Abstract
Self-efficacy – positive beliefs about one’s own competencies and mastery – is associated with better recovery outcomes for people using mental health services.
Aim:
To translate the Self-Efficacy for Personal Recovery Scale (SEPRS) into Arabic and evaluate the psychometric properties of the Arabic version.
Methods:
An established translation methodology was employed, involving back-translation, comparison, forward-translation, comparison, and piloting. The pre-final version of the Arabic translated scale was tested for clarity with young people with a primary diagnosis of mental health problem. The final Arabic version and standardized measures of hope and loneliness were administered to 119 young people in two rounds.
Results:
Internal consistency was adequate (Cronbach’s alpha = 0.87 in round 1, 0.91 in round 2). Consistent with the English version, a one-factor solution best fitted the data. The correlation between SEPRS and hope was R = 0.60 (round 1) and R = 0.61 (round 2), indicating convergent validity. The correlation between SEPRS and loneliness was R = −0.52 (round 1) and R = −0.60 (round 2). Correlation between test and retest was R = −0.998 indicated adequate test-retest reliability. Minimal floor and ceiling effects were detected.
Conclusion:
The use of the Arabic SEPRS with Arabic-speaking samples is supported. Further research to investigate divergent validity is warranted.
Background
Different understandings of recovery have emerged within mental health services (Slade, 2009). The term recovery has been used variably across countries and disciplines (Tew et al., 2012). One meaning of recovery has emerged from professional-led practice and involves symptoms’ control (Slade, Adams, et al., 2012) and enhancing vocational and social functioning of the person with mental health issue (Ibrahim et al., 2021). On the other hand, understanding recovery as a personal experience, and as a process not just an outcome, has structured mental health services in many countries (Slade et al., 2014; Slade, Leamy, et al., 2012; Slade, Williams, et al., 2012). This understanding – personal recovery – has been described as “a deeply personal, unique process of changing one’s attitudes, values, feelings, goals, skills, and/or roles. It is a way of living a satisfying, hopeful, and contributing life even with limitations caused by illness. Recovery involves the development of new meaning and purpose in one’s life as one grows beyond the catastrophic effects of mental illness” (Anthony, 1993).
Several studies have investigated the personal accounts of mental health service users in order to develop understanding of recovery process (Leonhardt et al., 2017; Nelson et al., 2001; Slade, 2009). A systematic review of this literature identified five key recovery processes: Connectedness; Hope and optimism about the future; Identity; Meaning in life; and Empowerment, giving the acronym CHIME (Leamy et al., 2011). The validity of the CHIME framework has been confirmed through consultation with current mental health service users (Bird et al., 2014) and for cross-cultural use (Slade, Leamy, et al., 2012), and it has become a widely-used approach to operationalizing personal recovery (van Weeghel et al., 2019).
Personal recovery is a dynamic and complex process dependent on several factors, including self-efficacy (Mancini, 2007), defined as confidence in one’s ability to execute tasks and deal with adversity (Luszczynska et al., 2005). Self-efficacy involves positive beliefs about one’s own competencies and mastery that enhances coping with and adaptation to mental health problems which in turn improve mental health outcomes (Davidson & Strauss, 1992; Szczebak, 2012) and facilitate the self-managed personal recovery journey (Silverstein & Bellack, 2008). Additionally, self-efficacy can buffer the disempowering treatment experiences (Hughes et al., 2009) that may be detrimental to self-integrity and impede recovery outcomes (McLeod et al., 2019).
The Self-Efficacy for Personal Recovery Scale (SEPRS) is an English-language scale developed in Australia and informed by the CHIME framework. The scale showed high internal consistency (Cronbach’s α = 0.961) and test-retest reliability(r = 0.855, 95% CI = 0.715, 0.929), and correlated with recovery related measures (hope, personal recovery and generalized self-efficacy), with independence from insight, positive symptoms and social desirability (Villagonzalo et al., 2018). Translations to other languages are needed to explore cross-cultural validity and ensure recovery research is not dominated by English-language studies. Arabic is one of the six official languages of the United Nations, and the official language of 24 countries in Africa and Asia., translating SEPRS to Arabic will facilitate recovery research with Arabic speaking populations.
Aim
The aim of this study was to translate SEPRS into Arabic and to test the psychometric properties of the Arabic version.
