Received: 2 April 2020 | Revised: 18 June 2020 | Accepted: 5 September 2020 DOI: 10.1111/hex.13135 ORIG INAL RE S E ARCH PAPER Cross-cultural validation of the patient-practitioner orientation scale among primary care professionals in Spain Lilisbeth Perestelo-Pérez1,2,3 | Amado Rivero-Santana2,3,4 | Ana Isabel González-González2,5,6 | Carlos Jesús Bermejo-Caja7,8 | 2,4 2,9 Vanesa Ramos-García | Débora Koatz | Alezandra Torres-Castaño2,4 | Marta Ballester2,9 | Marcos Muñoz-Balsa8 | Yolanda del Rey-Granado8 | Francisco Javier Pérez-Rivas10 | Yolanda Canellas-Criado11 | Ana Belén Ramírez-Puerta8 | Valeria Pacheco-Huergo12 | Carola Orrego2,9 1 Evaluation Unit of the Canary Islands Health Service (SESCS), Tenerife, Spain 2 Health Services Research on Chronic Patients Network (REDISSEC), Tenerife, Spain 3 Centre for Biomedical Research of the Canary Islands (CIBICAN), Tenerife, Spain 4 Canary Islands Foundation and Institute for Health Research (FIISC), Tenerife, Spain 5 Primary Care Centre Vicente Muzas, Community of Madrid Health Service, Madrid, Spain 6 Institute of General Medicine, Johann Wolfgang Goethe University, Frankfurt-am-Main, Germany 7 Support Unit for Primary Care, Community of Madrid Health Service, Madrid, Spain 8 Nursing Department, Autonomus University of Madrid, Madrid, Spain 9 Avedis Donabedian Institute, Autonomous University of Barcelona, Barcelona, Spain 10 Nursing Department, Complutense University of Madrid, Madrid, Spain 11 Primary Care Centre Monóvar, Community of Madrid Health Service, Madrid, Spain 12 EAP Turó, Barcelona, Spain Correspondence Lilisbeth Perestelo-Pérez, Servicio de Abstract Evaluación y Planificación, Dirección Background: In recent decades, many self-report instruments have been developed del Servicio Canario de la Salud. Camino Candelaria s/n. C/S El Chorrillo. 38109. S/C to assess the extent to which patients want to be informed and involved in decisions de Tenerife. Spain. about their health as part of the concept of person-centred care (PCC). The main ob- Email: lilisbethp@gmail.com jective of this research was to translate, adapt and validate the Patient-Practitioner Funding information Orientation Scale (PPOS) using a sample of primary care health-care professionals in This study has been funded by Instituto de Salud Carlos III through the project Spain. PI15/00164, PI15/00566 and PI15/00586. Methods: Baseline analysis of PPOS scores for 321 primary care professionals (gen- It was co-funded by European Regional Development Fund, (ERDF) ‘A way of eral practitioners and nurses) from 63 centres and 3 Spanish regions participating in shaping Europe’. a randomized controlled trial. We analysed missing values, distributions and descrip- tive statistics, item-to-scale correlations and internal consistency. Performed were confirmatory factor analysis (CFA) of the 2-factor model (sharing and caring dimen- sions), scale depuration and principal component analysis (PCA). This is an open access article under the terms of the Creative Commons Attribution License, which permits use, distribution and reproduction in any medium, provided the original work is properly cited. © 2020 The Authors Health Expectations published by John Wiley & Sons Ltd Health Expectations. 2021;24:33–41.  wileyonlinelibrary.com/journal/hex | 33 34 | PERESTELO-PÉREZ et al. Results: Low inter-item correlations were observed, and the CFA 2-factor model only obtained a good fit to the data after excluding 8 items. Internal consistency of the 10-item PPOS was acceptable (0.77), but low for individual subscales (0.70 and 0.55). PCA results suggest a possible 3-factor structure. Participants showed a patient-ori- ented style (mean = 4.46, SD = 0.73), with higher scores for caring than sharing. Conclusion: Although the 2-factor model obtained empirical support, measurement indicators of the PPOS (caring dimension) could be improved. Spanish primary care health-care professionals overall show a patient-oriented attitude, although less marked in issues such as patients’ need for and management of medical information. KEYWORDS health-care professionals, person-centred care, PPOS, primary care, validation 1 | I NTRO D U C TI O N develop certain communication and social skills—and an obvious prerequisite is to hold favourable beliefs and attitudes towards this Person-centred care (PCC) is currently advocated as the gold standard model of care. Consequently, assessing these attitudes becomes a of health care, as opposed to a disease-centred, paternalistic style.1 relevant issue in the research and implementation of PCC in routine PCC implies an egalitarian relationship between health service users practice and in academic curricula. While a number of instruments and professionals, in which patient values and preferences about their have been developed for that purpose, this is lesser extent than for health care are placed at the core of the decision-making process patients.14-16 about diagnostic, therapeutic and lifestyle modification procedures. One such instrument is the Patient-Practitioner Orientation Scale PCC aims to promote patient empowerment and involvement in the (PPOS), developed in the United States by Krupat et al.17 It consists of self-management of their diseases, in a collaborative work of shared 18 items grouped in 2 subscales of 9 items each: sharing assesses atti- decision making with their health-care providers, while respecting tudes about whether health-care professionals should share informa- their autonomy and personal values. Establishing an emphatic and tion, decisions and power with their patients, while caring assesses the trust-based