Cross-Cultural Adaptation and Psychometric Validation of the Chinese Version of the Attitude Scale Towards the Use of Artificial Intelligence Technologies in Nursing (ASUAITIN)

Shang Li¹, Xu Lanlan¹, Nie Xiaofei¹, Luo Yixue¹, Wang Dian¹, Li Ye²*

¹Hubei University of Medicine, Shiyan 442000, China
²Dongfang Hospital, Beijing University of Chinese Medicine, Beijing 100871, China

Abstract

Objective: To translate and culturally adapt the Attitude Scale Towards the Use of Artificial Intelligence Technologies in Nursing (ASUAITIN) into Chinese, and to evaluate its reliability and validity, thereby providing a psychometrically sound instrument for assessing clinical nurses' attitudes toward AI technology integration in nursing practice.

Methods: Following the Brislin translation model, the scale underwent forward translation, back-translation, cultural adaptation, and pilot testing to develop the Chinese version of ASUAITIN. Using convenience sampling, 396 nurses from a tertiary Grade A hospital in Beijing were surveyed in March 2025 to examine the scale's psychometric properties.

Results: The Chinese ASUAITIN comprised 15 items across two dimensions. The overall Cronbach's α coefficient was 0.939, split-half reliability was 0.738, and test-retest reliability was 0.935. Cronbach's α coefficients for the two dimensions were 0.945 and 0.956, respectively. The scale-level content validity index (S-CVI) was 0.981, with item-level CVIs (I-CVI) ranging from 0.875 to 1.000. Exploratory factor analysis extracted two common factors accounting for 77.402% of cumulative variance. Confirmatory factor analysis yielded the following fit indices: χ²/df = 2.242, CFI = 0.966, GFI = 0.882, NFI = 0.941, IFI = 0.967, TLI = 0.959, RFI = 0.929, and RMSEA = 0.079.

Conclusion: The Chinese version of ASUAITIN demonstrates excellent reliability and validity, making it a suitable tool for evaluating Chinese clinical nurses' attitudes toward AI technology application in nursing practice.

Keywords: Artificial Intelligence; Nurses; Scale; Cross-Cultural Adaptation; Reliability; Validity; Nursing

Introduction

The rapid advancement of Artificial Intelligence (AI) has significantly propelled progress in healthcare, with numerous AI technologies—including medication dispensing systems, data mining, speech recognition, and disease assessment tools—being actively integrated into nursing practice. Research indicates that AI offers substantial advantages in optimizing clinical decision-making, facilitating nurse-patient communication, reducing workload burden, and enhancing nursing efficiency. Furthermore, AI technologies are poised to substantially assist nursing personnel in delivering evidence-based and personalized care in the future. The National Health Commission's "Action Plan for Further Improving Nursing Services (2023-2025)" explicitly mandates that medical institutions should leverage information technologies such as AI and the Internet of Things to optimize nursing service processes and innovate care delivery models.

However, concerns regarding potential ethical risks and professional displacement have created divergent attitudes among nurses toward AI adoption in their work. Studies reveal that 43% of nurses worry that AI may threaten nursing practice, while 57% believe AI could precipitate a crisis for the nursing profession. As key providers of healthcare services, clinical nurses' acceptance of AI technology directly influences its clinical implementation effectiveness, making it critically important to understand their perspectives. Unfortunately, current research predominantly focuses on AI technology development, with a notable absence of validated tools for assessing nurses' attitudes toward AI. To address this gap, Dilek Yılmaz et al. developed the Attitude Scale Towards the Use of Artificial Intelligence Technologies in Nursing (ASUAITIN) in 2025—a psychometrically sound and easily applicable instrument. This study aims to translate and culturally adapt the ASUAITIN into Chinese, providing a reliable tool for evaluating clinical nurses' attitudes toward AI technology application.

1.1 Scale Introduction

The ASUAITIN was developed by Turkish scholars Dilek Yılmaz et al. in 2025 to measure clinical nurses' attitudes toward AI technology application. The scale contains 15 items distributed across two dimensions: positive attitudes (9 items) and negative attitudes (6 items). Each item uses a 5-point Likert scale ranging from 1 ("strongly disagree") to 5 ("strongly agree"). Total scores range from 15 to 75, with higher scores indicating more positive attitudes toward AI technology application. The original scale reported Cronbach's α values of 0.910 for the total scale, 0.933 for the positive attitudes dimension, and 0.917 for the negative attitudes dimension.

