Assessing patient-centeredness among medical students in Bahrain: the Arabic translation and validation of the Patient-Practitioner Orientation Scale (PPOS) Authors: Khalid Gaffer Mohamed, Amer Almarabheh, Aya M. Haiba, Leena Khalid, Sara Emad, Abdalrahman Alosaimi, Abdalwahab Alosaimi, Mohamed Hany Shehata, Hany Atwa Published: 2025 ABSTRACT BackgroundPerson-centered care (PCC) emphasizes a collaborative, empathetic approach to healthcare, which marks a significant shift from the traditional, hierarchical biomedical model. The Patient-Practitioner Orientation Scale (PPOS) is a validated tool used to assess attitudes toward PCC. However, no prior studies have evaluated such attitudes among medical students in Bahrain or adapted the PPOS to the Arabic language. This study aims to assess patient-centered attitudes among medical students at the Arabian Gulf University (AGU) in Bahrain, examine variations based on socio-demographic factors and academic year, and evaluate the psychometric properties of the Arabic translated PPOS (A-PPOS).MethodsA cross-sectional study was conducted at the College of Medicine and Health Sciences, Arabian Gulf University, using the PPOS. The scale was translated into the Arabic language following WHO guidelines. A comprehensive sampling technique was used, targeting medical students of all six medical school years. A total of 540 student questionnaires were included in the study. The internal consistency of the scale was assessed using Cronbach’s alpha, while the underlying structure was checked through exploratory and confirmatory factor analyses.ResultsThe mean PPOS score was 3.90 ± 0.47, suggesting moderate patient-centeredness. Female students scored significantly higher than males across all domains (p < 0.001). No significant differences were found based on academic year or parental background. The internal consistency of the A-PPOS was low (α = 0.517), and factor analyses did not confirm the scale’s original two-factor structure, showing six different underlying factors instead.ConclusionOur sample of medical students shows a moderate orientation toward PCC, with gender differences favoring females. The Arabic PPOS showed limited reliability and construct validity, indicating the need for cultural tailoring and improvement. Incorporating PCC principles into medical education remains vital for encouraging effective, patient-focused healthcare. Introduction Two models that define modern medicine are biomedical evidence-based medicine and bio-psycho-social patient-centered medicine [1]. The biomedical evidence-based medicine model emphasizes the integration of the best scientific evidence with clinical expertise to guide patient care. It relies on systematic research, particularly randomized controlled trials and meta-analyses, to ensure that medical decisions are objective, effective, and based on proven outcomes. On the other hand, the bio-psycho-social patient-centered medicine views health as a product of biological, psychological, and social factors. It emphasizes understanding patients’ experiences, values, and contexts, promoting empathy, shared decision-making, and holistic care [2, 3]. Person-centered care (PCC) is the current gold standard in healthcare, as opposed to a disease-centered, paternalistic approach [4]. This dictates that healthcare practitioners fulfill patients’ human needs, such as warmth, and comprehension, in addition to providing technical knowledge for illness treatment [5]. By using this approach, doctors utilize their communication skills to identify a shared understanding about the issue at hand and develop a management plan that both parties agree on. They can delve into patients’ concerns, information requirements, and main reasons for their visit; gain a holistic view of the patient’s circumstances, including emotional needs and life challenges; and enhance the doctor-patient relationship moving forward, while also promoting health and prevention [6]. In a cooperative effort of shared decision-making with their healthcare practitioners, PCC seeks to empower and involve patients in decision-making as far as their health is concerned, all while honoring their individuality and beliefs [7]. Under this method, patients are seen as the expert about their physical condition, allowing for discussions on psychosocial problems during the visit. It suggests that the doctor invites patients to talk about their medical experience and that the consultation is centered on the goals of both the patient and the doctor [5]. This fundamental idea opposes the overemphasis on biological causes of illness and the unbalanced power dynamics in patient-provider relationships that are frequently found in the healthcare system [8]. Adopting this way of interviewing patients has been shown to alter therapeutic activities in favor of patients, in addition to improving patient adherence to treatment, and satisfaction of both the patient and the physician [9–11]. Evidence