Kazuhiko Ohe

The University of Tokyo, 白山, Tōkyō, Japan

Are you Kazuhiko Ohe?

Claim your profile

Publications (120)140.78 Total impact

  • Source
    [Show abstract] [Hide abstract]
    ABSTRACT: Linked Data is a promising technology for knowledge integration on the web. Many research groups have developed ontologies and terminologies, and recently, they have published a wide variety of Linked Data in the biomedical domain. We have systematized an ontology of abnormal states in the definition of diseases. For effective use of existing biomedical data, one of the difficulties is a conceptual discrepancy rather than a superficial one since data are heterogeneous. This article focuses on knowledge integration with Linked Data in terms of abnormal states. First, we discuss ontological issues of reusing and integrating knowledge of abnormal states in existing biomedical resources. Next, we introduce our ontology of abnormal states. By using our ontology and making explicit the meaning of each concept, we show a solution for the integration. Then, applying a Linked Data technology, we introduce a prototype system to link our ontology as a hub of existing resources across species. In cooperation with disease ontology, we demonstrate finding commonality of causal relationships of abnormal states between diseases across clinical departments. Our approach will bring benefits to fill the gap between basic research and clinical medicine, and contribute to disease knowledge integration of good practice.
    Full-text · Article · Jan 2016 · Transactions of the Japanese Society for Artificial Intelligence
  • [Show abstract] [Hide abstract]
    ABSTRACT: Background: About 10% of patients with diabetes discontinue treatment, resulting in the progression of diabetes-related complications and reduced quality of life. Objective: The objective was to predict a missed clinical appointment (MA), which can lead to discontinued treatment for diabetes patients. Methods: A machine-learning algorithm was used to build a logistic regression model for MA predictions, with L2-norm regularization used to avoid over-fitting and 10-fold cross validation used to evaluate prediction performance. Data associated with patient MAs were extracted from electronic medical records and classified into two groups: one related to patients' clinical condition (X1) and the other related to previous findings (X2). The records used were those of the University of Tokyo Hospital, and they included the history of 16 026 clinical appointments scheduled by 879 patients whose initial clinical visit had been made after January 1, 2004, who had diagnostic codes indicating diabetes, and whose HbA1c had been tested within 3 months after their initial visit. Records between April 1, 2011, and June 30, 2014, were inspected for a history of MAs. Results: The best predictor of MAs proved to be X1 + X2 (AUC = 0.958); precision and recall rates were, respectively, 0.757 and 0.659. Among all the appointment data, the day of the week when an appointment was made was most strongly associated with MA predictions (weight = 2.22). Conclusions: Our findings may provide information to help clinicians make timely interventions to avoid MAs.
    No preview · Article · Nov 2015 · Journal of diabetes science and technology
  • [Show abstract] [Hide abstract]
    ABSTRACT: Background The accuracy of estimating nutritional intake and balance from photos of meals has not been well documented. However, DialBetics (DB)—our diabetes self-management support system, which is based on information and communication technologies—relies on the photos that type 2 diabetes patients take of their meals with smartphones. Therefore, we designed a study to evaluate this accuracy. Methods We prepared 61 dishes whose actual amount/value of total energy and each nutrient were known: protein, fat, carbohydrates, dietary fiber and salt. Their balance—the protein-fat-carbohydrate ratio—was also known, constituting the weighed food record (WFR). Smartphone photos of those dishes were taken, and three registered dietitians evaluated each dish from those photos, naming the dish and estimating the amount of each nutrient in