University of Economics and Business, Prague Faculty of Informatics and Statistics
Implementation of Advanced Analytical Methods into Healthcare Facilities in Turkey
MASTER THESIS
Study programme: Applied Informatics Field of study: Information Systems Management
Author: Bilgi Özdemir
Supervisor: Ing. Martin Potančok, PhD., June 2020, Prague
Declaration
I hereby declare that I am the sole author of the thesis entitled “Implementation of Advanced Analytical Methods into Healthcare Facilities in Turkey “. I duly identified all citations. The used literature and sources are stated in the attached list of references.
In Prague on June 2021 Bilgi Özdemir
Acknowledgement
I hereby wish to express my appreciation and gratitude to the supervisor of my thesis, Ing.
Martin Potančok, Ph.D. for his support during my master thesis studies.
I would also like to express my gratitude to the people I interviewed for allowing me to benefit from their knowledge and experience.
Finally, I would like to express my gratitude to my family for always giving me morale and support.
Abstract
Technology is developing rapidly, with the development of technology, it develops and changes rapidly in the field of health. There are many technologies that play a role in this change, one of which is advanced analytical methods. There are many benefits that advanced analytical methods can and will add to the health system. In this process, the implementation of advanced analytical methods in healthcare facilities plays an essential role. This
implementation process is quite complex. It has many components and all of them must be applied very carefully and effectively. The aim of this thesis is to discuss the process in detail and to create an exemplary implementation strategy.
Keywords
Advanced Analytical Methods, Artificial Intelligence, Health 4.0, Healthcare, Machine Learning, Health System of Turkey
Table of Contents
1. Introduction ... 7
1.1 Motivation ... 8
1.2 Research Objectives ... 8
1.3 Research Significance ... 8
1.4 Research Method... 9
1.5 Research Limitations ... 9
1.6 Research Structure ... 9
2. Health 4.0 Application-Management-Technologies ... 10
2.1 How Technology Transforming Health Sector? ... 10
2.2 Management of Health 4.0 ... 11
2.3 Health 4.0 Applications and Technologies ... 12
2.4 Benefits of Digital Technologies in Healthcare ... 14
3. Advance Analytical Methods and Artificial Intelligence in Healthcare ... 15
3.1 The Value of Data ... 15
3.2 Predictive Analysis ... 16
3.3 Artificial Intelligence in Healthcare ... 21
3.3.1 Artificial Intelligence Techniques in Health Field ... 22
3.4 Data Mining in Healthcare ... 23
3.4.1 Data Mining Algorithms in Healthcare ... 23
3.5 Machine Learning in Healthcare ... 24
3.5.1 Machine Learning Algorithms ... 25
3.5.2 Deep Learning ... 25
4. Ethical Reflections ... 26
5. Implementation of Advanced Analytics ... 29
5.1 Global Strategy of World Health Organization ... 29
5.2 Health Strategy of Turkey ... 31
5.2.1 Health Transformation Program ... 32
5.2.2 Digital Hospital ... 34
5.2.3 Strategic Plan of 2019-2023 ... 35
5.3 Comparison of Health Expenditures with Other Countries ... 41
5.4 Current Situation Evaluation of Turkey ... 43
5.5 Example Strategy for Turkey ... 60
5.5.1 Strategy Re-creation... 61
6. Conclusion ... 76
References ... 78 Annexes ... 82
1. Introduction
Everything has contributed to the development of the next process since the beginning of human history. The manpower, which led to the development, has been replaced by human- controlled machines in time. While these developments were taking place, the world population has increased rapidly, and developments have accelerated as a requirement of people's needs. The period in which unmanned production was made, the speed of the jet was reached in communication, the technology developed tremendously, and artificial intelligence and cloud databases started to be used were named “Industry 4.0”. Industry 4.0 has found its equivalent in the field of health as Health 4.0. (Kaya & Filiz, 2019)
Health 4.0 has been affected by changes in all areas. It has led to changes in many areas such as medicine, diagnostic services, and treatment services. It will continue to cause rapid changes in the future. The rapid development of information technologies causes the health field to develop and change rapidly. Thanks to these technologies, a more patient-oriented, more innovative, more sustainable, and more digitalized health system awaits us in the future.
The healthcare industry has an important potential in developing innovative products and technologies. The building blocks of digital transformation such as the Internet of things, machine-to-machine communication, cloud computing, big data, wearable, and portable technologies deeply affect the health system. The developments occurring in the world of science and technology every day cause the standards of health services to increase gradually.
(Güzel & Aslan, 2019)
The methods of Health 4.0 are interrelated. Thanks to the effective implementation and use of all of them, it will provide us with great advantages in the future. Advanced analytical
methods and artificial intelligence are essential parts of Health 4.0. This study examines of the Health 4.0, Advanced Analytical Methods and Artificial Intelligence to evaluate the implementation of the health institutions by addressing Turkey's health system and is researching how to implement in practice. An answer to the question of how to make an effective implementation process is sought by analyzing the current situation and strategies.
1.1 Motivation
The introduction of advanced analytical methods and artificial intelligence into our lives in many areas brings us many innovations. In addition to the advantages it brings, it also contains risks and ethical problems. However, especially in developing countries such as Turkey, it does not give much importance to these technologies.
1.2 Research Objectives
The aim of this thesis is to discuss the process in detail and to create an exemplary implementation strategy.
To achieve the purpose of the research, the following are aimed:
• Providing general information about health 4.0 by reviewing the resources, explaining advanced analytical methods and artificial intelligence, determining the risks and advantages.
• Evaluation of the government's policies and experts’ opinions. Conducting interviews with experts, collecting information, and data. Turkey's current situation analysis of the information with those obtained.
• Using what we have achieved in the analysis results, the creation of a strategy for Turkey. Identification of improvements and recommendations.
1.3 Research Significance
Because Turkey is a developing country, the importance of implementing these technologies is not fully understood. At the same time, there is no study on the subject in this context. The aim of the researcher is to explain the importance of technologies and offer solutions for the country.
1.4 Research Method
The research is conducted using both quantitative and qualitative methods. The numerical data we receive from government documents and various sources are quantitative data.
Qualitative data, on the other hand, are based on interviews and inferences.
