In general, medical records such as surgery consent forms include body part diagrams where doctors illustrate the positions, sizes, and conditions of disease. The information on which kinds of body part diagrams are in medical records is important and useful for the management and knowledge discovery of medical records. This study aims to propose and develop a method that automatically recognizes the kinds of body part diagrams. The past study applied a method consisting of the modified weighted direction index histogram and the modified quadratic discriminant function, and resulted in a recognition rate of 98.52 ［％］. However, there is room for performance improvement, and the processes and settings of the conventional method were complex and difficult to configure. The present study aims at better performance and simpler and easier processes and settings. We propose a method consisting of histograms of oriented gradients and a Mahalanobis-distance-based classifier, and conduct evaluation experiments for the recognition of 10 and 30 kinds of body part diagrams. The experimental results show that the proposed method achieves 100.00 ［％］ generalized recognition rates under almost all conditions, and is superior to the conventional method.

　Recently, frequent problems with patient safety have occurred in Japan due to overlooking radiogram interpretation reports (hereinafter, overlooking problem), and many hospitals have reported countermeasures. If the countermeasure information could be organized systematically, meaningful findings may be obtained and patient safety can be maintained. However, it is difficult to extract and organize written information from the literatures composed and described in a different manner. Nevertheless, the difficulty in organizing the literature is not specific to the overlooking problem and is common to most healthcare information systems. Therefore, in this study, we generalized the overlooking problem to “problems caused by user behavior” in medical practices with “hospital independent work flow” and assumed that there were many cases in the literature discussing these problems. Then, we proposed a general method for deriving (a) case causes, (b) case countermeasures, and (c) a general-purpose model to examine the countermeasure frameworks from the above cases.

　The usefulness of the proposed method was discussed by applying it to overlooked problems and examining whether meaningful findings could be obtained. Specifically, 62 literature cases from 48 hospitals extracted using Ichushi-Web and three academic journals containing many publications on overlooked problems were analyzed using the proposed method. As a result, the expected outcomes (a) to (c) could be obtained despite 74％ of the cases being abstracts described in ≤ 1,000 characters. The two key terms, “hospital independent work flow” and “problems caused by user behavior”, were important in that the proposed method worked effectively in this situation. Furthermore, our results suggested that the proposed method could be flexibly applied to similar problems.

　Recently, many services and applications have Web API for making use of their data from external systems. Web API allows external systems to access data with general HTTP requests and to receive them in machine readable format like JSON. Since this method is useful for retrieving data from EMR, HL7 FHIR, openEHR and other standards adopt Web API as their data exchange interfaces. However, in our hospital, we cannot use these standards or Web APIs on the EMR system because it doesn’t implement these standards and APIs. Moreover, the specifications of these standards are too large to implement by ourselves.

　Therefore, we implemented our own simple Web API to retrieve necessary data directly from the EMR’s database. As a result of implementing the Web API, it became easy to retrieve several EMR’s data from its database and we have been able to use them in some applications efficiently. In conclusion, Web API is useful as a data exchange interface and we would like to demand that future EMR systems implement standards like HL7 FHIR or OpenEHR, and the function that allows users to define which data can be retrieved with Web API.