has shown that the culture of medical education lays more emphasis on the biomedical aspects of illness treatment rather than cultivating a sense of understanding for patient preferences, worries, and emotions [12]. Integrating patient-centered approaches into medical school programs may help future healthcare professionals deliver high-quality care, enhance patient outcomes, and build effective health systems. Medical students represent future physicians, and examining their perspectives towards the doctor-patient relationship could help address much of the gaps present in today’s healthcare [8]. Medical students’ attitudes towards the doctor-patient relationship have been described and quantified in a number of studies globally. Factors influencing those attitudes have also been studied. In Brazil and the United States, students had strong patient-centered tendencies, whereas students in Pakistan had weaker ones [11–13]. In several studies, gender was found to influence these students’ dispositions, with females expressing mainly patient-centered beliefs [13–18]. In the framework of medical education and healthcare, evaluating such attitudes has grown in significance. Originally intended to gauge attitudes regarding medical professionals’ involvement in patient care, the Patient-Practitioner Orientation Scale (PPOS) is a tool used by physicians, medical students, and patients. Between the responses, it draws a contrast between an attitude focused on the patient and one focused on the doctor [19, 20]. The PPOS is an 18-item instrument that was developed in America in 1999. The tool assesses patient views on the doctor-patient relationship on two different dimensions: sharing and caring [19]. This scale is widely used in the United States and has been translated into other languages worldwide [11, 13, 17, 18, 21, 22]. The use of self-report instruments such as the PPOS must not be hampered by translation issues in non-English speaking societies. Therefore, psychometrically and linguistically valid questions must be developed to create an instrument with comparable validity and reliability to the source tool [23, 24]. This must be done while considering the language’s cultural context. To date, no study has attempted to measure and describe patient-centered inclinations among medical students in Bahrain. This addition to the literature should unmask pre-existing trends and perceptions associated with patient-centeredness, in an attempt to refine and elevate the delivery of healthcare. Our objectives were to evaluate patient-centered attitudes among medical students, investigate whether these attitudes differ based on academic year and socio-demographic factors, and examine the construct validity and internal consistency of the PPOS in this context. Materials & methods Study design and setting This is a cross-sectional descriptive study targeting medical students at the College of Medicine and Health Sciences, Arabian Gulf University (CMHS-AGU) in Manama, Bahrain. The CMHS-AGU adopts a six-year curriculum, divided into preclinical (years one to four), and clinical (years five and six). Basic communication skills are taught at year two, while advanced communication skills are taught as a longitudinal course in years two, three, and four. Almost all students are from the Gulf Cooperation Council (GCC)—mainly Saudi Arabia, Kuwait, Bahrain, and Oman. The mother language of all students is Arabic while the teaching language is English. Participants and data collection The sample size for this validation study was planned and justified based on established methodological guidelines for psychometric analyses. The minimum convenient sample size required for the validation process is at least 10 participants per scale item; as the PPOS has 18 items, the calculated minimum sample size was 180 participants [25]. The questionnaire was distributed in October 2024 in paper form to all 1300 medical students across all school years during their classroom sessions. A total of 619 students returned completed questionnaires. We excluded responses with missing data on any of the 18 PPOS items, as well as those that appeared to be filled in an irresponsible manner (e.g., selecting the same option for all questions). After these exclusions, the final dataset included 540 participants (79 responses were excluded). Data from the paper questionnaire sheets were manually entered into Excel sheets for subsequent statistical analysis. While the 10-participants-per-item rule is a widely accepted heuristic [25], contemporary research emphasizes that sample size adequacy should depend on the analytical objectives [26]. For factor analysis, a minimum of 200 participants is often considered acceptable, but larger samples are recommended to ensure stable parameter estimates and reliable factor structures [27]. Our final sample of 540 participants far exceeds these minimum thresholds, substantially increasing the statistical power, precision