Zhong Nan Da Xue Xue Bao Yi Xue Ban. 2026 Jun 28;51(6):1149-1159. doi: 10.11817/j.issn.1672-7347.2026.260006.
ABSTRACT
OBJECTIVES: With the rapid development of smart nursing, artificial intelligence (AI) systems for the dynamic assessment of venous thromboembolism (VTE) risk have gradually been introduced into clinical practice. As the direct users of such systems, nurses' acceptance substantially affects their effectiveness in clinical implementation. However, it remains unclear whether there is population heterogeneity in nurses' acceptance of AI-based dynamic VTE risk assessment. This study aims to identify latent profiles of nurses' acceptance of AI-based dynamic VTE risk assessment using latent profile analysis (LPA) and to explore the associated factors, thereby providing empirical evidence for development of targeted and stratified implementation and training strategies.
METHODS: Convenience sampling was used to survey nurses from tertiary hospitals in southern, central, and northern Anhui Province between April and June 2025. Data were collected using a general information questionnaire, the Clinical Nurses' Behavior Scale for the Prevention of VTE in Hospitalized Patients, the Nurses' Professional Attitude Scale, and the Questionnaire on Nurses' Acceptance of AI-Based Dynamic VTE Risk Assessment. LPA was performed to identify latent profiles of nurses' acceptance, and multinomial Logistic regression analysis was used to examine the factors associated with profile membership.
RESULTS: A total of 2 323 valid questionnaires were collected. Nurses' acceptance of AI-based dynamic VTE risk assessment was classified into 3 latent profiles: the low-acceptance-concerned resistance group [388 nurses (16.8%)], the moderate-acceptance-cautious observation group [1 177 nurses (50.5%)], and the high-acceptance-active adoption group [758 nurses (32.7%)]. Years of clinical experience, frequency of night shifts, scores on the knowledge dimension of the Clinical Nurses' Behavioral Scale for the Prevention of VTE in Hospitalized Patients, and scores on the Nurses' Professional Attitude Scale were factors associated with the latent profiles of nurses' acceptance of AI-based dynamic VTE risk assessment (all P<0.05). After adjustment for covariates, with the low-acceptance-concerned resistance group as the reference, higher scores on the Nurses' Professional Attitude Scale were associated with significantly greater likelihoods of membership in both the moderate-acceptance-cautious observation group and the high-acceptance-active adoption group, with odds ratios (OR) of 1.023 and 1.074, respectively (both P<0.001). Nurses who rarely or never work night shifts and those working 2-3 night shifts per month were more likely to be classified into the moderate-acceptance-cautious observation group (OR=1.773, OR=2.243, P<0.05) and the high-acceptance-active adoption group (OR=3.964, OR=4.300, P<0.001). Nurses with ≤5 years of clinical experience (OR=4.942, P<0.001) and those with 10 to 19 years of clinical experience (OR=3.361, P=0.006) were more likely to belong to the high-acceptance-active adoption group. Nurses with higher scores on the knowledge dimension of the Clinical Nurses' Behavior Scale for the Prevention of VTE in Hospitalized Patients were also more likely to belong to the high-acceptance-active adoption group (OR=1.042, P=0.009).
CONCLUSIONS: There is substantial population heterogeneity in nurses' acceptance of AI-based dynamic VTE risk assessment. Hospital administrators should pay particular attention to nurses in the low-acceptance-concerned resistance and moderate-acceptance-cautious observation groups and implement targeted measures to improve their acceptance of this new technology.
PMID:42702378 | DOI:10.11817/j.issn.1672-7347.2026.260006