Italian Nurses Get a New Tool to Predict Who Will Quit

Nursing is a profession under strain almost everywhere on Earth, and the numbers behind that strain are sobering. The World Health Organization and the International Council of Nurses estimate that by 2030 the world will be short roughly nine million nurses, a gap driven by aging populations, rising chronic disease, and working conditions that push experienced clinicians out of hospitals faster than they can be replaced. When a nurse leaves, the cost is not merely administrative. Losing seasoned professionals erodes intellectual capital, reduces productivity, and has been linked to poorer patient satisfaction, compromised staff safety, longer hospital stays, and lower quality of care. Organizations must then spend heavily to recruit, hire, and train replacements, while the nurses who remain often face heavier workloads, diminished job satisfaction, and weakened team cohesion. Against this backdrop, a research team in Italy has taken a step that could give hospital managers an early warning system: a carefully translated and preliminarily validated Italian version of the Anticipated Turnover Scale, a short questionnaire designed to measure whether nurses are thinking about walking out the door.

The Anticipated Turnover Scale, or ATS, was developed by Jan R. Atwood and Ada Sue Hinshaw in 1983 and has long been prized for its simplicity and reliability. The instrument consists of twelve items rated on a Likert scale, asking nurses how strongly they agree or disagree with statements about their intention to leave their current position, whether by moving to another hospital or transferring within the same organization. In its original form, responses ranged from 1, strongly disagree, to 7, strongly agree, producing total scores between 12 and 84, with higher scores signaling a greater intention to leave. The scale deliberately mixes positively and negatively worded items to blunt response bias, and it takes only about five minutes to complete. In the original validation, Cronbach’s alpha, a standard measure of internal consistency, was 0.84, and factor analysis revealed two main factors explaining 54.9 percent of the variance. A later study of newly graduated nurses using a five-point format achieved an alpha of 0.90. Until now, however, no reliable Italian-language version existed.

That gap mattered because intention to leave is widely regarded as the single best early indicator of actual turnover. The cognitive and emotional process of considering, planning, and deciding to quit precedes resignation, and research consistently shows a strong association between stated intention and eventual departure. Known drivers include job dissatisfaction, burnout, stress, work-life conflict, low pay, weak professional engagement, patient safety concerns, and leadership styles, including toxic leadership that breeds organizational mistrust. The COVID-19 pandemic intensified the problem dramatically, with roughly one-third of nurses reporting that they considered leaving the profession because of burnout and psychological distress. Generational patterns add another layer: Millennials change jobs more readily than Baby Boomers or Generation X, though across all generations, workload and staffing shortages dominate the reasons for leaving, while leadership support and salary anchor retention. Giving nurse managers a validated tool to detect these intentions early is therefore essential for planning retention strategies before resignations cascade.

The Italian team, publishing in Nursing Open, began by securing official authorization from Atwood and Hinshaw themselves, including permission to adopt a five-point Likert format instead of the original seven-point scale. The researchers chose the simplified format to improve usability for respondents, consistent with its successful use in prior studies. Translation and cultural adaptation then followed a rigorous four-step protocol aligned with internationally recommended procedures. Four translators with deliberately different profiles produced independent translations: a certified translator with no healthcare background, a non-healthcare translator with bicultural Italian-American experience, a nurse translator, and a nurse translator with bicultural Italian-English experience. Each was instructed to translate the instrument and its instructions while noting any difficulties. A bilingual committee, including a Director of Health Professions, a Departmental Nursing Manager, a Nursing Coordinator, a Labor Sociologist, a Psychologist, and an Associate Professor of Nursing, then reviewed the versions using a decentralized approach to preserve meaning in both languages.

The third and fourth steps closed the loop on linguistic fidelity. Two independent native English-speaking translators, one with and one without a healthcare background, back-translated the Italian version into English without ever seeing the original instrument or its translations. The research team then compared the two back-translated English versions, discussed discrepancies, and proposed a final version that the original author confirmed. With the translated scale in hand, the team moved to preliminary validation in a real-world setting: a public university hospital in northeastern Italy. All hospital nurses were invited to participate, with freelance nurses and those in managerial positions excluded. A self-administered questionnaire was distributed by email, and to minimize response bias the ATS was administered first, before any socio-demographic or occupational questions. Data collection ran from October 31 to November 20, 2023, through an anonymous online platform. Of 300 questionnaires distributed, 172 were returned and 168 were deemed valid for analysis.