relationship between patients and professionals and fa- degree to which professionals should show empathy and warmth and cilitating high-quality communications are thus necessary requisites treat patients as whole persons. Higher scores in the PPOS have been for PCC to take hold.2 PCC relies not only on ethical arguments about shown to be associated with more patient-centred behaviours in con- people's rights to autonomy and personal independence in a demo- sultations,18 while congruence between patients and physicians’ orien- cratic society, but also on its potential benefits for health and the sus- tations has been demonstrated to be associated with greater patient tainability of health systems. It is expected that more active, informed satisfaction,19-21 trust and endorsement of physicians,19 and fewer re- and empowered patients will be able to achieve better self-manage- ferrals.22 The PPOS, which has been translated from English to several ment of their conditions and will improve adherence to therapeutic other languages,23-25 has been widely used to assess preferences for plans discussed and agreed with their health-care providers. As well patient orientation among health-care professionals, medical students as improving coordination and continuity of services, this could result and even patients, since it is applicable to the general population.26-30 3-5 in improved health outcomes and greater resource use efficiency. However, few studies have reported the psychometric properties of In the last decades, many self-report instruments have been the instrument or, more specifically, its factorial validation.31-33 developed to assess the extent to which patients want to be in- Since (to our knowledge) no studies have been published on use formed and involved in decisions about their care, the most widely of the PPOS in Spain, our aim was to translate, adapt and carry out used of which are the Control Preference Scale6 and the Autonomy a psychometric validation of this scale using a sample of health-care Preference Index.7 Research has shown that most patients desire professionals. a collaborative or autonomous role in their medical decisions and that many of them do not feel as involved as they would want.3,8 Furthermore, less perceived involvement or a mismatch between 2 | M E TH O DS preferred and experienced involvement has been shown to be related to poorer satisfaction, adherence and quality of life.9-13 Although not This study analyses PPOS baseline data for participants, who were all variations in patients’ perceptions of being involved are caused by recruited during 2016, for a cluster randomized controlled trial an actual deficit in health-care professionals’ behaviour, the above that aimed to assess the impact of a virtual community practice results clearly reinforce the need to enhance PCC and improve pa- intervention on health-care professionals (general practitioners tient participation. To achieve this aim, health-care providers must and nurses) at the primary care level. 34 Primary care centres from PERESTELO-PÉREZ et al. | 35 3 Spanish autonomous communities (Canary Islands, Catalonia instrument and the semantic and content equivalence of the and Madrid) were contacted via their managers and invited to par- Spanish version with the English original. ticipate. Centres were randomly selected, while a balanced north/ 4. To ensure that meaning accurately reflected the English original, south geographical representation was maintained within each the Spanish version was back-translated by a different pair of bi- region. In-person meetings were held in each centre to explain lingual translators working blind to the original English version. the study in detail to interested professionals. Those who agreed 5. In a final equivalence testing step, the back-translation was com- to participate signed the informed consent and received a pass- pared with the original instrument by the study directors in Spain word to access a web interface where they could fill out the PPOS (LPP, AIGG, CJBG and CO) and, after some minor revisions, the questionnaire. Participants’ allocation to the intervention group Spanish version was considered ready to use. or control group was only disclosed after the questionnaire was completed. The final Spanish version of the PPOS was pre-tested on first twelve adult patients attended at the two primary care centre par- ticipants in this study and their responses were analysed to identify 2.1 | Measures whether any modifications were necessary, which resulted not to be the case. 2.1.1 | Patient-practitioner orientation scale This 18-item scale measures the orientations of patients and health- 2.4 | Statistical analyses care professionals regarding the patient-practitioner relationship. The scale is scored on a 1-6 Likert scale (totally disagree-totally The distribution and descriptive statistics for the items (missing agree). Items, except for 9, 13 and 17, are written in a physician-ori- values, frequencies, means, standard deviations, asymmetry and ented style; scoring is therefore reversed in such a way that a higher kurtosis) were calculated. Floor and ceiling effects for each item score indicates a patient-oriented style. Scores for the overall scale were defined as more than 85% of participants scoring 1 (totally (18 items) and sharing and caring subscales (9 items each) are divided disagree) and 6 (totally agree), respectively. Also calculated were by their corresponding number of items and thus range between 1 the mean inter-item correlations, corrected item-to-scale cor- and 6. relations and Cronbach's α after excluding each item. In order to assess whether the data fit the 2-factor model proposed for the scale (sharing and sharing), a confirmatory factor analysis 2.2 | Sociodemographic and professional data (CFA) was performed. Missing values were handled by using full information maximum likelihood estimation, which does not re- Data were collected on age, sex, profession (general practitioner or quire the imputation of missing values, but uses all the available nurse), years’ experience, tutorship of medical residents/medical or data to estimate population parameters. 