of estimates, and generalizability of findings [28]. Description of the instrument The PPOS is a self-report questionnaire that was developed by Krupat et al. [19] It aims to assess PCC, differentiating between a patient-centered approach and a doctor-centered approach. It is an eighteen-items Likert scale which is subdivided into two subscales. The sharing subscale (nine items) assesses the attitude towards sharing power, control, and information between the practitioner and the patient. The caring subscale (nine items) measures the attitude towards caring about patient’s preference, feelings, expectations, and perspectives. The Likert scale ranges from 1 (strongly agree) to 6 (strongly disagree). The total score ranges from 18 (most doctor-centered or paternalistic) to 108 (most patient-centered or egalitarian). The “mean” value is calculated by dividing the total score by 18 (thus ranging from 1 to 6). Internal consistency of the PPOS was measured by the developers (Krupat and colleagues) and was found to be satisfactory for the total score (α = 0.73), moderate for the sharing subscale (α = 0.67), and moderate for the caring subscale (α = 0.52). Originally, the PPOS showed a two-factor structure (caring and sharing) [29]. Along with the primary outcome variable (PPOS), the questionnaire included questions about sociodemographic traits (explanatory factors) that had been shown to be relevant in prior studies, such as age, gender, nationality, student year, whether one parent was a doctor, and whether one parent held a bachelor’s degree [14, 18, 30, 31]. The translation of the PPOS English version to Arabic The PPOS was translated following the international guidelines like Beaton et al. and the World Health Organization (WHO) approaches to questionnaire translation including conceptual equivalence, clarity, appropriateness to the targeted audience, and cultural relevance [23, 32]. The following steps were carried out to ensure the best possible translation: Forward translation was carried out separately by three independent bilingual experts from the original English version to the Arabic language, two of whom were professors who were experts in the field and had used the PPOS in a previous publication, while the third was an English Language instructor at the English Language Unit of the CMHS-AGU.Reconciliation: The three versions produced by the three independent bilingual experts through forward translation were revised by a panel of bilingual experts. The goal was to identify discrepancies and to merge the three versions in the most appropriate and accurate translation.Back translation from Arabic to English was carried out separately by two independent bilingual experts, one of them was a professor who had never used the PPOS but had experience in research related to linguistics and translation. The other was an expert in the field and had used the PPOS in previous research. The purpose of the back translation was to identify discrepancies that might have arisen unintentionally during the forward translation and reconciliation phases.A panel of experts including the forward and backward translators reviewed the translation steps and compared the final versions of translation and carried out the necessary modifications. Semantic, idiomatic, and conceptual equivalences were considered. The expert panel reached a consensus regarding the final Arabic version.The final Arabic version translation was piloted on 20 students of the targeted population in two rounds using cognitive interviewing to identify unclear sentences and to confirm that it was clear and understandable. Forward translation was carried out separately by three independent bilingual experts from the original English version to the Arabic language, two of whom were professors who were experts in the field and had used the PPOS in a previous publication, while the third was an English Language instructor at the English Language Unit of the CMHS-AGU. Forward translation was carried out separately by three independent bilingual experts from the original English version to the Arabic language, two of whom were professors who were experts in the field and had used the PPOS in a previous publication, while the third was an English Language instructor at the English Language Unit of the CMHS-AGU. Reconciliation: The three versions produced by the three independent bilingual experts through forward translation were revised by a panel of bilingual experts. The goal was to identify discrepancies and to merge the three versions in the most appropriate and accurate translation. Reconciliation: The three versions produced by the three independent bilingual experts through forward translation were revised by a panel of bilingual experts. The goal was to identify discrepancies and to merge the three versions in the most appropriate and accurate translation. Back translation from Arabic to English was carried out separately by two independent bilingual