The psychometric evaluation examined three pillars: face validity, content validity, and the internal structure of the scale. A panel of six experts judged that the translated instrument clearly and appropriately captured the phenomenon of nursing turnover, confirming its face validity without modifications and highlighting its potential for both research and practice. Content validity told a more nuanced story. Experts rated each item on a four-point relevance scale, with scores converted into binary relevant or not relevant categories and analyzed with a binomial test. Items 2 and 12 achieved unanimous relevance ratings with statistically significant agreement, while items 1, 5, 6, and 11 showed high though non-significant agreement. But items 7 and 9 were judged not relevant by five of the six evaluators, yielding item-level content validity index values of just 0.17. The scale-level index stood at 0.68, rising to 0.78 when the two problematic items were excluded. The authors caution that with only six evaluators, a single disagreement can substantially shift these indexes, so the estimates deserve careful interpretation.

The internal structure was explored through Principal Axis Factoring with Promax rotation, and the data proved highly suitable for factor analysis, with a Kaiser-Meyer-Olkin value of 0.891 and a significant Bartlett’s test of sphericity. For the full twelve-item version, the analysis extracted three factors explaining 57.82 percent of the variance. The dominant first factor, accounting for 44.59 percent, appeared to reflect a general attitude toward staying in or leaving the current job, primarily capturing internal turnover, meaning movement within the same organization. The second factor seemed tied to external turnover, the inclination to leave the organization entirely, with item 11, which expresses serious doubts about remaining, loading at a full 1.0. The third factor was associated almost exclusively with item 9, which conveys uncertainty about how long to stay in the current role. When items 7 and 9 were removed, a second factor analysis on the remaining ten items produced two factors explaining 59.16 percent of the variance, mirroring the internal-versus-external distinction.

Reliability results strengthened the case for the shortened version. The full twelve-item Italian scale achieved a Cronbach’s alpha of 0.828, already indicating good internal consistency, but excluding items 7 and 9 pushed the coefficient to 0.901, a level considered excellent. The first factor extracted from the twelve-item analysis showed an alpha of 0.912, while the first factor of the ten-item solution reached 0.809. These values sit comfortably within the range reported across four decades of ATS research: the original 1983 study reported 0.84, a Portuguese validation in Lisbon found 0.87 for twelve items rising to 0.91 after item exclusions, and a meta-analysis by Barlow and Zangaro documented alphas between 0.85 and 0.94 with a mean of 0.89. The Italian coefficients are slightly lower but largely acceptable, and the improvement after removing items 7 and 9 echoes the Portuguese findings, suggesting those items may pose cultural or linguistic difficulties beyond Italy alone.

Comparing validations across languages and forty years of healthcare change requires caution, and the authors are explicit about the limits of their work. The original development involved 1,525 professionals from fifteen rural Arizona hospitals, while the Portuguese study of 259 nurses is more comparable in scale. The Italian factor structure differs from previous validations even though it explains a similar proportion of variance, and no confirmatory factor analysis was performed, which would be needed to verify the observed structures in an independent sample. Criterion validity could not be assessed because no Italian gold standard for turnover measurement exists, and establishing a clinical cut-off score was infeasible due to privacy and timing constraints. Temporal stability, or test-retest reliability, remains untested, and the six-member expert panel may have been too small to fully capture content validity.

Even with these caveats, the implications for nursing management are concrete. A validated Italian instrument for measuring intention to leave would let nurse managers monitor turnover risk in real time and design targeted, personalized interventions to boost job satisfaction, engagement, and well-being before resignations occur. The authors conclude that the ten-item version, with its stronger content validity and internal consistency, is closest to practical use, though the scale is not yet ready for routine application. Future research should confirm the factor structure with independent samples, establish test-retest reliability, and link scores to actual turnover to define a meaningful threshold. In the meantime, the study adds an important piece to the global puzzle of nurse retention, offering Italian healthcare systems a scientifically grounded way to hear, earlier and more clearly, the quiet signal that precedes a resignation letter.

Subject of Research: Italian translation and preliminary psychometric validation of the Anticipated Turnover Scale for measuring nurses' intention to leave

Article Title: Evaluation of Nurses' Intent to Leave: Italian Translation, Cultural Adaptation, and Initial Psychometric Assessment of the Anticipated Turnover Scale

Article References: Cignola, S., Bembich, S., & Sanson, G. (2026). Evaluation of Nurses' Intent to Leave: Italian Translation, Cultural Adaptation, and Initial Psychometric Assessment of the Anticipated Turnover Scale. Nursing Open, 13(10), Article e70848. https://doi.org/10.1002/nop2.70848

Image Credits: AI Generated

DOI: 10.1002/nop2.70848

Keywords: nursing, turnover intention, Anticipated Turnover Scale, psychometric validation, cross-cultural adaptation, nurse retention, health workforce shortage, Cronbach's alpha, factor analysis, content validity, burnout, hospital management