36 However, in the pres- nursing students (yes/no), and patients attended per day. ence of non-normal data this technique can produce negatively biased standard errors, leading to an erroneous rejection of the null hypothesis. For this reason, we first assessed non-normality 2.3 | Questionnaire translation by means of Yuan, Lambert and Fouladi's normalized estimate of multivariate kurtosis, applicable to data with missing values 37 (a As the methodological model for Spanish translation of the PPOS, value outside the range −3,3 is indicative of multivariate kurtosis). we used the guidelines on cross-cultural adaptation of self-re- In the case of non-normality, standard errors were corrected using ported measure developed by Beaton et al, 35 based on five steps the robust method proposed by Yuan and Bentler. 38 as follows: The model was refined by repeating the analysis after exclud- ing items with non-significant coefficients or low R 2 values. Its fit 1. A pair of bilingual translators, competent in both English and was then assessed by means of the chi-square test (or the Yuan- Spanish, independently translated the original questionnaire Bentler's scaled chi-square,38 in the case of performing robust from English to Spanish. estimation); since this statistic is very sensitive to sample size, we 2. Working with the original questionnaire and both translation ver- calculated several other recommended goodness-of-fit indices39-41: sions, the translators synthesized the translation after reaching a χ2/df ratio, root mean square error of approximation (RMSEA) with a consensus on the translation of words, phrases and items. 90% confidence interval (CI), Tucker-Lewis index (TLI), comparative 3. Five primary care physicians and nurses independently tested the fit index (CFI) and Bollen's incremental fit index (IFI). We considered cultural appropriateness, representativeness and content validity the following thresholds for acceptable and good values, respec- of the translated instrument, rating the degree that each item re- tively39-41: under 3 and 2 for the χ2/df ratio; under 0.08 and 0.05 for flected the concept that it was intended to measure. The same the RMSEA; and above 0.90 and 0.95 for the TLI, CFI and IFI. If the professionals also rated the understandability of the translated model did not obtain an acceptable fit even after being refined, we 36 | PERESTELO-PÉREZ et al. carried out a principal component analysis (PCA) to explore other #9, written in patient-oriented terms (Patients should be treated as potential factorial models. if they were partners with the doctor, equal in power and status), disa- Finally, we assessed associations of the obtained scales with par- greement was high (80%). Favourable or less critical responses with ticipants’ sociodemographic and professional characteristics. Due to the physician-oriented style were obtained for 4 items, specifically, the clustered nature of the study design (ie professionals clustered item #5 (Patients should rely on their doctors’ knowledge and not try to into centres), we used a 2-level mixed multiple regression model, with find out their conditions on their own), item #8 (Many patients continue fixed effects for professionals (level 1) and random effects for centre asking questions even though they are not learning anything new), item (level 2), to adjust for correlated observations within the clusters. #10 (Patients generally want reassurance rather than information about Analyses were performed with SPSS 21.0 and EQS 6.2 software. their health) and item #18 (When patients look up medical information on their own, this usually confuses more than it helps). 3 | R E S U LT S 3.2 | Dimensionality and internal consistency Contacted were 113 primary care centres, 9 and 41 of which de- clined participation and failed to respond, respectively, leaving 63 Mean inter-item correlation was 0.15 (median 0.14). The 3 items writ- centres to be included (25 in the Canary Islands, 18 in Catalonia ten in a patient-oriented style (#9, #13 and #17) obtained the lowest and 20 in Madrid). These contributed 321 health-care professionals mean correlations (0.08, 0.07 and 0.09, respectively). Table 3 shows (mean 5.1, range 1-28). Table 1 shows sociodemographic and profes- the item-to-scale correlations and Cronbach's alphas after excluding sional characteristics of the professionals. Mean age was 47.7 years each item. The 3 mentioned items showed the lowest correlations (SD 8.8), and 76% were female. Over half (59%) were general practi- (under 0.14). The remaining value ranges were 0.26-0.51 (sharing), tioners, and the remaining 41% were nurses. Mean years’ experience 0.17-0.38 (caring) and 0.18-0.48 (overall). Excluding items #9, #13 was 22.0 (SD 8.84), mostly in primary care (mean 17.7; SD 8.92). The and #17 increased alphas from 0.72 to 0.77 (overall), from 0.66 to professionals attended a mean of 29 patients per day (SD 11.3), and 0.72 (sharing) and from 0.48 to 0.56 (caring); these items were there- 25.5% had tutored residents/medical and nursing students. fore excluded from