Collaborative duty rostering in health care professions

Abstract

For nurses the duty roster and its reliability has a significant impact on the compatibility of work and private life. In the research project GamOR (Game of Roster) ergonomists, designers and mathematicians cooperated with application partners to address this issue and developed a novel collaborative planning process for creating duty rosters.

The collaborative planning process consists of two parts. On the one hand, employees are informed about conflicts among their wishes for free time and are encouraged to solve these conflicts within the team. On the other hand, decision makers are supported in the creation of the final roster.

In this paper we present Constraint Programming (CP) approaches to support both of these parts. Based on a set of CP model components, which model for example staff requirements and legal regulations, we introduce a domain driven algorithm for detecting conflicts of wishes and argue that it outperforms approaches known from the literature. Moreover, we develop a backtracking search for generating complete rosters. This is done by appropriate variable and value selection strategies reflecting the objectives — balanced work time accounts, alternating free weekends, lengths of shift sequences (total and with the same shift definition) and forward rotation.

The presented approach was introduced for testing in various institutions and has been positively evaluated by both nurses and decision makers. Nurses particularly appreciate the transparency and timely feedback of conflicts. For decision makers, the time saved when creating the duty roster is a great benefit.