ISQUA18-2630Establishment of an Intelligence Decision Support System for Inpatient Quality and Efficiency to Enhance the Data-driven Management in a Medical Centre in Taiwan

Abstract

Objectives: Leaders in healthcare organizations are keen to run their service more efficient and responsive to the needs from client and resource of the facilities. We describe our experience in integrating inpatients capacity management into information technological application in a public university-affiliated medical centre in Taiwan, Asia. The hospital system has 5 branches located in different parts in Taiwan. The headquarter is located in northern parts with 7,929 employees and a medical staff of more than 1,400 physicians, has 2,600 beds and serves over 9,000 outpatients, 290 inpatients and 300 emergency patients daily. The 5 branches have various scales ranging from 31 beds, 50 beds, 350 beds, 812 beds, to 941 beds. Without a centralized dataset, leaders have limited information to makes comparison and evaluate the capacity. Method(s): Since October 2015, the leaders of headquarter proposed an action plan for intelligence decision support system of inpatient service and capacity including the following strategies. First, an organized task force group combined with the department for information technology and interdisciplinary resources developed the business intelligent (BI) system. Second, we established the centralized database with the same definition for the headquarter and 5 branches. Third, we established the inpatient dashboard with automatic data output for timely access. Fourth, we organized training courses and E-learning programmes to improve the understanding the software self-supported business intelligence tool. Fifth, in order to ensure the quality of data, we founded standardized process for indicator validation to correct the BI system. Finally, we involved the participation of the leaders from medical and administrative departments in our healthcare system. Result(s): We completed 4 operational indicators and 2 quality indicators, include the number of inpatients, the days of stay, the average length of stay, the bed occupancy rate, the mortality rate, and the net mortality rate. We used the statistical process control chart and stratification in the dashboard, the data can be drilled down from hospital-wide to the specialists, even to different wards and physicians. With the timely access, flexible self-service and userfriendly interface data analytics platform, leaders form clinical and administrative department in the hospital system can choose their affiliative branch and time period for data. It is emphasized early detection of managerial problems in utilization of the limited resource and quality outcome measures, rather than fixing the problems after they have occurred. We measured the efficiency of system use, the average dill down time of inpatient dashboard was 6.5 seconds. And we also monitored the indicator validation, the rate of the consistency of data were 90% above, ensuring the data correction. Conclusion(s): With the increasing demand for IT-supported application in hospital management, a dedicated team for BI application might be a better way to achieve decision making process innovation. Developing a centralized data infrastructure is beneficial for data comparison and evaluation in monitoring the inpatient service. The implementation our programme might improve timeliness for accessing data, with the better further management of data-driven policy making, which is a significant trend in the development of medical information technology.