Validity and reliability of the Turkish version of the general rehabilitation adherence scale in individuals with osteoarthritis Authors: Hasan Gercek, Sergen Ozturk, Furkan Cakir, Burcu Dursun, Mustafa Savas Torlak, Fatih Celik, Bayram Sonmez Unuvar, Aydan Aytar Published: 2026 ABSTRACT AbstractBackgroundTo evaluate the validity and reliability of the Turkish version of the General Rehabilitation Adherence Scale (GRAS) in individuals with osteoarthritis receiving physiotherapy.MethodsEighty-eight individuals diagnosed with osteoarthritis and undergoing physiotherapy participated in the study. The scale was translated using a forward-backward translation procedure and assessed for content validity. Construct validity was examined using exploratory factor analysis and confirmatory factor analysis to evaluate model fit. Internal consistency was evaluated using Cronbach’s alpha coefficients. Test-retest reliability was assessed using intraclass correlation coefficients (ICC).ResultsContent validity indices indicated adequate item relevance. Factor analyses supported a two-factor structure with acceptable model fit indices. Internal consistency was acceptable for the overall scale and subscales. Test-retest reliability demonstrated good to excellent stability over time (ICC > 0.80). The findings indicate that the Turkish GRAS is a psychometrically sound instrument for assessing rehabilitation adherence in individuals with osteoarthritis.ConclusionThe Turkish version of the GRAS is a valid and reliable tool for evaluating adherence behaviors in individuals with osteoarthritis receiving physiotherapy. Its use may assist clinicians in identifying patients at risk of poor rehabilitation engagement and may support individualized intervention planning to optimize functional outcomes.Clinical Trials RegistryClinicalTrials.gov Identifier: NCT07107412. Registry date: 30/07/2025. Registry link: https://clinicaltrials.gov/study/NCT07107412.Supplementary InformationThe online version contains supplementary material available at 10.1186/s12891-026-09922-7.Keywords: Osteoarthritis, Patient-reported outcome, Physiotherapy adherence, Reliability, Validity Background According to the World Report on Disability, nearly 200 million individuals worldwide experience significant functional difficulties as a result of disability. Neurological, orthopedic, and rheumatological musculoskeletal disorders contribute substantially to these functional limitations [1, 2]. Particularly when chronic, such conditions lead to long-term restrictions in activity and participation [2]. Osteoarthritis (OA) is a prominent example of these disorders and is characterized by pathological cartilage loss, remodeling of adjacent bone, and associated inflammation, resulting in functional impairments and reduced quality of life [3]. Evidence indicates that low levels of physical activity and exercise in individuals with OA exacerbate pain [4] and further impair physical function [5]. Consequently, physiotherapy interventions are strongly recommended in the management of OA and other chronic musculoskeletal conditions [3]. Physiotherapy is a therapeutic approach that aims to reduce disease-related pain, improve functional capacity, and enhance quality of life, and typically requires active patient participation [6, 7]. In addition, physiotherapy has been shown to be a potentially cost-effective intervention that improves activity and participation outcomes while reducing the duration of hospital stay [8]. The effectiveness of physiotherapy is largely dependent on patient adherence to the prescribed rehabilitation programme [9]. Adherence in physiotherapy is a multidimensional construct that encompasses active participation in treatment, compliance with the physiotherapist’s recommendations, and the regular performance of prescribed exercises [10]. Although the exact level of adherence to physiotherapy interventions is difficult to quantify, it varies widely depending on both patient-related and organizational factors [10]. Factors associated with increased adherence to physiotherapy include confidence in prescribed exercises, individually tailored treatment programmes that can be easily integrated into daily life, social support, and self-efficacy [10, 11]. Conversely, barriers to adherence include ineffective communication with healthcare providers, lack of motivation or a supportive environment, limited self-discipline or time constraints, forgetting to perform exercises, perceptions of exercises as difficult or ineffective, patient beliefs and attitudes, the quality of the therapist-patient relationship, treatment cost and