it, plus the dish’s balance. These estimated DB and WFR values were compared using the Wilcoxon matched-pairs rank-sum test; intraclass correlation coefficients (ICCs) were calculated. Agreement between the two values for each dish was assessed by Bland-Altman analysis. Results There were significant ICCs—0.84 for fat (95 % confidence interval 0.75–0.90) and 0.93 for carbohydrates (0.88, 0.96)—but no statistically significant differences between DB and WRF for other nutrients or balance. Bland-Altman analysis showed that differences between the two values were random and not biased against nutrient intake; 95 % limits of agreement were acceptable although wide (energy −198 to 210 kcal/dish; carbohydrates −22.7 to 25.8 g/dish). Conclusion DB’s diet evaluation by photos is reliable with apparent potential for assessing diets.
    No preview · Article · Oct 2015 · Diabetology International
  • [Show abstract] [Hide abstract]
    ABSTRACT: Diabetes self-management education is an essential element of diabetes care. Systems based on information and communication technology (ICT) for supporting lifestyle modification and self-management of diabetes are promising tools for helping patients better cope with diabetes. An earlier study had determined that diet improved and HbA1c declined for the patients who had used DialBetics during a 3-month randomized clinical trial. The objective of the current study was to test a more patient-friendly version of DialBetics, whose development was based on the original participants' feedback about the previous version of DialBetics. DialBetics comprises 4 modules: data transmission, evaluation, exercise input, and food recording and dietary evaluation. Food recording uses a multimedia food record, FoodLog. A 1-week pilot study was designed to determine if usability and compliance improved over the previous version, especially with the new meal-input function. In the earlier 3-month, diet-evaluation study, HbA1c had declined a significant 0.4% among those who used DialBetics compared with the control group. In the current 1-week study, input of meal photos was higher than with the previous version (84.8 ± 13.2% vs 77.1% ± 35.1% in the first 2 weeks of the 3-month trial). Interviews after the 1-week study showed that 4 of the 5 participants thought the meal-input function improved; the fifth found input easier, but did not consider the result an improvement. DialBetics with FoodLog was shown to be an effective and convenient tool, its new meal-photo input function helping provide patients with real-time support for diet modification. © 2015 Diabetes Technology Society.
    No preview · Article · Apr 2015 · Journal of diabetes science and technology
  • Source
    [Show abstract] [Hide abstract]
    ABSTRACT: Early reperfusion by percutaneous coronary intervention (PCI) is the current standard therapy for ST-elevation myocardial infarction (STEMI). To achieve better prognoses for these patients, reducing the door-to-balloon time is essential. As we reported previously, the Kitasato University Hospital Doctor Car (DC), an ambulance with a physician on board, is equipped with a novel mobile cloud 12-lead ECG system. Between September 2011 and August 2013, there were 260 emergency dispatches of our Doctor Car, of which 55 were for suspected acute myocardial infarction with chest pain and cold sweat. Among these 55 calls, 32 patients received emergent PCI due to STEMI (DC Group). We compared their data with those of 76 STEMI patients who were transported directly to our hospital by ambulance around the same period (Non-DC Group). There were no differences in patient age, gender, underlying diseases, or Killip classification between the two groups. The door-to-balloon time in the DC group was 56.1 ± 13.7 minutes and 74.0 ± 14.1 minutes in the Non-DC Group (P < 0.0001). Maximum levels of CPK were 2899 ± 308 and 2876 ± 269 IU/L (P = 0.703), and those of CK-MB were 292 ± 360 and 295 ± 284 ng/mL (P = 0.423), respectively, in the 2 groups. The Doctor Car system with the Mobile Cloud ECG was useful for reducing the door-to-balloon time.
    Full-text · Article · Feb 2015 · International Heart Journal
  • Source

