AI-assisted Shift Planning in Healthcare – Opportunities and Challenges

Shift planning is central to job satisfaction in healthcare and often triggers conflicts between work and private life. An interdisciplinary project by the Bern University of Applied Sciences shows the opportunities and challenges of AI and why fairness, transparency, and participation are crucial in designing duty rosters.

Shift scheduling plays a central role in the job satisfaction of nursing staff and is one of the main causes of conflicts between work and private life [1] . Shift work, rigid schedules, and limited opportunities for participation often lead to frustration, stress, and increased turnover [2] . Possible consequences include early career exits and a further aggravation of staff shortages in the healthcare sector. A flexible and employee-oriented approach to shift planning could therefore make an important contribution to reducing these negative effects while simultaneously promoting the long-term satisfaction and retention of nursing staff.

This is where the interdisciplinary research project “Participatory Shift Planning” by the Bern University of Applied Sciences (BFH) came in [3] . The project aimed to identify the central causes of conflict between work and private life and to shed light on the perspectives of nursing professionals, assistant staff, and management regarding fairness, transparency, and the meaningful use of Artificial Intelligence (AI) in shift planning. At the beginning of the project, a systematic literature review was conducted, examining over 230 scientific studies on the topic of shift planning and AI in healthcare. This foundation was supplemented by focus group interviews with 22 healthcare professionals from hospitals, home care organizations, and nursing homes [4] .

AI as an Opportunity – Potential for more Fairness and Transparency

The use of AI in shift scheduling was viewed predominantly positively by the surveyed professionals. Particularly highlighted was the potential to make planning decisions more objective and minimize personal influence. Many hope this will lead to a fairer distribution of shifts and a noticeable reduction in workload for planning teams.

Especially in planning processes that are strongly influenced by personnel or structural factors, AI is perceived as a neutral authority that creates more transparency and can reduce existing imbalances. From a practical perspective, the following advantages were particularly mentioned:

• Objectivity and neutrality: Planning decisions are based on transparent, predefined criteria. Subjective biases and personal preferences can thus be reduced.
• Time savings through automation: The planning effort is significantly reduced, relieving managers and giving them more freedom for management tasks or direct care work.

Limitations and Challenges in the Use of AI

Despite the potential shown, critical aspects related to the use of AI also became evident. A central challenge is the lack of emotional intelligence in AI systems. Personal stress situations or team conflicts cannot be automatically detected or adequately considered – human judgment and empathy remain indispensable here.

The quality of the underlying data was also critically questioned. The functionality and fairness of an AI system depend significantly on the timeliness, completeness, and correctness of the available data. Outdated or incomplete information can lead to erroneous or perceived unfair scheduling.

Another central aspect is trust in the technology. This only emerges when employees can understand the logic behind the shift scheduling. A transparent, comprehensible, and clearly communicated planning logic is therefore essential, not only for acceptance but also for active participation by the team.

Requirements for AI-supported Shift Planning

In dialogue with planning managers and nursing staff, four central requirements for AI-supported shift planning were formulated. It also became clear that expectations regarding fairness, transparency, and participation apply regardless of whether the planning is done manually or digitally.

1. Time savings and flexibility: Shift planning should be more efficient than existing systems and without additional administrative effort. At the same time, individual preferences and life situations must be taken into account. Short-term changes, such as shift swaps, should be easy and user-friendly to enter.
2. Fair and dynamic shift distribution: The distribution of early, late, night, and weekend shifts should not only be objectively fair but also subjectively perceived as equitable by considering individual needs and constraints. Intelligent absence management and automated overtime administration are considered additional important functions.
3. Transparency and co-creation: Nursing staff want to understand how duty rosters are created. A transparent, clearly communicated logic and user-friendly interfaces for entering requests and feedback promote trust in the technology and strengthen acceptance within the team.
4. AI as support – not as replacement: The technology should not completely replace human judgment. Especially in complex or emotionally challenging situations, human judgment and leadership skills remain essential.

The Technical Implementation

As part of the project, various AI methods such as Mixed-Integer Programming, Constraint Programming, Genetic Programming, and Reinforcement Learning were examined for their suitability regarding the requirements. The results show that especially the consideration of individual preferences presents a technical challenge and that too high a dependence on the technology poses a risk to practicality [5] .

Conclusion: AI-assisted Shift Planning Requires Participation and Transparency

AI-assisted shift planning offers great potential to improve the compatibility of work and private life in nursing. However, the key to successful implementation lies in a participatory, transparent, and fair planning process. Only when nursing staff are actively involved in the design and can understand the underlying logic of shift planning does the necessary trust develop – as a foundation for acceptance and sustainable use. Technology cannot replace human judgment but can meaningfully complement it. What’s crucial is a balanced interplay of technical efficiency and human experience. A market-ready AI solution that is transparent and enables participation still requires several development steps.

References

1 Peter, K. A., Voirol, C., Kunz, S., Gurtner, A., Renggli, F., Juvet, T., & Golz, C. (2024). Factors associated with health professionals’ stress reactions, job satisfaction, intention to leave and health-related outcomes in acute care, rehabilitation and psychiatric hospitals, nursing homes and home care organisations. BMC Health Services Research, 24(1), 269. https://doi.org/10.1186/s12913-024-10718-5

2 Webster, B., & Archibald, D. (2022). Self-rostering, work-life balance and job satisfaction in UK nursing: A literature review. Nursing Management (Harrow, London, England : 1994). https://doi.org/10.7748/nm.2022.e2048

3 Berner Fachhochschule (BFH). (o. J.). Partizipative Dienstplanung. Abgerufen 22. April 2025, von https://www.bfh.ch/de/forschung/forschungsprojekte/2024-700-173-738/

4 Gerlach, M., Renggli, F. J., Bieri, J. S., Sariyar, M., & Golz, C. (2024). Exploring Nurse Perspectives on AI-Based Shift Scheduling for Fairness, Transparency, and Work-Life Balance. In Review. https://doi.org/10.21203/rs.3.rs-5248710/v1

5 Renggli, F. J., Gerlach, M., Bieri, J. S., Golz, C., & Sariyar, M. (2024). Integrating Nurse Preferences into AI-Based Scheduling Systems: Qualitative Study (Preprint). JMIR Formative Research. https://doi.org/10.2196/67747