1.2 Scale Translation

After obtaining permission and authorization from the original authors, the research team translated the scale using the Brislin translation model. The process involved three steps: (1) Forward translation: Two nursing faculty members proficient in both Chinese and English (with CET-6 certification) independently translated the scale into Chinese, producing initial versions A and B. The research team then compared, revised, and integrated these versions to create Chinese version C. (2) Back-translation: A nursing PhD candidate and an English-for-nursing instructor independently back-translated version C into English versions D and E. Discrepancies were discussed and modified, resulting in integrated English version F. (3) Review: Version F was sent to the original authors for evaluation regarding fidelity to the original scale. The research team iteratively repeated the forward translation, back-translation, and review process until consensus with the original authors was achieved, yielding the final Chinese version G.

1.3 Cultural Adaptation

To ensure cultural appropriateness and content equivalence, an expert committee of eight members was convened via email or face-to-face meetings to evaluate item relevance, linguistic clarity, and cultural suitability for the Chinese context. The committee comprised two computer science researchers, three nursing informatics specialists from universities, and three clinical nursing experts from tertiary Grade A hospitals. All held associate professor or deputy chief nurse positions or higher, with three holding doctoral degrees, four holding master's degrees, and one holding a bachelor's degree.

1.4 Pilot Survey

In February 2025, a pilot survey was conducted using convenience sampling with 20 nursing staff members from a tertiary Grade A hospital in Beijing. The Chinese version G was administered, and detailed feedback regarding unclear or difficult-to-understand items and wording was collected. After discussion with the eight experts, modifications were made to produce the final version.

Methods

1.5 Participants and Data Collection

1.5.1 Participants: From March to April 2025, clinical nurses from a tertiary Grade A hospital in Beijing were recruited using convenience sampling. Inclusion criteria were: (1) minimum one year of clinical work experience, and (2) voluntary participation. Exclusion criteria included: (1) nurses from non-clinical departments, and (2) nurses on leave or absent during the survey period. Sample size was calculated using the 5-10 times item number method, with an additional 20% attrition rate, yielding a minimum requirement of 188 participants. Additionally, following Wu Minglong's recommendation that exploratory factor analysis requires at least 300 participants, this study aimed for a sample size exceeding 300. The study received ethical approval from the participating institution.

1.5.2 Data Collection Method: Electronic questionnaires were distributed online via Wenjuanxing platform. The introduction explained the study's purpose, significance, and completion instructions, emphasizing anonymous participation and research-only data use. Only fully completed questionnaires could be submitted. Consenting participants clicked "confirm" to proceed, while others could exit. Completion time ranged from 2-5 minutes. The questionnaire included self-designed demographic items (gender, age, education, department, work experience, professional title) and the Chinese ASUAITIN. Submitted questionnaires were individually screened; those with identical responses across most items or completion times under 2 minutes were considered invalid.

1.6 Statistical Analysis

Data were analyzed using SPSS 26.0 and AMOS 26.0. Categorical data were described using frequencies, proportions, or percentages. Item analysis employed the critical ratio method and correlation analysis. Content validity was evaluated using the scale-level content validity index (S-CVI/Ave) and item-level content validity index (I-CVI). Structural validity was assessed through exploratory factor analysis (EFA) and confirmatory factor analysis (CFA). Reliability was evaluated using Cronbach's α coefficient for internal consistency, Pearson correlation for test-retest reliability, and Spearman-Brown coefficient for split-half reliability. Statistical significance was set at P < 0.05.

Results

2.1 Cultural Adaptation and Pilot Survey Findings

Based on expert feedback and nurses' questions during the pilot survey, the following modifications were made to the Chinese ASUAITIN: Item 2 was revised from "When I think about how AI technology will be applied in nursing in the future, I feel uneasy" to "When I think about how AI technology will be applied in nursing in the future, I feel uneasy" (refined wording). Item 3 was changed from "I believe that if AI technology is used more in the future, the nursing profession will be harmed" to "I believe that if AI technology is used more extensively in the future, the nursing profession will be harmed." Item 5 was modified from "I believe that using AI technology in nursing work is incorrect" to "I believe that using AI technology in nursing work is inappropriate."

2.2 Respondent Characteristics

A total of 400 questionnaires were distributed, with 396 valid questionnaires returned (99% valid response rate). Respondent demographics are presented in Table 1 [TABLE:1].

2.3 Item Analysis

Critical Ratio Method: Scale scores from 396 participants were ranked in descending order, with the top 27% designated as the high group and bottom 27% as the low group. Independent samples t-tests compared item scores between groups, with items having critical ratios < 3.000 or non-significant differences slated for removal. Results showed critical ratios ranging from 14.513 to 28.641, with all differences statistically significant (P < 0.01), leading to retention of all items.

Correlation Method: Pearson correlations between each item and total scale score were calculated, with items correlating < 0.4 considered for removal due to low homogeneity. All item-total correlations ranged from 0.631 to 0.822 (P < 0.01), supporting retention of all items.

