Background
Unintended pregnancies are a significant global public health issue (Bellizzi et al., Reference Bellizzi, Mannava, Nagai and Sobel2020; United Nations, 2019). It is known that the main reason behind unintended pregnancies is inconsistent or non-use of modern contraceptive methods (Finner & Zolna, Reference Finer and Zolna2016). Several factors influence the use of contraceptive methods. Some of these are economic factors associated with obtaining contraceptive methods, cultural and social factors that conflict with method use, and widespread negative beliefs. Additionally, lack of knowledge about contraception, including how to use contraceptive methods effectively and where to access them, negatively impacts method use (D’Souza et al., Reference D’Souza, Bailey, Stephenson and Oliver2022; Shah et al., Reference Shah, Lee and Nisa Mir2021; Yeh et al., Reference Yeh, Kautsar, Kennedy and Gaffield2022).
Self-efficacy is important in the process of ending undesirable behaviours or promoting health-promoting behaviours (Bandura, Reference Bandura2005). For this reason, the use of contraceptive methods in a consistent manner requires behavioural changes and high self-efficacy (Lopez et al., Reference Lopez, Grey, Chen, Tolley and Stockton2016; Uysal et al., Reference Uysal, Boyce, Undie, Liambila, Wendoh and Pearson2023). Research has indicated that women with high contraceptive self-efficacy demonstrate greater acceptance and compliance with contraceptive methods (Hamidi et al., Reference Hamidi, Deimling, Lehman, Weisman and Chuang2018; Shah et al., Reference Shah, Lee and Nisa Mir2021). Evaluating contraceptive self-efficacy is crucial, as it is a significant predictor of modern contraceptive method use. There are few validated tools available for a comprehensive assessment of contraceptive self-efficacy (Brafford and Beck, Reference Brafford and Beck1991; Levinson et al., Reference Levinson, Wan and Beamer1998; Richardson et al., Reference Richardson, Allison, Gesink and Berry2016; Bhan et al., Reference Bhan, Thomas, McDougal, Nanda, Mahapatra and Das2024). The validated tools that do exist often focus on condom use or include items adapted from other existing tools (Burke et al., Reference Burke, Ridgeway, Murray, Mickler, Thomas and Williams2022; Hamidi et al., Reference Hamidi, Deimling, Lehman, Weisman and Chuang2018; Nelson et al., Reference Nelson, Borrero, Lehman, Velott and Chuang2017). Furthermore, these assessments often focus on ‘partner communication’, overlooking a more comprehensive perspective. Therefore, it is clear that using a comprehensive measurement tool is essential for assessing contraceptive self-efficacy accurately. Whiting-Collins et al. (Reference Whiting-Collins, Grenier, Winch, Tsui and Donohue2020) developed a measurement tool to evaluate the contraceptive self-efficacy of postpartum women and reported that it is valid and reliable in this population. It is considered that this short and easy-to-understand scale, which includes different dimensions of self-efficacy, can be used to measure the contraceptive self-efficacy of all women of reproductive age. To our knowledge, there is no measurement tool that evaluates the contraceptive self-efficacy of women with different dimensions. Therefore, examining the validity and reliability of this tool in Turkish is a guide for future studies. This study aimed to adapt the Contraceptive Self-Efficacy Among Women in Sub-Saharan Africa (CSESSA) scale into Turkish and to evaluate the validity and reliability.
Methods
This methodological study tested the validity and reliability of the Turkish version of CSESSA Scale. This study adhered to the ‘International Test Commission Guidelines for Translating and Adapting Tests’ (International Test Commission 2017). There are different approaches to determining the sample size of validity and reliability studies (Anthoine et al., Reference Anthoine, Moret, Regnault, Sébille and Hardouin2014). To evaluate the factor structure of the scale, it is recommended to aim for a large number of participants (Boateng et al., Reference Boateng, Neilands, Frongillo, Melgar-Quiñonez and Young2018; International Test Commission, 2017). In determining the sample size, Comrey and Lee (Reference Comrey and Lee2013) provided the following guidance: 100 = poor, 200 = fair, 300 = good, 500 = very good, ≥ 1000 = excellent. Accordingly, this study aimed to reach at least 500 women. This study was conducted with 510 women. Participants had to meet the following inclusion criteria: Reproductive age, having a sexual partner, being literate.
This study was conducted in three stages. (1) Translation and adaptation of the scale, (2) pilot study, (3) evaluation of psychometric properties (Figure 1).
Stages of adaptation of the scale into Turkish.