subsequent analyses. We carried out a CFA for the 2-(correlated) factor model. We used the maximum likelihood (with missing values) estimation 3.1 | Item analyses method, with robust standard errors due to the non-normal distri- bution of the data (multivariate kurtosis normalized estimate 25.5). For the 18 items, 25 participants (7.8%) missed between 1 and 3 All items obtained significant coefficients and, as shown in Table 4, items (20, 2 and 3 participants missed 1, 2 and 3 items, respectively). the χ2/df ratio (2.3) and RMSEA values (0.063, 90% CI: 0.052-0.075) Item #8 had 15 missing values (4.7%), whereas 9 more items had were acceptable, while the TLI, CFI and IFI values were unsatis- between 1 and 4 missing values (Table 2). There were no ceiling factory. We repeated the analysis excluding items #14, #15 and or floor effects for any item. Distributions were asymmetric, with #16, with R 2 values under 0.15; however, the improvement did not most participants stating some level of disagreement (slightly, mod- achieve the acceptability thresholds. Excluding 2 more items (#11 erately or totally disagreed) with the physician-oriented style. Item and #18, thus leaving 6 items for sharing and 4 for caring) yielded acceptable values for the χ2/df ratio (2.0), RMSEA (0.056, 90% CI: TA B L E 1 Sociodemographic and professional characteristics of 0.035-0.075), CFI and IFI (0.92 for both), while bringing TLI to the participants limit of that threshold (0.89). When errors for items #6 and #7 were allowed to covariate (based on their significant standardized residual N = 321 covariance), all indices except the chi-square indicated a good fit, Age (mean, SD) 47.7 (8.79) while the TLI was acceptable (0.93). A one-factor model retaining Age those 10 items yielded minimal differences with the 2-factor solu- 26-35 24 (7.5%) tion (Table 4). 36-45 114 (35.5%) The CFA only showed a good fit of the data after exclusion of 8 46-55 105 (32.7%) items: 3 sharing items (#9, Patients should be treated as if they were 56-66 78 (24.3%) partners with the doctor, equal in power and status, #15, The patient Female 243 (75.7%) must always be aware that the doctor is in charge and #18, When pa- tients look up medical information on their own, this usually confuses Physicians/nurses 190 (59.2%)/131 (40.8%) more than it helps) and 5 caring items (#11, If a doctor's primary tools Years’ experience 22.0 (8.84) are being open and warm, the doctor will not have a lot of success, #13, Years in primary care 17.7 (8.92) A treatment plan cannot succeed if it is in conflict with a patient's life- Tutor 82 (25.5%) style or values, #14, Most patients want to get in and out of the doc- Patients/day 29.4 (11.3) tor's office as quickly as possible, #16, It is not that important to know PERESTELO-PÉREZ et al. | 37 TA B L E 2 Missing values, distribution of responses and descriptive statistics of PPOS items Missing A/Da Meanb (%) (%) (SD) Asym. Kurt. 1. The doctor is the one who should decide what gets talked about during a 0 30/70 4.37 (1.4) 0.498 −0.915 visit. 2. Although health care is less personal these days, this is a small price to pay 0.3 24/76 4.63 (1.4) 0.780 −0.529 for medical advances. 3. The most important part of the standard medical visit is the physical exam. 0 31/69 4,12 (1.3) 0.413 −0.655 4. It is often best for patients if they do not have a full explanation of their 0 11/89 5.11 (1.2) 1.441 1.653 medical condition. 5. Patients should rely on their doctors’ knowledge and not try to find out 0 36/64 3.93 (1.5) 0.366 −0.957 their conditions on their own. 6. When doctors ask a lot of questions about a patient's background, they are 0 3/ 97 5.59 (0.8) 2.876 10.14 prying too much into personal matters. 7. If doctors are truly good at diagnosis and treatment, the way they relate to 0 4/96 5.54 (0.9) 2.578 7.665 patients is not that important. 8. Many patients continue asking questions even though they are not learning 4.7 52/48 3.51 (1.4) −0.035 −0.840 anything new. 9. Patients should be treated as if they were partners with the doctor, equal in 0.6 20/80 2.32 (1.4) 0.966 −0.060 power and status. 10. Patients generally want reassurance rather than information about their 0.9 44/56 3.67 (1.3) 0.051 −0.965 health. 11. If a doctor's primary tools are being open and warm, the doctor will not 0.6 13/87 4.89 (1.2) 1.089 0.644 have a lot of success. 12. When patients disagree with their doctor, this is a sign that the doctor 0.9 27/73 4.29 (1.4) 0.589 −0.610 does not have the patient's respect and trust. 13. A treatment plan cannot succeed if it is in conflict with a patient's lifestyle 0.3 72/28 4.40 (1.6) −0.746 −0.655 or values. 14. Most patients want to get in and out of the doctor's office as quickly as 0 21/79 4.51 (1.4) 0.909 −0.042 possible. 15. The patient must always be aware that the doctor is in charge. 0.3 12/88 5.10 (1.2) 1.420 1.454 16. It is not that important to know a patient's culture and background in 0 4/96 5.54 (0.9) 2.939 9.710 order to treat the person's illness. 17. Humor is a major ingredient in the doctor's treatment of the patient. 