    Full-text · Dataset · Jul 2014
  • Source
    [Show abstract] [Hide abstract]
    ABSTRACT: Background Recently, exchanging data and information has become a significant challenge in medicine. Such data include abnormal states. Establishing a unified representation framework of abnormal states can be a difficult task because of the diverse and heterogeneous nature of these states. Furthermore, in the definition of diseases found in several textbooks or dictionaries, abnormal states are not directly associated with the corresponding quantitative values of clinical test data, making the processing of such data by computers difficult. Results We focused on abnormal states in the definition of diseases and proposed a unified form to describe an abnormal state as a “property,” which can be decomposed into an “attribute” and a “value” in a qualitative representation. We have developed a three-layer ontological model of abnormal states from the generic to disease-specific level. By developing an is-a hierarchy and combining causal chains of diseases, 21,000 abnormal states from 6000 diseases have been captured as generic causal relations and commonalities have been found among diseases across 13 medical departments. Conclusions Our results showed that our representation framework promotes interoperability and flexibility of the quantitative raw data, qualitative information, and generic/conceptual knowledge of abnormal states. In addition, the results showed that our ontological model have found commonalities in abnormal states among diseases across 13 medical departments.
    Full-text · Article · May 2014 · Journal of Biomedical Semantics
  • Source
    Article: DialBetics
    [Show abstract] [Hide abstract]
    ABSTRACT: Numerous diabetes-management systems and programs for improving glycemic control to meet guideline targets have been proposed, using IT technology. But all of them allow only limited-or no-real-time interaction between patients and the system in terms of system response to patient input; few studies have effectively assessed the systems' usability and feasibility to determine how well patients understand and can adopt the technology involved. DialBetics is composed of 4 modules: (1) data transmission module, (2) evaluation module, (3) communication module, and (4) dietary evaluation module. A 3-month randomized study was designed to assess the safety and usability of a remote health-data monitoring system, and especially its impact on modifying patient lifestyles to improve diabetes self-management and, thus, clinical outcomes. Fifty-four type 2 diabetes patients were randomly divided into 2 groups, 27 in the DialBetics group and 27 in the non-DialBetics control group. HbA1c and fasting blood sugar (FBS) values declined significantly in the DialBetics group: HbA1c decreased an average of 0.4% (from 7.1 ± 1.0% to 6.7 ± 0.7%) compared with an average increase of 0.1% in the non-DialBetics group (from 7.0 ± 0.9% to 7.1 ± 1.1%) (P = .015); The DialBetics group FBS decreased an average of 5.5 mg/dl compared with a non-DialBetics group average increase of 16.9 mg/dl (P = .019). BMI improvement-although not statistically significant because of the small sample size-was greater in the DialBetics group. DialBetics was shown to be a feasible and an effective tool for improving HbA1c by providing patients with real-time support based on their measurements and inputs.
    Full-text · Article · Mar 2014 · Journal of diabetes science and technology
  • [Show abstract] [Hide abstract]
    ABSTRACT: Publishing open data as linked data is a significant trend in not only the Semantic Web community but also other domains such as life science, government, media, geographic research and publication. One feature of linked data is the instance-centric approach, which assumes that considerable linked instances can result in valuable knowledge. In the context of linked data, ontologies offer a common vocabulary and schema for RDF graphs. However, from an ontological engineering viewpoint, some ontologies offer systematized knowledge, developed under close cooperation between domain experts and ontology engineers. Such ontologies could be a valuable knowledge base for advanced information systems. Although ontologies in RDF formats using OWL or RDF(S) can be published as linked data, it is not always convenient to use other applications because of the complicated graph structures. Consequently, this paper discusses RDF data models for publishing ontologies as linked data. As a case study, we focus on a disease ontology in which diseases are defined as causal chains.
    No preview · Article · Jan 2014 · Transactions of the Japanese Society for Artificial Intelligence