experts, one of them was a professor who had never used the PPOS but had experience in research related to linguistics and translation. The other was an expert in the field and had used the PPOS in previous research. The purpose of the back translation was to identify discrepancies that might have arisen unintentionally during the forward translation and reconciliation phases. Back translation from Arabic to English was carried out separately by two independent bilingual experts, one of them was a professor who had never used the PPOS but had experience in research related to linguistics and translation. The other was an expert in the field and had used the PPOS in previous research. The purpose of the back translation was to identify discrepancies that might have arisen unintentionally during the forward translation and reconciliation phases. A panel of experts including the forward and backward translators reviewed the translation steps and compared the final versions of translation and carried out the necessary modifications. Semantic, idiomatic, and conceptual equivalences were considered. The expert panel reached a consensus regarding the final Arabic version. A panel of experts including the forward and backward translators reviewed the translation steps and compared the final versions of translation and carried out the necessary modifications. Semantic, idiomatic, and conceptual equivalences were considered. The expert panel reached a consensus regarding the final Arabic version. The final Arabic version translation was piloted on 20 students of the targeted population in two rounds using cognitive interviewing to identify unclear sentences and to confirm that it was clear and understandable. The final Arabic version translation was piloted on 20 students of the targeted population in two rounds using cognitive interviewing to identify unclear sentences and to confirm that it was clear and understandable. Statistical analysis The frequency and percentage were calculated for the qualitative variables. Cronbach’s α was used to evaluate the internal consistency of the PPOS and subscales. Construct validity was tested through exploratory factor analysis (EFA) using the principal component analysis method and the Varimax rotation. The suitability of the data for EFA was confirmed by the Kaiser normalized maximum variance method and Bartlett’s sphericity test (BT). The decision on the number of factors to extract was guided by several criteria: the Kaiser criterion (selecting factors with eigenvalues greater than 1, the Scree test (or Cattell criterion) to pinpoint the inflection point on the Scree plot, and the cumulative percentage of variance explained [33–35]. Confirmatory factor analysis (CFA) was performed, and maximum likelihood analysis was used. The main adjustment indicators of the validation model included the normed-fit index (NFI), incremental-fit index (IFI), comparative-fit index (CFI), goodness-of-fit index (GFI), adjusted goodness-of-fit index (AGFI), root mean square error of approximation (RMSEA), and χ2/df. GFI, AGFI, IFI, and CFI greater than 0.9 and an RMSEA value less than 0.08 suggested an ideal model fit. The recommended (χ2/df) ratio was 1 to 3. The statistical package for social sciences SPSS v29, and AMOS v23 were employed to perform data cleaning and analysis. Statistical significance was set at p < 0.05. Ethical consideration Permission to use the original PPOS tool was granted by its author, Professor Edward Krupat. The study was conducted in accordance with the ethical principles outlined in the Declaration of Helsinki for research involving human subjects. Participation was entirely voluntary, and informed consent was obtained from all participants following a clear explanation of the study’s objectives, procedures, and potential implications. Participants were assured of their right to withdraw at any stage without penalty. Confidentiality and anonymity were strictly maintained, with all responses securely stored and accessible only to the research team. Ethical approval for the study was obtained from the Institutional Review Board (IRB) at the CMHS-AGU (Reference No. E4-PI-04–24). Results Sample demographics The study aimed to validate the Arabic translated version of the PPOS through medical students’ responses to the translated questionnaire. Data was collected from a comprehensive sample. The response rate was 47.6%. The characteristics of the study participants (Table 1) revealed that the majority of respondents were female (70.2%). Participants were distributed across different academic years, with the highest representation from year two students (25.9%) and the lowest from year five students (8.5%). In terms of nationality, the largest proportion of participants were from Kuwait (33.3%). Only 8.6% of participants had a parent who was a doctor, and 86.1% of participants had at least one parent with a bachelor’s degree.Table 1Socio-demographic characteristics of the