1.5 Research Limitations
The interviews were conducted only with experts who work in Turkey. Meetings were held with academics, doctors, and IT department employees. Doctors working actively in
organizations such as labor unions were also interviewed.
1.6 Research Structure
This thesis study consists of six chapters. They include an introduction, theoretical section, practical section, and conclusion.
Chapter 1: In the first part, an introduction is made, how health 4.0 can change health and why it is important. Why it is important and why is it mentioned that advanced analytical methods benefit from the example of Turkey.
Chapter 2: In this chapter, the changes awaiting us are explained in more detail, and health 4.0 technologies and their benefits are described.
Chapter 3: In the third chapter, advanced analytical methods and artificial intelligence technologies to be used in healthcare facilities are explained.
Chapter 4: Ethical issues in this change are an important topic of discussion. Ethical reflections are described in this chapter.
Chapter 5: The fifth chapter includes the practical part of the thesis. The global strategy of the World Health Organization has been included. Turkey's current strategic plan was examined, done interviews with experts analyzing the existing situation and what needs to be done. The things that need to be done are defined for an effective strategy, designed an example strategy for Turkey.
Chapter 6: The last part is Conclusion.
2. Health 4.0 Application-Management-Technologies
The health sector is changing and transforming rapidly. Health systems, hospitals, treatment methods, and processes are developing rapidly with technology. The rapid development of technology is accelerating this change. We can say that advanced analytical methods, artificial intelligence, smart robots, sensors, and many more technologies play an important role in this transformation. These autonomous systems that can manage and control
themselves have started to take their place in our lives. In this thesis, I will focus on advanced analytical methods and artificial intelligence for healthcare.
2.1 How Technology Transforming Health Sector?
Digital health technologies and applications are the transformative power of modern information and communication technologies to improve health and health services (World Economic Forum, 2011). The purpose of healthcare services to achieve this transformation is to provide a safer, faster, more accessible service and to increase work efficiency. Nowadays, health institutions in most countries have started or are trying to keep up with this
transformation.
Digital health technologies and applications control one's own health, ensure compliance with the applied treatment protocols, and encourage preventive health activities; it is defined as the set of systems and tools that provide communication between the individual and the
healthcare worker (Lupton, 2013). The digital health environment is a different health environment than what was known and used in the past. By using data management,
analytical methods, artificial intelligence, social networks, internet, wearable technology; it is aimed to improve and develop human health and health practices. Digital health: It also takes the communication between its stakeholders, including patients, doctors, organizations, and hospitals, to another dimension. It improves and accelerates communication between individuals and institutions.
Biosensors, wearable technologies, wireless mobile devices placed on their bodies that allow individuals with disabilities, chronic illnesses, and any health problems to monitor their health conditions, including themselves in the treatment process; digital media tools
providing medical counseling; digital technologies that help to share health-related data and experiences with other individuals, and to solve the problems of individuals about therapy, treatment, health, and diseases; social media sites used by healthcare professionals to provide information about interacting with patients and the service they provide; Technologies that enable the gene mapping of the individual; All digital medical imaging devices constitute the scope of digital health applications and tools (Lupton, 2013).
Digitalization of health is considered the future of human health. These technologies will offer us many innovations. Real-time data about the health status of the individual is
collected and the health status is kept under constant control. When necessary, the health of the patient will be regularly monitored and kept under control by transmitting the mobile application data to the doctor. With the use of advanced analytical methods and artificial intelligence, diseases will be examined much faster. Even if it is not applied today, it will be possible to perform surgery with robotic technologies in the future. Success rates in treatment will increase significantly, and individuals will offer the opportunity to follow their health status at some point. Wearable technologies will continuously record and process an individual's health data. Service quality of hospitals will increase, diagnosis time will decrease, therefore hospitalization time will decrease.
The use of these technologies is increasing day by day. Due to the pandemic we are in, people have better understood what kind of facilities and benefits this technology can offer us in the future. But at the same time, digital health brings us new discussions and problems.
Whether artificial intelligence will replace healthcare professionals is an important debate. At the same time, issues such as data sharing and data security bring ethical issues and
discussions with them.
2.2 Management of Health 4.0
The implementation and management of digital transformation in healthcare is a very complex process. It requires very good strategic planning. In this process, governments, the private sector, health institutions, healthcare professionals, and universities working on this subject need to work in a planned and integrated manner.
According to the OECD’s publication written by Slawomirski, advances in digital health depend on an enabling policy environment. This requires the determination of governments on three main issues:
1. An overarching digital strategy: Most of the countries on the way to digitalization do not have such a strategy. There are many strategies, but very few are comprehensive strategies. At the same time, few contain a consolidated vision, plan and policy framework. An ideal strategy should also be broader and compatible with other sectors.
2. Strengthening the governance of health data: Governance ensures the safe use of data and digital technologies and also ensures respect for the privacy of the individual.
Legal barriers, a lack of trust in the use and protection of data among patients, the general public, data custodians, and other stakeholders, as well as a lack of agreement on data standards and exchange formats both within and across countries, are all major roadblocks.
3. Building institutional and operational capacity: The workforce must be equipped and prepared to take advantage of the opportunities of digital technology. The public must be empowered to take advantage of it. It also entails establishing systems and
institutional arrangements that allow for efficient data linking and analysis. This necessitates a policy environment that allows key actors to not only access data and extract knowledge from it, but also to use that knowledge to effect change and advance policy goals (Slawomirski, 2019).
2.3 Health 4.0 Applications and Technologies
Electronic tools, systems, devices, and resources that generate, store, process, and/or transmit data are referred to as digital technologies. These include tangible products such as computers and smartphones, as well as software, web-based platforms, and algorithms, such as Artificial Intelligence (OECD, 2020). According to the publication of the OECD, we can examine digital health technologies under 6 headings:
1) Data-driven automation, prediction & decision support: Analytics using data and technologies like Artificial Intelligence. Artificial intelligence is a machine-based system that can make predictions and make recommendations for a defined set of goals. Machine learning enables digital systems to achieve goals without being given explicit instructions on how to do so, instead of analyzing patterns in training data that must be properly prepared, including or excluding labeling. Deep learning is a
subfield of artificial intelligence in which the digital system achieves this goal by determining the distinguishing features of data sets in a hierarchical manner.