accessibility, as well as cultural and language factors [12, 13]. In chronic and long-term conditions such as osteoarthritis (OA), sustained adherence to prescribed exercise programmes is a key determinant of long-term clinical success [14, 15]. However, the extent to which patients maintain adherence to these exercises over time remains unclear. Lifelong engagement in exercise is considered essential for optimizing well-being and functional capacity in adults with OA [16]. Moreover, evidence indicates that low adherence to physiotherapy exercises among individuals with OA negatively affects both the effectiveness and clinical outcomes of prescribed exercise interventions [17, 18]. Consequently, there is a clear need for valid and comprehensive tools to assess patient adherence to physiotherapy programmes, enabling accurate evaluation of treatment outcomes and identification of barriers to adherence [17]. Such tools not only support patient-centered care but also have the potential to contribute positively to the assessment and improvement of healthcare service quality [19]. In response to this need, the General Rehabilitation Adherence Scale (GRAS), developed by Naqvi et al. [19], is a validated instrument designed to accurately assess adherence to physiotherapy through a threshold-based scoring system that categorises patients according to their level of adherence to exercise therapy. In addition, the GRAS enables clinicians and therapists to identify potential determinants of non-adherence and to target modifiable barriers during rehabilitation. Accordingly, the aim of this study was to examine the validity and reliability of the Turkish-translated version of the GRAS in individuals diagnosed with osteoarthritis. Methods This study comprised the translation of the instrument following the acquisition of the required permissions, evaluation of the internal consistency of the Turkish version, conduct of exploratory and confirmatory factor analyses, and assessment of test-retest reliability. Ethical approval for the study was obtained from the KTO Karatay University Faculty of Medicine Non-Invasive Clinical Research Ethics Committee (Protocol No: 2024/034), and the study was prospectively registered (ClinicalTrials.gov Identifier: NCT07107412). Data collection was carried out between August 2025 and November 2025. Translation procedure The translation of the GRAS-OA into Turkish was conducted in accordance with a widely accepted protocol described in the literature [20]. To ensure both linguistic and conceptual equivalence with the original version, the translation process consisted of multiple systematic steps. Initially, two bilingual (Turkish-English) translators with relevant academic backgrounds independently performed the forward translations. These versions were subsequently reviewed and synthesized by a two-member expert committee to resolve discrepancies and produce a single reconciled version. Next, two different native English-speaking translators, who were blinded to the original version, independently carried out the back-translation of the synthesized Turkish version. A separate expert panel then compared the back-translated versions with the original instrument in terms of semantic, grammatical, and conceptual equivalence. Following the implementation of the necessary revisions, the pre-final version was pilot-tested on a small sample from the target population to assess clarity and cultural appropriateness. Based on participant feedback, minor adjustments were made where necessary. Following the implementation of the necessary revisions, the final Turkish version was established. Upon final review, the scale was considered to be accurate and comprehensive, with no further modifications required. Participants and setting Outpatients aged 18–65 years who had been diagnosed with osteoarthritis and were scheduled to receive physiotherapy for at least two weeks at a private physiotherapy clinic were included in the study. Exclusion criteria comprised insufficient proficiency in Turkish, the presence of any cognitive or mental condition that could hinder participation, and receipt of any treatment other than the prescribed physiotherapy program during the study period. Prior to enrolment, all participants were informed about the study procedures, and written informed consent was obtained. The study was conducted in accordance with the principles of the Declaration of Helsinki at all stages. Data were collected on participants’ sociodemographic characteristics, osteoarthritis-related clinical history, previous physiotherapy experience, and pain levels during activity, rest, and