sphericity, and the Kaiser-Meyer-Olkin (KMO) test was used to test the sampling adequacy. Items with a factor loading ≥0.4 were assigned to the factor [14], and a new factor structure (CEQ2.0-R) was developed.

Using the validation set, we conducted a confirmatory factor analysis (CFA) to assess the factor structure suggested in the EFA. Factor structure was tested in the models using model fit statistics; χ2, the comparative fit index (CFI) [15], Tucker–Lewis index (TLI), standardized root mean square residual (SRMR), and root mean square error of approximation (RMSEA). A CFA model was considered as a relatively good fit if the CFI and TLI values are close to 0.95, the SRMR value is close to 0.08, and the RMSEA value is close to 0.06 [16]. Hu and Bentler suggested using the word “close to” because the fit indices did not work well equally under different conditions [16].

Internal construct validity

Item-total correlation was used to assess the internal construct validity in the whole sample. A correlation coefficient should be ≥0.2 to be remained in the scale [17].

Internal consistency

To assess internal consistency, Cronbach’s alpha coefficient (α) was calculated for each of the four factors in the whole sample, using the same item loadings as the CFA for CEQ2.0, α ≥0.70 to 0.90, was considered to have good internal consistency [18].

Known-group validity

We used a known-group validity assessment to test the ability of CEQ2.0-R to differentiate the groups. Independent t-tests were used to test the differences in the four factors and the total scores between parity, oxytocin augmentation, and mode of birth based on the previous studies [11, 19]. In addition, the use of pain medication and companionship during labour were also examined as studies show that there were differences in childbirth experiences between groups [20, 21]. Cohen’s d effect sizes (d) were calculated and considered as trivial (<0.2), small (≥0.2 and <0.5), moderate (≥0.5 and <0.8) or large (>0.8) [22].

Statistical analyses were conducted using Stata Statistical Software: Release 16 (StataCorp. 2013. College Station, TX: StataCorp LP). A .05 level of significance was used throughout the study.

Results

Overall, 1035 participants who were in the postpartum period completed the Chinese version of the survey. We excluded 60 participants because they had fixed or random response for CEQ2.0. Therefore, 975 participants were included in the analysis. The median time of completing the questionnaire is 4.1 months postpartum. The majority of participants were born in Hong Kong (88.5%), were in employment (66.7%), were married (97.1%), primiparous (71.1%), had a spontaneous vaginal birth (67.9%), and had their birth in public hospitals (74.5%) (Table 1).

Table 1 Descriptive characteristics of the study population
Full size table

Factor analysis

A CFA was conducted on the whole sample, and a poor fit of the data into the model proposed by the original authors [11] (CFI=0.656; TLI=0.608; SRMR=0.102; RMSEA=0.104) (see Table 2). The factor structure of CEQ2.0 is shown in Appendix I. Bartlett’s test of sphericity was significant (P<.001), and the KMO measure of sampling adequacy was 0.851, which supports the use of factor analysis for this data. Therefore, the data were randomly split into a training set for conducting the EFA (n=487) and a validation set for conducting the CFA (n=488).

Table 2 Goodness of fit of the indicators in the original CEQ2 factor model and CEQ2-R factor model
Full size table

The scree plot suggested a four-factor structure, and the corresponding rotated factor loadings showed the factors were “Own capacity”, “General support”, “Perceived safety” and “Professional support” (Table 3) . The four factors accounted for 59.8%, 22.7%, 9.64%, and 7.8% of the variances, respectively. Nine items (item 3, 5, 8, 9, 14, 15, 17, 20 and 22) had factor loadings <0.4 and were removed. The resulting 13 items formed CEQ2.0-R. The factors consist of 2 to 5 items, i.e. “Own capacity” (6 items), “General support” (3 items), “Perceived safety” (2 items), and “Professional support” (2 items). The factor structure of CEQ2.0-R is shown in Fig. 1. CEQ2.0-R improved the fit indices (CFI=0.912; TLI=0.884; SRMR=0.053; RMSEA=0.072) when compared with the original model (Table 2).

Table 3 Exploratory factor analysis (EFA) results
Full size table
Fig. 1
figure 1

Factor structure of CEQ2.0-R

Full size image

Internal construct validity and reliability

We used the item-total correlation to assess the internal construct validity. Satisfactory item-total correlation was shown in all the items (see Table 4). In addition, good reliability was shown in the four domains. The α was ≥0.7 in all the domains.

Table 4 Descriptive statistics, internal construct validity and reliability (n=975)
Full size table

Known-group validity

The domains were compared among primiparous and multiparous participants and participants who received different intrapartum interventions (see Table 5). Primiparous mothers were less likely to receive General support (d=-0.16; P=.02), Perceived safety (d=-0.22; P=.002), Professional support (d=-0.24; P<.001) and the total scale score (d=-0.30; P<.001). Participants who received oxytocin augmentation were more likely to perceive safety and receive professional support. When comparing the participants who had spontaneous vaginal birth and operative birth, statistical differences were found in the domains Own capacity (d=0.19; P<.001) and Professional support (d=-0.16; P=.02). Participants, who had companionship during labour, had higher scores in all the domains and the total scale.