  • No preview · Article · Jan 2014 · Studies in health technology and informatics
  • [Show abstract] [Hide abstract]
    ABSTRACT: Publishing open data as linked data is a significant trend in not only the Semantic Web community but also other domains such as life science, government, media, geographic research and publication. One feature of linked data is the instance-centric approach, which assumes that considerable linked instances can result in valuable knowledge. In the context of linked data, ontologies offer a common vocabulary and schema for RDF graphs. However, from an ontological engineering viewpoint, some ontologies offer systematized knowledge, developed under close cooperation between domain experts and ontology engineers. Such ontologies could be a valuable knowledge base for advanced information systems. Although ontologies in RDF formats using OWL or RDF(S) can be published as linked data, it is not always convenient to use other applications because of the complicated graph structures. Consequently, this paper discusses RDF data models for publishing ontologies as linked data. As a case study, we focus on a disease ontology in which diseases are defined as causal chains.
    No preview · Chapter · Jan 2014
  • Source
    [Show abstract] [Hide abstract]
    ABSTRACT: Objective: Complex regional pain syndrome (CRPS) describes a broad spectrum of symptoms that predominantly localize to the extremities. Although limb fracture is one of the most frequently reported triggering events, few large-scale studies have shown the occurrence of and factors associated with CRPS following limb fracture. This study aimed to show the occurrence and identify of those factors. Methods: Using the Japanese Diagnosis Procedure Combination database, we identified 39 patients diagnosed with CRPS immediately after open reduction and internal fixation (ORIF) for limb fracture from a cohort of 185 378 inpatients treated with ORIF between 1 July and 31 December of each year between 2007 and 2010. Patient and clinical characteristics such as age, gender, fracture site, duration of anaesthesia and use of regional anaesthesia were investigated by logistic regression analyses to examine associations between these factors and the in-hospital occurrence of CRPS after ORIF. Results: The occurrence of CRPS was relatively high in fractures of the distal forearm, but low in fractures of the lower limb and in patients with multiple fractures. Generally females are considered to be at high risk of CRPS; however, we found a comparable number of male and female patients suffering from CRPS after ORIF for limb fracture. In terms of perioperative factors, a longer duration of anaesthesia, but not regional anaesthesia, was significantly associated with a higher incidence of CRPS. Conclusion: Although a limited number of CRPS patients were analysed in this study, reduced operative time might help to prevent the development of acute CRPS following limb fracture.
    Full-text · Article · Dec 2013 · Rheumatology (Oxford, England)
  • [Show abstract] [Hide abstract]
    ABSTRACT: We are aware of only one report describing the relationship between operative volume and outcomes in musculoskeletal tumor surgery, although numerous studies have described such relationships in other surgical procedures. The aim of the present study was to use a nationally representative inpatient database to evaluate the impact of hospital volume on the rates of postoperative complications and in-hospital mortality after musculoskeletal tumor surgery. We used the Japanese Diagnostic Procedure Combination administrative database to retrospectively identify 4803 patients who had undergone musculoskeletal tumor surgery during 2007 to 2010. Patients were then divided into tertiles of approximately equal size on the basis of the annual hospital volume (number of patients undergoing musculoskeletal tumor surgery): low, twelve or fewer cases/year; medium, thirteen to thirty-one cases/year; and high, thirty-two or more cases/year. Logistic regression analyses were performed to examine the relationships between various factors and the rates of postoperative complications and in-hospital mortality adjusted for all patient demographic characteristics. The overall postoperative complication rate was 7.2% (348 of 4803), and the in-hospital mortality rate was 2.4% (116 of 4803). Postoperative complications included surgical site infections in 132 patients (2.7%), cardiac events in sixty-four (1.3%), respiratory complications in fifty-one (1.1%), sepsis in thirty-one (0.6%), pulmonary emboli in sixteen (0.3%), acute renal failure in eleven (0.2%), and cerebrovascular events in seven (0.1%). The postoperative complication rate was related to the duration of anesthesia (odds ratio [OR] for a duration of more than 240 compared with less than 120 minutes, 2.44; 95% confidence interval [CI], 1.68 to 3.53; p < 0.001) and to hospital volume (OR for high compared with low volume, 0.73; 95% CI, 0.55 to 0.96; p = 0.027). The mortality rate was related to the diagnosis (OR for a metastatic compared with a primary bone tumor, 3.67; 95% CI, 1.66 to 8.09; p = 0.001), type of surgery (OR for amputation compared with soft-tissue tumor resection without prosthetic reconstruction, 3.81; 95% CI, 1.42 to 10.20; p = 0.008), and hospital volume (OR for high compared with low volume, 0.26; 95% CI, 0.14 to 0.50; p < 0.001). We identified an independent effect of hospital volume on outcomes after adjusting for patient demographic characteristics. We recommend regionalization of musculoskeletal tumor surgery to high-volume hospitals in an attempt to improve patient outcomes. Prognostic Level II. See Instructions for Authors for a complete description of levels of evidence.
