Conference Agenda

An efficient use of medical staff, i.e., efficient scheduling, is vital for hospitals due to increasing economic pressure. However, the shortage of skilled workers in hospitals is a serious problem that (partly) results from unattractive working hours. Thus, creating attractive working conditions is crucial to counteract this shortage. The current state of the literature addressing quantitative healthcare scheduling approaches largely neglects behavioral findings regarding worker behavior while behavioral studies leave open how the behavioral insights gained are to be implemented in workforce schedules. This paper aims to close this gap by investigating whether incorporating behavioral insights into quantitative scheduling can provide benefits to both medical staff and hospital management. Specifically, the trade-off between schedule consistency (preferred by workers) and worker flexibility (favored by planners) is quantified. A MILP model is developed that complements common hospital regulations with constraints designed to model performance degradation and recovery due to scheduling inconsistencies. The model captures how inconsistent shift patterns impair worker performance over time, while consistent schedules allow adaptation and performance recovery. Due to the computational complexity, we decompose the compact model via the Dantzig-Wolfe decomposition and solve it using a column generation approach. Our findings offer insights into how accommodating human behavioral factors within optimization models can enhance both employee satisfaction and operational performance in healthcare staff scheduling. Consequently, this work provides a foundation for rethinking workforce scheduling practices to create schedules that are not only feasible but also desirable from both the employee and management perspectives.

Offering a healthy work-life-balance to its employees is a big challenge for healthcare institutions. This is particular true, for employees that work in shifts, like nurses. Negative effects of shift work can be reduced by giving employees better opportunities to shape their working schedule. Currently, there is a lack of tools for automated schedulers that consider the interests of all stakeholders.

We present a Decision Support Systems (DSS) that allows nurses to enter their needs and preferences in an app and contains advanced models and algorithms to create excellent and fair rosters. Planners usually only need to make small adjustments (usually due to undocumented and/or soft information) before generated rosters are finalized and released. We developed an Integer Programming model that uses stints (predefined sequences of shifts and rest days) as decision variables. They are generated in a preprocessing step via simple construction heuristics and/or employee preferences. Stints constitute a useful interface between employee and mathematical model and allow easy formulation of complex working time regulations, demands per skill patterns, and employees needs and preferences.

We have tested our DSS for several months in real-life settings in various departements of hospitals and nursing homes in Switzerland and Liechtenstein. Using our DSS, the shift planners find good schedules in short time. The participative approach led to great acceptance among employees, planners, and healthcare institutions. We present interesting numerical results and valuable best practices based on our experiences so far.

Personnel, particularly in the healthcare sector, represents an exceptionally rare and valuable asset. Consequently, maintaining employee satisfaction is paramount. One strategy healthcare institutions are adopting involves incorporating employees' preferences into the planning process. Our study aims to ascertain whether individual employees could potentially skew the final plans in their favor by misrepresenting their preferences. Furthermore, we seek to identify the variables on which this possibility depends. To this end, we will conduct a focused, small-scale study specifically addressing this issue. During our presentation, we will both outline the problem setting and delve into the variables, examining how they influence the scenario. Our primary interest lies in exploring the strategies employed by the employee, the nature of the planning process, and the characteristics of the problem instances, including their size and type.