Psychometric proprieties of the Arabic Cambridge worry scale among pregnant women

Dur ing pregnancy, many women experience multiple psychological changes, and may worry about various elements of pregnancy and birth (Osborne et al, 2021). Multiple factors may increase worry among pregnant women, such as a confirmed diagnosis of fetal malformation during the antenatal period or the risk of miscarriage (Thorsness et al, 2018). Pregnant women may also worry about their ability to mother, employment and career disruptions, having adequate savings or changes to the marital relationship (Mortazavi and Akaberi, 2016).

Worry adversely affects maternal and infant health (Gourounti et al, 2012). As a result of worry, pregnant women may experience cardiac palpitations, muscle weakness, insomnia, stomach pain and headaches (Costa et al, 2020). Enhanced worry can also decrease maternal blood flow, contributing to complicated perinatal outcomes, including premature birth (DeSocio, 2018). Given the importance and significance of maternal wellbeing, assessing worry during pregnancy is crucial.

Numerous tools have been used to examine perinatal stress, anxiety and worry, such as the Spielberger state-trait-anxiety inventory, which is used to measure anxiety caused by maternity-related conditions, including miscarriages and parenthood (Mortazavi and Akaberi, 2016). The inventory has been widely administered to measure cognitive, physical, affective and behavioural components of anxiety in pregnant women (Laux et al, 1981). Although it is valid and readily completed, it only measures the severity of general anxiety at a certain time point and provides no information on what causes women to be anxious over the course of their pregnancy (Öhman et al, 2003). Moreover, Petersen et al (2009) indicated that its reliability around the time of birth was questionable.

The Cambridge worry scale was developed by developed by Green et al (2003) to measure both the level and nature of women's worries during pregnancy. In a sample of 1072 childbearing women in the UK, the scale was used to measure pregnancy-related worries, including fetal health and death, childbearing, financial and social challenges and parenthood. The English version of the scale was found to be valid and reliable, and demonstrated satisfactory feasibility in early pregnancy (Green et al, 2003).

The Cambridge worry scale has been translated into several languages including German (Petersen et al, 2009), Swedish (Öhman et al, 2003), Spanish (Carmona Monge et al, 2012), Greek (Gourounti et al, 2012), Turkish (Gunay and Gul, 2015), and Farisi (Mortazavi and Akaberi, 2016). Translated versions of the scale have demonstrated satisfactory reliability and content validity. However, an Arabic version is not available, and thus its reliability and validity have not been established among Arabic-speaking women. This psychometric study aimed to validate the Arabic version of the Cambridge worry scale with a sample of pregnant women in Jordan.

Methods

To facilitate tool translation and validation, a cross-sectional design was used for this psychometric study. Recruitment took place in five maternal and child health centres in Irbid city in north Jordan. Using convenience sampling, women in their third trimester, expecting their first or subsequent baby, between 18 and 45 years of age, who were able to speak and read Arabic were recruited. Mothers with current or previous psychological illnesses were excluded from the study. Women who met the inclusion criteria were identified by staff and invited to speak with a member of the research team.

Sample size

The sample size was based on the subject to item ratio for exploratory factor analysis (Costello and Osborne, 2005). An acceptable ratio is 10:1 participants per item, while the best ratio is 20:1 (Costello and Osborne, 2005). As the scale has 17 items, a sample of over 340 women was considered acceptable (Costello and Osborne, 2005).

Data collection

The survey included a 6-item sociodemographic form to collect data on participants' age, education, total monthly income and parity. Participants also reported if their last pregnancy was planned or not and if they had financial difficulties.

The Cambridge worry scale consists of 17 items that aim to measure the severity of common worries during pregnancy, with four subscales: socioeconomic, sociomedical, own health, and relational (Green et al, 2003). It can be administered during pregnancy and postpartum, with women and men, and can be modified to more accurately measure worry in a specific population (Mohammad et al, 2011; Öhman et al, 2003). During pregnancy, Cronbach's alpha for the scale has been found to be 0.78 and it is better at discriminating between women with different reproductive histories than other anxiety scales (Green et al, 2003).

The scale was translated from English to Arabic by two professional translators (holding a Master's degree in translation) along with two bilingual, PhD-qualified academics (one in nursing and one in midwifery). The scale was then back-translated from Arabic to English by the two bilingual, PhD-qualified academics. The two versions were matched by the translators and academics and necessary changes were undertaken to ensure that terms were accurate and meanings were clear correct to produce a common Arabic version.

Participants were asked to respond to items on a 6-point Likert scale (0=‘not a worry’ to 5=‘a major worry’). Ratings for each item were summed to give a total score ranging from 0–85, with higher scores reflecting more worry.

