#DHA2020CHINA

Big data analytics and AI in digital health

Applying Big Data to Construct a Dynamic Medical Expenditure Forecast Model for the Taiwan's Surgical Department (DHA2019-12)

Author/s: Che-Wei Chang

Abstract: In the past 10 years, Taiwan implements Taiwan Diagnosis Related Groups (TW-DRGS) has affected the income of surgeons. In order to understand the impact of the surgeon's income in the total payment system, this study proposes a dynamic prediction model, first using the health data bureau to issue 2000-2012 big data, and using data mining technology to classify the data. Therefore, based on the Global Budgets System and classify the data, to develop the EWMA (Exponentially Weighted Moving Average, EWMA) dynamic prediction model can strike a balance between medical quality and physician income. Expected results: Use the 2013-2018 data to verify the accuracy of the dynamic model, providing health care bureaus, hospitals and surgeons with instant access to medical services and payment status.

 

Extracting Characteristics of Diabetes Mellitus Patients Using Machine Learning (DHA2019-25)

Author/s: Madurapperumage Anuradha Erandathi

Abstract: Due to the high prevalence rate of diabetes mellitus in all over the world, the diagnosis and prediction of diabetes are crucial. Since prevention is always better than cure, issuing early warnings for individuals who express the characteristics of the diabetic patient is beneficial in maintaining a healthy life. Extracting the characteristics of diabetes mellitus patients is the main objective of this research followed by a statistical model which will be able to classify the individual as diabetes or non-diabetes by considering their class probabilities. The descriptive statistical measurements were used in extracting characteristics of diabetes patients while checking the significance of each rule through one sample t-test. To make the combined characteristics, the correlation of features was visualized using a heatmap. Eight characteristics were extracted using the details of diabetes patients considering one feature at a time. The null hypothesis of all the features except glucose level and insulin level cannot be rejected because of the calculated p-values of them. The outcome of this research is beneficial to doctors, patients and health authorities to make decisions at early stages of diabetes.

 

Not everyone appreciates the benefits of social support: the effects of relational self and attachment styles (DHA2019-35)

Author/s: Weisha Wang

Abstract: Research on virtual communities and social support are well documented in psychology and marketing literature. There has been little research on how people react to the health-related information provided by other community members. Drawing from social support, attachment style, and relational self literatures, this research proposes a conceptual framework to examine these issues in the healthcare context, and seeking to understand how marketing messages can affect one’s attitude toward the healthcare community.

 

Multi-objective optimization for determining nursing staff demand in nursing home (DHA2019-33)

Author/s: Polly P.L. Leung, C.H. Wu, C.K. Kwong, W.H. Ip and W.K. Ching

Abstract: Due to the need for improving the utilization of existing nursing staff and better sharing their workload, nursing staff demand modelling (NSDM) plays a critical role. Nevertheless, NSDM relies on human experience and often leads to ineffective planning. In previous studies, various critical factors affecting NSDM were commonly ignored. In this study, the formulation of a multi-objective optimization model for the planning of optimal size and mix of nursing teams in nursing homes is described. Critical factors, such as heterogeneous workforce and resident case mix, are considered. A multi-objective evolutionary algorithm, namely Non-dominated Sorting Genetic Algorithm II is applied to solve the proposed model. A case study of a subvented nursing home in Hong Kong was conducted to illustrate the effectiveness of the proposed model and algorithm. Results showed that the total overtime work (in hours) were minimized while the nurse-to-resident ratio were significantly higher within the given budgetary boundary. The model permits for the study of changing overtime work and nursing staff demand as a consequence of changes in resident case mix and resident service requirement. Further, results suggest that it is cost beneficial to introduce temporary staff in the workforce and to increase the degree of skill mix for NSDM.

 

Dynamic Analysis of Medical System Efficiency in Guizhou Province in the 40 Years of Reform and Opening Up: Based on Game Cross-Efficiency DEA Model and Global Malmquist Index (DHA2019-24)

Author/s: Jinna Yu, Tingting Zhang and Zhen Liu

Abstract: Considering the competition of medical resource in the medical system, this paper uses the game cross-efficiency DEA model and global Malmquist index to dynamically analyze the input-output efficiency, total factor productivity (TFP) change and its decomposition of the medical system, based on the data of 40 years of Reform and Opening Up in Guizhou Province. The conclusions are as follows: (1) The efficiency of medical system in Guizhou Province has experienced a trend of “two declines and two rises”. (2) The TFP change and technical change in Guizhou Province have a similar evolution trend. During the first 20 years of Reform and Opening Up, both of them showed an increasing trend, from the 20th to 30th years of Reform and Opening Up, both of them showed a decreasing trend, and then showed an increasing trend again. The technical efficiency change showed a decreasing trend in the first 30 year of Reform and Opening Up, and then showed a strong increasing trend. (3) The progress of medical system in Guizahou has not been expected since the Reform and Opening Up, and there are structural differences in the medical system in Guizhou. In order to improve the efficiency and TFP of medical system, and improve the overall health level in Guizhou, it is suggested that Guizhou should actively learn from the advanced and mature medical reform model in the process of deepening the medical reform, strengthen the technical innovation in the medical field, make full use of Big Data to coordinate the development of the medical system in all prefectures or cities, and realize the sharing of medical resources.

 

Big data in health: Perceptions of Professionals across the Healthcare Sector (DHA2019-21)

Author/s: Kasuni Weerasinghe, Nazim Taskin, David J Pauleen and Shane L Scahill

Abstract: “Big data” is well defined in the literature to mean data with 5V characteristics: volume, variety, velocity, veracity and value. Within the healthcare context, while traditional health data has big data characteristics, new types of data such as genomics and patient generated data are accentuating this trend. The availability of such big healthcare data opens opportunities for modern technologies around big data analytics to be applied to the healthcare context. However, what is missing from the big data literature is research on the role of social dynamics around big data for successful implementation of big data technologies. Social dynamics refers to people’s perceptions and how they understand a technological phenomenon, such as big data. This paper presents a qualitative theory-based enquiry carried out to explore people’s perceptions about the term big data, within the New Zealand healthcare sector. Thirty-two professionals with roles across policy-making, planning, implementation and clinical utilisation were interviewed to understand the social dynamics around big data. The interview data were analysed applying Social Representation Theory (SRT). Findings show that while there was no clear understanding of what constitutes big data, there were demonstrated linkages to the 5Vs. The key implication of this research for policy and practice is understanding the need to initiate dialogue across the sector to clarify the notion of big data and its applicability to the NZ healthcare sector.

Digitalisation Adding Value to Healthcare

©2020 by DHA2020.