PREDICTIVE DATA MINING OF CHRONIC DISEASES USING DECISION TREE:
A CASE STUDY OF HEALTH INSURANCE COMPANY IN INDONESIA
BY
DINI HIDAYATUL QUDSI
A dissertation submitted in fulfilment of the requirement for the degree of Master of Information Technology
Kulliyyah of Information and Communication Technology International Islamic University Malaysia
JULY 2015
ii
ABSTRACT
The development of information and communication technology has rapidly penetrated to several sectors including health sector. A good data management has become necessity for a healthcare company since it will provide better control of the costs and mitigate risks. However, to develop a good quality data management is complex. Therefore, data mining as one of the advancements of science and technology development offers its technique (such as decision tree) to mine the hidden information from the large amounts of medical data that may improve the decision making. It is the aim of this study to identify the potential benefits that data mining can bring to the health sector, using Indonesian Health Insurance company data as case study. The most commonly data mining technique, decision tree, was used to generate the prediction model by visualizing the tree to perform predictive analysis of chronic diseases. All the steps in data mining process such as data collection, data preprocessing and data mining have been performed by a data mining tool, named WEKA. Additionally, WEKA also was utilized to evaluate the prediction performance by measuring the accuracy, the specificity and the sensitivity. Among the result found in this study shows some factors that the health insurance can take into account when predicting the treatment cost of a patient.
iii
ثحبلا صّخلم
لا نم ديدعلا لىإ عيرس لكشب تلااصتلااو تامولعلما ايجولونكت روطت لغوت دقل ق
،تاعاط
ورضلا نم ةديلجا تناايبلا ةرادإ تحبصأو ،يحصلا عاطقلا كلذ في ابم تاكرشلل تيار
نم نأ يرغ ،لضفأ لكشب رطاخلما نم دلحاو ،فيلاكتلا مكتح ىلع اتهردقل ،ةيحصلا إ ريوطت بعصلا ةيعونلا ةدولجا تاذ تناايبلا ةراد
تناايبلا نع ثحبلا وأ بيقنتلا ناك اذلهو ،
ثيح ،ايجولونكتلا ريوطتو مولعلا تلاامج نم ةمدقتلما تلاالمجا ىدحإ نأ
نكملما نم
ةيفخلما تناايبلا نع بيقنتلبا موقت نأ تارارقلا ةرجش لثم ،تناايبلا نع بيقنتلا تاينقتل بطلا تناايبلا نم ةلتكل ةي
، ةفرعم ةساردلا هذه نم فدلها .رارقلا عنص ينستح في ةهماسلماو
ةلمتلمحا دئاوفلا لل نكيم تيلا
ت ةناعتسلابا كلذو ،يحصلا عاطقلل اهبليج نأ تناايبلا نع بيقن
تارارقلا ةرجش دعُتو .ةسارد ةلاحك اهمدختساو ،ينمأتلل ةيسينودنإ ةيحص ةكرش تناايبب نقت نمض اعويش تاينقتلا رثكأ جارختسلا اهمادختسا َّتم كلذل ،تناايبلا نع بيقنتلا تاي
لتح ءارجإ ضرغل اهتنياعمو ةرجشلا راهظإ قيرط نع كلذو يؤبنتلا جذومنلا ؤبنت لي
ضارملأل ي
تناايبلا عجم :لثم ،تناايبلا نع بيقنتلا تايلمع ءانثأ تاوطلخا عيجم ءارجإ َّتم دقو .ةنمزلما
،
ساو ،اهتلجاعمو قيرط نع اهصلاخت
بيقنت ةادأ كيوِب ةامسلماو تناايبلا
ا (WEKA) ،
كيو مادختسا َّتم كلذ لىإ ةفاضلإبا ا
،ةقدلا :نم لك سايق للاخ نم ؤبنتلا ءادأ مييقتل
ينمأتلا تاكرشل نكيم تيلا لماوعلا ضعب ةساردلا جئاتن ترهظأ .ةيساسلحاو ،ةيصوصلخاو
.ضيرلما جلاع فيلاكتب ؤبنتلا ءانثأ رابتعلاا ينعب اهعضو يحصلا
iv
APPROVAL PAGE
I certify that I have supervised and read this study and that in my opinion, it conforms to acceptable standards of scholarly presentation and is fully adequate, in scope and quality, as a dissertation for the degree of Master of Information Technology
...