Content validity

The expert panel, which consisted of three PhD-qualified academics (one in nursing and two in midwifery), a psychologist and a gynaecologist, evaluated each item and discussed their relevance to Jordanian culture. The expert panel recommended three additional items, which were used in a previous validation study of the scale in Iran (which has a similar culture to Jordan) (Mortazavi and Akaberi, 2016). These items were the baby's sex, provision of quality labour care and whether the pregnancy was unwanted or unplanned. The panel recommended the item ‘whether your partner will be with you for the birth’ be changed to ‘whether your husband will be with you at the time of admission to labour’. This is because men are not allowed in the labour room in government-funded hospitals in Jordan.

A content validity ratio was used to assess validity. According to the Lawshe (1975) method, the panel of experts assessed the necessity of the items on a 3-point rating scale: (1) not necessary, (2) useful, but not essential, and (3) essential. The ratio for each item was calculated, and no item had a ratio of less than 0.62 (an acceptable value) (Lawshe, 1975). After that, a content validity index was calculated to show the relevance, clarity and simplicity of each item. The panel assessed the simplicity, comprehensibility, relevance and clarity of each item on a 4-point Likert scale (1=not relevant, not simple and not clear, 4=very relevant, very simple and very clear). The content validity index was calculated by dividing the total number of items by the number of experts who rated a particular item 3 or 4. No item had an index of less than 0.8 (an acceptable value) (Polit and Beck, 2011).

The draft 20-item Arabic version of the scale was piloted with 20 pregnant women (all excluded from the main study) to assess simplicity, clarity and time taken to complete the scale, as well as the relevance of the items to pregnant women's concerns. The importance of each item was rated by participants on a 5-point Likert scale (1=not important, 5=very important). The score for each item was calculated (Juniper et al, 1997) and found to be ≥1.5. Participants reported that the scale was easy to complete and understand and no changes were required.

Data collection procedure

An interview to complete the survey was conducted in private either before or after the woman's scheduled antenatal appointment. This enabled women to clarify the meaning of any question if needed. The interview lasted approximately 25 minutes and consisted of the demographic data form and Cambridge worry scale. Data collection took place between September and December 2020.

Data analysis

Data analysis was undertaken using the Statistical Package for Social Sciences (version 23) for all analysis except confirmatory factor analysis, which was conducted using LISREL 8.80 for Windows. The suitability of the data for factor analysis was investigated using the Kaiser-Meyer-Olkin sample adequacy measure (Kaiser, 1974) and Bartlett's (1954) test of sphericity with P<0.001. Principle component analysis with varimax rotation was used for exploratory factor analysis. Factor loadings ≥3 were considered appropriate (Tsang et al, 2017).

Confirmatory factor analysis was used to examine the fitness of the model. Chi-square/df, comparative fit index, incremental fit index, normed fit index, non-normed fit index, root mean square error of approximation and standardised root mean square residual were used as fitness indices. To identify the reliability of the scale, Cronbach's alpha and inter-class correlation coefficients were calculated for each factor and for the total scale. A coefficient ≥0.6 was considered acceptable (Houser, 2008); and the acceptable level for Cronbach's alpha coefficients was set at 0.7 (Terwee et al, 2010).

Ethical considerations

Ethical approval was obtained from the institutional review board of the Jordan University of Science and Technology and the Jordan Ministry of Health (reference number: 238/2019). Staff working in participating centres were briefed about the study. A research assistant provided written and verbal information about the study and obtained written consent. Women were advised about confidentiality, anonymity and their right to not provide a response and withdraw at any time without explanation.

Results

A convenience sample of 369 women participated in the study, with 185 involved in exploratory factor analysis and 184 in confirmatory factor analysis. Participants' characteristics are shown in Table 1. Around half (48.2%) were aged 20–29 years old. Gestation ranged from 29–38 weeks (mean=34.66 weeks, standard deviation=1.91). The majority of participants (65.6%) were housewives and had a total monthly income of 301–500 Jordanian Dollars (58.9%).

Exploratory factor analysis

Tests of assumptions revealed good sampling adequacy and suitability for component factor analysis (Kaiser-Meyer-Olkin=0.86; Bartlett's test of sphericity=5822.87, P<0.001). The initial analysis indicated a four-factor structure for the scale. A final 15-item scale was loaded on four distinct constructs that jointly accounted for 83.7% of the variance observed (Table 2). The highest explained variance was factor 1 (45.0%), while the lowest was factor 4 (7.5%). The factors were labelled sociomedical, socioeconomic, health of the baby and health of mother/others and relationships. The rotated matrix showed that item loadings ranged between 0.55–0.97. Items 10, 11, 12, 13 and 19 were grouped together to represent factor one (sociomedical); factor two (socioeconomic) included four items (1, 2, 18 and 20), factor three (health of the baby) consisted of three items (9, 16 and 17) and factor four (health of mother/others and relationships) consisted of three items (4, 6 and 7).

