Personnel Planning in Health Care: Optimization of Nurse and Physician Rosters with real-life applications
Doctoral thesis
Permanent lenke
https://hdl.handle.net/11250/2980011Utgivelsesdato
2021Metadata
Vis full innførselSamlinger
Sammendrag
Personnel rostering is challenging in multiple ways. Firstly, creating rosters is an inherently complex task from a combinatorial perspective. Secondly, as rosters reflect operations at a ward and affect the lives of staff, many different versions of real-life rostering problems exist. This means modelling, analyzing and implementation of decision-support are also very complicated tasks. This thesis attempts to answer relevant questions related to all these challenges.
In the first article a comprehensive framework for robust personnel planning is developed at the Department of Neonatal Intensive Care at St. Olav's Hospital. We firstly perform nurse rostering in a detailed way. When a roster is established, uncertainty is realized through simulation using extensive historical data obtained at the department. We perform daily simulations of the supply of staff and the demand for health care, and perform rerostering to take into account any disruptions from the uncertainty realization. This enables analyses of robust rostering strategies such as strategic overstaffing, implementing shadow shifts to cover absent nurses and trading extra weekend work for time off.
The second article deals with physician rostering at the Clinic of Surgery at St. Olav's Hospital. Here, surgeons must work emergency shifts in a cyclic structure, while also ensuring an even and robust staffing level at the sections. The mathematical structure of this problem is novel and difficult to solve, and we develop a matheuristic to produce robust rosters of high quality.
The third article presents a generic ward with 24 hour staff demand, where we minimize nurse fatigue. We incorporate a model of human sleep in the Nurse Rostering Problem, and define biological profiles to analyze how rosters should be individualized to minimize fatigue. The approach is theoretical, but insights and a large potential for future research and implementation exists.
The fourth article presented is a formalization of the experiences from performing pilot projects of implementing an optimization-based rostering tool. Creating the tool entailed development of a detailed model customized to Maternity Ward West at St. Olav's Hospital in Trondheim. We discuss visions for how the implementation of decision support systems for rostering will affect future work life.
Består av
Paper I: K. K. Klyve, I. N. Løyning, L. M. H. Melby, H. Andersson, A. N. Gullhav: A modelling framework for evaluating proactive and reactive nurse rostering strategies - A case study from a Neonatal Intensive Care UnitPaper 2: Klyve, Kjartan Kastet; Andersson, Henrik; Gullhav, Anders Nordby; Endreseth, Birger Henning. Semi-cyclic rostering of ranked surgeons — A real-life case with stability and flexibility measures. Operations Research for Health Care 2021 ;Volum 28. https://doi.org/10.1016/j.orhc.2021.100286
Paper 3: K. K. Klyve, I. Senthooran, M. Wallace: Nurse Rostering with Fatigue Modelling - Incorporating a Validated Sleep Model with Biological Variations in Nurse Rostering. The final published version is available in Health Care Manag Science (2022). https://doi.org/10.1007/s10729-022-09613-4 - This article is licensed under a Creative Commons Attribution 4.0 International License (CC BY 4.0)
Paper 4: K. K. Klyve and A. N. Gullhav: Future hospital rostering – experiences with new technology / Norsk arbeidsliv mot 2050. Muligheter og trusler.. Fagbokforlaget 2021