2) Assested-living technologies: Assisted-living technologies are combinations of digital health apps and other software. For example, sensors, robots that help the elderly and patients, robots used in home-based treatments are some of them. The main purpose of these technologies is to enable patients to stay in their own homes, to live independently for longer, or to return to their homes faster after treatment.
3) Mobile Health: It refers to the use of digital health applications and sensors on commonly available mobile devices such as smartphones and tablet computers, as well as wearable devices such as smartwatches. It also includes portable monitoring systems and other mobile devices designed specifically for use by health care providers for service delivery and data collection.
4) Telehealth: Mobile health devices and digital health apps are frequently used in telehealth. Telehealth includes a combination of digital solutions that allow clinical services to be delivered and care and treatment to be monitored remotely with relevant caregivers and patient.
5) Electronic Medical Records and Electronic Health Records (EMR and EHR):
Electronic Medical and Health Records (EMR and EHR) are digital records that store a variety of health information about a person. EMR is established in a system or organization that delivers healthcare.
6) ePrescribing, eAppointment: ePrescribing between prescribers and dispensing pharmacies, as well as eAppointments for booking consultations online, are common
2.4 Benefits of Digital Technologies in Healthcare
From the Patient Perspective:
• Speeding up diagnosis and treatment processes
• Holistic or personalized healthcare approach
• Increasing health literacy
• Reducing the severity of the disease
• Equal service delivery to the whole community
• 24/7 access to healthcare
• Improving treatment services
From the Perspective of Healthcare Professionals:
• Developing health technologies for diagnosis and treatment
• Providing easy access to the latest clinical information
• Increasing job satisfaction
• Increasing time for knowledge and skill development
From the Health Service System Perspective:
• Increasing efficiency and speed against situations that threaten public health
• Improving Public-Private Partnership
• Realizing health policies
• Improving the content and scope of healthcare
• Developing own resources
• Capacity building for medical procedures
• Integration of health services
From Stakeholders Perspective
• Development of devices that will facilitate health protection, treatment, and diagnosis
• Reduced management and treatment costs (Timmis & Timmis, 2017)
3. Advance Analytical Methods and Artificial Intelligence in Healthcare
Advanced analytical methods help us better understand why something is happening,
generate predictive insights, identify trends, and optimize for a specific result. It does this by researching the data thoroughly. Some advanced analytical methods are data mining, machine learning, cohort analysis, cluster analysis, retention analysis, complex event analysis,
predictive analysis.
The use of advanced analytical methods in health provides us many benefits. Although there are not many technologies implemented in healthcare organizations yet, their importance has begun to be understood more and more. The combined and effective use of advanced
analytical methods and artificial intelligence will take us to a very different dimension in the field of health. Artificial intelligence uses advanced analytical methods, so it is not possible to separate these two technologies from each other. Therefore, this study investigates the use of advanced analytical methods and artificial intelligence in the field of health and healthcare facilities. Today, there are many studies in the field of health, especially artificial
intelligence, data mining, predictive analysis, and machine learning. Each of these methods actually benefits from each other and are interdependent.
3.1 The Value of Data
The process of analyzing all of the data collected from various sources is known as healthcare data management. This helps healthcare organizations in treating patients holistically,
providing personalized treatments, and improving health outcomes. To extract value from the data, effective tools and methods are required. Technologies help to make decisions that will improve the quality of healthcare services using data. Reviewing the analytics attained provides a larger and more detailed picture of the patient's condition. As a result, more precision-driven care and treatment can be offered to the patient. These technologies increase the competitiveness of healthcare organizations and provide process optimization. (Menon, 2018)
Figure 1: The Value of Data
Source: Importance of Advanced Analytics in Healthcare, Menon, 2018
3.2 Predictive Analysis
Predictive analysis: Analyzes using artificial intelligence, along with modeling of historical data, data mining, and statistics to make predictions for future events or decisions. Predictive analysis in health is the development of personalized treatment methods with software tools and predictions about the future of a person from a person's past information. The experience of healthcare professionals is made by predicting a person's medical history, demographic information, and behavior, as well as her/his future situation.
Organizations may use predictive analytics to enhance anything from chronic disease
management to patient experience and precision medicine. Predictive analytics tools provide a viable way to forecast future outcomes before they arise for those trying to remain one step ahead of the confusion that pervades healthcare. (Kent, 2020)
Predictive analytics is based on logically derived from human-created hypotheses to fit a hypothesis. An algorithm is a set of rules and procedures that are combined into a formula that performs calculations. Unsupervised learning, which does not have a guiding theory and uses an algorithm to find patterns and structure in data and cluster them into groups or insights, can also be used in predictive analytics. The machine may not know what it is looking for in unsupervised learning, but as it processes the data, it begins to notice complex
processes and patterns that a human would never notice, adding significant value to researchers looking for something new. Both supervised predictive modeling and
unsupervised predictive modeling are analytical tools to be used in the versatile use of these technologies. (Deloitte, 2019)
Figure 2: Benefits and risks of predictive analysis
Source: Predictive Analytics in Healthcare: Emerging Value and Risks, Deloitte, 2019
Benefits of Predictive Analysis:
Efficiencies for Operational Management: Predictive analytics enables improved operational efficiency. Today, integral parts of business intelligence strategies are big data and predictive analytics. Real-time reporting can provide timely insights on data. This technology allows the scrutinization of historical and real-time patient admittance rates to determine and flow, while also providing a capability to evaluate and analyze staff
performance in real-time. (Deloitte, 2019)
Predictive models can help optimize staffing levels. The patient-to-staff ratio is known, and managers know how many staff they should have in the facility. The calendar variables (weather, day of the week, official holidays, etc.) can be evaluated to benefit. Predictive
be developed to determine the risk profile of aged care services based on data such as print injuries, staff-to-patient ratios, qualified staff, wages, patient turnover, and profitability statistics. This information can help predict what training is required in the future. Predictive analytics in health ensures effective treatment of ailments after an accurate diagnosis. It can allow it to prevent readmissions and emergency room visits and other adverse events.
Predictive technologies like remote patient tracking and machine learning will work together to help hospitals make better decisions by using risk scoring and threshold warnings.
(Deloitte, 2019)
In other words, the use of predictive analysis will bring many benefits to healthcare organizations. Service quality will increase significantly, more accurate diagnosis and treatment will be provided for patients.