Phase 1 Translation and adaptation
The first phase of the cross-cultural adaptation process is translation. Experts who are familiar with the terminology of the translated scale and who have experience collecting data on the subject take part in the translation phase (World Health Organization, 2017). Based on this recommendation, the scale was independently translated into Turkish by two experts with PhD degrees in gynaecology and obstetrics nursing. The two translations obtained were synthesized by the researchers, and an independent linguist checked the translations and analysed the sentence structures. In the subsequent stage, back-translation was performed by two independent translators with no knowledge of the original scale. After this stage, feedback was received from Whiting-Collins regarding the back-translation of the scale. It was decided that the translation of the scale items was appropriate and no items were removed. In the next stage, expert opinions regarding the scale items were obtained from 10 experts (gynaecology and obstetrics nursing, public health nursing, measurement, and evaluation experts). Modifications were made according to their feedback and the final form for the pilot study was obtained (Borsa et al., Reference Borsa, Damásio and Bandeira2012; Coster & Mancini, Reference Coster and Mancini2015).
Phase 2 Pilot study
A pilot study was conducted to assess the validity of the scale. The scale was applied to 50 women who were similar to the target population (Borsa et al., Reference Borsa, Damásio and Bandeira2012; World Health Organization, 2017). The results of the pilot study indicated that all scale items were understandable, and no revisions were required.
Phase 3 Evaluation of psychometric properties
Data were collected face-to-face from women admitted to primary health care centres in a province in Türkiye between May and December 2022. The data were collected by using a descriptive information form and the Turkish version of the CSESSA Scale. The baseline characteristics of 510 participants were analysed. Two weeks later, a retest was performed with 30 participants to evaluate test-retest reliability (Alpar, Reference Alpar2014). Confirmatory factor analysis (CFA) was performed to assess the construct validity of the scale, while test-retest reliability, item-total score correlations, and internal consistency analyses were carried out to evaluate its reliability.
Data collection tools
Descriptive information form
The researchers developed the form based on existing literature. It includes 15 items related to the participants’ demographic, obstetric, and contraceptive characteristics (Bellizzi et al., Reference Bellizzi, Mannava, Nagai and Sobel2020; Mutumba et al., Reference Mutumba, Wekesa and Stephenson2018; Upadhyay et al., Reference Upadhyay, Dworkin, Weitz and Foster2014).
Contraceptive self-efficacy scale
The scale, developed in Kenya and Nigeria, measures the contraceptive self-efficacy of postpartum women. It consists of 11 items divided into three subscales: ‘husband/partner communication’, ‘provider communication’, and ‘choosing and managing a method’. Each item is scored on a visual analog scale from 0 (‘cannot do at all’) to 10 (‘highly certain can do’). Although the subscales can be scored independently, the total score is the sum of the subscale scores and ranges from 0 to 110. Higher scores indicate a higher contraceptive self-efficacy level. The original scale has Cronbach’s alpha of 0.90. The subscales have Cronbach’s alpha coefficients of 0.89, 0.89, and 0.88, respectively (Whiting-Collins et al., Reference Whiting-Collins, Grenier, Winch, Tsui and Donohue2020).
Data analysis
The data were analysed using SPSS (v.22) and AMOS (v.21). For categorical data, we calculated the number and percentage. For quantitative data, we calculated the mean and standard deviation. Content and construct validity indices were examined to assess the scale’s validity. To assess content validity, the Davis technique was used, and 10 experts evaluated each item of the scale as ‘very appropriate’ (4 points), ‘appropriate but minor changes are needed’ (3 points), ‘the item needs to be transformed into an appropriate form’ (2 points) and ‘not appropriate’ (1 point). Item-Content Validity Index (I-CVI) and Scale Content Validity Index (S-CVI/Ave) were calculated (Polit and Beck, Reference Polit and Beck2017). The I-CVI was calculated by dividing the sum of the items rated with three and four points by the total number of experts. S-CVI-Ave was calculated by dividing the sum of the I-CVI by the total number of items. The construct validity of the scale was tested by calculating Confirmatory Factor Analysis (CFA) goodness-of-fit indices. The following fit indices were evaluated to assess model fit: χ2/df (Chi-Square/Degree of Freedom), Comparative Fit Index (CFI), Goodness-of-Fit Index (GFI), Adjusted Goodness-of-Fit Index (AGFI), Tucker-Lewis Index (TLI), Normed Fit Index (NFI), and Root Mean Square Error of Approximation (RMSEA). The acceptable fit values for these indices are as follows: χ²/df <5, CFI ≥ 0.90, GFI ≥ 0.90, AGFI ≥ 0.85, TLI ≥ 0.90, NFI ≥ 0.90, and RMSEA ≤ 0.08. To assess reliability, we used Cronbach’s alpha coefficient to evaluate the internal consistency of the scale and its subscales (Polit and Beck, Reference Polit and Beck2017). We also performed an item-total score analysis to examine the relationship between item scores and the total score. To determine time invariance, we carried out a test-retest and calculated the correlation coefficient (Streiner et al., Reference Streiner, Norman and Cairney2015). In addition, Hotelling’s T2 test was used to evaluate response bias and Tukey’s Test of Additivity was used to evaluate additivity. A p-value of less than 0.05 was considered to show a statistically significant result.