1.2 86/14 4.66 (1.2) −0.893 0.683 18. When patients look up medical information on their own, this usually 0.3 67/33 3.02 (1.3) −0.343 −0.478 confuses more than it helps. Note: Asym., asymmetry; Kurt., kurtosis. a Agree/disagree, collapsing the 3 categories for each (slightly, moderately, strongly). b Higher scores indicate more patient-oriented style (reversed scores for all items except #9, #13, #17). a patient's culture and background in order to treat the person's illness favouring technical aspects of medicine and an asymmetric relation- and #17, Humour is a major ingredient in the doctor's treatment of the ship between patient and professional, respectively. Items #4, #11 patient). The resulting scales had Cronbach's alphas of 0.70 (sharing), and #14 saturated above 0.30 in 2 or 3 components. When 2 com- 0.55 (caring) and 0.77 (overall) and means (SDs) of 4.12 (0.88), 4.97 ponents were extracted, the above-mentioned first and third compo- (0.74) and 4.46 (0.73), respectively. nents collapsed into a single dimension (not shown in the table). We performed several PCA with varimax rotation to explore other potential latent structures, including only the 15 items phrased in a physician-oriented style. The analysis yielded 4 components with ei- 3.3 | Associations with sociodemographic and genvalues greater than one. When a fifth factor was extracted, only professional variables a single item showed high loading (#14). In the 4-factor solution, the last factor included only 2 items with high loadings (#16, #15), not Multilevel mixed regression models did not point to any significant clearly related semantically speaking. The 3-factor solution is shown association between the 10-item PPOS overall or its subscales and in Table 5. The first, second and third components include 5 items on age, sex, profession, years’ experience, tutorship, or patients at- information, 6 items on the patient-physician relationship and 4 items tended per day. 38 | PERESTELO-PÉREZ et al. TA B L E 3 Item-to-scale correlations Sharing (α = 0.660) Caring (α = 0.477) Total (α = 0.722) and PPOS overall and subscale internal Item to α without Item to α without Item to α without consistency Items scale item scale item scale item #1 0.41 0.615 0.45 0.694 #4 0.39 0.623 0.45 0.698 #5 0.51 0.587 0.48 0.689 #8 0.44 0.608 0.41 0.699 #9 −0.08 0.722 −0.03 0.743 #10 0.44 0.611 0.46 0.695 #12 0.37 0.625 0.37 0.703 #15 0.26 0.648 0.29 0.711 #18 0.35 0.631 0.31 0.710 #2 0.27 0.418 0.37 0.703 #3 0.26 0.425 0.36 0.704 #6 0.33 0.425 0.35 0.710 #7 0.38 0.405 0.41 0.705 #11 0.24 0.434 0.25 0.715 #13 0.10 0.499 0.13 0.731 #14 0.19 0.457 0.18 0.723 #16 0.17 0.458 0.24 0.715 #17 0.04 0.501 0.11 0.726 TA B L E 4 Fit statistics from PPOS confirmatory factor analysis N = 321 Χ2; df (p)a Χ2/df RMSEA (90% CI) TLI CFI IFI 2-factor (15 items) 202.6; 87 (<0.001) 2.3 0.063 (0.052 − 0.075) 0.777 0.815 0.821 2-factor (12 items) 128.6; 51 (<0.001) 2.5 0.068 (0.053 − 0.083) 0.815 0.857 0.861 2-factor (10 items) 65.5; 32 (<0.001) 2.0 0.056 (0.035 − 0.075) 0.888 0.920 0.923 b 2-factor modified (10 50.3; 31 (0.016) 1.6 0.043 (0.017 − 0.064) 0.934 0.955 0.956 items) One-factor (10 items) 70.3; 34 (<0.001) 2.1 0.056 (0.037 − 0.075) 0.886 0.914 0.916 b One-factor modified 53.4; 33 (0.014) 1.6 0.042 (0.018 − 0.063) 0.935 0.952 0.954 (10 items) Note: CFI, comparative fix index; df, degrees of freedom; IFI, incremental fit index; RMSEA, root mean square error of approximation; TLI, Tucker- Lewis fit index. a Yuan-Bentler's scaled chi-squared statistic. b Allowing the covariance of errors of items 6 and 7. 4 | D I S CU S S I O N respectively, to sharing and caring. The main difference between the CFA and PCA results was the functioning of items #2 (Although health For our sample of Spanish health-care professionals in the primary care is less personal these days, this is a small price to pay for medical care sector, the proposed structure of the PPOS with 2 correlated advances) and #3 (The most important part of the standard medical visit factors (representing the sharing and caring dimensions) only obtained is the physical exam). These items were retained in the caring 4-item an acceptable fit to the data after excluding 8 items (almost half of the subscale in the CFA; in the PCA, they formed the third component total scale with 18 items). Inter-item correlations were observed to be in the 3-factor solution and the first component in the 2-factor solu- low overall; the 3 items phrased in a patient-oriented style were the tion, thus seeming to some extent to be independent of the relational most uncorrelated, pointing out to a possible methodological effect aspects of the patient-physician interaction represented by C2. From 31,32 that has also been observed in other studies for 2 of these items. this perspective, the items may be mere indicators of the importance The PCA also pointed to poor functioning of several of the remaining attributed to technical aspects of medicine and, therefore, may not items, which loaded similarly on more than a single component. Of necessarily be incompatible with the interpersonal or socio-affective the extracted components, 2 (C1 and C2) can be partially assimilated, aspects of health care as represented by caring. PERESTELO-PÉREZ et al. | 39 TA B L E 5 Principal component analysis of the PPOS items identified in those studies from different countries around the world (excluding items #9, #13 and #17) are similar to those found by us, suggesting that our results are not C1 C2 C3 h2 sample-specific but due in fact to non-optimal functioning of the questionnaire. #18 0.77 −0.10 −0.06 0.37 Regarding the scores obtained, most participants showed a #5 0.68 0.15 0.21 0.50 patient-oriented style, with mean values for the 10-item version #8 0.58 −0.00 0.30 0.45 similar to or higher than observed for other published samples of #10 0.53 0.18 0.23 0.44 health-care professionals. The caring dimension obtained a higher #12 0.47 0.19 0.23 0.52 score, indicating a greater preference for the socio-affective side #16 0.09 0.72 −0.15 0.53 of PCC than for sharing information and power. This result, while #6 0.05 0.70 0.20 0.52 corroborated by several studies, 29,31,32,44,45 