Mira Kartiwi Supervisor
I certify that I have read this study and that in my opinion it conforms to acceptable standards of scholarly presentation and is fully adequate, in scope and quality, as a dissertation for the degree of Master of Information Technology
...
Jamaludin Ibrahim Examiner
This dissertation was submitted to the Department of Information Systems and is accepted as a fulfilment of the requirement for the degree of Master of Information Technology
...
Siti Rohimi Bt. Hamedon
Head, Department of Information Systems
This dissertation was submitted to the Kulliyyah of Information and Communication Technology and is accepted as a fulfilment of the requirement for the degree of Master of Information Technology
...
Abdul Wahab Abdul Rahman Dean, Kulliyyah of Information and Communication Technology
v
DECLARATION
I hereby declare that this dissertation is the result of my own investigations, except where otherwise stated. I also declare that it has not been previously or concurrently submitted as a whole for any other degrees at IIUM or other institutions.
Dini Hidayatul Qudsi
Signature ……… Date ………
vi
Copyright Page
INTERNATIONAL ISLAMIC UNIVERSITY MALAYSIA
DECLARATION OF COPYRIGHT AND AFFIRMATION OF FAIR USE OF UNPUBLISHED RESEARCH
Copyright © 2015 by Dini Hidayatul Qudsi. All rights reserved.
PREDICTIVE DATA MINING OF CHRONIC DISEASES USING DECISION TREE: A CASE STUDY OF HEALTH INSURANCE COMPANY IN
INDONESIA
No part of this unpublished research may be reproduced, stored in a retrieval system, or transmitted, in any form or by any means, electronic, mechanical, photocopying, recording, or otherwise without prior written permission of the copyright holder except as provided below.
1. Any material contained in or derived from this unpublished research may only be used by others in their writing with due acknowledgement.
2. IIUM or its library will have the right to make and transmit copies (print or electronic) for institutional and academic purposes.
3. The IIUM library will have the right to make, store in a retrieval system and supply copies of this unpublished research if requested by other universities and research libraries.
Affirmed by Dini Hidayatul Qudsi
………..….. ………
Signature Date
vii
Dedication
Dedication to:
My beloved parents, brother, sisters and friends Thank you for your prayers, support, and believe in me
viii
ACKNOWLEDGEMENTS
In the name of Allah, the Most Gracious and the Most Merciful, along with Salawat and Salam to our role model, the Prophet Muhammad SAW.
Alhamdulillah, praise to Allah for his blessings and guidance of His grace, and also for giving me strength, idea, ability and patience, so I could complete this dissertation.
I would like to take this opportunity to express my gratitude to my supervisor, Assistant Professor Dr. Mira Kartiwi who has given me guidance, advice, supervision and idea throughout the study. There is nothing more pleasant than having a supervisor who is very kind and understanding like her.
I would like to thank Mr. Jamaludin Ibrahim and Dr. Izzuddin Mohd Thamrin for their helpful feedback and suggestions.
Enormous thanks to my beloved parents, my dearest brother and sisters. Thank you for always being by my side and never stop believing in me that I would be able to get here.
Finally, I would like to thank my lovely friends who have given me support, strength and smiles during the completion of this dissertation.
