Human-Machine task allocation: Impact on patient safety in hospital settings

The paper aims at evaluating the impact of the tasks' allocation between human and machine on the efficiency of work systems in the healthcare domain. Usability testing was performed to compare two different tasks' allocations depending on two different systems (semi-automated system vs. manual system) for the tasks of preparing and administering the drugs in hospital settings. The results show that the allocation of tasks with the semi-automated system seems to produce a better performance in terms of task durations compared to the manual cabinet. Concerning safety goals, the semi-automated system seems to fix known common errors generated by the manual system. But it seems also to bring new errors more numerous that can be easily fixed. The main problem highlighted by the study lies in the risks of automation exposure over time. After a one-hour simulation, participants already have the feeling of loss of control and present a disengagement. The tasks' allocation with the semi-automated system changes in fact the nature of the feedback provided to nurses and then under-specifies the cognitive operations of nurses. Design recommendations were given in order to give back part of the control to nurses through the reallocation of tasks. (C) 2014 Societe francaise de psychologie. Published by Elsevier Masson SAS. All rights reserved.