could depend on the #7 −0.01 0.55 0.47 0.43 cultural background of professionals or their medical specialty.46,47 #15 0.12 0.53 0.15 0.36 In our sample, sharing subscale items that obtained more physi- #14 0.34 0.41 −0.32 0.31 cian-oriented responses were related to the patients’ need for #11 −0.02 0.39 0.38 0.32 information and searches for themselves. These results may indi- #2 0.24 0.07 0.66 0.55 cate professionals’ concerns about patients’ exposure to unreliable information, a risk that is not negligible nowadays with the great #3 0.21 −0.05 0.64 0.61 amount of information available online, but also reflects a possible #1 0.28 0.22 0.49 0.39 underestimation of patients’ real needs for accurate information #4 0.33 0.33 0.47 0.30 about their conditions (and not just reassurance). Explained 15.5% 14.5% 13.9% We found no significant associations for the overall or subscale variance scores with sociodemographic or professional variables. Studies that Cronbach's α 0.68 0.60* 0.63 have used the PPOS to identify predictors of patient orientation Note: Kaiser-Meyer-Olkin 0.82; Bartlett's sphericity test: χ2 = 799.2, in health-care professionals mostly observed no significant asso- P < .001. ciations with age or sex, 26,31,48,49 although a Greek study did find Loadings above 0.30 in bold. * that younger physicians were more patient-oriented in the sharing Items #16, #6, #7, #15. dimension.50 As for differences between physicians and nurses, like us, Zhumadilova et al30 observed no significant differences in Few studies in recent years have reported on the psychometric Kazakhstan, whereas Chan et al51 found higher scores for physicians properties of the PPOS since its initial development and psycho- in Malaysia. Given the wide geographic distribution of countries metric testing, mainly focused on criterion validity.19,20,26 Those in which the PPOS has been used, future studies should compare studies have also pointed to certain psychometric limitations of the scores and correlates of the PPOS for different countries and geo- instrument. Wang et al31 in reporting results with a mixed sample of graphic regions, since differences in cultural backgrounds and health physicians (20%) and patients in China, documented a very poor fit systems may be influencing the attitudes of health-care profession- to the data in a CFA with the 18 items; depuration of the scales re- als to PCC and its correlates. tained 11 items, 6 for sharing and 5 for caring, although the excluded Assessing these attitudes is an important issue for research, items were not the same as in our study. In a German study, the scale clinical practice and educational purposes, since PCC cannot be was also reduced, in that case to 12 items.42 In a study conducted successfully implemented if health-care professionals are not well in Mali, Hurley et al (2018) 43 also obtained a very poor model fit for disposed to the principles of this model of care. Consequently, the the 18 items for a sample of medical students, while depuration of availability of measurement instruments with good psychometric the scale by means of exploratory factor analysis did not identify properties is a basic requisite to be able to appropriately assess the 33 an interpretable structure that fit well to the data. Pereira et al in level of patient-centredness in our health systems, identify areas for a validation study in Brazil, observed low internal consistency val- improvement and evaluate the effect of interventions to promote ues for the overall score (0.61) and for the sharing (0.49) and caring PCC and shared decision making. This study adds new evidence on (0.46) scores for a sample of medical students/residents (33%) and the psychometric characteristics of the PPOS, an instrument widely patients; a CFA with the 18 items yielded unsatisfactory values for used worldwide. the CFI (0.84) and TLI (0.81), although the χ2/df ratio, RMSEA and SRMR values were acceptable. Another study with medical stu- dents/health-care professionals (29%) and patients in Sri Lanka, al- 5 | LI M ITATI O N S though it did not analyse the factor structure of the scale, found low internal consistency values (0.50-0.63) and item-to-scale correlation A limitation of this study is that, although the sample was not small, values.32 Although the comparisons with our study are not straight- it was not large enough to be randomly split into 2 subsamples: forward due to differences in samples, the psychometric limitations to explore the latent structure and to act as a validation sample 40 | PERESTELO-PÉREZ et al. (recommended rules of thumb such as a ratio of items/participant of REFERENCES at least 1/10 or a minimum of 200 participants would not have been 1. Santana MJ, Manalili K, Jolley RJ, Zelinsky S, Quan H, Lu M. How met). Another possible limitation is that we did not impute missed to practice person-centred care: A conceptual framework. Heal Expect. 2018;21(2):429-440. data. However, the rate of missing values was low and we used ap- 2. Care P-C. A definition and essential elements. J Am Geriatr. propriate techniques to deal with them. Yet another limitation is 2016;64(1):15-18. that we did not apply empirical methods to determine the number 3. Singh JA, Sloan JA, Atherton PJ, et al. Preferred roles in treatment of components to extract in the PCA; nonetheless, the inspection decision making among patients with cancer: a pooled analysis of studies using the Control Preferences Scale. Am J Manag Care. of the different alternatives showed that solutions based on 2 or 3 2010;16(9):688-696. components were the most plausible, given the low number of items 4. Meterko M, Wright S, Lin H, Lowy E, Cleary PD. Mortality among included and the low factor loadings in the fourth and fifth extracted patients with acute myocardial infarction: the influences of pa- components. tient-centered care and evidence-based medicine. Health Serv Res. 2010;45(5p1):1188-1204. 