Alhamdulillah
ix
TABLE OF CONTENTS
Abstract ... ii
Abstract in Arabic ... iii
Approval Page ... iv
Declaration ... v
Copyright Page ... vi
Dedication ... vii
Acknowledgements ... viii
List of Tables ... xi
List of Figures ... xii
CHAPTER 1: INTRODUCTION ... 1
1.1 Background of the Study ... 1
1.2 Problem Statement ... 4
1.3 Research Objectives ... 5
1.4 Research Questions ... 5
1.5 Research Scope ... 5
1.6 Significance of the Study ... 6
1.7 Organization of the Study ... 7
1.8 Chapter Summary... 7
CHAPTER 2: LITERATURE REVIEW ... 8
2.1 Introduction ... 8
2.2 Chronic Disease ... 8
2.3 Use of IT in the Health Sector... 9
2.3.1 Electronic Health (E-Health)... 10
2.3.2 Telemedicine ... 11
2.3.3 IT Decision Support ... 12
2.4 Data Management Challenge in Health Sector ... 14
2.5 Data Mining Implementation in the Health Sector ... 16
2.5.1 An Overview of Data Mining ... 16
2.5.2 Knowledge Discovery in Databases (KDD) ... 19
2.5.3 Data Mining Benefit to Health Sector ... 20
2.5.4 Previous Research of Predictive Medical Data Mining ... 23
2.5.5 Data Mining Challenges in the Health Sector ... 25
2.6 Data Mining Techniques ... 27
2.6.1 Decision Tree ... 29
2.7 WEKA ... 31
2.8 Chapter Summary... 34
CHAPTER 3: METHODOLOGY ... 36
3.1 Introduction ... 36
3.2 CRISP-DM Methodology ... 36
3.3 Research Design ... 38
3.4 Data Preprocessing ... 42
3.4.1 Raw Data ... 43
x
3.4.2 Data Integration ... 43
3.4.3 Data Cleaning ... 44
3.5 Prediction Model ... 44
3.5.1 WEKA ... 45
3.5.2 Decision Tree ... 45
3.5.3 J48 Algorithm ... 45
3.5.4 Visualize the Tree ... 48
3.5.4.1 How to Analyze the Tree Visualizer ... 49
3.6 Validate Prediction Model Performance ... 50
3.7 Instrument/Measures ... 51
3.8 Chapter Summary... 52
CHAPTER 4: RESULT AND FINDINGS ... 53
4.1 Introduction ... 53
4.2 Data Collection and Data Understanding ... 53
4.3 Data Preprocessing ... 54
4.4 Modelling and Experiments ... 55
4.4.1 Data Preprocessing in WEKA ... 56
4.4.2 Data Mining Process ... 58
4.4.3 Decision Tree ... 60
4.4.4 Outpatients Analysis ... 62
4.4.5 Inpatients Analysis ... 64
4.5 The Summary Evaluation ... 66
4.5.1 The Decision Tree Summary Analysis ... 67
4.5.2 The Accuracy of Prediction Performance Analysis ... 67
4.5.3 The Confusion Matrix ... 75
4.5.3.1 Sensitivity and Specificity ... 77
4.6 Data Mining Performance ... 78
4.7 Chapter Summary... 79
CHAPTER 5: CONCLUSION AND SUGGESTIONS ... 80
5.1 Introduction ... 80
5.2 Summary of Findings ... 80
5.2.1 Data Mining Benefit For Health Insurance Company ... 81
5.2.2 The Importance of Data Quality in Health Sector ... 83
5.3 Limitations and Recommendations ... 85
5.4 Chapter Summary... 86
BIBLIOGRAPHY ... 87
APPENDIX I: PREPROCESS PANEL OF WEKA ... 94
APPENDIX II: CLASSIFY PANEL OF WEKA ... 95
APPENDIX III: CLASSIFIER OPTIONS IN WEKA ... 96
APPENDIX IV: THE CLASSIFICATION DATA IN WEKA ... 97
xi
LIST OF TABLES
Table no. Page no.
2.1 The Medical Data Mining Previous Research For Predictive Modeling 24
3.1 WEKA Confusion Matrix Description 51
4.1 The Description of the Attributes in the Dataset 55
4.2 The Attributes of the Sample Set 55
4.3 The Classification Rules of Outpatients 62
4.4 The Prediction of Chronic Diseases For Outpatients Based On Age Group 63
4.5 The Classification Rules of Inpatients 64
4.6 The Prediction of Chronic Diseases For Inpatients 66
4.7 The Factors that Influence Chronic Diseases 67
4.8 The Reliability of Data Validation Using SPSS 75
4.9 The Confusion Matrix 76
xii
LIST OF FIGURES
Figure No. Page No.
