Nurse-perceived workload, staffing, and patient turnover: An analysis of longitudinal routine data.

Background: Nursing workload is a critical factor in ensuring safe, high-quality patient care and maintaining nurse well-being. While high workload has been linked to adverse outcomes for both patients and staff, its assessment remains challenging. Common approaches rely on activity- or dependency-based measures or professional judgement, yet these often overlook important subjective and contextual factors. Nurse-perceived workload, particularly at the shift level, offers valuable insights but has received limited attention in research and is rarely used to guide staffing decisions. Objective: This retrospective longitudinal study explored nurse-perceived workload on shift level and its associations with nurse staffing, patient comorbidity score and turnover. Design: Retrospective observational study using longitudinal shift-level routine data. Settings: Five internal medicine units in a large Swiss acute care hospital. Participants: 131 registered nurses (RN). Methods: We analysed five years (2015-2019) of routinely collected data from five internal medicine units of a Swiss tertiary hospital. Nurse-perceived workload was documented at the end of each shift using a single-item rating. Staffing system data (e.g., shift assignments, working time) were merged with patient discharge data (e.g., diagnoses, demographics, admission). Descriptive statistics, multilevel logistic regression, and a gradient boosting model were applied. Results: High workload was classified in 9,646 of 42,035 nurse-shift observations (22.9%). The number of incoming patients showed the strongest positive association with high workload perception (OR 1.30; 95% CI: 1.26-1.33). Higher RN staffing relative to expected staffing was associated with lower odds of high workload (OR 0.84; 95% CI: 0.78-0.91). Patient comorbidities also showed a positive association (OR 1.07; 95% CI: 1.06-1.08), as well as the number of outgoing patients (OR 1.02; 95% CI: 1.004-1.04). The gradient boosting model showed limited predictive performance. Conclusions: This study is among the first to investigate nurse-perceived workload using extensive longitudinal routine data. The findings suggest that single-item ratings are a practical means of capturing workload perception at scale. High workload perception was common and was positively associated with patient inflow and comorbidity, while higher staffing levels were associated with lower odds of high workload. These results underscore the potential of integrating subjective workload data into staffing policy and workforce planning to better support nurses and maintain care quality. Registration: not registered. Social media abstract: Nurse workload isn't just personal. It's part of the system: 23% of nurse-shift observations were classified as high workload; odds were higher with admissions, complexity, and lower with better staffing.