2.4 Validity Analysis

2.4.1 Content Validity: Eight experts rated item relevance using a 4-point Likert scale (1 = completely irrelevant, 4 = highly relevant). Acceptable content validity was defined as I-CVI ≥ 0.78 and S-CVI/Ave ≥ 0.90. The Chinese ASUAITIN achieved I-CVI values of 0.875-1.000 and S-CVI/Ave of 0.981.

2.4.2 Structural Validity:

Exploratory Factor Analysis: The 396 questionnaires were randomly split into two groups of 198 for EFA and CFA. The KMO value was 0.918, and Bartlett's sphericity test yielded χ² = 3301.116 (P < 0.01), indicating suitability for principal component analysis. Two factors with eigenvalues > 1 were extracted, accounting for 77.402% of cumulative variance. The factor loading matrix is presented in Table 2 [TABLE:2].

Confirmatory Factor Analysis: Using AMOS 26.0, CFA results showed: χ²/df = 2.242, CFI = 0.966, GFI = 0.882, NFI = 0.941, IFI = 0.967, TLI = 0.959, RFI = 0.929, and RMSEA = 0.079, all within acceptable ranges.

2.5 Reliability Analysis

The Chinese ASUAITIN demonstrated a Cronbach's α coefficient of 0.939, split-half reliability of 0.738, and test-retest reliability of 0.935. Cronbach's α coefficients for the two dimensions were 0.945 and 0.956, respectively.

Discussion

3.1 The Chinese ASUAITIN Demonstrates Strong Practical Value

Recent breakthroughs in AI technology promise paradigm shifts in healthcare delivery, patient outcomes, and nursing roles. While AI shows remarkable advantages in patient monitoring and health management, nursing remains a highly personalized, human-centered profession characterized by unique emotional interactions, complex decision-making, and humanistic caring attributes that are irreplaceable. Consequently, systematically assessing nurses' willingness to adopt AI technology is essential. This study's cross-cultural adaptation of ASUAITIN provides a standardized measurement tool tailored to China's nursing context. The scale can evaluate clinical nurses' cognitive attitudes, usage motivations, and potential concerns regarding AI technology, offering empirical evidence for intelligent equipment allocation, stratified staff training, and human-machine collaboration optimization. In practice, assessment results can directly inform decisions about introducing AI nursing tools, guide interventions for occupational stress, and foster innovation in patient-centered smart nursing service models.

3.2 The Chinese ASUAITIN Exhibits Excellent Discriminant Validity

Item analysis revealed critical ratios ranging from 14.513 to 28.641, with all item-total correlations exceeding 0.40 (P < 0.01). These results indicate strong content consistency between individual items and the overall scale, meeting homogeneity requirements for scale development. The Chinese ASUAITIN thus demonstrates robust discriminant validity, enabling effective identification of varying attitude levels among nursing personnel toward AI technology application in their work.

3.3 The Chinese ASUAITIN Shows Strong Reliability

Reliability assesses a measurement tool's consistency, stability, and dependability. The Chinese ASUAITIN achieved a Cronbach's α coefficient of 0.939, split-half reliability of 0.738, and dimension-specific Cronbach's α coefficients of 0.945 and 0.956, indicating excellent internal consistency. Test-retest reliability analysis revealed a correlation coefficient of 0.935 over a two-week interval, demonstrating strong temporal stability. These findings confirm that the Chinese ASUAITIN possesses ideal reliability for consistently capturing clinical nurses' true attitudes toward AI technology.

3.4 The Chinese ASUAITIN Demonstrates Strong Validity

Validity refers to the degree to which an instrument accurately measures its intended psychological or behavioral constructs. Content validity results showed I-CVI values of 0.875-1.000 and S-CVI/Ave of 0.981, meeting established criteria and confirming high consistency with the original scale. Structural validity analysis revealed that exploratory factor analysis extracted two factors with eigenvalues ≥ 1, explaining 77.402% of variance, with all factor loadings > 0.5 (exceeding the 0.4 minimum threshold). Confirmatory factor analysis demonstrated acceptable model fit indices. These results indicate that the Chinese ASUAITIN exhibits ideal validity and meets psychometric standards.

Conclusion

This study successfully translated and culturally adapted the ASUAITIN into a Chinese version comprising 15 items across two dimensions. The scale demonstrates excellent reliability, validity, brevity, clarity, and comprehensibility, making it a robust tool for evaluating Chinese nursing staff attitudes toward AI technology application. However, this study was limited to a single tertiary hospital in Beijing. Future research should expand sample size and conduct multi-center investigations to provide stronger evidence for clinical nursing administrators developing AI integration strategies.