2.1 The Data Quality Improvement Activities 15
2.2 The Data Mining Process 17
2.3 KDD Process 20
2.4 The Classification of Data Mining Methods 29
2.5 A Simple Decision Tree For Mammalia Classification 31
2.6 WEKA GUI Chooser Interface 32
3.1 The Research Framework 40
3.2 The Visualization of Decision Tree with One Leaf 49
3.3 How to Analyze the Visualization Tree 50
4.1 Preprocess the Data 56
4.2 The Histograms Based On Gender, LOS, Chronic Disease Attribute 57
4.3 The Classify Tab of WEKA 59
4.4 The Decision Tree 60
4.5 The Decision Tree where LOS as the Most Critical Factor 61
4.6 The Classifier Output of Decision Tree 68
4.7 The Classifier Error Graph of Prediction Model 69
4.8 The Classifier Error Information 69
4.9 The Accuracy of Prediction Performance After Removing the Errors 71
4.10 The Decision Tree Based On Age Grouping 72
4.11 The Accuracy of Prediction Performance For Outpatient Only 74 4.12 The Accuracy of Prediction Performance For Inpatient Only 74
4.13 Sensitivity And Specificity Proportion 78
1
CHAPTER ONE INTRODUCTION
1.1 BACKGROUND OF THE STUDY
In the last few years, Information Technology (IT) has been developing very fast, which has changed the way we live and work. The utilization of Information Technology (IT) wisely will greatly assist in the work field. Therefore, the development of Information Technology (IT) has been penetrated into the health sector.
Information Technology (IT) plays an important role in the health sector. Some examples of the IT implementation in the health sector; namely e-health, telemedicine, and IT as decision support are successful support business processes and decision making processes. It is agreed that the use of IT in the health sector will not only provide benefit for the users, but also for the health organizations, such as hospitals, health centers, clinics and insurance health company.
Badan Penyelenggara Jaminan Sosial (BPJS), Healthcare and Social Security Agency, is one of the health insurance companies in Indonesia. BPJS Health is a State Owned Enterprise that is specifically assigned by the government to provide health insurance for civil servants, Pension Recipients army / police, Veterans, Independence Pioneers (and their families) and other business entities. This company handles health insurance for all kinds of diseases starting from mild to serious illnesses. An example of the latter is their handling of chronic diseases.
Chronic disease is a disease that lasts for a long duration of time. Currently, the numbers of diseases that belong to the category of chronic disease have increased.
2
The following is a list of chronic diseases that are common in this modern era, such as ALS (Lou Gehrig’s Disease), Alzheimer’s disease (and other Dementias), Arthritis, Asthma, Cancer, Chronic Obstructive Pulmonary Disease (COPD), Cystic Fibrosis, Diabetes, Heart Disease, Oral Health, Osteoporosis, Reflex Sympathetic Dystrophy (RSD) Syndrome, etc. In addition, nowadays, the majority of chronic diseases are caused by an unhealthy and wrong lifestyle. Therefore, health counseling is needed to provide an early warning to the young people to start living healthy.
BPJS Health insurance company has been implementing IT system for many years, to perform several business tasks, such as to manage company’s participants database which is about 16.8 million people from around Indonesia, to manage some applications which are spread in several hospitals and to manage medicines and medical services. Thus, it is necessary for BPJS Health insurance company to have a good data management. Since a good data management will generate revenue, control costs and mitigate risks (American Institute of of CPAs, 2013).
Money and people have long been considered to be assets, but nowadays, many organizations rely on their data to make more informed and effective decisions which help the organizations to achieve their goals. Hence, data needs to be managed seriously (Searchdatamanagement.techtarget.com, 2013). But developing a good quality data management is not easy. Still sometimes, the organization meets some challenges during data management process, especially in the health sector, huge amounts of data need to be organized and stored.
Price et al. (2013) stated that various efforts such as case management implementation, utilization review, and disease management, have been made by the health care data management practitioners to control the cost of the healthcare and handle the utilization of services. However, all of these programs do not appear to
3
work in controlling the cost. They suggested different methods to identify patients with chronic disease (since it has higher risk for readmission) or to predict disease progression and health status have been considered by the health insurers and health systems to control cost in medical professional manner (Price et al., 2013). Therefore, predictive models built by data mining could be one of the solutions.