5. Chiang C-Y, Choi K-C, Ho K-M, Yu S-F. Effectiveness of nurse-led patient-centered care behavioral risk modification on secondary 6 | CO N C LU S I O N prevention of coronary heart disease: A systematic review. Int J Nurs Stud. 2018;84:28-39. The similarity of our results with those of recent validation studies 6. Degner LF, Sloan JA, Venkatesh P. The Control Preferences Scale. Can J Nurs Res. 1997;29(3):21-43. supports the conclusion that the PPOS has psychometric limitations. 7. Ende J, Kazis L, Ash A, Moskowitz MA. Measuring patients’ desire Although the 2-factor dimensions of sharing and caring obtained sup- for autonomy. J Gen Int Med. 1989;4(1):23-30. port from the data, measurement indicators could be improved, es- 8. Lechner S, Herzog W, Boehlen F, et al. Control preferences in treat- pecially those of the caring factor. ment decisions among older adults — Results of a large popula- tion-based study. J Psychosom Res. 2016;86:28-33. 9. Vogel BA, Leonhart R, Helmes AW. Communication matters: The AC K N OW L E D G E M E N T S impact of communication and participation in decision making on We thank health-care professionals who participated in this study breast cancer patients’ depression and quality of life. Patient Educ for their dedication and contributions to testing the PPOS. Couns. 2009;77(3):391-397. 10. De las Cuevas C, De Rivera L, Peñate W. To what extent is treat- ment adherence of psychiatric patients influenced by their par- C O N FL I C T O F I N T E R E S T ticipation in shared decision making? Patient Prefer Adherence. The authors state that there is no conflict of interests. 2014;8:1547-1553. 11. Schinkel S, Schouten BC, Street RL, van den Putte B, van Weert JCM. Enhancing health communication outcomes among ethnic AU T H O R C O N T R I B U T I O N S minority patients: the effects of the match between participation LPP conceived the study, participated in its design, is the PI of the preferences and perceptions and doctor-patient concordance. J Canary Islands region and collaborated in writing this manuscript. Health Commun. 2016;21(12):1251-1259. AIGG conceived the study, participated in its design, is the PI for 12. Atherton PJ, Smith T, Singh JA, et al. The relation between cancer patient treatment decision-making roles and quality of life. Cancer the Madrid region and wrote up the final manuscript. CO conceived 2013;119(12):2342-2349. the study, participated in its design, coordinates the trial, is the prin- 13. Ghane A, Huynh HP, Andrews SE, Legg AM, Tabuenca A, Sweeny K. cipal investigator (PI) for the Catalonia region and collaborated in The relative importance of patients’ decisional control preferences writing this manuscript. CBCJ conceived the study, participated in and experiences. Psychol Health. 2014;29(10):1105-1118. its design, is co-PI for the Madrid region and collaborated in writing 14. Puschner B, Neumann P, Jordan H, et al. Development and psy- chometric properties of a five-language multiperspective instru- this manuscript. LPP and ARS analysed data and wrote drew up the ment to assess clinical decision making style in the treatment initial manuscript draft. VRG, DK, ATC, MB, MMB, YRG, YCC and of people with severe mental illness (CDMS). BMC Psychiatry. ABRP contributed to writing the manuscript. All authors made sub- 2013;13(1):48. stantial contributions to the revising of the manuscript and approved 15. Knapp P, Raynor DK, Thistlethwaite JE, Jones MB. A questionnaire to measure health practitioners’ attitudes to partnership in medi- the final version. cine taking: LATCon II. Heal Expect. 2009;12(2):175-186. 16. Pollard S, Bansback N, Bryan S. Physician attitudes toward DATA AVA I L A B I L I T Y S TAT E M E N T shared decision making: A systematic review. Patient Educ Couns. All data generated or analysed during this study are included in this 2015;98(9):1046-1057. 17. Krupat E, Hiam CM, Fleming MZ, Freeman P. Patient-centeredness published article. and its correlates among first year medical students. Int J Psychiatry Med. 1999;29(3):347-356. ORCID 18. Shaw WS, Woiszwillo MJ, Krupat E. Further validation of the Lilisbeth Perestelo-Pérez https://orcid. Patient-Practitioner Orientation Scale (PPOS) from recorded visits for back pain. Patient Educ Couns. 2012;89(2):288-291. org/0000-0002-6113-7313 19. Krupat E, Bell RA, Kravitz RL, Thom D, Azari R. When physicians Carlos Jesús Bermejo-Caja https://orcid. and patients think alike: patient-centered beliefs and their impact org/0000-0003-1425-6691 on satisfaction and trust. J Fam Pract. 2001;50(12):1057-1062. PERESTELO-PÉREZ et al. | 41 20. Krupat E, Yeager CM, Putnam S. Patient role orientations, doctor-pa- 38. Yuan K-H, Bentler PM. 5. Three likelihood-based methods for mean tient fit, and visit satisfaction. Psychol Health. 2000;15(5):707-719. and covariance structure analysis with nonnormal missing data. 21. Chan CMH, Azman WA. Attitudes and role orientations on doc- Sociol Methodol. 2000;30(1):165-200. tor-patient fit and patient satisfaction in cancer care. Singapore Med 39. Hu L, Bentler PM. Cutoff criteria for fit indexes in covariance struc- J. 2012;53(1):52-56. ture analysis: Conventional criteria versus new alternatives. Struct 22. Carlsen B, Aakvik A, Norheim OF. Variation in practice: A ques- Equ Model A Multidiscip J. 1999;6(1):1-55. tionnaire survey of how congruence in attitudes between doc- 40. Fidell LSTB. Using Multivariate Statistics. New York: New York tors and patients influences referral decisions. Med Decis Mak. Harper and Row; 2006. 