participants (n = 540)Socio-demographic variablesN (%)Gender Male161 (29.8) Female372 (70.2)Study year Year 175 (13.9) Year 2140 (25.9) Year 3108 (20) Year 4124 (23) Year 546 (8.5) Year 647 (8.7)Nationality Kuwait175 (33.3) Bahrain130 (24.8) Saudi124 (23.6) Oman92 (17.5) Other4 (0.8)Parent Doctor Yes46 (8.6) No488 (91.4)Parent bachelor Yes459 (86.1) No74 (13.9) Table 1Socio-demographic characteristics of the participants (n = 540)Socio-demographic variablesN (%)Gender Male161 (29.8) Female372 (70.2)Study year Year 175 (13.9) Year 2140 (25.9) Year 3108 (20) Year 4124 (23) Year 546 (8.5) Year 647 (8.7)Nationality Kuwait175 (33.3) Bahrain130 (24.8) Saudi124 (23.6) Oman92 (17.5) Other4 (0.8)Parent Doctor Yes46 (8.6) No488 (91.4)Parent bachelor Yes459 (86.1) No74 (13.9) Socio-demographic characteristics of the participants (n = 540) A-PPOS psychometrics The test of reliability yielded a low Cronbach’s alpha of (α = 0.517) for the full scale, the caring subscale (α = 0.318), and the sharing subscale (α = 0.439). In assessing construct validity, the CFA for the scale indicates that the model fit needs improvement. The CFA did not support the scale’s original two-factor structure (Fig. 1). While some indices (GFI = 0.904, AGFI = 0.877, RMSEA = 0.069) suggest an acceptable fit, others (CFI = 0.550, TLI = 0.486, IFI = 0.560) are below the recommended threshold, indicating poor fit (Table 2). The chi-square statistics also point to a significant difference between observed and model-implied data. Overall, the model may require modifications for better alignment with the data.Fig. 1Confirmatory factor analysis of the PPOSTable 2Model fit indices for the CFA of the PPOSFit IndexValueRecommended ThresholdInterpretationChi-Square (χ²), P = 0.000480.972Non-significant (p > 0.05)Significant (poor fit)Degrees of Freedom (df)134CMIN/DF (χ²/df)3.589≤ 3 (acceptable)Slightly above the thresholdGoodness-of-Fit Index (GFI)0.904≥ 0.90 (good fit)Acceptable fitAdjusted Goodness-of-Fit Index (AGFI)0.877≥ 0.90 (good fit)Slightly below idealComparative Fit Index (CFI)0.550≥ 0.90 (good fit)Poor fitTucker-Lewis Index (TLI)0.486≥ 0.90 (good fit)Poor fitIncremental Fit Index (IFI)0.560≥ 0.90 (good fit)Poor fitRoot Mean Square Error of Approximation (RMSEA)0.069≤ 0.08 (acceptable)Moderate fit Confirmatory factor analysis of the PPOS Table 2Model fit indices for the CFA of the PPOSFit IndexValueRecommended ThresholdInterpretationChi-Square (χ²), P = 0.000480.972Non-significant (p > 0.05)Significant (poor fit)Degrees of Freedom (df)134CMIN/DF (χ²/df)3.589≤ 3 (acceptable)Slightly above the thresholdGoodness-of-Fit Index (GFI)0.904≥ 0.90 (good fit)Acceptable fitAdjusted Goodness-of-Fit Index (AGFI)0.877≥ 0.90 (good fit)Slightly below idealComparative Fit Index (CFI)0.550≥ 0.90 (good fit)Poor fitTucker-Lewis Index (TLI)0.486≥ 0.90 (good fit)Poor fitIncremental Fit Index (IFI)0.560≥ 0.90 (good fit)Poor fitRoot Mean Square Error of Approximation (RMSEA)0.069≤ 0.08 (acceptable)Moderate fit Model fit indices for the CFA of the PPOS EFA identified six factors with eigenvalues greater than 1 (Fig. 2), ranging from 2.625 (Factor 1) to 1.042 (Factor 6), indicating that these factors explain meaningful variance in the data, accounting for 50.738% of the total variance. The six factors explained 14.585%, 10.189%, 7.087%, 7.012%, 6.075%, and 5.790% of the total variance, respectively. All items had factor loadings greater than 0.4, with two factors containing the fewest items (Table 3). The KMO measure of sampling adequacy was 0.673 (p < 0.001). Overall, the EFA did not support the scale’s original two-factor structure. The factors were labelled as follows:Factor 1: Information GatekeepingFactor 2: Collaborative PartnershipFactor 3: Focus on CureFactor 4: Authority & DependenceFactor 5: Reassurance PreferenceFactor 6: Empathy & TrustFig. 2Scree Plot of the EFA for the PPOSTable 3EFA factor loading matrix of the PPOSItemsSix factor loadingFactor 1:Information GatekeepingFactor 2: Collaborative PartnershipFactor 3: Focus on CureFactor 4: Authority & DependenceFactor 5: Reassurance PreferenceFactor 6: Empathy & TrustPPOS Item 40.556PPOS Item 60.697PPOS Item 70.732PPOS Item 160.563PPOS Item 90.615PPOS Item 130.590PPOS Item 14−0.449PPOS Item 170.592PPOS Item 10.799PPOS Item 20.571PPOS Item 30.499PPOS Item 50.781PPOS Item 150.454PPOS Item 180.670PPOS Item 80.632PPOS Item 100.732PPOS Item 110.508PPOS Item 120.757Eigenvalue2.6251.8341.2761.2621.0941.042Variance explained (%)14.58510.1897.0877.0126.0755.790Cronbach’s alpha0.5830.0590.4070.4140.4760.280 Factor 1: Information Gatekeeping Factor 1: Information Gatekeeping Factor 2: Collaborative Partnership Factor 2: Collaborative Partnership Factor 3: Focus on Cure Factor 3: Focus on Cure Factor 4: Authority & Dependence Factor 4: Authority & Dependence Factor 5: Reassurance Preference Factor 5: Reassurance Preference Factor 6: Empathy & Trust Factor 6: Empathy & Trust Scree Plot of the EFA for the PPOS Table 3EFA factor