Personal Medicine: One type of treatment for a particular disease may be beneficial for one person, but not for other patients. Predictive analysis can enable the use of big data to provide suggestions for personalized medicine, including preliminary information for diseases that doctors cannot predict. Personalized medicine is about getting the right treatment for a person's disease in the right amount and time. Identifying possible disease risks before the person is sick is also part of personalized medicine. Using the gene mapping method, mutations in the genome are detected beforehand and the appropriate treatment method is available for the person. In this way, they can detect the potential carrier of many diseases in advance. Personalized medicine has begun to change the way we think, define, and manage health problems. The same treatment approach for each disease has begun to be replaced by a treatment approach suitable for each patient's molecular and genetic characteristics. Equipped with more precise and specific tools, healthcare professionals can choose a treatment method or protocol that will minimize harmful side effects for their patients and guarantee a more successful result.
Personalized medicine is used for the detection of mutations known to cause cancer in cancer patients, diagnosis, treatment, and prognosis of cancer, as well as the determination of
susceptibility to cancer. Between 5-10% of cancers are hereditary. (Anadolu Hospitals, 2020) Especially for these hereditary cancers, determining the risk of getting cancer and preventive monitoring according to this risk allows the disease to be diagnosed at a very early stage and therefore the health of the patient.
One of the most important uses of personalized medicine is preventive medicine applications against cardiovascular diseases. There are many different genes associated with familial predispositions to cardiovascular diseases, and the detection of variants in these genes, which are known to be risk factors, is very useful in regulating the living conditions of susceptible people to reduce this risk and in preventive treatment for some cases.
Cohort Treatment: Digitizing electronic health records and legal performance reporting for healthcare organizations provide large amounts of data to gain insight into the health of a population. Access to this data is closely monitored and legalized to protect individuals and avoid the risk of personal information being stolen. Predictive analytics on wide population studies can produce profiles of community and other cohort health patterns using volumes of health system data including regional, demographic, and medical condition information. It can inform health organizations and government agencies about better target interventions.
Estimates are made about the probability of illness and chronic illness based on historical data. These projections may create early interventions aimed at reducing the financial and resource burden on the public health system in the future. Data from the pharmaceutical industry can be used to more accurately target demand, anticipate and reorient supply chain needs and resources, thereby preventing drug shortages. Predictive analysis can provide insights into future health problems. This will help proactively identify groups of people at risk in the future. (Deloitte, 2019)
Risks of Predictive Analysis:
Privacy Issues: The rapid collection of electronic health records and other data increases the available data rapidly every day. Predictive analytics also contains some problems in terms of security and reliability of data, as it uses data. Given the increasing amount of data that is usually stored in the cloud or otherwise accessible over the internet, malicious individuals will be the target for the constant hacking threat. Cloud technology has some ethical
drawbacks to predictive analysis. The security of health data in the cloud contains significant risks. Privacy is an important right for the patient. Countries set laws and rules for the
protection of these data.
The General Data Protection Regulation (GDPR) stands out as the most important reform in Europe's privacy laws for the last 20 years. The regulation requires all companies that collect or store data from EU citizens to comply with this regulation. The regulation states that otherwise, companies have to bear legal consequences.
Moral Hazard: Various ethicists argue that the human touch is essential to healing and that entrusting healthcare decision-making to machines is not considered respectful. The
importance of adhering to recognized ethical norms and points of intervention where human touch or an empathetic human decision is more important than that of a machine should be considered when using predictive analytics in healthcare. The ethics of predictive analytics in the healthcare industry goes down to the role of different stakeholders in the industry. The doctor's role is also one of the areas that can pose a moral hazard. (Deloitte, 2019)
Fast Pace of Technology: Change is happening faster than ever before. The term "digital disruption" means to describe how quickly everything is changing as a result of new technologies. (Deloitte, 2019) The way we do things has begun to completely change. We can attend classes online, shop online, and even use the internet for socializing. The health system is changing too. Due to the pace of change, it can be difficult to keep up.
Lack of Regulation and Algorithm Bias: Bias or unbiased representation are potential problems with predictive analytics. Extrapolative analytical models require a fairly large amount of data representing the entire population. While creating predictive models, the bias developed should also be addressed by developing responsible algorithms that can be traced in an analytical model based on the prediction of decision-making processes. Algorithmic bias happens when technology represents the behaviors and values of the humans who are coding, gathering, choosing, or using the data to train the algorithm, whether they are conscious of it or not. People sometimes put their faith in algorithms, believing them to be neutral and impartial but it is incorrect. Most of the algorithms that drive predictive analytics, whether consciously or unconsciously, have been developed by biased and fallible people.
There should be a continuous feedback loop and efforts should be made to reduce bias, otherwise statistical errors may occur. (Deloitte, 2019)
Figure 3: Risk matrix
Source: Predictive Analytics in Healthcare: Emerging Value and Risks, Deloitte, 2019
The development and validation of the model to be used in the predictive analysis is another ethical consideration. It is critical that the whole project is patient-centered and has patient- centered perspectives; otherwise, it may be considered unethical. Patients' choices, needs, and values are respected and valued in patient-centered care. (Deloitte, 2019) That is why it is very important to establish ethical committees.
3.3 Artificial Intelligence in Healthcare
Artificial intelligence can be defined as computers and computer software capable of
intelligent behavior such as analysis and learning. Artificial intelligence produces solutions to complex problems and does this by imitating human intelligence. It is a wide category that grows and changes day by day and is the pioneer of technological development. Equipped with artificial intelligence technology in automation systems, the decision-making power of the computer is used. More and more commercial systems are emerging day by day and the functional features of the systems are increasing. Artificial intelligence does what people do with fewer errors, at a lower cost, and faster. It offers us many benefits and innovations in the field of health. That is why its popularity is increasing day by day.
3.3.1 Artificial Intelligence Techniques in Health Field
Some of the artificial intelligence techniques we use in the field of health are:
Expert Systems: Expert systems is the one of the most important application areas of artificial intelligence. They are equipped with expert knowledge and look for solutions to problems like an expert. They establish relationships between information, make inferences and make decisions.
Fuzzy Logic: The basic point of fuzzy logic program; to form the knowledge, experience, intuition, and control results of a specialist system operator as a knowledge base.