Therefore, data mining as one of the advancements of science and technology development offers its technique (such as decision tree) to extract information from the huge amount of data that may improve the quality of data decision making management (Milovic & Milovic, 2012). Data mining can be greatly beneficial for the healthcare industry.
Data mining is becoming more well-known day by day, since it strengthens the companies to discover profitable patterns and trends from their existing databases (Larose, 2005). The crucial objective of data mining is prediction. Predictive data mining is the most common type of data mining and one that has the most straight business applications (Statsoft.com, 2014). Data mining uses a technique to build a model and to validate the predictive performance. Decision tree, as one of the data mining techniques has proven to become the most accurate predictor among other techniques, namely artificial neural network and regression model (Delen, Walker, &
Kadam, 2005).
As can be seen from the benefits of data mining above, it is the aim of this study to identify the potential benefits that data mining can bring to the health sector in Indonesia. The study will utilize the health insurance data owned by BPJS Health insurance company to predict factors that influence chronic diseases by using decision tree as the data mining technique.
4 1.2 PROBLEM STATEMENT
In many studies done previously, data mining has been proven to be a very useful tool to predict a disease from medical data records and has been applied in many health organizations. Data mining executes a large amount of data that can be used to make better decisions in an organization (Taylor, 2012). However, implementing data mining in medical sector is a challenging task which requires time and efforts. As (Cios & Moore, 2002) stated that it is very challenging yet fascinating to apply data mining, knowledge discovery and machine learning techniques to medical data.
The quantity and the quality of the data play an important role to generate an accurate predictive model of data mining (Cios & Moore, 2002). However, it is not easy to obtain accurate and comprehensive medical data. Generally, the dataset to process data mining are very large, heterogeneous, complicated and differ in quality (Hosseinkhah, Ashktorab, Veen, & Owrang-Ojaboni, 2008). Data cleaning and data preprocessing are needed to eliminate the redundant and incomplete data before processing data mining. But, even after carrying out those two methods, the result achieved is not useful if the quality data is poor. Thus, poor data quality is one of the major obstacles to generate a successful data mining (Thorat & Kute, 2014).
In addition, the lack of integration can also be the cause of the data mining failure. The dataset which is used to load into data mining process comes from heterogeneous system (distributed between hospitals, health insurance and government departments) which increases the requirement to uniform standard for integrating dataset (Thorat & Kute, 2014).
5 1.3 RESEARCH OBJECTIVES
The objectives of this research are the following:
1. To identify the capabilities of data mining in health sector, especially using medical dataset.
2. To predict the factors that influence chronic disease and to identify length of treatment.
3. To evaluate the knowledge derived from patterns generated by the data mining technique for BPJS Health insurance company.
1.4 RESEARCH QUESTIONS
The research questions of this research are the following:
1. What kind of knowledge emerged from patterns generated by the classification process and how it may benefit health insurance company.
2. From the data mining process, what are the conditions that the company / organization has to handle in producing optimum model.
1.5 RESEARCH SCOPE
The scopes and limitations of this research are:
1. This research will focus on how data mining can predict the factors which influence chronic diseases based on 4 criteria, namely age, gender, los (length of stay) and disease.
2. All of the data mining process, such as data preprocessing and data classifying will be executed in WEKA, a data mining tool.
3. Decision tree technique is used in this research for data mining process.
6
4. C4.5, known as J4.8 in WEKA, is used as the algorithm to generate the decision tree.
5. Medical dataset that is utilized in this research has come from BPJS Health insurance company in Indonesia.
1.6 SIGNIFICANCE OF THE STUDY
The findings of this study would provide some potential benefits that data mining can bring to the health sector in Indonesia, especially to BPJS Health insurance company.
Also, some information can be provided on how a good data management could be beneficial for the company in order to produce an optimum model of predictive data mining.
