| Title |
Exploring the predictors of nurses’ turnover intentions through neural network modeling: a national cross-sectional study in Lithuania |
| Authors |
Žiedelis, Arūnas ; Lazauskaitė-Zabielskė, Jurgita ; Istomina, Natalja ; Urbanavičė, Rita ; Stanislavovienė, Jelena |
| DOI |
10.3390/healthcare14070831 |
| Full Text |
|
| Is Part of |
Healtcare.. Basel : MDPI. 2026, vol. 14, iss. 7, art. no. 831, p. [1-19].. eISSN 2227-9032 |
| Keywords [eng] |
turnover intentions ; nurse retention ; neural network modeling |
| Abstract [eng] |
Background/Objectives: Nurses’ turnover intentions are strong predictors of actual turnover, which increases costs, reduces care quality, and destabilies the workforce. This study aimed to identify the key predictors of nurses’ turnover intentions using advanced machine learning methods and to explore how demographic, well-being, and work environment factors contribute to these intentions. Methods: Cross-sectional data were collected from 2459 nurses employed across various healthcare institutions. We used multichannel invitation and snowball sampling. An artificial neural network regression model was applied, combined with iterative feature selection and SHAP analysis, to identify the most important predictors of turnover intentions and to examine nonlinear and context-dependent relationships among variables. Results: Seven predictors explained 49.8% of the variance in turnover intentions, outperforming traditional linear models. Age was the strongest predictor, with younger nurses demonstrating a substantially higher likelihood of intending to leave; this association was nonlinear, with intentions decreasing more sharply at older ages. Job satisfaction and burnout were also strong predictors, particularly among younger nurses. Four work environment factors further contributed to turnover intentions: managerial support functioned as a protective factor, interpersonal conflict increased intentions to leave, limited professional development opportunities were strongly associated with higher turnover intentions, and role conflict showed heterogeneous effects. Conclusions: Machine learning approaches enhance understanding of complex workforce dynamics and enable more precise identification of high-risk groups. The findings support age-sensitive retention strategies, proactive monitoring of nurse well-being, and organizational interventions to strengthen managerial support and professional development, ensuring workforce stability and sustainable healthcare service delivery. |
| Published |
Basel : MDPI |
| Type |
Journal article |
| Language |
English |
| Publication date |
2026 |
| CC license |
|
