Validation of the Chinese version of the telehealth attitudes scale for nurses: a cross-sectional study Authors: Jinlong Zheng, Wenhao Tang, Bo Zheng, Haihang Xu, Jun Zhou, Mingzhu Chen Published: 2026 ABSTRACT BackgroundTelehealth helps overcome temporal and spatial constraints, enabling remote disease monitoring and rehabilitation guidance for patients. Nurse-led telehealth services optimize resource allocation by significantly reducing non-essential hospital visits and readmissions, while providing personalized care and rehabilitation, thereby lowering the economic burden on patients. However, effective tools for assessing nurses’ attitudes towards telehealth adoption remain lacking.AimThis study primarily aimed to translate and culturally adapt the Nurses’ Attitudes Towards Telehealth Scale (NATUTS) for the Chinese context and to subsequently validate its psychometric properties.MethodsAfter securing authorization from the original authors, the NATUTS was systematically adapted through translation, back-translation, cultural refinement, and pilot testing. The culturally adapted Chinese version was then psychometrically validated with 452 nurses from tertiary hospitals in Hubei Province, selected via convenience sampling.ResultsThe Chinese NATUTS comprises 19 items across three dimensions: Satisfaction, Rejection, and Development. The scale demonstrated satisfactory reliability (Cronbach’s α = 0.893; McDonald’s ω = 0.928; test-retest = 0.879) and validity (I-CVI: 0.778-1.00; S-CVI/Ave = 0.971). Confirmatory factor analysis yielded a good model fit (χ²/df = 1.335, RMSEA = 0.058,SRMR = 0.045, CFI = 0.983, TLI = 0.981).ConclusionThe Chinese Nurses’ Telehealth Attitudes Scale exhibits strong reliability and validity, demonstrating its effectiveness as an instrument for measuring nurses’ attitudes towards telehealth technology adoption in China. Introduction Telehealth has emerged as a transformative force in global healthcare systems since its inception in the 1970s [1]. Defined as a technology-enabled service model that uses electronic devices, sensors, and internet platforms to transmit patient data and support remote diagnosis, monitoring, and care delivery [1], telehealth has gained substantial momentum worldwide. In the United States, the European Union, and several Asian countries [2], it has been integrated into mainstream care, especially during the COVID-19 pandemic, helping to mitigate geographical and resource barriers while maintaining continuity of care [3]. Nurses play an indispensable role in the implementation and coordination of telehealth services globally [4]. Studies from Australia, Canada [5], and the United Kingdom have documented nurse-led telehealth initiatives across chronic disease management, maternal and geriatric care [6], rehabilitation, and palliative support. Through structured telephone follow-ups, virtual consultations, and digital monitoring systems [4], nurses have improved health outcomes while reducing hospital readmissions and costs [7]. These international experiences underscore the importance of nursing engagement as a determinant of telehealth success. Theoretical models such as the Technology Acceptance Model (TAM) [8]and the Unified Theory of Acceptance and Use of Technology (UTAUT) [9]offer useful frameworks for understanding nurses’ adoption of telehealth. These models posit that technology use is influenced by factors such as perceived usefulness, ease of use, and social influence. Such constructs closely align with nursing perceptions of telehealth, including satisfaction, rejection, and developmental perspectives, highlighting the need for validated instruments that capture these dimensions across cultural settings [10]. In China, the Ministry of Industry and Information Technology and the National Health Commission issued a policy in 2020 to strengthen telehealth network infrastructure and promote “Internet Plus” healthcare [11]. With more than 5.6 million registered nurses [12], the country has seen growing adoption of telehealth in areas such as home-based palliative care [13], COPD management [14], and pelvic floor rehabilitation [15]. Despite this rapid expansion, a critical gap remains: no validated instrument exists to assess Chinese nurses’ attitudes toward telehealth. Most existing studies in China have relied on ad hoc questionnaires or scales developed for other health professions, which lack documented psychometric properties and cross-cultural relevance. Consequently, researchers and policymakers cannot reliably evaluate how nurses perceive telehealth, identify barriers to adoption, or design targeted training interventions—all of which are essential for successful implementation. The Nurses’ Attitudes Toward Telehealth Scale (NATUTS), originally developed and validated in Turkey, demonstrates strong psychometric properties (Cronbach’s α = 0.92; subscale α = 0.87–0.96) [16]. However, its direct application in China is problematic for several reasons. First, Chinese nurses operate within a distinct