The findings of this study are developed through data mining process which utilized the health insurance data owned by BPJS Health insurance company. Such findings can be used to predict factors that influence chronic diseases where it can assist in the implementation of a new policies / wisdoms for the company. The new information is extracted from patterns generated after data mining process by using decision tree as the data mining technique. The reason for using decision tree is the predictive result of decision tree is easier to read and interpret, especially for people who are not familiar with data mining, so they can directly draw the information from the visualization of decision tree.
7 1.7 ORGANIZATION OF THE STUDY
Chapter One: The introduction of the study, the background of the research, problem statement, research objectives, research scope, and significance of the study are elaborated in this chapter.
Chapter Two: This chapter describes the literature review which starts from the introduction of chronic diseases, continued with the use of IT in the health sector, followed by the challenges of data management in health sector and how data mining could offer its benefit. Also, a brief overview of data mining, the previous researches of predictive medical data mining, the benefit and the challenges of data mining to health sector have been elaborated. Lastly, in the last section, it explains the decision tree as a data mining technique and WEKA as a data mining tool
Chapter Three: This chapter describes the methodology that has been implemented in the research. This is followed by the research framework, which explains the process taken in every stage to carry out the study. In the last section, the instrument or measurement of the research is elaborated.
Chapter Four: This chapter presents the finding and result from data analysis after implementing the research methodology of this study.
Chapter Five: This chapter provides the conclusions, limitations and recommendations for further study.
1.8 CHAPTER SUMMARY
This chapter is an introduction to the background of this study and why it was conducted. Followed by the problem statement; the scope, the objectives, and the questions of the research are discussed. In the end, the significance of the study and the organization of this report have been described.
8
CHAPTER TWO LITERATURE REVIEW
2.1 INTRODUCTION
This chapter would elaborate on the chronic diseases in the first section, followed by the use of IT in the health sector, for the second section. Section three explains the challenges of data management in the health sector. Meanwhile, in the fourth section, data mining implementation in the health sector would be elaborated, starting from a brief overview of data mining and Knowledge Discovery in Databases (KDD), followed by the previous researches of predictive medical data mining, ending with the benefits and challenges of data mining to the health sector. Data mining techniques will be introduced in the fifth section. Finally, WEKA, data mining software used in this research would be introduced in the last section.
2.2 CHRONIC DISEASE
Chronic Disease, also known as Non Communicable Disease (NCDs) is a disease that continues for a long time with the slow progress which is not distributed from person to person (Who.int, 2014). There are 4 main types of NCDs, which are cardiovascular diseases (heart attack and stroke), cancers, chronic respiratory disease (chronic obstructed pulmonary disease and asthma) and diabetes. Moreover, it is elaborated by Health.ny.gov (2014) that chronic diseases also include ALS (Lou Gehrig’s Disease), Alzheimer’s Disease and other Dementias, Arthritis, Cystic Fibrosis, Oral Health, Osteoporosis, Reflex Sympathetic Dystrophy (RSD) Syndrome.
9
Cmcd.sph.umich.edu (2014) stated that chronic diseases have a major impact on the people around the world. Chronic diseases are not only the most deadly disease but also the cause of disability and premature death all over the world. But at the same time, chronic diseases can be efficiently prevented and avoided.
The risk factors of chronic diseases are widely known which include unhealthy diet, physical inactivity and tobacco use. These factors have been proven as the causes that increase blood pressure, elevate glucose levels, abnormal blood lipids, overweight, and obesity (World Health Organization, 2014). Furthermore, chronic diseases do not always happen to old people but may also occur to all age groups, such as type 1 of diabetes and childhood asthma; these are the examples of chronic diseases that start in early life (Aihw.gov.au, 2014).
2.3 USE OF IT IN THE HEALTH SECTOR
There are several benefits of using IT in the health sector which create efficiency among patients, doctors, and practitioners. IT has become an important role of information management in several hospitals in Indonesia. It has been stated by Oberty (2012), that in Indonesia, information system has assisted health practitioners in performing their duties related to decision making (Decision Support System). Not only in decision making, but the benefits of IT can also be seen in the implementation of e-health, tele nursing, etc. that can improve Indonesian public health services (Murdiyanti, 2012). Below are the several benefits of IT implementation in the health sector.
