Biomathematical models hold promise as tools that can be used to manage fatigue risks in an operational setting. There are numerous models available to predict fatigue-related performance impairment arising from sleep loss, circadian misalignment, and sleep inertia. It is imperative that any model considered for use during space flight missions be validated in an operational setting in order to ensure that predictions are reliable and consistent. Given the complexity of the underlying causes of fatigue, there are wide individual differences in response to mission stressors. However, the majority of candidate models provide predictions based on average human responses, making it difficult to use the models to make operational decisions for individuals. In addition, there are no published reports on the acceptability, usability, and feasibility of any of the current models. Effective incorporation in a space flight environment requires model-based software systems that are easy to use by a wide-range of operators and are integrated with other scheduling constraints that are relevant in space flight operations.
Investigators aimed to identify which models are the best candidates for inclusion in four Human Exploration Research Analog (HERA) missions. Crewmembers were trained on how to use the modeling software for the selected models and evaluated the usability of the current model software. Investigators asked crewmembers to collect measures of cognitive impairment and simulated operational data throughout the mission using actigraphy, sleep logs, the Psychomotor Vigilance Task (PVT), User Experience Questionnaire (UEQ), Samn-Perelli fatigue Scale (SPS), and the Robotic On-Board Trainer (ROBoT). Investigators used this objective data to validate the predictions generated by the selected models. In an effort to identify whether the utility of the models might be enhanced by integration with crew scheduling tools, they evaluated the selected models after integrating the models into Playbook, the scheduling tool that is currently being evaluated on the International Space Station (ISS) in preparation for deep space missions.
Investigators studied crews of four individuals during 4.5 missions in the HERA (n = 20). Crewmembers were allowed five hours of sleep during the week and eight hours of sleep on weekends. Crewmembers completed a reaction time test and fatigue ratings five times a day approximately every three days over 45 days. The investigators compared objective performance on the reaction time test to the predictions of biomathematical models designed to predict fatigue. They found that performance deteriorated over the course of the mission on average, however, there were wide individual differences in performance trajectories between crewmembers. Self-reported fatigue did not change over the mission. Biomathematical models were able to predict average performance, but not performance at the individual level.
This evaluation provided operational personnel, including flight surgeons, mission directors, mission operations planners, and flight controllers, with an understanding of 1) the validity of the model predictions in an operational setting, including how model predictions may vary by individual characteristics; 2) the acceptability, usability, and feasibility of the incorporation of the models as is into an operational setting, and; 3) a preliminary, qualitative assessment of the utility of integrating the model predictions into existing scheduling tools.