at night. The GRAS-OA was administered to 88 patients attending the clinic for physiotherapy to assess internal consistency and to conduct exploratory factor analysis. The sample size was considered adequate based on commonly recommended item-to-participant ratios (approximately 10 participants per item) for factor analysis. For the evaluation of test-retest reliability, the same 88 participants completed the GRAS-OA at the end of the first week of their initial physiotherapy sessions, and the scale was repeated two weeks later. Validation measure (GRAS) The General Rehabilitation Adherence Scale (GRAS) was developed in 2020 to assess adherence to rehabilitation sessions among physiotherapy patients. The scale consists of eight items and has been reported to exhibit a two-factor structure. Each item is rated on a four-point Likert scale, with response options scored as 0, 1, 2, and 3, corresponding to always, mostly, sometimes, and never, respectively [19]. Based on findings from the population in which the scale was originally developed, total scores ranging from 20 to 24 indicate high adherence, 17 to 19 indicate good adherence, 12 to 16 indicate partial adherence, 8 to 11 indicate low adherence, and scores of 7 or below indicate poor adherence. The original study reported a Cronbach’s alpha coefficient of 0.63 and a test-retest reliability coefficient of 0.88 [19]. Content Validity Content validity was assessed by integrating the perspectives of both academic experts and clinicians. A content validity analysis was performed with a panel of 10 experts (comprising two physiatrists, two physiotherapists, two rheumatologists, two occupational therapists, and two academicians) to evaluate the relevance, adequacy, pertinence, coherence, and clarity of the GRAS-OA items. Each item was rated using a four-point Likert scale (1 = not relevant or not comprehensive; 4 = highly relevant or highly comprehensive), and the Content Validity Index (CVI) was calculated in accordance with established recommendations [21]. Factor analyses Construct validity was firstly evaluated using exploratory factor analysis (EFA). As the items capture related yet distinct aspects of rehabilitation adherence, including individual and contextual barriers, an oblique rotation method (Direct Oblimin) was employed. The number of factors was determined by considering parallel analysis results in conjunction with scree plot inspection and eigenvalues greater than one. The suitability of the data for factor analysis was examined using the Kaiser-Meyer-Olkin (KMO) measure of sampling adequacy and Bartlett’s test of sphericity. A KMO value exceeding 0.60 and a statistically significant Bartlett’s test were considered indicative of adequate factorability. Factor loadings of 0.35 or higher were accepted as sufficient for assigning items to their respective factors. In addition, item communalities were examined to assess the adequacy of item representation within the factor structure, with values around 0.30 generally considered acceptable [22]. Confirmatory factor analysis (CFA) was subsequently performed on the same sample to verify the factor structure identified by the EFA. Model fit was evaluated using the Root Mean Square Error of Approximation (RMSEA), the Comparative Fit Index (CFI), and the Tucker-Lewis Index (TLI). Model fit for continuous data was interpreted as acceptable when the following criteria were met: CFI > 0.95, TLI > 0.90, and RMSEA < 0.08 [23]. Internal consistency and reliability analyses Internal consistency of the scale and its subscales was assessed using Cronbach’s alpha coefficients. For each participant, mean scores for the relevant subscales were calculated, and descriptive statistics (means and standard deviations) were reported. Reliability was interpreted according to established thresholds: excellent (α ≥ 0.90), very good (0.80–0.89), respectable (0.70–0.79), acceptable (0.65–0.69), and poor (< 0.65) [24]. The test-retest reliability between two time-points was assessed after a gap of 2 weeks through Test-retest reliability between two time points was assessed after a 2-week interval using intraclass correlation coefficients (ICC). ICC values greater than 0.75 were interpreted as indicating good reliability [25]. Data analysis The planned data analysis comprised descriptive statistics, evaluation of the factorial structure and confirmatory factor analysis fit indices, assessment of internal consistency using Cronbach’s alpha, and examination of test-retest reliability through intraclass correlation coefficients (ICC [1, 2]) with corresponding confidence intervals. All statistical analyses