Transactions are carried out with rules based on knowledge and experience. Experiences are used effectively in fuzzy logic.
Artificial Intelligence Networks: Artificial neural networks collect information about examples, make generalizations, and then make decisions about those examples by using the information they have learned when mixed with examples that they have never seen. Because of these learning and generalization features, artificial neural networks find wide application possibilities in many scientific fields today and reveal the ability to successfully solve complex problems. (Ergezer et al., 2003)
Genetic Algorithms: Genetic algorithms are a search and optimization method that works similarly to the evolutionary process observed in nature. He seeks the best solution based on the principle of survival of the best in the complex multidimensional search space.(Elen &
Turan, 2013)
Neural Fuzzy Systems: Neural fuzzy systems are a combination of neural networks and fuzzy systems. These two models have an independent field in the first place. However, the combination of the two provides benefits for solving most problems. (Serhatlıoğlu &
Hardalaç, 2009)
3.4 Data Mining in Healthcare
The development of tools and procedures for making use of data has alarmed knowledge discovery in databases (KDD). Data mining is the one of the important steps of the KDD.
(Neesha et al., 2015) Data mining is used to reveal confidential, valuable, usable information from large amounts of data and to provide strategic decision support. Data mining has also created a new perspective in the areas of the use of health data. Its use in the field of health has become widespread.
The accuracy of health decisions and policies depends on up-to-date and accurate data.
Generating useful information from vast amounts of health data is the main goal. Health data are collected by many organizations such as hospitals, insurance companies, and public institutions. The information collected is used for many purposes. Better health service delivery, better management of institutions, establishing health policies, efficient use of resources are some of them.
3.4.1 Data Mining Algorithms in Healthcare
1) Anomality Detection: The most significant changes in the data set are discovered using anomaly detection. (Fayyad et al., 1996) Anomaly detection finds and identifies nonnormal data points, events of a data set.
2) Clustering: Clustering divides data into clusters or categories for identification. The aim is to divide similar data into clusters. For example, types of disease can be categorized in this way.
3) Classification: In the healthcare industry, classification is one of the most widely used data mining techniques. It assigns target classes to data samples. For each data point, the classification methodology forecasts the target class. By evaluating patients' disease trends, a risk factor can be associated with them using a classification approach. (Ahmad et al., 2015) Classification techniques used in the field of health:
Decision Tree, K-Nearest Neighbor, Support Vector Machine, Bayesian Classifier, Logistic Regression, Swarm Intelligence, Statistical, Discriminant Analysis.
3.5 Machine Learning in Healthcare
Machine learning is a popular sub-discipline of artificial intelligence. It is the development of algorithms that enable learning according to data types. They are programmed for computers to learn information without human intervention. We can define systems that imitate human thinking and decision-making mechanisms as artificial intelligence. Machine learning algorithms learn to see certain patterns, just like humans. However, algorithms need a lot of data and instances to learn. Artificial intelligence uses large data sets and defines the
interaction patterns between variables.
The field of health is a rapidly developing field. Machine learning has many potentials uses in healthcare, and its usage rate is increasing every day. Like the technologies we mentioned above, machine learning also offers faster diagnosis, lower costs, and less possibility of making mistakes.
We can give an example of three common applications where artificial intelligence is used in the field of health. The first is the application of machine learning to medical images such as MRIs, computerized axial tomography (CAT) scans, ultrasound imaging, and positron emission tomography (PET) scans. The result of these imaging modalities is a sequence or series of photographs that must be interpreted and diagnosed by a radiologist. The computer quickly predicts and finds images that could indicate a disease or serious problem. (Toh &
Brody, 2021)
The second is medical record natural language processing. It is the systematized collection of patient's data electronically stored in digital format. Many healthcare practitioners agree that the drive toward electronic medical records (EMR) in many countries is sluggish, boring, and, in many cases, totally botched. This is due to the amount of physical medical records and documents that currently exist in many hospitals and clinics. (Toh & Brody, 2021) The use of human genetics to predict illness and discover sources of disease is the third machine learning application. A lot of research is now being done on discriminating information about how genetics can affect human health. This is due to the emergence of gene generation sequencing (NGS) techniques and the explosion of genetic data, including large databases of population-wide genetic information. Understanding how genetics affects a
person's risk of developing a disease and how complex diseases can arise can play a
substantial role in preventive healthcare. It can help doctors how to implement the patient's care plan and how to reduce the risks. (Toh & Brody, 2021)
3.5.1 Machine Learning Algorithms
The two main algorithms of machine learning are supervised learning and unsupervised learning. Reinforcement learning is the combination of these two algorithms. It uses trial and error to maximize accuracy. (Noorbakhsh-Sabet et al., 2019)
Figure 4: Machine learning algorithms
Source: Artificial Intelligence Transforms the Future of Health Care, Noorbakhsh-Sabet et al., 2019
3.5.2 Deep Learning
Deep learning is a subset of machine learning that uses multiple layers of artificial neural networks to generate automated predictions from training datasets, similar to how the human brain works. Multiple parameters and layers are common in models focused on deep learning strategies. (Noorbakhsh-Sabet et al., 2019) Deep learning, unlike traditional machine learning methods, instead of learning with coded rules; it can automatically learn from the symbols of the data of pictures, videos, sounds, and texts. Estimation accuracy can increase according to the size of the data.
4. Ethical Reflections
Evaluating medical practices and attitudes as good or bad, right or wrong, and basing these evaluations on certain principles is called Medical Ethics. The development of technology is changing the health field. These changes lead to new ethical debates. Especially the ethics of artificial intelligence and big data have led to controversy.
Ethics, which is a guide in human life; fundamentally, examines moral problems such as what to do, how to behave, and what to pursue. Not only does it propose some values for the problems it discusses, but it also defines the values from scratch and tries to base them consistently with principles. With this feature, ethics is divided into two as theoretical and practical. The theoretical approach of ethics is valid for theoretical ethics, which has a history of approximately two thousand five hundred years. However, the theoretical approach of ethics does not apply to the applied ethics that emerged in the last quarter of the 20th century.