Methods
Design
The five stages of the guidelines for translating, adapting and validating scales for cross cultural use were followed (Sousa & Rojjanasrirat, 2011). The stages comprised: forward translation of the original scale to Arabic, comparison of the translated versions of the target language, blind back translation to the original source language, comparison of the blind back-translated versions and pilot testing of the pre-final version of the scale with a sample of the target population.
In Step 1 (Forward translation), three forward translations of SEPRS from English to Arabic were produced by three independent bilingual researchers, two of whom have exposure to the English culture. Translators were from different backgrounds (health sciences, medicine, and social sciences). The translated drafts were in Egyptian colloquial language spoken by contemporary Egyptians, because modern standard Arabic is used only in formal speech, newspapers and books and not spoken in Egypt.
In Step 2 (Comparison of forward translations), the three forward translated versions were checked and compared with the original scale by a bilingual person not involved in the forward translation stage. The main comment addressed in this stage was to maintain the pronoun “you” in statements coming after item 4 as the heading of item 4 in the original English scale used the pronoun “you” then the underlying statements used reflexive pronoun “myself” or the personal pronoun “me”. This was agreed on by all researchers from stage (1). A preliminary Arabic version of the scale was generated.
In Step 3 (back translation), three independent bilingual individuals who were blind to the original English version of the scale back translated the preliminary Arabic version to English. One translators had knowledge of medical and health care terminology, the second had an English cultural exposure, and all had knowledge of colloquial language.
In Step 4 (Comparison of back translations), a member of the English-language development team and two members of the Arabic research team compared the three back translated English drafts. The panel agreed on refining item (7) “tell services my views on how to manage my mental health” as the intended meaning was not addressed. Item (7) was forward translated from English to Arabic by three individuals, the Arabic translations were compared, and blind translation from Arabic to English was conducted and agreed on by the panel in stage (4). This process produced the pre-final Arabic version.
In step 5 (Pilot testing), the pre-final Arabic scale was piloted with young people with a primary mental health diagnosis. A total number of ten participants (young people with a primary diagnosis of mental health problem and living in the community) identified through on-line advertisement were asked to dichotomously rate each item of the pre-final Arabic version of the Self-Efficacy for Personal Recovery Scale as either “Clear” or “Unclear” and to provide suggestions for statements rated as Unclear. All participants rated items of SEPRS as clear.
Evaluation of the Arabic SEPRS
Participants fulfilling the following criteria were recruited for the purpose of Arabic SEPRS evaluation; living in the community; aged 18 to 24 years; using outpatient mental health service; primary clinical mental health diagnosis. Exclusion criteria: comorbid substance use disorders; developmental disorders. Participants completed a demographics form and the Arabic versions of SEPRS. In order to test how closely the Arabic version of SEPRS relates with personal recovery measures; convergent validity was tested through investigating the correlation between Adult Hope Scale (AHS) and SEPRS. AHS is a 12-item measure that was originally developed by Snyder et al. (1991) and was translated to Arabic language by Abdel-Khalek and Snyder (2007). The Cronbach alpha of the Arabic version of the scale was 0.86.
Additionally, divergent validity was investigated by testing the correlation of the SEPRS with a measure that does not represent personal recovery (The University of California, Los Anglos loneliness scale (UCLA-LS)). The scale developed by Russell (1996) and was translated to Arabic by Al-Desoki 1998 (2008,
). UCLA-LS is a 20-item measure with two domains; emotional loneliness domain and the social loneliness domain. The Cronbach Alpha of the Arabic version of emotional loneliness domain of the UCLA-LS was 0.70 and for the social loneliness domain was 0.78.
Both convergent and divergent validity were performed at two different points in the same day with a distractive activity in-between.
Faculty Ethics Committee approval was obtained prior to commencing with the current study (Ref. No. 0191).
Analysis
To determine internal consistency, Cronbach’s alpha of all Arabic SEPRS items was calculated for round 1 and round 2. To establish if the same latent structure was visible in the Arabic translation of SEPRS, confirmatory factor analysis with a single factor solution was carried out at both rounds. This was followed by an exploratory factor analysis using varimax rotation to determine if other factor solutions existed in the translated version for both rounds. Hypothesis testing to determine the least number of significant factors needed was used to determine the number of factors, and a factor loading of greater than 0.4 was used to indicate an item is substantially loading on to a factor. To assess convergent validity, we correlated with AHS, since hope is expected to correlate with personal recovery. To assess divergent validity, we correlated with UCLA-LS, since loneliness is not expected to correlate with personal recovery. Pearson’s correlation and a pre-defined criteria of r > 0.50 was used for both validity assessments. To assess test re-test reliability, we performed Pearson’s correlation and paired t-tests for round 1 and 2 for the total SEPRS score. Floor and ceiling effects were analyzed for each item in each round, with a predefined threshold of >50% indicating scale attenuation.