healthcare system characterized by high patient volumes, hierarchical hospital structures, and rapidly digitizing but unevenly distributed resources. Second, cultural factors—such as attitudes toward technology, professional autonomy, and hierarchical decision-making—may shape nurses’ acceptance of telehealth differently than in Western or Middle Eastern contexts. Third, linguistic and conceptual equivalence cannot be assumed without rigorous translation and cultural adaptation. Therefore, a validated Chinese version of the NATUTS is urgently needed to fill this methodological void. Given this gap, the present study aims to translate, culturally adapt, and psychometrically validate the Chinese version of the NATUTS. By providing a reliable and valid instrument, this work will enable future research to assess nurses’ attitudes systematically, inform policy development, facilitate cross-cultural comparisons, and ultimately support the sustainable integration of telehealth into China’s nursing practice. Methods Participants and Ethical Statements We recruited registered nurses through convenience sampling from three tertiary hospitals across central, eastern, and northern Hubei Province from June to August 2024. Surveys were administered both electronically (Questionnaire Star app) and on paper, with participant eligibility criteria mirroring those from the pilot survey.The sample size was determined with reference to methodological standards for factor analysis, which recommend a subject-to-item ratio of 10:1 [17] and an absolute sample size of over 200 for CFA [18]. With 19 scale items and an anticipated 20% attrition rate the minimum target was 354 participants. Our final sample of 452 nurses comfortably meets these requirements.To ensure independent cross-validation, the total sample (N = 452) was randomly partitioned into two subsamples using SPSS 25.0: an EFA subsample (n = 244) for initial structure exploration and a CFA subsample (n = 208) for subsequent model confirmation. Ethical approval was obtained from the Ethics Committee of Xiangyang Central Hospital (KY202436). Introduction to the original scale The NATUTS scale, developed by Sümeyra et al. [16], is a validated instrument designed to assess nurses’ attitudes toward telehealth. Comprising 19 items across three dimensions—satisfaction, rejection (reverse-scored), and development—it employs a 5-point Likert scale (1 = strongly disagree to 5 = strongly agree). Total scores range from 19 to 95, with higher scores indicating more positive attitudes. Psychometric evaluation demonstrated strong reliability, with an overall Cronbach’s α of 0.92 in the original validation study. Scale translation and adaptation Authorization for the Chinese adaptation was obtained from the original author via email. The translation and cultural adaptation followed Brislin’s framework [19, 20], through four steps: (1) forward translation by two bilingual nursing postgraduates (CET-6 level); (2) back-translation by a nursing faculty member with overseas experience and a university English instructor; (3) cross-cultural adaptation by seven experts (master’s degree or higher, associate senior rank or above, ≥ 10 years of clinical nursing and research experience); (4) synthesis and finalization of the preliminary Chinese version after review by the original author and translation team. The flowchart of the translation and cultural adaptation process is presented in Fig. 1. Fig. 1The flowchart of the translation and cultural adaptation process The flowchart of the translation and cultural adaptation process Pilot testing A preliminary survey sample of 10 ~ 40 participants is recommended for scale refinement [21]. A pilot test was conducted in May 2024 using a convenience sample of 30 nurses from a tertiary hospital in central Hubei Province. Inclusion criteria required participants to: (1) be aged ≥ 18 years, (2) registered and employed at a hospital, and (3) provide informed consent. Nurses absent from clinical duties due to internships, rotations, study leave, or extended absence were excluded. Detailed feedback was collected from participants regarding any difficulties in completing the scale and ambiguities in item wording. Based on this feedback, the research team adjusted and modified the scale, resulting in the final Chinese version of the NATUTS. Data collection procedure Following approval by the hospital ethics committee, two systematically trained investigators administered the questionnaires face-to-face using standardized instructions. Before completing the survey, the purpose and significance of the study were explained to participants. After obtaining informed consent, participants were asked to scan a QR code to fill out the questionnaire independently. Upon completion, each questionnaire was checked for completeness, and participants were promptly reminded to supplement any missing items. Within 24 h after data collection, two researchers independently performed statistics, screening, and data entry of