were performed using IBM SPSS Statistics version 29.0. Demographic characteristics were summarized as frequencies (N) and percentages (%). Statistical significance was set at a p value of less than 0.05. Results A total of 88 participants were included in the study. The majority of the sample were female (n = 59, 67.0%), while 29 participants (33.0%) were male. Regarding educational attainment, 34 participants (38.6%) had completed high school, 22 (25.0%) had primary school education, 14 (15.9%) had secondary school education, and 18 (20.5%) had a university or postgraduate degree. Most participants were married (n = 67, 76.1%), whereas 21 (23.9%) were single. In terms of occupational status, 41 participants (46.6%) were unemployed, 25 (28.4%) were employed, and 22 (25.0%) were retired. Comorbid conditions were reported by 22 participants (25.0%), while the remaining 66 participants (75.0%) reported no comorbidities. With respect to physiotherapy session participation, the majority of participants attended sessions independently (n = 62, 78.4%), whereas 26 participants (21.6%) attended sessions accompanied by a caregiver or family member (Table 1). Table 1Participants informationParticipants information (N = 88) N %M (SD)Age (years)56.4 (11.7)Height (cm)167.0 (8.63)Weight (kg)77.6 (13.2)BMI (kg/m²)27.9 (4.59)GenderFemale5967.0Male2933.0Education Primary school2225.0 Secondary school1415.9 High school3438.6 University and postgraduate1820.5Marital status Single2123.9 Married6776.1Occupation Unemployed4146.6 Retired2225.0 Employed2528.4Comorbidity Yes2225.0 No comorbidity6675.0Session participation Alone6278.4 With caregiver/family member2621.6Pain at rest6.58 (1.96)GRAS-OA total21.4 (2.24)GRAS-OA adherence score interpretation High adherence (20–24 points)7079.5 Good adherence (17–19 points)1314.8 Partial adherence (12–16 points)55.7 Low adherence (8–11 points)00M Mean, SD Standard deviation Participants information M Mean, SD Standard deviation The translated Turkish version of the GRAS-OA, consisting of 8 items, was evaluated for content validity by a panel of 10 experts. Experts assessed each item in terms of relevance, adequacy, clarity, coherence, and comprehensibility. Based on expert ratings, no items were removed or modified, as all items met the predefined criteria for content validity. Item-level CVI (I-CVI) values ranged from 0.880 to 0.952, and the scale-level CVI (S-CVI/Ave) was calculated as 0.912, indicating acceptable content validity. No items fell below the recommended threshold. The finalized Turkish version of the GRAS-OA therefore retained all 8 items. Factorial structure of the GRAS-OA The factor structure of the GRAS-OA was examined using exploratory factor analysis (EFA) and Direct Oblimin rotation. Sampling adequacy was confirmed with a Kaiser-Meyer-Olkin (KMO) value of 0.656, and Bartlett’s test of sphericity was statistically significant (χ² = 142, df = 28, p < 0.001), indicating that the data were suitable for factor analysis. Based on eigenvalues greater than 1.0 and inspection of the scree plot, a two-factor solution was identified. Items with factor loadings ≥ 0.35 and without salient cross-loadings were retained. Factor loadings ranged from 0.487 to 0.854, demonstrating a clear and interpretable factor structure. The two-factor solution accounted for 49.8% of the total variance, with the first factor explaining 29.2% and the second factor explaining 20.6% of the variance. Based on item content, the first factor appears to reflect individual-related barriers to rehabilitation adherence (e.g., time management, pain), whereas the second factor appears to capture external or contextual barriers (e.g., cost, availability of caregiver or therapist) (Table 2). Table 2Exploratory factor analysis of GRAS-OA and factor loadings (N = 88)GRAS-OA ItemsFactors12CommunalityI1 - Other commitments0.6860.464I2 - Unable to manage time0.5830.332I3 - Feel well0.6940.639I4 - Excessive pain0.4870.325I5 - Treatment cost0.8120.691I6 - Not worth the money spent0.8540.724I7 - Unavailability of caregiver0.5160.426I8 - Unavailability of therapist0.7120.479Percentage of variance%29,2%20,6Cumulative of variances%49,8‘oblimin’ rotation was used, Loadings < 0.35 are deleted Exploratory factor analysis of GRAS-OA and factor loadings (N = 88) ‘oblimin’ rotation was used, Loadings < 0.35 are deleted Following the EFA, confirmatory factor analysis (CFA) was conducted to examine the extent to which the hypothesised two-factor structure was consistent with the observed data. Model fit was evaluated using multiple fit indices. The results indicated an acceptable to good fit of the