Results
A total of 119 participants completed round 1 measures, of whom 96 (81%) completed round 2 measures. Table 1 describes demographic characteristics for participants at both rounds.
Participants’ characteristics in round 1 and 2.
No demographic characteristic differed between the two rounds.
Cronbach’s alpha at round 1 was 0.87 (CI 0.84–0.91) and at round 2 was 0.91 (CI 0.88–0.93), indicating adequate internal consistency.
Confirmatory factor analysis with varimax rotation of the single factor solution for SEPRS is shown in Table 2. The models accounted for 35% of variance at round 1 and 42% of variance at round 2.
Confirmatory factor analysis of Arabic SEPRS for single factor solution.
Exploratory factor analysis found a two-factor solution for round 1 accounting for 42% of variance and a three-factor solution for round 2 accounting for 52% of variance, as shown in Table 3.
Exploratory factor analysis of Arabic SEPRS.
Note. Grey = factor loading <0.40.
The Pearson correlation of SEPRS to hope assessed using AHS was r = 0.60 (p < .0001) at round 1 and r = 0.61 (p < .0001) at round 2, indicating adequate convergent validity. The Pearson correlation of SEPRS to loneliness was r = −0.52 (p < .0001) at round 1 and r = −0.60 (p < .0001) at stage 2, not demonstrating adequate discriminant validity.
Test-retest reliability is shown in Table 4.
Test-retest reliability for SEPRS.
The Pearson correlation between test and retest was r = 0.998 (p < .0001). The mean difference in total score between rounds was 0.037 (p = .76) indicating adequate test-retest reliability. Test-retest reliability for individual items was adequate, with the exception of item 4.
Floor and ceiling effect analysis is shown in Table 5.
Floor and ceiling effects for each SEPRS item.
Note. Bold = >50%.
Only one item (round 2 item 10) had evidence of a ceiling effect, indicating overall that scores on the scale are adequately distributed.
Discussion
The current study aimed to translate the English version of SEPRS to Arabic and evaluate the psychometric properties of the Arabic version. Adequate evidence was found for internal consistency, convergent validity and test-retest reliability. A single-factor solution consistent with the English SEPRS was found, and minimal floor or ceiling effects were detected. Evidence for divergent validity was not demonstrated, and exploratory factor analysis also found two-factor and three-factor solutions.
Cronbach’s alpha was high for both rounds, indicating internal consistency of the Arabic version of SEPRS. This reflects the inter-relatedness of individual items in the Arabic SEPRS and the extent to which items in SEPRS measure the same construct (Tavakol & Dennick, 2011). Given that the purpose of reliability tests is to examine the extent to which repeated measurements agree over time (Aldridge et al., 2017), the Arabic SEPRS was not significantly different between both rounds for all items and the average score showed consistency between test and retest.
All factor solutions for the SEPRS accounted for only a small amount of variation. This may be due to the way the questionnaire is implemented, with scoring responses involving three categories rather than the continuous scale 0 to 100 used in the original version. The one-factor solution was consistent at both rounds 1 and 2 indicating the same internal consistency with the English version of SEPRS. Exploratory factor solutions were not consistent across rounds, with round 1 having a two-factor solution and round 2 having a three-factor solution, indicating that a one-factor solution was the most consistent across the two rounds. Finally, further research to investigate divergent validity is needed. UCLA-LS was chosen as a comparison domain due to the limited options available in Arabic, but the closely related construct of social connectedness forms part of the CHIME model on which the SEPRS is based, so a comparison domain which is less related to personal recovery might show better evidence for divergent validity.
Limitations
The participating sample in the evaluation of Arabic SEPRS was young people, which may represent an issue in the generalizability of the scale properties to an older sector of the population.
Conclusion
Evaluation of the Arabic SEPRS supports the use of the scale among Arabic-speaking samples. Positive self-efficacy is associated with an increased mental health recovery trajectory (Wu et al., 2021), so the Arabic SEPRS can be used as a personal recovery measure in future recovery research among Arabic-speaking populations.