2008;28(2):262-268. 41. Schermelleh-Engel K, Müller H. Evaluating the fit of structural 23. Tsimtsiou Z, Kerasidou O, Efstathiou N, Papaharitou S, equation models: Tests of significance and descriptive goodness- Hatzimouratidis K, Hatzichristou D. Medical students’ attitudes of-fit measures. Methods Psychol Res. 2003;8(2):28-74. toward patient-centred care: a longitudinal survey. Med Educ. 42. Kiessling C, Fabry G, Rudolf Fischer M, Steiner C, Langewitz W. 2007;41(2):146-153. Deutsche Übersetzung und Konstruktvalidierung des “Patient- 24. Haidet P, Dains JE, Paterniti DA, et al. Medical student at- Provider-Orientation Scale” (PPOS-D12). Psychother Psychosom titudes toward the doctor-patient relationship. Med Educ. Med Psychol. 2013;64(3/4):122-127. 2002;36(6):568-574. 43. Hurley EA, Doumbia S, Kennedy CE, et al. Patient-centred attitudes 25. Carlsen B, Aakvik A. Patient involvement in clinical decision mak- among medical students in Mali, West Africa: a cross-sectional ing: the effect of GP attitude on patient satisfaction. Heal Expect. study. BMJ Open. 2018;8(1):e019224. 2006;9(2):148-157. 44. Ahmad W, Ashraf H, Talat A, et al. Association of burnout with doc- 26. Street RL, Krupat E, Bell RA, Kravitz RL, Haidet P. Beliefs about con- tor–patient relationship and common stressors among postgradu- trol in the physician-patient relationship: effect on communication ate trainees and house officers in Lahore—a cross-sectional study. in medical encounters. J Gen Intern Med. 2003;18(8):609-616. PeerJ. 2018;6:e5519. 27. Ting X, Yong B, Yin L, Mi T. Patient perception and the barriers to 45. Akkafi M, Sajadi HS, Sajadi ZS, Krupat E. Attitudes toward pa- practicing patient-centered communication: A survey and in-depth tient-centered care in the mental care services in Isfahan. Iran. interview of Chinese patients and physicians. Patient Educ Couns. Community Ment Health J. 2019;55(3):548-552. 2016;99(3):364-369. 46. Laplante-Lévesque A, Hickson L, Grenness C. An Australian survey 28. Wahlqvist M, Gunnarsson RK, Dahlgren G, Nordgren S. Patient-centred of audiologists’ preferences for patient-centredness. Int J Audiol. attitudes among medical students: Gender and work experience in 2014;53(sup1):S76-S82. health care make a difference. Med Teach. 2010;32(4):e191-e198. 47. Manchaiah V, Gomersall PA, Tomé D, Ahmadi T, Krishna R. 29. Ishikawa H, Son D, Eto M, Kitamura K, Kiuchi T. Changes in pa- Audiologists’ preferences for patient-centredness: a cross-sec- tient-centered attitude and confidence in communicating with pa- tional questionnaire study of cross-cultural differences and simi- tients: a longitudinal study of resident physicians. BMC Med Educ. larities among professionals in Portugal, India and Iran. BMJ Open. 2018;18(1):20. 2014;4(10):e005915. 30. Zhumadilova A, Craig BJ, Bobak M. Patient-centered beliefs among 48. Zhumadilova A, Craig BJ, Tsoy A, Gabdrakhmanova A, Bobak M. patients and providers in Kazakhstan. Ochsner J. 2018;18(1):46-52. Evaluation of communication between physicians and patients in 31. Wang J, Zou R, Fu H, Qian H, Yan Y, Wang F. Measuring the pref- Astana City Hospital No 1. Cent Asian J Glob Heal. 2014;2(Suppl):82. erence towards patient-centred communication with the Chinese- https://doi.org/10.5195/cajgh.2013.82 revised Patient-Practitioner Orientation Scale: a cross-sectional 49. Manchaiah V, Dockens AL, Bellon-Harn M, Burns ES. study among physicians and patients in clinical settings in Shanghai, Noncongruence between audiologist and patient prefer- China. BMJ Open. 2017;7(9):e016902. ences for patient-centeredness. J Am Acad Audiol. 2017;28(7): 32. Mudiyanse R, Pallegama R, Jayalath T, Dharmaratne S, Krupat E. 636-643. Translation and validation of patient-practitioner orientation scale 50. Tsimtsiou Z, Benos A, Garyfallos AA, Hatzichristou D. Predictors of in Sri Lanka. Educ Heal. 2015;28(1):35. physicians’ attitudes toward sharing information with patients and 33. Pereira CMAS, Amaral CFS, Ribeiro MMF, et al. Cross-cultural vali- addressing psychosocial needs: a cross-sectional study in Greece. dation of the Patient-Practitioner Orientation Scale (PPOS). Patient Health Commun. 2012;27(3):257-263. Educ Couns. 2013;91(1):37-43. 51. Chan CMH, Ahmad WAW, Yusof MM, Ho GF, Patient-Centredness 34. González-González AI, Orrego C, Perestelo-Perez L, et al. KE. Job satisfaction and psychological distress: a brief survey com- Effectiveness of a virtual intervention for primary healthcare profes- paring oncology nurses and doctors. Asian Pacific J Cancer Prev. sionals aimed at improving attitudes towards the empowerment of 2015;16(16):6895-6898. patients with chronic diseases: study protocol for a cluster random- ized controlled trial (e-MPODERA project). Trials. 2017;18(1):505. 35. Beaton DE, Bombardier C, Guillemin F, Ferraz MB. Guidelines for How to cite this article: Perestelo-Pérez L, Rivero-Santana A, the process of cross-cultural adaptation of self-report measures. González-González AI, et al. Cross-cultural validation of the Spine 2000;25(24):3186-3191. 36. Enders CK. A primer on maximum likelihood algorithms avail- patient-practitioner orientation scale among primary care able for use with missing data. Struct Equ Model A Multidiscip J. professionals in Spain. Health Expect.2021;24:33–41. https:// 2001;8(1):128-141. doi.org/10.1111/hex.13135 37. Yuan KH, Lambert PLFR. Mardia’s multivariate kurtosis with miss- ing data. Multivariate Behav Res. 2004;39(3):413-437.