proposed model, with a Comparative Fit Index (CFI) of 0.966, Tucker-Lewis Index (TLI) of 0.950, Standardized Root Mean Square Residual (SRMR) of 0.0562, and Root Mean Square Error of Approximation (RMSEA) of 0.049 (90% CI: 0.00-0.111). These values were within commonly accepted thresholds for adequate model-data fit. It should be noted that CFA was performed on the same sample used for the exploratory factor analysis; therefore, these findings should be interpreted as preliminary evidence of model fit rather than a formal cross-validation of the factor structure. Internal consistency and reliability analysis The Cronbach’s alpha value of the first test for GRAS-OA factors ranged from 0.718 to 0.734 and for the total score it was 0.638 (Table 3). In the retest assessment, the Cronbach’s alpha values for the GRAS-OA factors ranged from 0.726 to 0.758, and the alpha for the total score was 0.656 (Table 3). Table 3Cronbach’s alpha for each subheading of the GRAS-OA (N = 88)FactorsCronbach’s Alpha1st Assessment2nd AssessmentFactor 1 (items 5–8)Range: 0–120.7340.758Factor 2 (items 1–4)Range: 0–120.7180.726Total GRAS-OA0.6380.656 Cronbach’s alpha for each subheading of the GRAS-OA (N = 88) Factor 1 (items 5–8) Range: 0–12 Factor 2 (items 1–4) Range: 0–12 Test-retest reliability of the GRAS-OA was evaluated in 88 participants who completed the scale again two weeks after the initial assessment. Intraclass correlation coefficients (ICC) indicated good to excellent reliability at the item level, with ICC values ranging from 0.668 to 0.908 (all p < 0.001). The highest stability was observed for Item 3 (ICC = 0.908; 95% CI: 0.863–0.939), whereas the lowest ICC was noted for Item 2 (ICC = 0.668; 95% CI: 0.535–0.769). The total GRAS-OA score demonstrated excellent test-retest reliability (ICC = 0.873; 95% CI: 0.813–0.915; p < 0.001), confirming the temporal stability of the scale over a two-week interval. Measurement error indices (SEM and MDC) were also calculated (Table 4). Table 4Test-retest reliability and measurement error of the GRAS-OA (ICC, SEM, and MDC)GRAS-OAICC (%95 CI) p SEMMDC95I1 - Other commitments0.846 (0.823 − 0.920)< 0.001***0.2210.61I2 - Unable to manage time0.668 (0.535 − 0.769)< 0.001***0.3871.07I3 - Feel well0.908 (0.863 − 0.939)< 0.001***0.2700.75I4 - Excessive pain0.698 (0.574 − 0.791)< 0.001***0.2570.71I5 - Treatment cost0.801 (0.711 − 0.865)< 0.001***0.2810.78I6 - Not worth the money spent0.765 (0.663 − 0.840)< 0.001***0.2490.69I7 - Unavailability of caregiver0.824 (0.744 − 0.881)< 0.001***0.2050.57I8 - Unavailability of therapist0.884 (0.829 − 0.922)< 0.001***0.0790.22GRAS-OA Total0.873 (0.813 − 0.915)< 0.001***0.8862.45CI Confidence Interval, MDC Minimal Detectable Change, SEM Standard Error of Measurement ***: p < 0.001 Test-retest reliability and measurement error of the GRAS-OA (ICC, SEM, and MDC) CI Confidence Interval, MDC Minimal Detectable Change, SEM Standard Error of Measurement ***: p < 0.001 Discussion The present study evaluated the psychometric properties of the Turkish version of the General Rehabilitation Adherence Scale adapted for individuals with osteoarthritis (GRAS-OA). The findings demonstrated that the GRAS-OA is a valid and reliable instrument for assessing adherence to physiotherapy among patients with osteoarthritis. The GRAS-OA may serve as a complementary instrument in quality of life research by enabling a more accurate interpretation of patient-reported outcomes in osteoarthritis rehabilitation. The demographic characteristics of the sample were consistent with the epidemiological profile of osteoarthritis, with a predominance of female participants and a largely middle-to-older-aged population. This distribution aligns with previous literature reporting higher prevalence and treatment-seeking behaviour among women with osteoarthritis [26]. Additionally, the relatively high proportion of unemployed and retired participants reflects the functional limitations and reduced work participation commonly observed in this population [27]. These characteristics suggest that the sample was appropriate for evaluating rehabilitation adherence in a real-world osteoarthritis context. Exploratory factor analysis supported a two-factor structure for the GRAS-OA, accounting for nearly half of the total variance. The identified factors were conceptually coherent and demonstrated satisfactory factor loadings, indicating that the scale captures distinct but related dimensions of rehabilitation adherence. The two-factor structure identified in the present study is consistent with findings from previous cross-cultural adaptations, which have similarly reported