Results
The study sample consisted of 510 participants aged between 20 and 53 years (mean=35.25 ± 6.58). All participants were married, and 37.1% held a university or master’s degree. Of all participants, 69.0% were unemployed, and 53.5% had a middle perceived economic level. Additionally, 91.4% of participants had at least one child, and 15.9% had a history of unintended pregnancy. While 70.8% reported not wanting to get pregnant, 32.7% did not use any modern contraceptive methods. Table 1 summarizes the baseline characteristics of the participants.
Characteristics of participants (n = 510)
* Condom, pill, injections, spiral, subcutaneous implant, tubal ligation.
** Withdrawal, calendar.
The scale has total scores ranging from 0 to 110 points. The overall mean score, calculated by dividing the total score by the number of scale items, is 9.03 with a standard deviation (SD) of 1.51. The mean scores in sub-dimensions are also high: 9.14 (SD 1.87) for ‘husband/partner communication’, 8.92 (SD 1.91) for ‘provider communication’ and 9.00 (SD 1.86) for ‘choosing and managing a method’ (Table 1).
The content validity was examined and revealed that all items had an I-CVI higher than 0.80 and the S-CVI Ave was 0.90. No items were removed from the scale. The CFA was performed using the three-factor structure of the original scale. CFA fit indices were found as χ2/df = 4.09, GFI = 0.95, CFI = 0.95, AGFI = 0.913, NFI = 0.940, TLI = 0.933, and RMSEA = 0.078 (Table 2). The results supported the construct validity of the three-factor model (Figure 2).
Goodness-of-fit indices values of the scale (n = 510)
Abbreviations: GFI, Goodness-of-fit index; CFI, Comparative fit index; NFI, Normed fit index; TLI, Tucker-Lewis Index; AGFI, Adjusted goodness-of-fit index; RMSEA, Root mean square error of approximation.
Confirmatory factor analysis of the Turkish version of the CSESSA scale.
Cronbach’s alpha coefficients of the subscales were as follows: ‘husband/partner communication’ subscale (0.842), ‘provider communication’ subscale (0.770), and ‘choosing and managing a method’ subscale (0.810). Cronbach’s alpha of the total scale was found to be 0.867. The internal consistency of the total scale and subscales were adequate (Table 3).
Item total correlation and Cronbach’s α coefficients of the scale
The item-total score correlations of the scale ranged from 0.495 to 0.646. There was no significant increase in Cronbach’s alpha values when an item was deleted (Table 3). The test-retest correlation of the scale was found to be 0.83. The relationship between the total scale score and the subscales was investigated. A significant, positive and strong relationship was found between the total score of the scale and all its subscales (‘husband/partner communication’ p < 0.001, r = 0.729; ‘provider communication’ p < 0.001, r = 0.759; ‘choosing and managing a method’ p < 0.001, r = 0.797).
Hotelling’s T2 test was used to evaluate the response bias, and the Tukey Nonadditivity test was used to evaluate whether the scale had additivity. Hotelling’s T2 score was 92.349 (p = 0.001). Tukey’s non-additivity test was found to be p > 0.05, and the scale additive.
Discussion
This study evaluated the psychometric properties of the Turkish version of CSESSA scale. Currently, no measurement tools are available for women’s contraceptive self-efficacy. The results confirm the validity and reliability of the scale. Furthermore, the results suggest that the CSESSA can be translated into other languages. This is significant, considering the limited number of validated international measures that comprehensively assess contraceptive self-efficacy.
The validity of a measurement tool means that it adequately measures its aims. In this study, independent translation and back-translation were performed, expert opinions were obtained, and a pilot test was conducted on a sample similar to the target population (International Test Commission, 2017). The results showed that the scale performed similarly to the original scale and provided cross-cultural validity (Souza et al., Reference Souza, Alexandre and Guirardello2017). The content validity of the scale was examined using the Davis technique, and I-CVI and S-CVI/Ave were calculated. It is recommended that the minimum I-CVI value should be above 0.78 and the S-CVI/Ave value should be above 0.90 in the number of 6-10 experts (Polit and Beck, Reference Polit and Beck2017). In this study, the I-CVI values for the items ranged from 0.80 and 1.00. The S-CVI/Ave value was determined as 0.90. Consequently, the Turkish version of the scale met the criteria for content validity (Almanasreh et al., Reference Almanasreh, Moles and Chen2019; Esin, Reference Esin, Erdogan, Nahcivan and Esin2014). In this study, the CFA was performed to test the construct validity of the scale. CFA evaluates how many factors the scale has, which items are related to which factors, and whether the model fit index values are appropriate. These results indicate that the model had an acceptable fit (Boateng et al., Reference Boateng, Neilands, Frongillo, Melgar-Quiñonez and Young2018; Byrne, Reference Byrne2016; Yaşlıoğlu, Reference Yaşlıoğlu2017).