Table 5 Known-group comparison (n=975)
Full size table

Childbirth experiences in Chinese women

The mean of CEQ2.0-R was 2.65 (SD=0.34). Participants had received adequate general support. There were 91.5% of the participants agreed that their impression of the team’s medical skills made them feel secure. However, a high proportion of participants reported that they would prefer the midwives to be more present during labour and birth (80.8%) and having more encouragement from the midwives (87.6%).

Discussion

This is the first study to examine the psychometric properties of the Chinese translated version of CEQ2.0. This study provides evidence for the reliability and validity of the translated CEQ2.0. The original model of CEQ2.0 showed a poor fit [11]. Therefore, a revised model CEQ2.0-R was constructed, which showed adequate validity and reliability. CEQ2.0-R is a self-report instrument, and it is easy to use. It took approximately five minutes to administer and can be used to evaluate childbirth experiences.

CEQ2.0-R measured the childbirth experiences in four domains (Own capacity, General support, Perceived safety and Professional support). Nine items were deleted from the original CEQ2.0. Not many variations were shown in the deleted items. Furthermore, three items were removed from the domain Perceived safety in CEQ2.0, and one item was shifted to another domain. The infant mortality rate was 1.4 per 1000 live birth, and the maternal mortality ratio was 0 in 2019 [23]. Therefore in Hong Kong, the infant mortality rate has been among the lowest in the world [24]. With this, participants in this population group may not perceive that safety or security in labour was a concern in Hong Kong. In addition, the domain Participation from the original model was removed, and one new domain General support was constructed in CEQ2.0-R. This domain measured the support from the general experiences during labour. It consists of the impression related to medical team skills, the support provided by the staff, and how the participants were treated in the hospital.

To our knowledge, CEQ2.0 was only validated in two Western countries, Sweden [11] and the United Kingdom [19]. Although the participants indicated that they had positive birthing experiences, the mean score was relatively low. The experiences were merely positive. Studies show that negative childbirth experiences are associated with adverse postnatal outcomes, such as postnatal depression [25] and post-traumatic stress symptoms [26]. Approximately 17% of postpartum women suffered from postpartum depression globally [27]. Therefore, it is important to understand how childbirth experiences would affect postpartum mental health outcomes. In addition, many participants indicated that they would like to receive more support from midwives. One study conducted in Hong Kong shows that women preferred nurses to meet their informational and individual needs, perform competent nursing skills, be approachable and have positive attitudes, and demonstrate cultural competence [28]. Effective interventions on improving childbirth experiences need to be developed.

Oxytocin augmentation and parity had small effects on the total childbirth experiences. However, having companions during labour can significantly improve childbirth experiences. A qualitative evidence synthesis shows that companionship is associated with positive childbirth experiences [29]. During the COVID-19 pandemic, childbirth companionship was suspended in public hospitals in Hong Kong [1]. A recent study conducted in Hong Kong shows that after the suspension of childbirth companionship, the proportion of women who received childbirth massages decreased while there was an increased prevalence in opioid pain medication use during the postpartum period [1]. In addition, a higher proportion of women developed depressive symptoms in the early postpartum period after the announcement of the coronavirus alert [1]. However, it is unclear whether the increase in the depressive symptoms was related to the change in hospital practices or the COVID-19 pandemic. Therefore, further studies are needed with the validated CEQ2.0-R is needed to deepen our understanding of the experience during the COVID-19 pandemic. This can also differentiate the effect of different government restrictions worldwide due to the COVID-19 pandemic on maternal health outcomes.

This study has several strengths and limitations. This study was the first study to assess the ability of CEQ2.0 in examining the childbirth experiences of Chinese women. The translated and validated CEQ2.0-R may be useful for measuring the childbirth experiences of Chinese women. Secondly, this study has a large sample size, which is sufficient for performing CFA [30]. However, there are also some limitations of this study. All questions were self-reported, instead of retrieved from the medical records. Data was collected based on participants’ recall of the use of intrapartum interventions, therefore recall bias may occur. Secondly, the psychometric properties of the Chinese version of CEQ2.0 are based on a convenience sample of Hong Kong mothers. The mothers who have positive childbirth experiences may be more in favor of participating in the study. In addition, our sample has a higher educational level and monthly family income when compared to the general population [31]. Therefore, the study results may not be generalizable to other populations within Hong Kong.

Conclusions

CEQ2.0-R shows adequate psychometric performance in assessing women’s childbirth experiences. The questionnaire is easy to be administered and can be used to assess several domains of the childbirth experiences. This questionnaire may be useful for midwives to assess women's childbirth experiences in the postpartum period and evaluate the aspects of additional support needed during childbirth when providing health care to the general population.

Availability of data and materials

The data analyzed in this study are not publicly available due to privacy policy, but are available from the corresponding author on reasonable request. For further information about data access, please contact the corresponding author, Kris Lok by email.