Design of a Personalized Caffeine and Sleep Management Application for Shift Workers

Shift workers frequently struggle to maintain optimal alertness and healthy sleep patterns due to irregular and demanding schedules. Although caffeine is widely used to counteract fatigue, most individuals rely on habitual consumption patterns that often lead to suboptimal alertness and disrupted sleep. Existing caffeine recommendation tools do not adequately account for the unique physiological and behavioral needs of shift workers. To address this gap, we developed WAKEMATE, a system that delivers personalized caffeine and sleep schedule recommendations based on users' physiology, sleep history, work schedules, and real-time alertness data. The system incorporates the PS-Alert model, a highly adaptable algorithm that uses machine learning to capture individual caffeine responses and manage alertness across varying work shifts. In parallel with the model, we designed a user interface focused on enhancing usability and engagement. To ensure practical relevance and a user-centered approach, we engaged two shift workers-one novice and one with over 20 years of experience-to evaluate the system and highlight the differing needs across the shift work population.