A human data entry and notification system to monitor the delivery of meals in a hospital setting

To combat the risk of hyperglycemia while in the hospital, a patient's insulin delivery should be correlated with blood glucose levels and meal consumption. At present, there is a particularly acute disconnect between insulin delivery and meal consumption. There are simply no formal processes for communicating the status of meal delivery, and thereby too much variation in the timing of meal delivery. To address this problem, a needs assessment was performed which included observations, surveys, and interviews with various stakeholders from nursing and dietary as well as information technology. As a result, a new process was developed, where 1) catering associates enter time estimates upon imminent departure from the kitchen, and 2) nurses are notified of the time estimates prior to meal arrival. Graphical user interfaces were designed to afford data entry in limited time, with little required training. A database is used to capture and convey information between the interfaces. As well, there is a means whereby catering associates, as well as managers, can review past performance and compliance. The system was preliminarily evaluated in terms of heuristic evaluation, color contrast, GOMS, and user scenarios (N=5). Then, initial field tests were performed with catering associates (N=2). The results indicate that human data entry is done in less than 30 seconds for basic scenarios, and 60 seconds for more complex scenarios; and that data entry is intuitive to learn. By only the fourth scenario, the duration of task completion was near theoretical estimates.