loading matrix of the PPOSItemsSix factor loadingFactor 1:Information GatekeepingFactor 2: Collaborative PartnershipFactor 3: Focus on CureFactor 4: Authority & DependenceFactor 5: Reassurance PreferenceFactor 6: Empathy & TrustPPOS Item 40.556PPOS Item 60.697PPOS Item 70.732PPOS Item 160.563PPOS Item 90.615PPOS Item 130.590PPOS Item 14−0.449PPOS Item 170.592PPOS Item 10.799PPOS Item 20.571PPOS Item 30.499PPOS Item 50.781PPOS Item 150.454PPOS Item 180.670PPOS Item 80.632PPOS Item 100.732PPOS Item 110.508PPOS Item 120.757Eigenvalue2.6251.8341.2761.2621.0941.042Variance explained (%)14.58510.1897.0877.0126.0755.790Cronbach’s alpha0.5830.0590.4070.4140.4760.280 EFA factor loading matrix of the PPOS PPOS scoring According to the original scale’s coding, mean PPOS score for the entire sample was 3.90 ± 0.47. Male and female mean PPOS scores were different, as were the scores for the caring and sharing subscales. Compared to their male peers, female students had a higher overall PPOS score (3.79 ± 0.53 and 4.02 ± 0.41, respectively). Female gender was found to be significantly associated with total PPOS score (p < 0.001), the caring subscale score (p < 0.001), and the sharing subscale score (p < 0.001). Pre-clinical and clinical study-year status and having parents who were doctors had no bearing on patient-centeredness (p > 0.05) (Table 4).Table 4Comparison of PPOS scores according to socio-demographic variablesSharingCaringOverallGender (Mean ± SD) Male3.46 ± 0.664.11 ± 0.633.79 ± 0.53 Female3.72 ± 0.604.32 ± 0.474.02 ± 0.41P-Value< 0.001< 0.001< 0.001Clinical phase (Mean ± SD) Pre-Clinical3.62 ± 0.614.23 ± 0.553.93 ± 0.46 Clinical3.67 ± 0.694.34 ± 0.454.01 ± 0.48P-Value0.3960.0660.104Parent bachelor (Mean ± SD) Yes3.64 ± 0.624.26 ± 0.543.95 ± 0.47 No3.61 ± 0.664.22 ± 0.493.92 ± 0.45P-Value0.7130.6200.594Parent physician (Mean ± SD) Yes3.77 ± 0.654.23 ± 0.674.00 ± 0.58 No3.62 ± 0.624.26 ± 0.523.94 ± 0.46P-Value0.1230.7350.400Nationality (Mean ± SD) Bahrain3.75 ± 0.684.21 ± 0.573.98 ± 0.51 Saudi3.55 ± 0.624.23 ± 0.513.89 ± 0.44 Kuwait3.61 ± 0.634.26 ± 0.573.93 ± 0.48 Oman3.65 ± 0.534.33 ± 0.453.99 ± 0.40 Other3.25 ± 0.944.39 ± 0.643.82 ± 0.75P-Value0.0730.5000.449 Table 4Comparison of PPOS scores according to socio-demographic variablesSharingCaringOverallGender (Mean ± SD) Male3.46 ± 0.664.11 ± 0.633.79 ± 0.53 Female3.72 ± 0.604.32 ± 0.474.02 ± 0.41P-Value< 0.001< 0.001< 0.001Clinical phase (Mean ± SD) Pre-Clinical3.62 ± 0.614.23 ± 0.553.93 ± 0.46 Clinical3.67 ± 0.694.34 ± 0.454.01 ± 0.48P-Value0.3960.0660.104Parent bachelor (Mean ± SD) Yes3.64 ± 0.624.26 ± 0.543.95 ± 0.47 No3.61 ± 0.664.22 ± 0.493.92 ± 0.45P-Value0.7130.6200.594Parent physician (Mean ± SD) Yes3.77 ± 0.654.23 ± 0.674.00 ± 0.58 No3.62 ± 0.624.26 ± 0.523.94 ± 0.46P-Value0.1230.7350.400Nationality (Mean ± SD) Bahrain3.75 ± 0.684.21 ± 0.573.98 ± 0.51 Saudi3.55 ± 0.624.23 ± 0.513.89 ± 0.44 Kuwait3.61 ± 0.634.26 ± 0.573.93 ± 0.48 Oman3.65 ± 0.534.33 ± 0.453.99 ± 0.40 Other3.25 ± 0.944.39 ± 0.643.82 ± 0.75P-Value0.0730.5000.449 Comparison of PPOS scores according to socio-demographic variables Discussion This study aimed to assess patient-centered attitudes among undergraduate medical students in Bahrain, in addition to the translation and adaptation of the PPOS to the Arabic language. In spite of the PPOS being a widely used and validated instrument in different cultures, its validity and reliability have shown to be controversial in several other cultures [36]. The psychometric assessment of the A-PPOS in our sample revealed suboptimal reliability and model fit. The Cronbach’s alpha for the full scale (α = 0.517) was below the conventional threshold of 0.70, indicating low internal consistency and potential measurement error [25, 37]. Similarly, the caring (α = 0.318) and sharing (α = 0.439) subscales also exhibited inadequate internal consistency, suggesting that the items within each domain did not reliably measure a unified construct in this context. Although the Cronbach’s alpha value of the full PPOS scale (0.517), and its subscales (caring: α = 0.318; sharing: α = 0.439), is below acceptable limits, this aligns with previous validation studies in other languages, such as the Korean adaptation, where low reliability necessitated item removal [38]. Similarly, the reported Cronbach’s alpha for the Spanish PPOS was 0.65, which was attributed to cultural nuances in patient-centeredness [39]. This finding mirrors those reported by investigators in South Africa (α = 0.51), Saudi Arabia (α = 0.56), and Mali (α = 0.58) [8, 40, 41]. In Brazil, the reported alpha value of 0.605 was deemed acceptable despite it falling short of the commonly accepted standard of 0.70 as an indication of adequate internal consistency [42]. Other studies have attempted translation and adaptation but did not report on internal consistency [11, 17]. The low alpha values in the current study may reflect cultural differences in interpreting items or the heterogeneity of the student