Because, unlike theoretical ethics, applied ethics aims to solve problems and develop opinions towards more current issues. Applied ethics approaches events by taking a more holistic approach to problems. In other words, applied ethics does not only focus on the events with philosophical knowledge but also creates thoughts on problems with psychology, sociology, and biological perspectives. (Cevizci, 2015)
Practical ethics, which is today's ethical discourse, offers a solution for people to encounter a crisis or an epistemological problem. Although the development of technology and science has provided convenience in human life, the progress of technology also brings concerns about the future. The problems and debates this situation will cause ultimately require its inclusion in the issue of ethics. The fact that the computer and artificial intelligence have a place in human life will limit thinking and business skills today; The idea that there may be more for the human race in the future shows that artificial intelligence needs ethical
discourse. In this sense, it is expected that applied ethics will guide the problems regarding the future of artificial intelligence, especially against the issues that are expected to cause ethical problems. Applied ethics, in terms of forming the basis of ethical action regarding the future of artificial intelligence; "Can artificial intelligence be responsible, an ethical artificial intelligence possible, can it have the power to choose?" creates a perspective for such
questions. (Öztürk, 2019)
The increase in commercialization and corporatization in the world has caused artificial intelligence studies to be included in a corporate structure. Increasing artificial intelligence practices by corporatization raise concerns that they will bring about moral, legal and social change in the future. It is thought that the rapid growth of the artificial intelligence industry and having more knowledge and workforce than humans will lead to moral change. Due to these considerations, the inability of artificial intelligence to take responsibility, will, moral action and the increase of human production necessitate the existence of ethical studies. (Dağ, 2018)
The ethics of data, that is, big data, is also an important dimension of ethical debates. Thanks to big data, the collection and size of data have reached a very different dimension. The use and storage of these data for different purposes in different fields have created new problems in the context of privacy and victimization.
While the fundamental changes in the formation of information in the context of big data and the differentiation of research methods in the field of health lead to the emergence of new and better treatment opportunities, on the other hand, the existing problems become complex in terms of ethics and at the same time, new questions and problems arise. Among all these developments, there are many stakeholders including the directly affected patient, the health system where big data facilities are used for diagnosis, treatment, or disease prevention, the science world where research are carried out and data is transformed into information, and companies that provide and manage the technical infrastructure that collects data. (Hand, 2018)
The biggest problem caused by big data is data security. The process of transferring
information about the health of the person to the digital environment and protecting them and using them with the consent of the person has led to serious discussions. Laws and
regulations have been drawn up for this process to proceed in an ethical manner.
One of the classic problems in this regard is that the person's health-related information can be accessed by others and there are some consequences against the person. For example, knowing this information by private health insurance or life insurance may lead to an increase in insurance premiums to be paid. The unauthorized and uncontrolled spread of this
stigmatization, or exclusion for that person according to the type of information. The classic example for these is the presence of sexually transmitted diseases or private examination and test results in a virtual environment and their unauthorized use. (Ucar & İlkilic, 2019)
Anonymization of data is offered as a solution to privacy-related issues in both treatment services and clinical research. Although anonymization provides some benefits, it contains paradoxes in itself. When a person's data is processed independently of his / her identity information, the determination of who the data belongs to becomes easier with the possibilities provided by big data. In this context, it is a difficult method to apply perfect anonymization. (Ucar & İlkilic, 2019)
5. Implementation of Advanced Analytics
Implementation of advanced analytical methods in healthcare facilities is a complex process.
The speed of change in technology, epidemics and the rapid increase of the population increase our need for these technologies, especially in the future.
Having a sustainable health system and healthy citizens is important for every country.
Therefore, they need to give importance to these technologies, allocate sufficient funds, and create a comprehensive strategy. Some countries and companies have been successful in this process, while others are still not fully aware of its importance. Developed countries are way ahead of others in this implementation.
In this part of the thesis, the strategic plan of the World Health Organization will be examined shortly, then Turkey will be examined to better understand the process. Current health policy, strategies, and real-life situations will be analyzed. Suggestions will be made for an effective strategy in line with the results.
5.1 Global Strategy of World Health Organization
The World Health Organization published a strategic plan called "Global strategy on digital health 2020-2025" in 2020. This strategy is summarized below:
The vision of this strategy is to accelerate the development and adoption of appropriate, accessible, affordable, scalable, and sustainable person-centered digital health solutions to detect, prevent, and respond to outbreaks. The aim of this global strategy is to improve health systems by using digital health technology to empower patients and achieve the vision of health for all. Digital health will be adopted in the following situations is accessible and fair, affordable and improves the efficiency and sustainability of health systems, respects the privacy and security of patient information. The vision aims to further develop research and development, innovation, and collaboration across sectors. (WHO, 2020)
Health systems and services increasingly recognize that digital health can fundamentally change health outcomes if they are supported by adequate investment in governance,
institutional and workforce capacity to make the necessary changes in digital systems and data use education, planning and management. This strategy is designed for use by all member states. (WHO, 2020)
Digital health broadens the concept of e-Health. The Internet of Things also encompasses other uses of digital technologies for health, such as advanced computing, big data analytics, artificial intelligence, including machine learning, and robotics. (WHO, 2020)
The global digital strategy states that health data should be classified as sensitive personal data that requires a high standard of security. It emphasizes the need for a legal and
regulatory basis to protect the security of data. Furthermore, it highlights the importance of communicating effectively and maintaining transparency about data security strategies.
(WHO, 2020)
It aims to establish a common understanding among all Member States and to establish an approach towards creating an interoperable digital health ecosystem. (WHO, 2020) 4 guiding principles have been defined. These:
“
I. Acknowledge that institutionalization of digital health in the national health system requires a decision and commitment by countries
II. Recognize that successful digital health initiatives require an integrated strategy
III. Promote the appropriate use of digital technologies for health
IV. Recognize the urgent need to address the major impediments faced by least- developed countries implementing digital health technologies” (WHO, 2020)
Furthermore, 4 Strategic Objectives were determined. These:
“
I. Promote global collaboration and advance the transfer of knowledge to digital health
II. Advance the implementation of national digital health strategies
III. Strengthen governance for digital health at global, regional, and national levels
IV. Advocate people-centered health systems that are enabled by digital health”
(WHO, 2020)
The Strategy Action Plan has been created and summarized in the following figure.