the questionnaires. Exclusion criteria included: (1) Completion time < 300 s, (2) responses that did not meet the requirements, or (3) answers showing obvious patterned responses. A total of 28 questionnaires were excluded based on these criteria: 12 for completion time < 300 s, 10 for incomplete data, and 6 for patterned responses. Statistical analysis Upon completion of data collection, all entries underwent dual verification before being imported into Excel. Statistical analyses were performed using SPSS 25.0, while confirmatory factor analysis was conducted with Mplus 8.7. Categorical variables were expressed as frequencies (percentages, %), and normally distributed continuous variables were presented as mean \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:\pm\:\:$$\end{document}standard deviation. All tests were two-tailed, with the statistical significance level set at α = 0.05. Reliability and validity testing Following the COSMIN (Consensus-based Standards for the selection of health Measurement Instruments) guideline for patient-reported outcome measures [22], the following criteria were adopted for evaluating psychometric properties: Item analysis of the scale Critical Ratio (CR) Screening: Total scale scores were calculated and sorted in descending order. Participants were stratified into upper 27% (high-score group, assigned to 1) and lower 27% (low-score group, assigned to 2). Independent samples t-tests were performed to compare group differences on each item. Items with CR < 3.0 or non-significant t-test results (P > 0.05) were excluded [22, 23].Correlation Analysis: Perform Pearson correlation analysis between each item of the scale and the total score of the scale, as well as between each dimension. Delete items with a correlation coefficient < 0.30 [22, 24]. Critical Ratio (CR) Screening: Total scale scores were calculated and sorted in descending order. Participants were stratified into upper 27% (high-score group, assigned to 1) and lower 27% (low-score group, assigned to 2). Independent samples t-tests were performed to compare group differences on each item. Items with CR < 3.0 or non-significant t-test results (P > 0.05) were excluded [22, 23]. Critical Ratio (CR) Screening: Total scale scores were calculated and sorted in descending order. Participants were stratified into upper 27% (high-score group, assigned to 1) and lower 27% (low-score group, assigned to 2). Independent samples t-tests were performed to compare group differences on each item. Items with CR < 3.0 or non-significant t-test results (P > 0.05) were excluded [22, 23]. Correlation Analysis: Perform Pearson correlation analysis between each item of the scale and the total score of the scale, as well as between each dimension. Delete items with a correlation coefficient < 0.30 [22, 24]. Correlation Analysis: Perform Pearson correlation analysis between each item of the scale and the total score of the scale, as well as between each dimension. Delete items with a correlation coefficient < 0.30 [22, 24]. Reliability testing Internal Consistency: Cronbach’s α coefficients exceeding 0.70 and McDonald’s ω coefficients surpassing 0.80 demonstrate robust internal consistency reliability of the scale [22, 25].Test-Retest Reliability: This was assessed by re-administering the scale to a randomly selected subsample of 30 nurses after a two-week interval. The intraclass correlation coefficient (ICC) between administrations was calculated using SPSS 25.0 to evaluate measurement stability [22].Split-Half Reliability: A split-half reliability coefficient > 0.60 is generally considered acceptable for demonstrating internal consistency [22]. Internal Consistency: Cronbach’s α coefficients exceeding 0.70 and McDonald’s ω coefficients surpassing 0.80 demonstrate robust internal consistency reliability of the scale [22, 25]. Internal Consistency: Cronbach’s α coefficients exceeding 0.70 and McDonald’s ω coefficients surpassing 0.80 demonstrate robust internal consistency reliability of the scale [22, 25]. Test-Retest Reliability: This was assessed by re-administering the scale to a randomly selected subsample of 30 nurses after a two-week interval. The intraclass correlation coefficient (ICC) between administrations was calculated using SPSS 25.0 to evaluate measurement stability [22]. Test-Retest Reliability: This was assessed by re-administering the scale to a randomly selected subsample of 30 nurses after a two-week interval. The intraclass correlation coefficient (ICC) between administrations was calculated using SPSS 25.0 to evaluate measurement stability [22]. Split-Half Reliability: A split-half reliability coefficient > 0.60 is generally considered acceptable for demonstrating internal consistency [22]. Split-Half Reliability: A split-half reliability coefficient > 0.60 is generally considered acceptable for demonstrating internal consistency [22]. Validity testing Content Validity Through an expert consultation process, nine specialists in cross-cultural adaptation—drawn