sample, which differs from the practicing clinicians in the original PPOS validation [29]. The results of validity testing performed in our study demonstrated that the factors extracted in the EFA did not support the scale’s original two-factor structure, namely ‘caring’ and ‘sharing’ dimensions. The results of EFA revealed six factors accounting for 50.74% of the variance. This contrasts with the original two-factor PPOS but parallels findings in multicultural validations. For example, a Brazilian study identified four factors, while investigators in Spain reported three [39, 43]. The CFA results further challenge the validity of the original two-factor (caring and sharing) structure in this population. While some absolute fit indices such as GFI (0.904) and RMSEA (0.069) met acceptable thresholds, the comparative fit indices (CFI = 0.550, TLI = 0.486, IFI = 0.560) were far below the recommended level of 0.90 [44]. This pattern suggests that the a priori two-factor model does not adequately reproduce the observed covariance structure in our data, reinforcing the need to reconsider the factorial structure of the scale. The emergence of six factors in this study, along with the poor CFA fit indices, may reflect finer cultural distinctions in patient-centeredness and further supports the notion that the original two-factor model may not be universally applicable. Taken together, these psychometric findings indicate that the use of the overall PPOS or subscale scores as reliable and valid measures of patient-centeredness in this population is not supported. They highlight the critical importance of re-validating and possibly revising instruments before applying them in new cultural and educational contexts. Based on these results, the fit statistics of the CFA indicate that the model may need to be adjusted to better capture the underlying structure of the PPOS scale. Our findings contrast those of studies that successfully replicated the two-factor structure, such as the Dutch validation [45]. However, similar discrepancies were observed in a Chinese study, where CFA also yielded poor fit, prompting authors to propose a modified three-factor model [46]. The same challenge arose in Brazil, Mali, and an additional study in China, where the two-factor structure could not be reliably reproduced [8, 42, 47]. The divergent results underscore the impact of cultural context on conceptualizing patient-centered care. For instance, hierarchical doctor-patient relationships in Gulf countries may influence students’ perceptions of sharing decision-making, reducing item coherence [48]. The recognized low performance of the A-PPOS in our experiments might have been caused by several factors. First, there is much ambiguity surrounding the exact meaning of patient-centeredness and the most accurate method for determining its efficacy. Establishing a common definition of “patient-centeredness” is essential to the construct validity of the PPOS. Numerous pertinent research has reiterated this [41, 49]. Second, there is the issue of the population’s perception of the items on the PPOS. It is possible that the PPOS’s adaptation to foreign context will not prove satisfactory due in large part to the disparate viewpoints on the patient-provider relationship and the incongruity of themes raised with values espoused [8, 41]. This is further affirmed by a lack of international agreement on what constitutes “good communication” as well as clear indications of varying emotional engagement depending on the local socio-cultural setting in which the study population is located [49]. Additionally, the two-factor distinction (caring and sharing) might not align with how patient-centeredness manifests in this setting, or some items may be cross-loaded or ambiguously phrased. Finally, the use of self-administered questionnaires in classroom settings may have introduced social desirability bias, influencing students to provide institutionally favorable responses. It is worth noting that the PPOS was used to assess the impact of communication skills courses on medical students’ orientation toward patient-centeredness in Sudan and Saudi Arabia [50, 51]. The PPOS tool was used before and after the communication skills courses. Results from these studies showed no significant change in the overall PPOS score. Once again, this raises the question of the suitability of the PPOS in measuring patient-centeredness in these contexts. The mean PPOS score of 3.90 ± 0.47 for the total cohort of medical students indicates that our sample of medical students has a modest level of patient-centeredness based on the original PPOS scoring. The mean PPOS score was greater than that of medical students in China (3.63), Nepal (3.7), and Pakistan (3.40) [12, 16, 52]. However, they were lower than those recorded among students in the USA (4.57) and Brazil (4.66), but equivalent to those found among students in Saudi Arabia (4.00) and Sudan (4.08) [11, 13, 53, 54]. This