Figure 5: Summary implementation of the action plan
Source: WHO, Global Strategy on Digital Health 2020-2025
5.2 Health Strategy of Turkey
The initiative of digitalization of health services in Turkey is based on the 1990s. In the 2000s, it started to be actively implemented. Turkey has started to implement the Health Transformation Program in 2003. In this section, advanced analytical methods, artificial intelligence, and information technology strategies and implementations will be considered.
5.2.1 Health Transformation Program
In the evaluation report Türkiye Sağlıkta Dönüşüm Programı: Değerlendirme Raporu 2003- 2011 published in 2012 there is no information about advanced analytical methods and artificial intelligence. However, information about data analysis and decision support systems is included. The systems where data is stored and used are defined.
Health Transformation Program, while reaching the World Health Organization's "21. It also takes into account the Health for Everyone in the 21st Century" policy, the "Accession Partnership Document" announced by the European Union, and other international
experiences. Human is at the center of the Health Transformation Program. It is essential to protect the health of the person together with society. For this reason, “accessible, qualified, and sustainable health services for everyone” is the main idea of this program. (Akdağ, 2012) The scopes that may concern digital transformation in the 2012 Turkey Transformation of Health Program Evaluation Report are summarized below:
• The system called MHRS (Central Physician Appointment System) started to become widespread in 2011. The rate of being examined with MHRS by appointment is 24%.
• Home Health services have been mentioned, but this is not within the scope of technology. The services provided by the health teams going to the patient's home were discussed.
• "Drug Tracking System" describes an infrastructure established for the monitoring of each unit of medicine in Turkey. Drug Tracking System is the application of the structure defined as "Mark and Follow" in the literature on drugs. While the products are marked with the data matrix that enables the singularization of the products; It is also ensured that the product is tracked by notifications to the central database from every point it passes.
• National Health Information System (USBS), one of the main components of the Health Transformation Program, is one of the important stages of the reforms put into practice. With this program, national standards in health information systems were developed and thus an effective information system infrastructure was established.
This system has an infrastructure where citizens can access health information
individually. The records cover all stages of life starting from the prenatal period. The system also enables medical image transfer. On the other hand, it is possible to record other data such as manpower and financial data of institutions and organizations providing health services with this system.
• The main objectives of electronic health projects are defined in the report as follows:
“Ensuring health data standardization”
“Data analysis support and establishment of decision support systems”
“Accelerating data flow among e-Health stakeholders”
“Creating electronic personal health records”
“Saving resources and increasing efficiency”
“Supporting scientific studies”
“Accelerating the national adoption of the e-Health concept”
• Sağlık-NET is an integrated, secure, fast, expandable information and communication platform that collects the data produced in the electronic environment in health institutions and organizations in accordance with the standards. It collects and processes sufficient data in determining the problems and priorities in the health sector, taking measures, planning sector resources and investments, and evaluating the quality of health services provided.
• Health Coding Reference Server (SKRS) is a reference and sharing system that brings together health information system standards and coding systems, shares them with open technologies (XML web services), and easily updates them.
• Telemedicine is the provision of healthcare services remotely. Telemedicine: It is the provision of services such as radiology, pathology, biochemistry and
electrocardiography (ECG) by specialized healthcare institutions through information and communication technologies.
• Electronic Health Records is a secure information system where the health data of citizens are kept in health institutions and can be shared with certain criteria.
Electronic Health Records covers a wide range from demographic information to examination, from chronic diseases to immunization. (Akdağ, 2012)
When we look at this report published in 2012, we see that steps towards digitization have been taken. However, no information is given about the use of advanced analytical methods in diagnosis and treatment. Nevertheless, data digitization, the mention of the data analysis and decision support systems, one indication that conscious about the future of health care.
5.2.2 Digital Hospital
The digital hospital is a hospital model in which information systems in the hospital work in an integrated manner, all kinds of medical devices can send data to the information
management system through networks and sensors, and employees and patients can access information if they have their authorization.
In the 2013-2017 Strategic Plan of the Ministry of Health, the goal of "creating and spreading the concept of the digital hospital in facilities belonging to the Ministry and its affiliates" is included. (T.C Ministry of Health, 2012) The Ministry of Health decided to carry out this project in cooperation and consultancy with HIMSS (Healthcare Information and
Management Systems Society).
The digital hospital project first started in 2012 in pilot hospitals. Later, it was started to be applied in other hospitals. There are 7 levels in the digital hospital model. In Turkey, there are currently 4 level-7 hospitals. 179 hospitals are currently at level 6. In 2017, according to Ministry of Health data, there are 1518 hospitals in Turkey. (T.C Ministry of Health, 2020)
The concept of the digital hospital is actually the digitization of data and information. It is a concept in which systems are integrated with each other and exchange information. Includes telemedicine and mobile medicine applications. When we look at it actually, it is a big step for the future. Digitization of health data and radiological images will facilitate the
integration of advanced analytical methods and artificial intelligence.
5.2.3 Strategic Plan of 2019-2023
SWOT analysis of Turkey's Health System is made in the Strategic Plan of 2019-2023. In this SWOT analysis, the inefficient use of information technologies is considered a weakness.
Providing more effective health services with the changes in information system technologies is considered as an opportunity. The strong technology infrastructure has been one of the strengths.
There are objectives in the strategic plan. Information technologies are included in the
"Objective 5: Increasing the satisfaction of citizens and healthcare workers and ensuring the sustainability of the health system".
Information technologies are mentioned as follows:
• In recent years, a rapid digitalization has been experienced in the processes within the scope of health services. As a result, the central responsibility of health information systems for implementing the strategies of the Ministry of Health is increasing. The basic requirements for the use of health information systems within the Ministry are divided into two groups: I) Innovative electronic health applications in order to increase the efficiency of health service delivery to citizens; II) Technology and data support to make health management processes more efficient.
• These requirements include the creation of new functions in the software architecture, as well as the consistent collection of larger amounts of data from more points, the further duplication of integrations that enable data flows between systems, and the creation of new functions and competencies for data processing and use. It is aimed to accelerate the studies on health information systems.
• Health information systems new era strategy: The information systems service model consists of objectives in the areas of hardware infrastructure and management, the development of innovative e-health applications and the development of health management systems technologies.
• Improvements will be made in the information technology service model by focusing on the areas of information systems management, internal processes, external
stakeholder management, and data-software governance mechanism.
• It is planned to develop innovative electronic health applications in order to ensure sustainability in health service delivery and to provide citizen-based health services.