from the fields of nursing management, psychology, specialized nursing, and clinical medicine—were invited to evaluate each item of the Chinese version of the NATUTS. These experts, who had not participated in the prior adaptation phase, rated item relevance using a 4-point scale (1 = not relevant, 2 = somewhat relevant, 3 = quite relevant, 4 = highly relevant). The item-level content validity index (I-CVI), scale-level content validity index/universal agreement (S-CVI/UA), and the average scale-level content validity index (S-CVI/Ave) were calculated to assess content validity [26, 27]. The content validity was considered acceptable if all I-CVIs, S-CVI/UA, and S-CVI/Ave achieved values of 0.80 or higher [22]. Structural Validity Exploratory factor analysis (EFA) was conducted using principal axis factoring (PAF) with Promax oblique rotation, as the underlying factors were theoretically expected to correlate [28]. The suitability for factor analysis was assessed through: (1) Kaiser-Meyer-Olkin (KMO) measure of sampling adequacy (> 0.8); (2) Bartlett’s test of sphericity (statistically significant chi-square value, P < 0.05) [26]. Construct validity was assessed through both exploratory factor analysis (EFA) and confirmatory factor analysis (CFA) [29]. In the EFA, the KMO measure and Bartlett’s test of sphericity were first conducted. The data were considered suitable for factor analysis given a KMO value greater than 0.80 and a significant Bartlett’s test result (P < 0.05). Principal axis factoring (PAF) with Promax oblique rotation was then applied to extract common factors with eigenvalues greater than or equal to 1. Items were removed if they exhibited factor loadings below 0.50 or showed significant cross-loadings. Given the categorical nature of the indicator variables, the CFA was conducted using the robust weighted least squares (WLSMV) estimator, which provides accurate parameter estimates, standard errors, and fit statistics for ordinal data. The model was considered to have a good fit when the following criteria were met: chi-square/degrees of freedom (χ²/df) < 5.00, Comparative Fit Index (CFI) > 0.90, Tucker-Lewis Index(TLI) > 0.90, root mean square error of approximation (RMSEA) < 0.08, Standardized Root Mean Square Residual(SRMR) < 0.10 [22, 29]. Convergent validity Convergent validity refers to the degree to which observed variables measuring the same latent construct converge or load strongly on that construct. In this study, a confirmatory factor analysis model was constructed using Mplus 8.7, and the convergent validity was assessed through factor loadings, average variance extracted (AVE), and composite reliability (CR). It is generally accepted that convergent validity is adequate when the AVE exceeds 0.50 and the CR exceeds 0.60 [22]. Discriminant validity The Heterotrait-Monotrait Ratio (HTMT) serves as a critical metric for evaluating discriminant validity, which reflects the extent to which distinct constructs can be empirically distinguished. In this study, we employed HTMT to assess whether the NATUTS scale demonstrates sufficient discriminant validity. According to established criteria [22, 30], if all HTMT values between constructs are below the stringent threshold of 0.85, this indicates adequate discriminant validity. Introduction to the original scale The NATUTS scale, developed by Sümeyra et al. [16], is a validated instrument designed to assess nurses’ attitudes toward telehealth. Comprising 19 items across three dimensions—satisfaction, rejection (reverse-scored), and development—it employs a 5-point Likert scale (1 = strongly disagree to 5 = strongly agree). Total scores range from 19 to 95, with higher scores indicating more positive attitudes. Psychometric evaluation demonstrated strong reliability, with an overall Cronbach’s α of 0.92 in the original validation study. Results Demographic characteristics of participants Of the 480 questionnaires distributed, 452 valid responses were returned, yielding a valid response rate of 94.17%. The participants (35 males, 7.74%; 417 females, 92.26%) ranged in age from 20 to 55 years (mean 30.45 ± 7.96 years). Regarding professional characteristics, 52.21% had less than five years of work experience, 25.88% had 5–10 years, and 21.91% had more than 10 years. Educational attainment included 283 participants (62.61%) with a bachelor’s degree and 80 (16.81%) with a master’s degree or higher. Notably, 97 participants (21.46%) reported prior telehealth experience. Further analysis revealed that nurses with telehealth experience scored significantly higher on the NATUTS than those without (76.24 ± 9.31 vs. 73.00 ± 9.52, P = 0.003).Our known-groups validity analysis revealed that nurses with prior telehealth experience scored significantly higher on both the total NATUTS and the Satisfaction subscale than those without such experience (P < 0.003), indicating that the scale possesses good discriminant validity. Complete demographic characteristics are presented in Table 1. Table 1General