With the aim of designing an interface that supports troubleshooting of a dialysis machine, a medical-engineer (ME) cognitive task analysis was conducted in this study, with the error messages currently provided by a hemodialysis machine also being analyzed and evaluated. First, we developed the “error-message mechanism diagram” for the given problem, indicating the relationship between the error message and the notifying conditions of this message (corresponding to the candidate for the cause of the problem). Next, we developed the “cognitive task flow diagram,” which shows the cause candidates generated by the ME until the source of the problem was detected. This diagram also clarifies the manner in which the ME verifies the cause candidates and the information or knowledge employed by the ME. Then, for the given problem, we compared the cognitive task flow diagram of an ME who successfully detected the problem cause and corresponding error-message mechanism diagram to evaluate the efficacy of the error messages currently provided by the device.