(T.C Ministry of Health, 2018)
Goals that ensure the implementation of every objective in the strategic plan have been determined. The 4th goal of the objective 5 is: To strengthen the health system by increasing the use of information technologies in health service delivery and decision-making processes.
Performance indicators have been determined for the goals. The table was created, year by year goals were determined. Risks, strategies, needs, findings have been determined and cost estimation has been made. This table is given below:
Objective A5: To increase the satisfaction of citizens and healthcare workers and to ensure the sustainability of the health system
Goal H5.4: Strengthening the health system by increasing the use of information technologies in health service delivery and decision- making processes
Performance Indicators
Effect on Goal(%)
Plan Period Start Value
2019 2020 2021 2022 2023 Frequency of
Monitoring
Reporting Frequency
PG5.4.1:
e-Nabız Personal Health System Profiles
10 9,4166 (2018)
14 15 16 18 20 6 months 6 months
(Million) (Cumulative) PG5.4.2:
The ratio of Indicator Group Requested from
International Databases and Meeting International Standards from Ministry of Health Electronic Data Systems (%)
10 70167
(2018)
75 80 82 85 86 1 year 1 year
PG5.4.3:
Annual Service Continuity Rate in e-Nabız Excluding Planned Interruptions (%)
10 98,2168 (2018)
98,2 98,5 98,7 99 99 1 year 1 year
PG5.4.4:
The Number of Training Activities Conducted to Strengthen the Organizational Bond
10 1169
(2018)
2 2 3 3 3 1 year 1 year
PG5.4.5:
Compliance of Hospitals with Information Systems Requirements and Number of Data
Quality Audits
10 5170
(2018)
30 40 50 60 70 1 year 1 year
PG5.4.6:
Prepared in Electronic Environment
10 5171
(2018)
6 3 1 1 1 1 year 1 year
Number of Reports Ready to Share with Relevant Ministries (According to Type)
PG5.4.7:
Number of Radiological Examination Types Analyzed Using Artificial Intelligence Technology
10 0172
(2018)
0 1 1 1 1 1 year 1 year
PG5.4.8:
Completion Rate of the Inspection Process in One of the First Three Branches Recommended by the "Neyim Var Sistemi"
(%)
10 0173
(2018)
0 70 75 85 90 1 year 1 year
PG5.4.9:
The Ratio of Deaths Registered Through the Death Notification System (ÖBS) to the Total Number of Deaths in Healthcare Facilities (%)
10 0174
(2018)
5 50 80 90 95 1 year 1 year
PG5.4.10:
Monthly Average Number of Births Registered Through the e- Report / e-
10 0175
(2018)
500 1000 3000 5000 10000 1 year 1 year
Delivery System Responsible Department
General Directorate of Health Information Systems Departments
to Cooperate
General Directorate of Public Hospitals, General Directorate of Public Health, General Directorate of Health Services
Risks • The health authority's policy change in the standard used in the digitalization process
• Finding the possibility of a cyber attack
• Health facilities do not send the data they produce to the Ministry or adversely affect decision-making processes due to failure to comply with standards
Strategies 5.4.1 Quality, availability, availability, and security of health data and software will be improved through stronger governance
5.4.2 A metadata dictionary (data identity) will be created for the indicators requested by both international databases and included in the health statistics yearbook in accordance with the determined standards 5.4.3 The quality of statistical data prepared in accordance with international standards will be evaluated, collected, and shared in a common database.
5.4.4 Relevant external stakeholder management processes will be
organized in order to strengthen the health information systems ecosystem in our country.
5.4.5 Health information systems' internal processes will be regulated on the basis of efficiency, security, and agility.
5.4.6 Health information technologies, hardware infrastructure and hardware management processes will be developed in order to strengthen sustainability and service continuity.
5.4.7 Ministry management processes will be improved and strengthened by using information technology infrastructure.
5.4.8 Coordination and integration structure regarding e-Government services offered / to be provided to citizens, private sector, and public will be strengthened.
Cost Estimation (TL)
528.319.000 TL
Findings • Difficulties in bureaucratic procedures within the scope of other services provided by citizens related to health services
• Increase in the workload of physicians as a result of citizens' being directed to the wrong polyclinics during their application to health facilities or when making appointments.
• The decision support system mechanism established for the decision- making processes of the central and provincial organization managers of the Ministry has not yet become widespread.
• Loss of time and workload and cost increase due to duplication of examinations performed in health facilities
• Difficulties in the integration of the data structure in accordance with international standards to the field and this increases the workload of the physician.
• The need to increase communication and cooperation between Central and Provincial units
Needs • Making data entries in the field from all health institutions in accordance with international definitions, in an accurate, high quality, and
comprehensive manner, ensuring sustainability, regularly checking data entries of units, and informing in the field
• Reviewing all data dictionaries, creating a data dictionary in accordance with international standards
• Removal of all data collection processes (Basic Health Statistics Module, excel, paper form) except e-Nabız
• Gathering the software development teams in all institutions related to the Ministry under SBSGM and strengthening the organizational ties of central and provincial informatics employees
• Meeting the budget requests for the hardware and software needs of healthcare facilities in the digitalization process if approved by the Ministry.
• Raising awareness of indicators requested by international institutions, which are under the responsibility of the Ministry of Health and Affiliated Organizations.
Table 1: Performance Indicators
Source: T.C Ministry of Health 2019-2023 Stratejik Planı
Judging by the success of Turkey's 2013-2017 strategic plan: 117 performance indicators were located in the strategic plan. The rate of reaching the target was announced as 25%.
The number of examinations planned to use artificial intelligence is 1 per year. 4 radiological examinations in 2023 are targeted. The target for 2019 was 0. In other words, the first
radiological examination was planned for 2020. This situation shows us that when this strategic plan was prepared, no action was taken on this issue before. However, there is nothing other than advanced analytics and artificial intelligence implementation other than this goal. The country does not have a separate target or policy in this regard.
However, steps such as collecting data in a single system have been taken. This is a big step for the future because advanced analytical methods and artificial intelligence help us in the diagnosis and treatment process using data. The healthier data you have, the healthier the result will be. The goal of transferring all data in hospitals to electronic media and the goal of a completely digital hospital is an important point in the implementation of this process.