characteristics of the participants (N = 452)VariablesCategory N %GenderMale357.74Female41792.26Professional TitlePrimary27360.40Senior15033.19Deputy Chief Nurse or higher296.41Education LevelDiploma Degree8919.69Bachelor’s Degree28362.61Master’s Degree7616.81Doctoral Degree40.89Work years<5 years23652.215 ~ 10 years11725.88>10 years9921.91Prior telehealth experienceYes9721.46No35578.54 Table 1General characteristics of the participants (N = 452)VariablesCategory N %GenderMale357.74Female41792.26Professional TitlePrimary27360.40Senior15033.19Deputy Chief Nurse or higher296.41Education LevelDiploma Degree8919.69Bachelor’s Degree28362.61Master’s Degree7616.81Doctoral Degree40.89Work years<5 years23652.215 ~ 10 years11725.88>10 years9921.91Prior telehealth experienceYes9721.46No35578.54 General characteristics of the participants (N = 452) Item analysis Critical Ratio (CR) Method: Independent samples t-test results demonstrated CR values ranging from 4.726 to 15.977, all exceeding 3.0 with statistical significance (P < 0.05); Correlation Coefficient Method: Pearson correlation analysis revealed item-total correlations ranging from 0.391 to 0.785, all exceeding 0.30 with statistical significance (P < 0.05). No items were excluded based on these criteria. Reliability analysis The Chinese NATUTS demonstrated strong reliability, with a Cronbach’s α of 0.893 for the total scale and 0.960, 0.843, and 0.763 for the three subscales, respectively. McDonald’s ω coefficients were 0.928 for the total scale and 0.961, 0.846, and 0.786 for the subscales. Test-retest reliability was 0.879, and the split-half reliability was 0.644, which may be attributable to the scale’s multidimensional structure affecting random-split equivalence. Validity analysis Content validity For the Chinese version of the NATUTS, the I-CVI ranged from 0.778 to 1.00, the S-CVI/UA was 0.789, and the S-CVI/Ave was 0.971, indicating good content validity. Structural validity The Kaiser-Meyer-Olkin (KMO) measure verified the sampling adequacy for the analysis (KMO = 0.923), and Bartlett’s test of sphericity (χ² =3569.053, P < 0.001) indicated that correlations between items were sufficiently large for factor analysis. A Principal axis factoring (PAF) with Promax rotation extracted three factors that accounted for 63.33% of the cumulative variance (Fig. 2). The three factors were labelled as follows: Satisfaction (11 items, Q1-Q11), Rejection (5 items, Q12-Q16), and Development (3 items, Q17-Q19). All items loaded above 0.50 on their respective factors (Table 2), meeting retention criteria without requiring deletion. Fig. 2Scree plot of factor analysis Scree plot of factor analysis Table 2Exploratory factor analysis of the Chinese version of the NATUTSItemsFactor 1Factor 2Factor 3Q1. Telehealth reduces nurses’ workload.0.762Q2. Telehealth is effective in improving patients’ quality of life.0.899Q3. Telehealth is a reliable mode of care delivery.0.890Q4. Telehealth streamlines communication between nurses and patients.0.737Q5. Telehealth enables more efficient delivery of nursing care.0.838Q6. Telehealth offers advantages for my professional practice.0.736Q7. Telehealth makes nursing practice more convenient.0.849Q8. Telehealth is an integral component of nursing services.0.737Q9. Telehealth plays a vital role in nursing practice.0.903Q10. Telehealth should be integrated into standard clinical nursing practice.0.857Q11. Telehealth is crucial to the future of nursing care.0.905Q12. I am uncomfortable with delivering care via telehealth.0.796Q13. I feel anxious when using telehealth platforms.0.759Q14. I am resistant to adopting telehealth methodologies.0.700Q15. I have to use telehealth for my work.0.695Q16. I would not recommend telehealth to my colleagues.0.662Q17. I stay updated on the latest advancements in telehealth.0.828Q18. I intend to conduct in-depth research in the field of telehealth.0.865Q19. I am committed to enhancing my skills and knowledge in telehealth.0.742Eigenvalues7.682.651.70Percentage of variance (%)40.4213.968.95Note: ‘Principal axis factoring’ extraction method was used in combination with a ‘promax’ rotation Table 2Exploratory factor analysis of the Chinese version of the NATUTSItemsFactor 1Factor 2Factor 3Q1. Telehealth reduces nurses’ workload.0.762Q2. Telehealth is effective in improving patients’ quality of life.0.899Q3. Telehealth is a reliable mode of care delivery.0.890Q4. Telehealth streamlines communication between nurses and patients.0.737Q5. Telehealth enables more efficient delivery of nursing care.0.838Q6. Telehealth offers advantages for my professional practice.0.736Q7. Telehealth makes nursing practice more convenient.0.849Q8. Telehealth is an integral component of nursing services.0.737Q9. Telehealth plays a vital role in nursing practice.0.903Q10. Telehealth should be integrated into standard clinical nursing practice.0.857Q11. Telehealth is crucial to the future of nursing care.0.905Q12. I am uncomfortable with delivering care via telehealth.0.796Q13. I feel