multidimensional structures, supporting the stability of the scale’s underlying construct across different cultural contexts [28, 29]. From a conceptual perspective, this structure appears to reflect both individual-related barriers (e.g., time management, perceived well-being, pain) and external or contextual barriers (e.g., cost, availability of caregiver or therapist), suggesting that rehabilitation adherence is influenced by both personal and environmental factors. In the context of osteoarthritis, adherence behaviour is influenced by multiple components, including attendance behaviour, engagement with therapy, and perceived responsibility toward rehabilitation. Therefore, the emergence of a two-factor structure may reflect the multidimensional nature of adherence in this clinical population. The adequacy of the factor structure was further supported by confirmatory factor analysis, which demonstrated good model fit across multiple indices (CFI, TLI, SRMR, and RMSEA). These findings suggest that the proposed two-factor model is consistent with the observed data and provides an acceptable representation of the underlying construct. The consistency between EFA and CFA results indicates that the identified structure is stable within the current sample. However, as both analyses were conducted on the same dataset, these findings should be interpreted as preliminary evidence rather than a formal validation of the factor structure. Internal consistency analysis revealed acceptable Cronbach’s alpha values for the subscales, while the total score demonstrated moderate internal consistency, with values close to the commonly accepted threshold. Although the retest value exceeded 0.65, the baseline value was slightly below this threshold, indicating that the internal consistency of the total score should be interpreted with caution. This finding is comparable to the original GRAS development study, which also reported modest alpha values for the total scale [19]. Given the small number of items and the behavioural heterogeneity inherent in adherence constructs, moderate internal consistency may be expected. In particular, Cronbach’s alpha is known to be influenced by the number of items, and shorter scales tend to yield lower alpha values. Adherence encompasses a range of behaviours that may not be highly intercorrelated, such as session attendance, compliance with therapist instructions, and personal motivation. Therefore, while the observed Cronbach’s alpha values can be considered acceptable for subscale interpretation, the use of the total score may require cautious interpretation. Test-retest reliability analyses demonstrated good to excellent temporal stability at both the item and total score levels. The high ICC for the total GRAS-OA score indicates that the scale provides stable measurements over time, supporting its reliability for both clinical follow-up and research applications. The variability observed across individual items may reflect fluctuations in short-term behavioural factors or situational influences, which are inherent to adherence-related constructs. Nevertheless, the overall stability of the scale confirms its robustness. Conclusion Adherence to rehabilitation programs represents a multidimensional construct influenced by cognitive, behavioral, and contextual factors. Within the framework of chronic musculoskeletal conditions such as osteoarthritis, rehabilitation engagement is shaped not only by symptom severity but also by patients’ beliefs, self-efficacy, perceived benefits, and environmental support. From a theoretical perspective, adherence can be conceptualized within health behavior models emphasizing motivation, behavioral regulation, and patient activation. The availability of a culturally adapted and psychometrically robust instrument such as the Turkish GRAS allows for a more systematic operationalization of this construct in clinical rehabilitation research. By quantifying adherence-related behaviors, clinicians and researchers may better examine the interaction between rehabilitation engagement and functional recovery trajectories. Furthermore, standardized adherence assessment may contribute to more precise modeling of treatment responsiveness and facilitate the development of theory-informed, patient-centered rehabilitation strategies in osteoarthritis management. The findings of this study have important clinical implications. Reliable assessment of rehabilitation adherence in osteoarthritis is essential for optimizing treatment outcomes, identifying patients at risk of poor engagement, and tailoring intervention strategies accordingly. The GRAS-OA