Five items (giving up work, problems with the law, employment problems, whether your husband will be with you at the time of admission to labour and relationship with family and friends) did not load on any factor and were removed.

Confirmatory factor analysis

The 15-item scale was used for confirmatory factor analysis to test the model fit. All comparative indices of the model, including the comparative fit index, incremental fit index, and normed and non-normed fit index, were more than 0.9 (0.91, 0.92, 0.97 and 0.94 respectively) showing the goodness of fit for the data. The root mean square error of approximation of the model was 0.073 with a lower bound of 0.059. The standardised root mean square residual was <0.08, confirming an adequate fit for the model. The final model, shown in Figure 1, outlines the results of the analysis.

Reliability assessment

Cronbach's alpha for the total scale was 0.92, ranging from 0.86–0.97 for the four subscales and the suitability of scale, as examined by inter-class correlation coefficients, varied from 0.71–0.83. The results are shown in Table 3.

Women's responses to scale items

Participants' most important worry was about giving birth (mean score=3.29, standard deviation=1.57), followed by whether they would be provided with good care in labour (mean score=3.23, standard deviation=1.87). They worried least about the health of someone close to them (mean score=2.41, standard deviation=1.90) followed by housing (mean score=2.43, standard deviation=1.95). Table 4 presents the mean scores of the scale items.

Discussion

The present study showed the adequate psychometric proprieties of the Arabic version of the Cambridge worry scale. The scale assessed possible worries that women may experience during pregnancy, either specifically related to their pregnancy or to their daily life. Five items were removed and the 15-item Arabic scale was found to be reliable and valid.

The internal consistency of the scale (0.91) and its subscales was satisfactory and somewhat consistent with previous studies (Petersen et al, 2009; Carmona Monge et al, 2012; Gourounti et al, 2014; Mortazavi and Akaberi, 2016). The four factors were sociomedical, socioeconomic, the health of the baby and the health of the mother/others and relationships.

The structure obtained from the current study differed slightly from the original structure, which included sociomedical, socioeconomic, health and relationship factors (Green et al, 2003). This is consistent with the structure reported in other studies, which also differed from the original. Petersen et al (2009) reported sociomedical, socioeconomic and relationships, health of the baby and health of the mother and others. Mortazavi and Akaberi (2016) reported sociomedical, health of the mother/others and relationships, health of the baby and socioeconomic factors. Jomeen and Martin (2005) reported a structure similar to Green et al (2003), but with two items related to more than one latent variable. Three other studies also reported different factor orders (Peñacoba-Puente et al, 2011; Carmona Monge et al, 2012; Gourounti et al, 2014). Gunay and Gul (2015) found the Turkish form of the scale to be suitable in terms of content validity and language, but no factor structure was confirmed. The present study found a structure similar to an Iranian study but with different factor orders (Mortazavi and Akaberi, 2016).

Across studies, the sociomedical factor has been found to explain a large percentage of variance (27–30%). In the present study, this factor explained about 45% of the variance. The most prevalent worries for the participants were giving birth, the provision of good care, possible adverse outcomes for the baby and financial problems. Similar concerns have been reported by others (Green et al, 2003; Petersen et al, 2009; Gourount et al, 2012; Gunay and Gul, 2015; Mortazavi and Akaberi, 2016). This suggests similar worries but different priorities among pregnant women from different countries.

The focus on birth, quality of care and the baby's wellbeing by participants in the present study may relate to a risk-adverse culture and the medically-dominated model of care in Jordan. Intrapartum care in Jordan is characterised by high rates of obstetric intervention including induction, augmentation, numerous vaginal examinations, episiotomy, continuous external fetal monitoring and use of the lithotomy position (Shaban et al, 2011; Mohammad et al, 2014a; 2014b). The focus on technology and intervention is to the detriment of respect for maternal autonomy, one-to-one supportive care and the wellbeing of the mother and her family (Hatamleh et al, 2013; Mohammad et al, 2014a; 2014b). One study reported that 32% of Jordanian women reported feeling neglected and 38% were verbally abused during childbirth, with most labouring women not receiving any information about childbirth or their healthcare rights (Alzyoud et al, 2018). Dissatisfaction with intrapartum care among Jordanian women increases their risk of psychological problems, such as postnatal depression and anxiety (Mohammad et al, 2011; 2019). It is therefore not surprising that Jordanian women's major worries were about giving birth and the provision of good care.