anxious when using telehealth platforms.0.759Q14. I am resistant to adopting telehealth methodologies.0.700Q15. I have to use telehealth for my work.0.695Q16. I would not recommend telehealth to my colleagues.0.662Q17. I stay updated on the latest advancements in telehealth.0.828Q18. I intend to conduct in-depth research in the field of telehealth.0.865Q19. I am committed to enhancing my skills and knowledge in telehealth.0.742Eigenvalues7.682.651.70Percentage of variance (%)40.4213.968.95Note: ‘Principal axis factoring’ extraction method was used in combination with a ‘promax’ rotation Exploratory factor analysis of the Chinese version of the NATUTS Note: ‘Principal axis factoring’ extraction method was used in combination with a ‘promax’ rotation Convergent validity Convergent validity refers to the extent to which items measuring the same construct share a high proportion of variance. It was assessed using two standard indicators: average variance extracted (AVE) and composite reliability (CR). AVE measures the amount of variance captured by a construct relative to measurement error, with a recommended threshold of > 0.50 indicating adequate convergence. CR evaluates the internal consistency of construct indicators, with > 0.70 considered acceptable. In this study, all 19 items demonstrated standardized factor loadings above 0.60, with the majority exceeding 0.70. The AVE for each construct exceeded 0.50, and CR values ranged from 0.794 to 0.928, with most surpassing 0.70. These results collectively demonstrate satisfactory convergent validity of the measurement model. The specific AVE and CR values for the confirmatory factor analysis model are presented in Table 3. Table 3Convergent validity: Average variance extracted (AVE), square root of AVE (√AVE), and composite reliability (CR)FactorAVE√AVECRSatisfaction0.5430.7360.928Rejection0.5850.7650.875Development0.5630.7500.794Note: AVE values all exceed the recommended threshold of 0.50, and CR values exceed 0.70, indicating adequate convergent validity. The √AVE values are presented for reference in discriminant validity assessment (Fornell–Larcker criterion) Table 3Convergent validity: Average variance extracted (AVE), square root of AVE (√AVE), and composite reliability (CR)FactorAVE√AVECRSatisfaction0.5430.7360.928Rejection0.5850.7650.875Development0.5630.7500.794Note: AVE values all exceed the recommended threshold of 0.50, and CR values exceed 0.70, indicating adequate convergent validity. The √AVE values are presented for reference in discriminant validity assessment (Fornell–Larcker criterion) Convergent validity: Average variance extracted (AVE), square root of AVE (√AVE), and composite reliability (CR) Note: AVE values all exceed the recommended threshold of 0.50, and CR values exceed 0.70, indicating adequate convergent validity. The √AVE values are presented for reference in discriminant validity assessment (Fornell–Larcker criterion) Discriminant validity Discriminant validity refers to the degree to which distinct constructs are empirically distinct from one another. It was evaluated using the heterotrait-monotrait (HTMT) ratio of correlations, which estimates the true correlation between two constructs after accounting for measurement error. A HTMT value below 0.85 provides adequate evidence that two constructs are empirically distinguishable. In this study, all HTMT ratios remained below the stringent threshold of 0.85 (Satisfaction–Rejection = 0.384; Satisfaction–Development = 0.326; Rejection–Development = 0.329), thereby supporting the discriminant validity of the measurement model. Confirmatory factor analysis The confirmatory factor analysis (CFA) yielded a good model fit, with all indices meeting the recommended thresholds: χ²=277.68, df = 208, χ²/df = 1.335(< 3 ideal), RMSEA = 0.058, SRMR = 0.045(< 0.10 ideal), CFI = 0.983, TLI = 0.981(all > 0.90 threshold). These results support the adequacy of the proposed factor structure, which is presented in Fig. 3. Fig. 3Confirmatory factor analysis Confirmatory factor analysis Discussion The Chinese version of the NATUTS scale has good reliability and validity The Chinese version of the Nurses’ Attitudes Toward Telehealth Scale (NATUTS) was developed following Brislin’s translation model [19], which included forward translation, back-translation, and cultural adaptation. Psychometric evaluation showed acceptable internal consistency (Cronbach’s α and McDonald’s ω). However, the split-half reliability coefficient (0.644) fell below the recommended threshold of 0.70, which we acknowledge as a limitation. Split-half reliability estimates the correlation between two randomly divided halves of a scale. For a unidimensional scale, the two halves tend to be content-equivalent, yielding a high correlation. For a multidimensional scale such as the NATUTS—comprising three distinct dimensions (Satisfaction, Rejection, and Development)—random splitting can result in an uneven distribution of items across halves, compromising content equivalence and thus