JMIR Publications

ABSTRACT

Background:

Inequitable and time-consuming shift scheduling contributes to nurse burnout, dissatisfaction, and turnover. In Taiwan, annual nurse turnover exceeds 11%, and rigid 3-shift systems combined with perceived unfairness in workload distribution are frequently cited concerns. Although AI scheduling tools exist, most lack transparency and do not adequately address nurses’ concerns about fairness and trust, limiting their adoption in practice.

Objective:

This study aimed to develop and evaluate a transparent, nurse-centered scheduling decision support system designed to reduce administrative burden, improve workload equity, and enhance staff acceptance in routine clinical settings.

Methods:

We conducted a pragmatic before-and-after implementation study at a 677-bed teaching hospital in Taiwan, involving 8 nursing departments and 156 nurses. A 6-month manual scheduling period was compared with a 6-month period using the new AI scheduling system. The system supported nurse managers by providing predictive workload insights, transparent explanations for scheduling decisions, and real-time equity monitoring. Outcomes included scheduling time, scheduling errors, workload variation, preference satisfaction, and user acceptance. Statistical analyses included linear mixed-effects and generalized estimating models.

Results:

Implementation reduced monthly scheduling time by 81.2% (32.0±8.0 to 6.0±2.0 hours; p<.001) and decreased scheduling errors by 73.8% (18.3% to 4.8%; p<.001). Nurse satisfaction increased significantly (3.2±0.8 to 4.4±0.6; p<.001), and routine adoption reached 94% by Month 3. Workload distribution became substantially more equitable, with reduced variation in shift allocation and elimination of experience-related disparities. Preference satisfaction was evenly distributed across staff levels. Greater engagement with schedule explanations was associated with higher satisfaction (r=0.456; p<.001).

Conclusions:

A transparent and fairness-oriented scheduling system can meaningfully reduce managerial workload, enhance perceived equity, and improve nurse acceptance in real-world practice. These findings suggest that explainable AI tools may support nurse well-being and promote more sustainable workforce management in hospital settings.