Yuko Tsumoto

Risk Mining in Medicine: Application of Data Mining to Medical Risk Management

Organizations in our modern society grow larger and more complex to provide advanced services due to the varieties of social demands. Such organizations are highly efficient for routine work processes but known to be not robust to unexpected situations. According to this observation, the importance of the organizational risk management has been noticed in recent years. On the other hand, a large amount of data on the work processes has been automatically stored since information technology was introduced to the organizations. Thus, it has been expected that reuse of collected data should contribute to risk management for large-scale organizations. This paper proposes risk mining, where data mining techniques were applied to detection and analysis of risks potentially existing in the organizations and to usage of risk information for better organizational management. We applied this technique to the following three medical domains: risk aversion of nurse incidents, infection control and hospital management. The results show that data mining methods were effective to detection of risk factors.

Exploratory Univariate Analysis on the Characterization of a University Hospital: A Preliminary Step to Data-Mining-Based Hospital Management Using an Exploratory Univariate Analysis of a University Hospital

Abstract.Rapid progress in information technology allows us to store all the information of a hospital in a hospital information system, including management data, patient records, discharge summaries and laboratory data. Although those data have not been reused to date, the stored data can contribute to an analysis of hospital management. In this paper, the discharge summaries of Chiba University Hospital, which has been stored since 1980’s, were analyzed to characterize the university hospital. This paper focuses on using exploratory univariate analysis to show that the major reason for admissions in this university hospital is neoplasm and the distribution on length of stay and NHI (National Health Insurance, Japan) points to total cases and neoplasm followed logarithmic normal distributions, which are supported by descriptive statistics.

Similarity-based behavior and process mining of medical practices

This paper presents data mining results in which the temporal behavior of global hospital activities is visualized. The results show that the reuse of stored data will provide a powerful tool for hospital management and lead to improvement of hospital services.

Clustering-based analysis in hospital information systems

Rapid progress in electronization of hospital information gives an opportunity to realize evidence-based hospital management and services. This paper proposes a clustering-based data mining approach to temporal data in hospital information. The process consists of repetitions of clustering for grouping records and decision tree inducion for feature selection. It will terminate if the clustering results are converged. We evaluated this method to data on order history stored in hospital information system. The results show that the reuse of stored data will give a powerful tool for hospital management and lead to improvement of hospital services.

Correlation and Regression Analysis for Characterizations of a University Hospital

One of the most important issues in hospital management is how to determine the indicators for revenue of a large-scale hospital. Although it has been pointed out that the averaged value of length of stay in a hospital is an indicator of its revenue, quantitative evaluation has not been conducted. In this paper, we propose an analytic process based on methods for correlation and regression analysis to evaluate this indicator by using stored data in hospital information systems. We applied this method to data combined with discharge summaries and medical treatment fees, collected from 1997 to 2000 in Chiba University Hospital. The result shows that the length of stay can explain around 60-90% of the revenue during this period.

Information reuse in hospital information systems: A data mining approach

This paper presents application of data mining to data stored in a hospital information system in which temporal behavior of global hospital activities are visualized. The results show that the reuse of stored data will give a powerful tool for hospital management and lead to improvement of hospital services.

Towards Data-Oriented Hospital Services: Data Mining-Based Hospital Management

It has passed about twenty years since clinical information are stored electronically as a hospital information system since 1980s. Stored data include from accounting information to laboratory data and even patient records are now started to be accumulated: in other words, a hospital cannot function without the information system, where almost all the pieces of medical information are stored as multimedia databases. In this paper, we applied temporal data mining and exploratory data analysis techniques to hospital management data. The results show several interesting results, which suggests that the reuse of stored data will give a powerful tool for hospital management.

Characterizing Hospital Services Using Temporal Data Mining

Computerization of hospital information enables us to visualize and analyze temporal characteristics of hospital services, which can be viewed as a first step to improve and innovate clinical services. This paper proposes a temporal data mining process which consists of decision tree, clustering, MDS and three-dimensional trajectories mining and applied the method to datasets extracted from hospital information systems. The results show that the reuse of stored data will give a powerful tool to characterize medical services in the following ways: (1) Statistics and temporal characteristics of clinincal orders were visualized. (2) Divisions were classified in terms of temporal patterns of orders. (3) The temporal interval important to characterize the behavior of the divisions were evaluated. (4) Characterization of nursing orders showed the classification of nursing orders into disease-specific ones and patient- specific ones.

Temporal data mining in history data of hospital information systems

Rapid progress in electronization of hospital information gives an opportunity to realize evidence-based hospital management and services. This paper presents a preliminary approach to service innovation in a hospital by using data mining in which temporal behavior of global hospital activities are visualized. The results show that the reuse of stored data will give a powerful tool for hospital management and lead to improvement of hospital services.

Exploratory temporal data mining process in hospital information systems

This paper proposes an exploratory temporal data mining process which aims at capturing behavior of medical staff. The process consists of the following four process. First, datasets will be extracted from hospital information systems through double-step datawarehousing. Second, similarities between temporal sequences are calculated from datasets. Third, data mining methods such as clustering, multidimensional scaling are applied for obtaining the class labels. Finally, other data mining methods, such as decision tree and correspondence analysis are applied to original data sets with the class labels.