attenuating the coefficient. Methodological literature indicates that split-half reliability is less appropriate for multidimensional instruments, where Cronbach’s α or McDonald’s ω are preferred [22, 25, 31]. Accordingly, despite the lower split-half coefficient, the overall reliability of the scale is supported by the other indicators (Cronbach’s α, ω, and test-retest reliability). The split-half result should not be overinterpreted, and future studies should assess reliability using multiple estimators. The three-factor structure of the NATUTS—Satisfaction, Rejection, and Development—aligns meaningfully with core constructs of the Technology Acceptance Model (TAM) [8] and the Unified Theory of Acceptance and Use of Technology (UTAUT) [9]. Specifically, the Satisfaction dimension corresponds to TAM’s perceived usefulness and perceived ease of use, reflecting nurses’ cognitive appraisal of telehealth benefits and usability. The Rejection dimension aligns with UTAUT’s facilitating conditions and effort expectancy, capturing barriers related to organizational support, technical infrastructure, and personal confidence. The Development dimension is consistent with UTAUT’s performance expectancy and social influence, as well as Rogers’ diffusion of innovation theory, emphasising nurses’ orientation toward professional growth and system-level improvements. This theoretical alignment supports the construct validity of the Chinese NATUTS and suggests that the scale adequately captures key dimensions of technology acceptance in nursing practice. The reliability coefficients obtained in this study closely align with those reported for the original Turkish version (Cronbach’s α = 0.920) [16], suggesting acceptable cross-cultural psychometric performance. Subtle differences in factor loadings and dimension-specific reliability may reflect unique aspects of Chinese nursing practice. Notably, the higher reliability observed for the Development dimension may indicate that Chinese nurses place particular emphasis on professional development opportunities associated with telehealth—a finding potentially influenced by China’s rapidly evolving healthcare digitalization policies. From a practical perspective, this suggests that policy initiatives aimed at expanding telehealth should explicitly incorporate career development pathways for nurses, such as formal training programmes, certification schemes, and leadership roles in digital health teams, to sustain positive attitudes. The Chinese version of the NATUTS scale has good application value Telehealth is an innovative healthcare delivery model that utilizes modern communication technologies—such as computer systems, remote sensing, telemetry, and remote control—to facilitate diagnostic and therapeutic interactions between healthcare providers and patients beyond geographical constraints [32]. This model not only reduces the time and cost associated with patient care but also significantly enhances the overall healthcare experience [33]. Nurses play an indispensable role in healthcare delivery [34]; however, there is currently a lack of validated instruments in China to assess nurses’ attitudes toward telehealth. The Chinese version of the NATUTS comprises three dimensions—satisfaction, rejection, and development—with a total of 19 items. The scale features clear and comprehensible items, demonstrates strong practicality, and can be completed within 5 ~ 8 min. It effectively captures nurses’ attitudes toward telehealth without imposing additional workload, making it a feasible tool for routine assessment. The validated scale has several implications for nursing policy in China. First, policymakers at the national and provincial levels may use the NATUTS to monitor attitudes toward telehealth among different regions and hospital types. Such data can inform the development of telehealth training standards and competency frameworks for nurses. Second, the three‑dimensional structure allows policy-makers to identify whether nurses’ concerns are rooted in low satisfaction (e.g., usability issues), high rejection (e.g., resistance to change), or weak development expectations (e.g., lack of career pathways). This differentiation enables targeted policy interventions—for example, investing in user‑friendly telehealth platforms to address satisfaction, launching public awareness campaigns to reduce rejection, or integrating telehealth skills into nursing career ladders to support development. Third, the scale can be used as a pre-and post-evaluation tool for policy pilot projects (e.g., telehealth reimbursement schemes or rural telehealth rollouts) to assess their impact on nurses’ attitudes. Fourth, the known‑groups finding that nurses with prior telehealth experience report higher satisfaction supports continued investment in hands-on telehealth opportunities through clinical rotations and dedicated training funds. Finally, the absence of significant differences in the Rejection and Development dimensions between experienced