places our sample of students in a middle ground globally and regionally, potentially reflecting a blend of traditional and modern influences within the local healthcare context and medical education system. Females have outperformed males in all domains of the PPOS. This corresponds to what was reported by studies in America, Singapore, China, Greece, Sweden, Brazil, and Sudan [11, 13, 15–18, 54]. Gender-based differences in PPOS domains were not observed among students in Pakistan and Nepal [12, 52]. The tendency of females to exhibit more patient-centered inclinations has been attributed to their perceived superiority in communication [55]. Our medical students were moderately oriented towards the caring aspect of patient-centeredness and slightly less orientated towards the sharing aspect (4.2 and 3.59, respectively). There was no discernible trend in the mean PPOS or caring and sharing subscale scores throughout consecutive academic years. Pre-clinical and clinical status had no influence on patient-centeredness, which shows that students’ inclinations did not change with advancing study year. This was consistent with studies from Pakistan, Singapore and Sweden [12, 15, 18]. The lack of change across academic years may indicate that the curriculum does not adequately emphasize experimental or reflective learning components that foster patient-centered growth. This reinforces the need for curricula that foster patient-centeredness among medical students so as to ensure better healthcare delivery, patient satisfaction, adherence to treatment, and potentially improved therapeutic outcomes [56]. It is encouraging that our sample of medical students has not deviated from the idealistic mindset that prevails when students first enter medical school. This is typically caused by the psychological distress that medical students experience, which has been linked to growing stress levels brought on by heavy workloads and responsibilities, a loss of empathy, and ultimately a loss of patient-centeredness [57, 58]. This study has several limitations. First, we collected data from only one medical school, which naturally restricts the application of our findings to other institutions in Bahrain or the wider GCC region. Second, fewer fifth- and sixth-year students participated. These senior students were primarily engaged in clinical practice at hospitals, meaning our recruitment method favored the inclusion of pre-clinical students. Third, the cross-sectional design—a single snapshot in time—prevents us from analyzing how attitudes change within the same group of students over time. Fourth, given the A-PPOS’s low reliability and poor model fit, the resulting attitude scores require cautious interpretation. It is possible that the instrument’s translation, or the fundamental concept of patient-centeredness itself, doesn’t quite capture the intended meaning in this specific cultural setting. Finally, the PPOS measures self-reported attitudes, and these scores may not directly reflect actual behavior when students are treating patients. Conclusion This study investigated the translation and cultural adaptation of the PPOS into the Arabic language, in addition to the evaluation of patient-centered attitudes among medical students in Bahrain. Our data indicate that students at the AGU were moderately oriented towards patient-centeredness. Female students consistently scored better on the scale than their male colleagues, while advancing through the academic years of the medical program did not show a corresponding connection to any major change in these attitudes towards patient care. This suggests that simply moving through the established medical curriculum may not be adequate to cause a significant reshaping of their views on patient care. Psychometric analyses of the Arabic PPOS demonstrated suboptimal internal consistency, and the clustering of items did not replicate the factor structure of the original instrument. This implies that some items may not possess full cultural or contextual relevance for this particular student population. These observations highlight a compelling need for scale refinement, which will require a more rigorous and meticulous adaptation process. Subsequent academic endeavors should consider employing expert consensus methodologies, such as the Delphi technique, to correctly identify aspects of patient-centeredness that are most pertinent to the local culture and to modify or eliminate items that show poor statistical performance. Despite the identified limitations in the scale’s current form, this study provides essential preliminary insight into the nature and prevalence of patient-centered attitudes among future physicians in Bahrain. The conclusions affirm the vital necessity of reinforcing patient-centered principles within medical education and the ongoing mandate to develop reliable, culturally suitable assessment tools for these essential attitudes within Arabic-speaking contexts.