In the previous study, we developed Preeclampsia Ontology (PEO) from research articles of international journals on Preeclampsia (PE). PEO provides wide range of terminology, concepts and their relationships about PE in hierarchical structure and expected to be used for PE-specific knowledge discovery. However, PEO was not able to be applied to secondary use of clinical data and literature in Japanese because PEO was not localized to Japanese. In this study, we localized terminology of PEO from English to Japanese to develop Japanese version of PEO. Localization was performed with a combination of dictionary-based using multiple dictionaries and expert-based localization. The developed Japanese version of PEO provides formal naming of preeclampsia knowledge domain in both Japanese and English. The terminology of the Japanese version of PEO consist mainly of molecules and phenotypes include gene names, chemical compounds, outcomes of pregnancy, complications, laboratory tests and diagnoses. The Japanese version of PEO is expected to apply various medical informatics tasks such as information retrieval and interoperation of clinical data in international cooperation.
Patient safety problems arising from overlooking radiogram interpretation reports (hereinafter “overlooking problem”) can in principle occur at any hospital. In fact, many hospitals throughout Japan have reported the causes of and countermeasures to the problems, but such valuable information, knowledge and experience are dispersed. Therefore, it would be highly important to organize the dispersed reports systematically for hospitals with or without established forms of countermeasure. In previous research, the author has proposed a method to obtain wisdom by systematically consolidating the information contained in the reports. Then, the proposed method was applied to 62 reports discussed about the overlooking problem to obtain “ (1) a causal events casebook” and (2) a general-purpose model for consideration of the hospital countermeasure framework”. Here, the latter consists of “basic elements of countermeasure” which obtained by coarse-graining each hospital countermeasures from the viewpoint of 6W1H, and aggregating similar elements.
The above product (2) can not only provide support for countermeasure design by hospitals, but can also facilitate derivation of a support system encompassing the countermeasures of all target hospitals. Therefore, using the above products (1) and (2), we proceed to derive the functional requirements for a support system (hereinafter, “comprehensive support system”) for prevention of overlooking problem, encompassing various types of hospital countermeasures from small-to large-scale.
The functional requirements were obtained using four procedures. (a) In order to reduce un-interpreted reports as much as possible, we extracted system-based “basic elements of countermeasure” from the product (2) and defined it as the “core function” of the comprehensive support system. (b) For hospitals that use just part of the core function and combine it with human-based countermeasures, we extracted from the product (2) the human-based “basic elements of countermeasure” and added them to the functional requirements. (c) We added functional requirements to deal with the final un-interpreted reports after implementing countermeasures using functions (a) and/or (b). (d) In the procedures (a) to (c), the functional requirements were configured so that various problems of the product (1) could be comprehensively improved. At that time, the functional requirements were adjusted depending on whether the relationship of the basic elements was “competition, coexistence, or cooperation”.
As a result of the above procedures, we derived the functional requirements of the comprehensive support system comprising 25 system-based segments that constitute the 6 core functions and 6 related human-based segments, plus 2 other segments, resulting in a total of 33 segments. These functional requirements not only comprehensively include the countermeasures implemented by the various hospitals, but can also suggest an answer to the question of “how should one design a system that makes it possible to flexibly modify the countermeasure pattern of a hospital in accordance with its own operational state”.
We have established a system that in a short period of time we can accurately input the medical information of patients with ulcerative colitis and Crohn’s disease. This is achieved using a template function of the electronic medical records. Indicator scores of these diseases, such as CDAI (Crohn’s Disease Activity Index), CAI (Clinical activity index) and Mayo score are very complex and time consuming to calculate. To solve this problem, we created specialized template forms characteristic of these diseases and set up scripts to automatically calculate scores based on particular inspection values. Those inspection values being the number of defecations per day and the patient’s subjective condition, abdominal pain and presence of complications for the last seven days. Score calculation is time consuming due to the wide variety these inputs. To solve this problem, we created a smartphone application allowing patients to input personal medical data on an outpatient basis. The smartphone application converts the personal medical data entered into a QR code. By scanning the QR code into the template, data input is completed immediately. This method of template input has greatly shortened input time, decreased input and omission errors caused by manual input.
[Introduction] To create an accessible medical institution website, it is necessary to mark up a content of webpage correctly so that important information can be understood along with the context of webpage. As heading information is particularly important for web accessibility and web usability, we investigate whether the placement of heading information based on properties of human short-term memory does not hinder browsing by the visually-impaired and sighted people or not. [Methods] The authors prepared two top pages of medical institution website with little and much (4.7 links per heading) heading information, and conducted a questionnaire survey asking visually-impaired and sighted people to answer which was better via e-mail. [Results] There was no tendency to choose the top page with little heading information for both the visually-impaired and sighted. [Discussion] Considering the way of visually-impaired browsing webpages, the authors think that the more heading information will improve web accessibility and web usability of medical institution webpage.