and inexperienced nurses suggests that systemic policies—such as professional recognition for telehealth competence and national guidelines for nurse‑led telehealth services—may be more effective than relying solely on individual experience to shape these attitudes. In summary, the Chinese NATUTS provides a practical, evidence-based tool for aligning nursing workforce policies with the country’s rapid digital health transformation. Conclusion The Chinese version of the NATUTS was developed following a systematic localization procedure and underwent preliminary validation among registered nurses in mainland China. Results demonstrated that the scale consists of 19 items across three dimensions and exhibits sound psychometric properties, supporting its utility in effectively assessing nurses’ attitudes toward telehealth adoption. Limitation Several limitations of this study should be acknowledged. First, the generalizability of our findings may be limited by the exclusive recruitment of participants from three Grade A tertiary hospitals located in central, eastern, and northern regions of Hubei Province. Future studies should validate the scale in broader and more diverse nursing populations across China. Second, this study employed a cross-sectional design, which precludes assessment of the scale’s test-retest reliability over extended periods and its predictive validity—i.e., whether baseline attitudes can predict future telehealth adoption behaviors. Longitudinal studies are warranted to track the evolving trajectory of nurses’ attitudes as their telehealth experience accumulates and to establish the scale’s sensitivity to change over time. Third, all data were collected via self-reported questionnaires, which may introduce common method bias and social desirability bias, potentially inflating observed relationships among constructs. Future research could incorporate objective behavioral measures or multi-source evaluations (e.g., supervisor ratings, actual telehealth usage logs) to mitigate this concern. Fourth, the lack of external criterion validation—such as comparison with objective indicators of telehealth adoption or job performance—limits the assessment of criterion-related validity. Future studies should establish such criteria to further validate the scale. Fifth, although convergent and discriminant validity were evaluated using average variance extracted (AVE) and the heterotrait–monotrait (HTMT) ratio—both widely accepted statistical techniques—these indices were calculated within the same scale structure. As noted in the methodological literature, such internal validity estimates may be prone to showing adequate results even when the scale’s relationship with external measures (e.g., theoretically related constructs, established attitude scales, or observable telehealth behaviors) has not been empirically established [27, 28]. Therefore, the present findings should be interpreted as preliminary evidence of construct validity, rather than definitive confirmation. Sixth, the scale-level content validity index based on the universal agreement method (S-CVI/UA = 0.789) fell slightly below the conventional threshold of 0.80. This reflects the stringent nature of the universal agreement method with a panel of nine experts and is mitigated by the acceptable S-CVI/Ave (0.971) and the fact that all item-level I-CVIs met the acceptable criterion (≥ 0.78). Nevertheless, it constitutes a minor limitation. Future studies may consider using the average method as the primary content validity index. Seventh, the split-half reliability coefficient was below the recommended benchmark of 0.70. Although this may be partly attributable to the scale’s multidimensional structure (which reduces the equivalence of randomly split halves), we explicitly acknowledge this as a limitation. Therefore, we rely primarily on Cronbach’s α and McDonald’s ω, which are more appropriate for multidimensional instruments, as the main indicators of internal consistency. Future research should assess reliability using alternative methods (e.g., greatest lower bound) and in larger, more diverse samples. Based on the current findings and limitations, we propose the following directions for future research: (1) cross-cultural adaptation and validation of the Chinese NATUTS in different cultural contexts to examine measurement equivalence and compare nurses’ attitudes across various healthcare systems; (2) longitudinal studies to assess the scale’s predictive validity and sensitivity to change; (3) multi-method validation incorporating objective behavioral indicators to reduce self-report bias; (4) external validation using criterion measures (e.g., other validated telehealth attitude scales or actual usage data) to complement the internal AVE/HTMT evidence; (5) further content validity assessment using the average method as the primary index; and (6) reliability studies employing multiple estimators, including the greatest lower bound, to complement the traditional split-half approach.