Factors associated with the frequency of turnover intention among hospital nurses in Japan: an ordinal logistic regression analysis

Background Nurse turnover significantly affects healthcare systems worldwide, particularly in aging societies such as Japan. Although workplace harassment, workload, and sickness presenteeism are known contributors to nurses’ turnover intention, most studies have conceptualized turnover intention as a binary outcome. This study examined factors associated with turnover intention frequency among hospital nurses in Japan using an ordinal approach. Methods A nationwide cross-sectional, web-based survey was conducted in March 2025 among nurses employed in hospitals with 20 or more inpatient beds. Data from 625 responses were analyzed to assess turnover intention, defined as the frequency of considering leaving the current hospital during the past year (five-level ordinal scale). Explanatory variables included experiences of workplace harassment (from patients/families, coworkers, supervisors, and other staff), perceived busyness, sickness presenteeism, workplace safety measures, psychological safety, and socio-demographic and work-related characteristics. Sickness presenteeism was assessed using a validated Japanese presenteeism scale and converted to a 0%–100% productivity loss score. Factors associated with higher frequencies of turnover intention were identified by multivariable ordinal logistic regression analysis. Results Higher sickness presenteeism (odds ratio [OR] = 1.019, 95% confidence interval [CI] = 1.011–1.027) and greater perceived busyness (OR = 1.278, 95% CI = 1.179–1.385) were independently associated with higher turnover intention frequency. A dose–response relationship was observed between turnover intention and experiences of workplace harassment. Compared with nurses reporting no harassment, those who experienced harassment on two to four instances had significantly higher odds of frequent turnover intention (ORs = 1.94–2.27). Psychological safety was not significantly associated with turnover intention after adjustment. Nurses without caregiving responsibilities were more likely to report frequent turnover intention (OR = 1.97, 95% CI = 1.10–3.53). Conclusions Turnover intention among hospital nurses should be examined as a graded process rather than a binary outcome. Sickness presenteeism, perceived busyness, and cumulative workplace harassment were significantly associated with more frequent consideration of leaving. Early identification of productivity loss and excessive workload, along with comprehensive strategies to prevent repeated harassment, may be essential for retaining the nursing workforce.