10 2.3.1 Electronic Health (E-Health)
Eysenbach (2001) defined E-Health as a developing area of using information and communication technology not only in technical development which are the business of public health, medical informatics and health services improved through the internet and technology, but also a state-of-mind, a way of thinking, an attitude, a pledge for networked, global thinking, to develop health care locally, regionally and universally.
E-health has been implemented in many healthcare enterprises in order to support the transformation of the healthcare organizations. For example, the Tasmanian e-Health Collaborative Project which cooperated with private insurers, private hospitals, Veterans Affair Department, the Australian Centre for Health Research and Tasmanian Department of Health and Human Services. The objective of this project was to develop an ICT infrastructure that could share the electronic discharge summaries over public and private sector or from private hospitals to the primary General Practitioner (GP) and other legal care providers. Meanwhile the information was delivered in a safe way through a broadband network, so it would be possible to share the information with GP, other legal healthcare providers and consumers. The fast, reliable, and appropriate e-Health system was expected to improve the quality, safety, coordination and steadiness of care for patients to exchange the information between private hospitals and GPs. Moreover, e-Health system performed a business and technology solution that assisted to support the sustainability of the health sector and the range of healthcare providers as well as specialists, community services and federated health (Georgeff, 2007).
Furthermore, the existence of the internet allows HIT (Health Information Technology) system to utilize cloud computing that can provide a robust infrastructure
11
and better services. HIT enable health organizations to act more efficiently and cost- effectively. This encourages the development of a model, "E-health Cloud" that can help the healthcare industry to overcome the recent and future demands with minimum cost. Even though HIT, as a Cloud Computing model, has a great potential to increase the quality of healthcare industry, still not much literature discusses about this integrated HIT (AbuKhousa, Mohamed, & Al-Jaroodi, 2012).
In the future, all of the relevant stakeholders of e-health (healthcare professionals, hospitals, insurance companies and drug productions) will concentrate on digitization of their process and assist the practice of these processes by others. For instances, pharmaceuticals will concentrate on e-prescriptions and e-refills, and also will assist doctors and other practitioners to obtain and put their orders of several medicines. Meanwhile, the insurance companies will concentrate only in e-billing and e-payments among hospitals, doctors, and pharmacies, whereas the patients will get the information of results, diagnoses, prescriptions, appointments, and insurance (Varshney, 2009a).
2.3.2 Telemedicine
Telemedicine can generally be defined as the utilization of telecommunication technologies that offers medical healthcare services to the patients to share the information over distances (Varshney, 2009b). It is not difficult to find the telemedicine technology nowadays. Strong telecommunications network and video equipment are available extensively with many options. Although currently the technologies in the health field are still underdeveloped, the manufacturers have already offered several products that meet industry standards and ensure interoperability with other devices. Therefore, the development of data integration
12
between different systems continues to be pursued. Since it is important to health care considering the patient’s data must be available whenever needed (Harnett, 2006).
Telemedicine technologies have been implemented in many health industries.
For example, tele-psychiatry, videoconferencing and tele-psychology were utilized in the field of mental health in UK. Videoconferencing has been proved that it could improve the psychiatric services, particularly for those patients who lived in rural areas, which can be an effective method to help the patients (Norman, 2006).
Another example of the telemedicine implementation is in South Africa. The potential advantage of Information and Communication Technology (ICT) in delivering healthcare to the rural areas has been well-known for the South African National Department of Health (DoH). Since half of the population of the country lives in rural areas, telemedicine becomes a good strategy to overcome the disproportionate distribution of healthcare resources. The telemedicine maturity model was proposed to measure, manage, and enhance all the components of a telemedicine system to generate an improvement process that would suite an enterprise (Van Dyk, Fortuin, & Schutte, 2012).
2.3.3 IT Decision Support
Berner & Lande (2007) stated that the implementation of IT in health sector to support decision making, called Clinical Decision Support (CDS) System, is a computer system that is built to influence physicians’ decision-making for the patients. The decision is made at that moment in time.
Some of the benefits of the clinical decision support system are to increase the quality of the medical diagnosis and to decrease diagnostic errors. Graber & Mathew (2008) have developed a new web-based clinical decision support system that can