    No preview · Article · Sep 2013 · The Journal of Bone and Joint Surgery
  • Yuichi Yoshida · Takeshi Imai · Kazuhiko Ohe
    [Show abstract] [Hide abstract]
    ABSTRACT: We evaluate the status of health information system (HIS) adoption (In this paper, "HIS" means electronic medical record system (EMR) and computerized provider order entry system (CPOE)). We also evaluate the affect of the policies of Japanese government. The status of HIS adoption in Japan from 2002 to 2011 was investigated using reports from complete surveys of all medical institutions conducted by the Ministry of Health, Labour and Welfare (MHLW). HIS-related budgets invested by the Japanese government from 2000 to 2008 were surveyed mainly using literatures and administrative documents of the Japanese government (MHLW and Ministry of Economy, Trade and Industry). The rates of HIS adoption in Japan in 2011 were: 20.9% for the rate of EMR adoption in clinics, 20.1% for the rate of EMR adoption and 36.6% for the rate of CPOE adoption in hospitals. In hospitals, the rate of EMR and CPOE adoption were 51.5% and 78.6% in 822 large hospitals (400 or more beds), 27.3% and 52.1% in 1832 medium hospitals (200-399 beds), and 13.5% and 26.0% in 5951 small hospitals (less than 200 beds), respectively. Japan has a large number of medical institutions (99,547 clinics and 8605 hospitals) with a low rate of EMR adoption in clinics and a high rate of HIS adoption in hospitals. The national budget to expand HIS use was implemented for medium and large hospitals mainly. The policy target of New IT Reform Strategy was not achieved. The rate of HIS adoption in Japanese medium and large hospitals is high compared to small hospitals and clinics, and this is attributable to the fact that the Japanese government placed the target for HIS adoption on key hospitals with a large number of beds and concentrated budget investment in those hospitals. Besides, legal approval of EMR and the introduction of Diagnostic Procedure Combination system facilitated EMR adoption. There is less financial support for small hospitals than medium and large hospitals. The low rate of EMR adoption in clinics stems from the facts that there was little subsidies or incentives in the national remuneration for medical services, lack of cooperation from medical associations, and a failed attempt to mandate computerization of medical accounting (medical billing). Giving financial incentives is an effective means of raising EMR adoption rate. For wide usage of HIS, more financial support and incentive may be necessary for small hospitals and clinics.
    No preview · Article · Aug 2013 · International Journal of Medical Informatics
  • [Show abstract] [Hide abstract]
    ABSTRACT: To improve emergency services for accurate diagnosis of cardiac emergency, we developed a low-cost new mobile electrocardiography system "Cloud Cardiology®" based upon cloud computing for prehospital diagnosis. This comprises a compact 12-lead ECG unit equipped with Bluetooth and Android Smartphone with an application for transmission. Cloud server enables us to share ECG simultaneously inside and outside the hospital. We evaluated the clinical effectiveness by conducting a clinical trial with historical comparison to evaluate this system in a rapid response car in the real emergency service settings. We found that this system has an ability to shorten the onset to balloon time of patients with acute myocardial infarction, resulting in better clinical outcome. Here we propose that cloud-computing based simultaneous data sharing could be powerful solution for emergency service for cardiology, along with its significant clinical outcome.
    No preview · Article · Aug 2013 · Studies in health technology and informatics
  • [Show abstract] [Hide abstract]
    ABSTRACT: Physiological knowledge is often described in terms of mathematical models in the domain of bioinformatics, and some ontologies have been developed to integrate these models. However, such models do not explicitly describe clinicians' qualitative knowledge, which is required for clinical applications including decision support and counseling of patients to help them understand their clinical situation. This paper proposes a description framework for a qualitative and context-independent ontology of physiology, QliP, which has three features: 1) It models physiological knowledge qualitatively without mathematical knowledge; 2) The described knowledge is independent of surrounding anatomical entities and abnormality; and 3) It targets physiological components in varying degrees of granularity, from cells to organ systems. An ontology based on this proposed model enables automatic generation of a physiological state transition, starting and ending with a given state.
