When final-year nursing students step onto a hospital ward for the first time, the gap between classroom knowledge and clinical reality can be jarring. A new study from China suggests that this transition shock is not evenly distributed: most nursing interns are entering the workforce with low work readiness, and the factors that separate them from their better-prepared peers include something distinctly modern — how literate they are in generative artificial intelligence.
Researchers led by Yanfang Zhang of Bengbu Medical University surveyed 712 final-year nursing interns from three nursing schools in Anhui Province between November 2025 and June 2026. All participants had completed at least eight months of clinical practicum and were approaching graduation in 2026. The team published its findings in Frontiers in Medicine, using a statistical technique called Latent Profile Analysis to sort interns into distinct readiness groups rather than treating work readiness as a single continuous score.
Work readiness is a multidimensional construct. It encompasses professional knowledge, clinical skills, attitudes, psychosocial characteristics, and workplace adaptability — everything a person needs to enter and adapt to the workforce. For nursing interns, it is the bridge between academic learning and professional practice, a stage marked by role transformation and employment challenges. Traditional academic achievement alone, the researchers note, may not fully capture a student’s ability to adapt to complex clinical environments.
The results were striking. Three latent profiles emerged from the data: a low work readiness group of 415 interns (58.3 percent of the sample), a moderate group of 227 (31.9 percent), and a high group of just 70 (9.8 percent). In other words, nearly six in ten nursing interns about to graduate showed consistently low scores across all five dimensions of the Nursing Practice Readiness Scale, which measures clinical judgment and nursing performance, professional attitude, patient-centeredness, self-regulation, and collaborative interpersonal relationships.
The statistical machinery behind the classification was rigorous. The team entered the five dimension scores of the readiness scale as continuous indicators into models with one to five profiles, estimated using Mplus version 8.3. Fit was evaluated with the Akaike Information Criterion, the Bayesian Information Criterion, entropy, and likelihood ratio tests. Although a five-profile model had the lowest information criteria values, the Lo–Mendell–Rubin test was not statistically significant, and a four-profile solution contained a subgroup of only 6.5 percent of the sample, raising concerns about stability. The three-profile model, with favorable entropy of 0.864 and a more balanced distribution, was retained as the most appropriate solution.
What separated the groups? Guided by Social Cognitive Career Theory, which frames career development as an interaction between personal factors and contextual resources, the researchers examined the clinical learning environment and generative AI literacy alongside academic performance, clinical learning outcomes, and sense of professional belonging. The overall multinomial logistic regression model was highly significant (χ² = 467.253, P < 0.001), and all five factors were significantly associated with profile membership.
Generative AI literacy proved to be a particularly powerful predictor. Interns with higher scores on the Generative Artificial Intelligence Literacy Scale — which covers awareness of GenAI, knowledge construction, competency development, mindset, and ethical cognition — had significantly greater odds of belonging to the moderate profile rather than the low one (OR = 1.161 per unit, 95% CI: 1.128–1.195), and even stronger odds for the high profile (OR = 1.321, 95% CI: 1.259–1.387). The mean GenAI literacy score climbed steadily across profiles, from 124.88 in the low group to 153.16 in the high group. The researchers suggest that interns with greater GenAI literacy may be better able to access clinical information, organize learning resources, and use digital tools, which may strengthen digital self-efficacy and positive outcome expectations.
The clinical learning environment also mattered. Each one-unit increase in Clinical Learning Environment Questionnaire scores — covering clinical case exposure, patient care organization, preceptor supervision, and learning motivation — raised the odds of moderate (OR = 1.035) or high (OR = 1.043) profile membership. Better clinical learning outcomes and a stronger sense of professional belonging were likewise associated with higher readiness profiles, with the belonging effect especially pronounced in the high-profile comparison (fair belonging: OR = 56.106, though with wide confidence intervals reflecting the small subgroup). Academic performance showed a differentiated pattern: below-average and average records were significant negative predictors of higher readiness classification, but the researchers caution that work readiness encompasses non-cognitive competencies such as communication, teamwork, and situational decision-making that grades alone cannot capture.
The study has limitations the authors acknowledge. Participants were recruited by convenience sampling from only three institutions in Anhui Province, data were self-reported, and the cross-sectional design cannot establish causality. The Nursing Practice Readiness Scale was originally developed for newly employed nurses, and classification uncertainty in latent profile analysis may have affected some estimates, particularly where subgroups were small.
Still, the findings carry practical weight. Nearly 60 percent of soon-to-graduate nursing interns fell into the low readiness profile, suggesting that standardized training approaches may not address the diverse developmental needs revealed by the analysis. The researchers propose individualized educational support strategies: strengthening foundational clinical training and simulation for the low-readiness majority, offering case-based learning and supervised clinical decision-making for the moderate group, and providing peer mentoring opportunities for the high-readiness minority. Incorporating GenAI-related activities into curricula, they suggest, may support students’ digital learning competencies at a time when intelligent healthcare is reshaping the profession. Future longitudinal and intervention studies, the team concludes, are needed to clarify temporal relationships and evaluate which strategies actually work.
Subject of Research: A cross-sectional study of 712 Chinese nursing interns used latent profile analysis to identify low, moderate, and high work readiness groups and found that generative AI literacy, the clinical learning environment, academic performance, clinical learning outcomes, and professional belonging were associated with profile membership.
Article Title: Latent profiles of work readiness and their associated factors among nursing interns: a cross-sectional study
Article References: Zhang, Y., Xie, X., Wu, M., Lv, Q., Zhao, L., & Yang, X. (2026). Latent profiles of work readiness and their associated factors among nursing interns: a cross-sectional study. Frontiers in Medicine, 13, Article 1937251. https://doi.org/10.3389/fmed.2026.1937251
Image Credits: AI Generated
DOI: 10.3389/fmed.2026.1937251
Keywords: nursing interns, work readiness, latent profile analysis, generative AI literacy, clinical learning environment, Social Cognitive Career Theory, nursing education, cross-sectional study, professional belonging, academic performance, Latent, profiles
News Source: Ophelia Keating. (October 16, 2026). Most nursing interns are not ready for the ward, and AI literacy may help decide who is. Scienmag.