References

[1] Booth RG, Strudwick G, McBride S, O'Connor S, López ALS. How the nursing profession should adapt for a digital future[J]. BMJ. 2021;373:1–5.

[2] Topaz M, Murga L, Gaddis KM, et al. Mining fall-related information in clinical notes: Comparison of rule-based and novel word embedding-based machine learning approaches[J]. J Biomed Inform. 2019;90:103103.

[3] Ronquillo CE, Peltonen LM, Pruinelli L, et al. Artificial intelligence in nursing: Priorities and opportunities from an international think-tank of the Nursing and Artificial Intelligence Leadership Collaborative[J]. J Adv Nurs. 2021;77(9):3707-3717.

[4] Hobensack M, von Gerich H, Vyas P, et al. A rapid review on current and potential uses of large language models in nursing[J]. Int J Nurs Stud. 2024;154:104753.

[5] Rony MKK, Parvin MR, Ferdousi S. Advancing nursing practice with artificial intelligence: Enhancing preparedness for the future[J]. Nurs Open. 2024;11(1):10.1002/nop2.2070.

[6] Abuzaid MM, Elshami W, Fadden SM. Integration of artificial intelligence into nursing practice[J]. Health Technol (Berl). 2022;12(6):1109-1115.

[7] Topaz M, Murga L, Gaddis KM, et al. Mining fall-related information in clinical notes: Comparison of rule-based and novel word embedding-based machine learning approaches[J]. J Biomed Inform. 2019;90:103103.

[8] National Health Commission. Notice on Issuing the Action Plan for Further Improving Nursing Services (2023-2025)[Z]. Gazette of the National Health Commission of the People's Republic of China, 2023(6):5-8.

[9] von Gerich H, Moen H, Block LJ, et al. Artificial Intelligence-based technologies in nursing: A scoping literature review of the evidence[J]. Int J Nurs Stud. 2022;127:104153.

[10] Rony MKK, Parvin MR, Ferdousi S. Advancing nursing practice with artificial intelligence: Enhancing preparedness for the future[J]. Nurs Open. 2024;11(1):10.1002/nop2.2070.

[11] Abuzaid MM, Elshami W, Fadden SM. Integration of artificial intelligence into nursing practice[J]. Health Technol (Berl). 2022;12(6):1109-1115.

[12] Yılmaz D, Uzelli D, Dikmen Y. Psychometrics of the Attitude Scale towards the use of Artificial Intelligence Technologies in Nursing[J]. BMC Nurs. 2025;24(1):151.

[13] Jones PS, Lee JW, Phillips LR, Zhang XE, Jaceldo KB. An adaptation of Brislin's translation model for cross-cultural research[J]. Nurs Res. 2001;50(5):300-304.

[14] Guo JY, Li Z. Process and evaluation criteria for scale introduction[J]. Chinese Journal of Nursing, 2012,47(3):283-285.

[15] Hu Y, Wang ZW. Nursing Research[M]. 6th ed. Beijing: People's Medical Publishing House, 2017:123-126.

[16] Wu ML. Structural Equation Modeling: AMOS Operation and Application[M]. 2nd ed. Chongqing: Chongqing University Press, 2017:241-243.

[17] Zhang C, Zhou YX. Analysis of Misapplications of Content Validity Index in Chinese Nursing Measurement Tool Literature[J]. Journal of Nursing Science, 2020,35(4):86-88,92.

[18] Li Z, Zhou M, Chen LN, et al. Cross-cultural adaptation and psychometric validation of the Work Quality of Life Scale for cancer patients[J]. Chinese Journal of Nursing, 2023,58(6):695-700.

[19] Cary MP Jr, De Gagne JC, Kauschinger ED, Carter BM. Advancing Health Equity Through Artificial Intelligence: An Educational Framework for Preparing Nurses in Clinical Practice and Research[J]. Creat Nurs. 2024;30(2):154-164.

[20] Ibuki T, Ibuki A, Nakazawa E. Possibilities and ethical issues of entrusting nursing tasks to artificial intelligence[J]. Nurs Ethics. 2024;31(6):1010-1020.

[21] Wang XT, Ren JX, Wang HX. Cross-cultural adaptation and psychometric validation of the Middle-aged Women's Aging Anxiety Scale[J]. Nursing Research, 2025,39(7):1114-1119.

Author Contributions

Shang Li and Li Ye: Conceptualization, study design, manuscript writing and revision. Xu Lanlan and Nie Xiaofei: Scale translation and cultural adaptation. Luo Yixue and Wang Dian: Data analysis and organization.

Preprint Declaration

1. This manuscript is a preprint archive with sensitive information desensitized.
2. Display of scale items has been authorized by the original developer.
3. The final published version should be considered authoritative; please cite using the final DOI.