offers a brief, easy-to-administer tool that can be readily integrated into routine physiotherapy practice. Moreover, its demonstrated validity and reliability support its use in clinical research focusing on adherence-related outcomes. Limitations This study has several limitations that should be acknowledged. First, the sample size was relatively modest and participants were recruited from a single outpatient physiotherapy clinic, which may limit the generalizability of the findings. Relatively high adherence levels observed in the sample may indicate a potential ceiling effect, which could reduce the discriminative ability of the instrument across different adherence levels. Although the sample size was considered adequate for the conducted psychometric analyses, future multicenter studies with larger and more heterogeneous samples are needed to confirm the factor structure and reliability of the GRAS-OA. The responsiveness of the instrument should also be evaluated in longitudinal designs to determine its ability to detect clinically meaningful changes over time. Second, exploratory and confirmatory factor analyses were performed on the same sample, which may increase the risk of model overfitting. This approach was adopted due to sample size constraints and the preliminary nature of this validation study; therefore, replication using independent samples is recommended. Third, the internal consistency of the total GRAS-OA score was moderate. This finding is consistent with the original version of the scale and may reflect the brief and behavioral nature of adherence constructs, which involve diverse patient behaviors rather than a single homogeneous dimension. Another limitation of the present study is the absence of hypothesis testing for construct validity through correlations with external measures. Evaluating predefined relationships with related constructs (e.g., clinical outcomes or behavioral measures) would have provided stronger evidence for construct validity. Future longitudinal studies should investigate the relationship between rehabilitation adherence measured by the GRAS-OA and rehabilitation outcomes. Finally, as the validation was conducted exclusively in individuals with osteoarthritis, further research is required to evaluate the applicability of the GRAS in other rehabilitation populations. In conclusion, the Turkish version of the GRAS-OA demonstrates satisfactory validity, acceptable internal consistency, and excellent test-retest reliability. These results indicate that the GRAS-OA is a psychometrically sound instrument for assessing adherence to physiotherapy in individuals with osteoarthritis and may serve as a valuable tool in both clinical and research settings. Supplementary Information Supplementary Material 1. Acknowledgments Disclosure statement No potential conflict of interest was reported by the authors. Abbreviation General Rehabilitation Adherence Scale adapted for individuals with osteoarthritis Authors’ contributions H.G.: Conceptualization, Investigation, Methodology, Project administration, Writing – original draft, Writing – review & editing; S.O.: Conceptualization, Resources, Supervision, Writing – review & editing; F.C.: Conceptualization, Resources, Supervision, Writing – review & editing; B.D.: Methodology, Writing – review & editing; M.S.T.: Conceptualization, Writing – review & editing, F.C.: Resources, Investigation, Writing – original draft, B.S.U.: Conceptualization, Methodology, Writing – review & editing, A.A.: Conceptualization, Methodology, Writing – review & editing. Funding Authors did not receive funding to carry out this research. Data availability Data can be provided by the corresponding author upon request. Declarations Ethics approval and consent to participate The approval of KTO Karatay University Non-Pharmaceutical and Medical Device Studies ethics committee (Protocol No: 2024/034) was obtained for this study. All procedures performed in studies involving human participants were in accordance with the ethical standards of the institutional and/or national research committee and with the Helsinki Declaration. Informed consent was obtained from all individual participants included in the study. Consent for publication All participants included in this study provided their written consent for publication. Competing interests The authors declare no competing interests. Footnotes Publisher’s Note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. References Associated Data Supplementary Materials Supplementary Material 1. Data Availability Statement Data can be provided by the corresponding author upon request.