Cronbach’s alpha coefficient, item-total score analysis, and test-retest analysis were used to evaluate the reliability of the scale. Cronbach’s alpha coefficient, which is used to evaluate the internal consistency of a scale, was found to be >0.70 in this study, which is considered acceptable reliability (Polit and Beck, Reference Polit and Beck2017). The scale was found to be highly reliable with good internal consistency (Cronbach’s alpha = 0.867). All subscales were found to be either reliable or highly reliable. The item-total score correlation provides the discrimination index of each item (DeVellis, Reference DeVellis2012). In this study, item-total correlation values were higher than 0.30 and varied between 0.495 and 0.646. These results show that the items accurately discriminated between participants. (Çokluk et al., Reference Çokluk, Şekercioğlu and Büyüköztürk2018; Streiner et al., Reference Streiner, Norman and Cairney2015). The test-retest correlation method was used to demonstrate the time invariance of the scale. This method is important because it helps to establish the reliability of the scale over time. The test-retest correlation of the scale is expected to be 0.80 (Nunnally & Bernstein, Reference Nunnally and Bernstein1994; Weir, Reference Weir2005). In this study, the test-retest correlation was calculated and found to be 0.83, indicating an acceptable limit. Hotelling’s T2 test was used to assess whether the scale items were perceived as identical by the participants and whether the responses were biased. When determining self-efficacy levels, the participants may have had biased responses. Therefore, evaluation of response bias is important for scale reliability (Seçer, Reference Seçer2018; Özdamar, Reference Özdamar2016). The test outcomes indicated that the mean item scores differed and there was no response bias (Hotelling’s T2= 92.349, p = 0.001). These findings suggest that women do not exhibit bias when completing the scale. This scale comprises of three subscales. Tukey’s additivity test was used to assess the additivity of scale items. The test results showed that the scale items could be summed to obtain a total score (p > 0.05) (Özdamar, Reference Özdamar2016).
The overall mean score of this scale was 9.03 (SD 1.51). In the original scale, the overall mean score was 8.72 (SD 1.72). Scores on the scale were high and skewed for both sample populations. This may be due to the high prevalence of modern contraceptives in the samples (73.3% for the Kenya sample and 67.3% for this sample) (Whiting-Collins et al., Reference Whiting-Collins, Grenier, Winch, Tsui and Donohue2020). The CSESSA scale assesses women’s self-efficacy in ‘husband/partner communication’, ‘provider communication’, and ‘choosing and managing a method’. This allows for a more comprehensive assessment of contraceptive self-efficacy. This scale is an appropriate tool for use by health professionals to evaluate contraceptive self-efficacy in women of reproductive age.
Study limitations
This study also has some limitations. The data of the study were collected in a single region, and convenience sampling method was used. This may have affected the generalizability of the study.
Conclusion
The Turkish version of the scale, a tool measuring contraceptive self-efficacy, is valid and reliable. This scale can assist health professionals in planning tailored interventions based on an individual’s specific needs and self-efficacy levels, with the aim of preventing unintended pregnancies.
Acknowledgements
We are grateful to all the women who o participated in this research.
Author contributions
Conceptualization: [Aliye Dogan Gangal; Ayten Senturk Erenel]; Methodology: [Aliye Dogan Gangal; Ayten Senturk Erenel], Formal analysis and investigation: [Aliye Dogan Gangal; Ayten Senturk Erenel], Writing – original draft preparation: [Aliye Dogan Gangal]; Writing – review and editing: [Ayten Senturk Erenel], Resources: [Aliye Dogan Gangal]; Supervision: [Ayten Senturk Erenel]
Funding statement
This research did not receive any specific grant from funding agencies in the public, commercial, or not for-profit sectors.
Competing interests
The authors report there are no competing interests to declare.
Ethical standards
Gazi University Ethics Committee (No: E-77082166-604.01.02-355909 & Date: 11.05.22). The institution’s permit Issue: E-51381736-771. The research was conducted according to the principles of the Declaration of Helsinki. Written informed consent was obtained from all women.
Open access