Women's worries regarding the baby's health may relate to high levels of prenatal screening and intrapartum monitoring in Jordan and is consistent with the literature (Petersen et al, 2009; Carmona Monge et al, 2012; Gourount et al, 2012). Rather than being reassured by screening and monitoring, these interventions (in the absence of information and a supportive care relationship) may serve to increase women's concerns about the baby.

Financial concerns were ranked fourth by participants, which is consistent with a study in Greece where they were ranked third (Gourount et al, 2012). Financial concerns were not found to rank highly for women in European countries such as Spain and Germany (Petersen et al, 2009; Carmona Monge et al, 2012). However, the majority of pregnant women in the present study reported a low monthly household income, with nearly a third being unemployed, which may have affected their priorities. While it is often assumed that women will focus on worries about pregnancy, birth and the baby, social factors can also impact women's wellbeing and birth outcomes.

Participants reported that they were worried about the gender of the baby, and there is a high demand for prenatal screening of fetal gender in Jordan (Mohammad et al, 2019). Culturally-based gender preferences for a male child in Arab countries continue to have a significant impact on maternal wellbeing (Mohammad et al, 2011; 2019). Male children are perceived to be important in Arab countries because sons are expected to financially support their extended family, continue the family legacy and improve the social status of their mothers in their families (Shrestha et al, 2014; Mohammad et al, 2018; 2019). However, in an Iranian study, a baby's gender was found to be a low source of worry (Mortazavi and Akaberi, 2016). This may be because data collection took place later in this study, when baby gender may already have been known. Midwives need to be positive if a woman expresses concern about the baby's gender prior to childbirth and encourage parental acceptance of the baby regardless of gender.

Five items were removed from the Arabic Cambridge worry scale. Unlike previous studies (Green et al, 2003; Petersen et al, 2009; Carmona Monge et al, 2012; Mortazavi and Akaberi, 2016), item 3 (problems with the law) did not load on any factor. This may reflect cultural sensibilities, as it is perceived to be unacceptable for Jordanian women to have problems with the law, and it would be culturally unacceptable to acknowledge such concerns. Items 8 (employment problems) and 14 (giving up work) did not load either, despite participants expressing financial concerns. As most pregnant women in the present study were unemployed at the time of data collection, these items may have seemed irrelevant to them. These findings concur with other studies which suggested excluding the item ‘giving up work’ (Green et al, 2003; Petersen et al, 2009; Carmona Monge et al, 2012; Mortazavi and Akaberi, 2016).

The revised item ‘whether your husband will be with you at the time of admission to labour’ also did not load on any factor, indicating its lack of importance to pregnant Jordanian women. In Jordan, it is common for husbands to take their wives to hospital for labour and birth, but not stay with them. Item 5 (relationships with family and friends) was removed from the Arabic Cambridge worry scale, because of its low communality in the exploratory factor analysis.

Strengths and limitations

The large sample size enabled assessment of construct validity and the systematic approach used enhanced rigour. However, there are limitations associated with cross-sectional designs and a single point of data collection (in the third trimester). The psychometric proprieties of the Arabic Cambridge worry scale are based on this study's sample of pregnant women with a mean gestational age of 34.6 weeks. During pregnancy, the extent and nature of worries may differ, with women experiencing more worries in early pregnancy, fewer in mid-pregnancy and increasing concerns in late pregnancy (Öhman et al, 2003; Petersen et al, 2009). Assessing the psychometric proprieties of the scale with women in their first or second trimester may produce different results. Enabling women to complete the form anonymously, without assistance, may produce different findings on items with cultural sensitivities.

Implications

The Arabic 15-item version of the Cambridge worry scale is a short, flexible and simple scale for measuring women's worries in the third trimester of pregnancy in Jordan, and has applicability to other Arab-speaking countries. Understanding women's concerns will enable healthcare professionals to assist, support and better meet pregnancy-related and psychosocial needs of pregnant women.

Conclusions

The present study confirmed the content and construct validity and reliability of the Arabic Cambridge worry scale. The four-factor structure produced was similar to findings of earlier studies. Pregnant women's major concerns related to giving birth, the provision of good care in labour, the possibility of adverse neonatal outcomes and financial problems, which reflected the sociocultural context. This highlights the ongoing need to translate and validate measures in different contexts.

Key points

• Translated versions of the Cambridge worry scale have demonstrated satisfactory reliability and content validity in multiple settings.
• An Arabic version of the scale is not available, and thus its reliability and validity have not been established among Arabic-speaking women.
• The present study confirmed the content and construct validity and reliability of the Arabic Cambridge worry scale.
• Understanding women's concerns will enable healthcare professionals to support and better meet the pregnancy-related and psychosocial needs of pregnant women.