    No preview · Article · Aug 2013 · Studies in health technology and informatics
  • [Show abstract] [Hide abstract]
    ABSTRACT: Medication alert systems have been implemented worldwide. The purpose of this study is evaluation of the medication alert systems from a clinical perspective. We surveyed physicians with regard to their reactions to the medication alerts. We collected the revised prescription information and assessed risk avoidance in all cases. The system reviewed 51,006 prescriptions and produced 16,718 physician alerts related to 13,823 prescriptions over the course of 1 month. We identified 45 prescriptions that were revised following the alert and four cases in which patient treatment may have been discontinued or adverse drug events (ADEs) may have occurred if the alerts had not been issued. We demonstrated that the system prevented these potential medication errors. This study adopted a clinical perspective and demonstrated that a real-time alert system can contribute to prevention of ADEs.
    No preview · Article · Aug 2013 · Studies in health technology and informatics
  • [Show abstract] [Hide abstract]
    ABSTRACT: This study presents a prediction-based approach to determine thresholds for a medication alert in a computerized physician order entry. Traditional static thresholds can sometimes lead to physician's alert fatigue or overlook potentially excessive medication even if the doses are belowthe configured threshold. To address this problem, we applied a random forest algorithm to develop a prediction model for medication doses, and applied a boxplot to determine the thresholds based on the prediction results. An evaluation of the eight drugs most frequently causing alerts in our hospital showed that the performances of the prediction were high, except for two drugs. It was also found that using the thresholds based on the predictions would reduce the alerts to a half of those when using the static thresholds. Notably, some cases were detected only by the prediction thresholds. The significance of the thresholds should be discussed in terms of the trade-offs between gains and losses; however, our approach, which relies on physicians' collective experiences, has practical advantages.
    No preview · Article · Aug 2013 · Studies in health technology and informatics
  • Takeshi Imai · Masayo Hayakawa · Kazuhiko Ohe
    [Show abstract] [Hide abstract]
    ABSTRACT: Prediction of synergistic or antagonistic effects of drug-drug interaction (DDI) in vivo has been of considerable interest over the years. Formal representation of pharmacological knowledge such as ontology is indispensable for machine reasoning of possible DDIs. However, current pharmacology knowledge bases are not sufficient to provide formal representation of DDI information. With this background, this paper presents: (1) a description framework of pharmacodynamics ontology; and (2) a methodology to utilize pharmacodynamics ontology to detect different types of possible DDI pairs with supporting information such as underlying pharmacodynamics mechanisms. We also evaluated our methodology in the field of drugs related to noradrenaline signal transduction process and 11 different types of possible DDI pairs were detected. The main features of our methodology are the explanation capability of the reason for possible DDIs and the distinguishability of different types of DDIs. These features will not only be useful for providing supporting information to prescribers, but also for large-scale monitoring of drug safety.
    No preview · Article · Aug 2013 · Studies in health technology and informatics
  • [Show abstract] [Hide abstract]
    ABSTRACT: The openEHR has adopted the dual model architecture consisting of Reference Model and Archetype. The specification, however, lacks formal definitions of archetype semantics, so that its behaviors have remained ambiguous. The objective of this poster is to analyze semantics of the openEHR archetypes: its variance and mutability. We use a typed lambda calculus as an analyzing tool. As a result, we have reached the conclusion that archetypes should be 1) covariant and 2) immutable schema.
    No preview · Article · Aug 2013 · Studies in health technology and informatics

Publication Stats

821 Citations
140.78 Total Impact Points

Institutions

  • 2004-2015
    • The University of Tokyo
      • • Faculty & Graduate School of Medicine
      • • Department of Health Science and Nursing
      白山, Tōkyō, Japan
  • 2014
    • Japan Advanced Institute of Science and Technology
      • School of Knowledge Science
      KMQ, Ishikawa, Japan
  • 1996-2014
    • Tokyo Medical University
      Edo, Tōkyō, Japan