Akademik Veri Yönetim Sistemi
İzmir Sosyal Bilimler Dergisi, cilt.6, sa.2, ss.90-115, 2024 (Hakemli Dergi)
Sağlık Sektöründe Karar Destek Araçları: İş Zekâsı, Makine Öğrenmesi, Derin Öğrenme ve Yapay Zeka Uygulamaları
Bilgi ve iletişim teknolojileri tüm sektörleri olduğu gibi sağlık sektörünü de dönüştürmekte ve şekillendirmektedir. Bu muazzam dönüşüm içinde her geçen gün sağlık sektörü yönetim süreçlerinden günlük operasyonel süreçlerine kadar bilgi ve iletişim teknolojilerinden faydalanmakta ve karar süreçlerinde teknolojinin imkanlarından faydalanmaktadır. Çalışmamız kapsamında son yıllarda sağlık sektöründe önemi gittikçe artan iki farklı teknolojik gelişmeyi karar destek aracı olarak kapsamlı bir şekilde değerlendirmekteyiz. Yapay zeka ve iş zekası teknolojileri merkeze alınarak bu iki önemli kavramın kavramsal boyutları, sağlık sektörü için oluşturduğu değer kapsamlı bir şekilde değerlendirilmektedir. Yapay zeka içerisinde, makine öğrenmesi ve derin öğrenme gibi iki kritik kavram da değerlendirilmektedir. Makine öğrenmesi, yapay zeka, derin öğrenme ve iş zekası konuları pek çok farklı çalışmada farklı başlıklarda değerlendirmiştir. Fakat literatürde ilgili teknolojileri toplu olarak kapsamlı bir şekilde değerlendiren bir çalışmaya rastlanmamıştır. Aynı zamanda ilgili konu başlıklarının sağlık bilimleri alanında tartışıldığı bir çalışmaya da rastlanmamıştır. Çalışmamız bu boşluğu gidermeyi hedeflemektedir. Özellikle son yıllarda pek çok ülkenin yapay zeka konusunda önemli yatırımlar yaptığı günümüz koşullarında Türkiye’de bir ekonomik çıktı olarak yapay zeka uygulamaları konusunda ne tür kazanımlar elde edebileceğimizi konu kapsamında değerlendirilmektedir. Geleceğe dönük sağlık politikaları için kural koyucular ve politika yürütücüleri için çözüm önerileri ve örnek uygulama önerileri ortaya konmaktadır.
Decision Support Tools in the Health Sector: Business Intelligence, Machine Learning, Deep Learning, and Artificial Intelligence Applications
Information and communication technologies (ICT) are transforming and shaping the healthcare sector, as they are in all industries. In this immense transformation, the healthcare sector is increasingly utilizing ICT in management processes, daily operational procedures, and decision-making processes. This study comprehensively evaluates two significant technological advancements, which have gained increasing importance in recent years within the healthcare sector, as decision support tools. Artificial intelligence (AI) and business intelligence (BI) technologies are at the center of this evaluation, focusing on their conceptual dimensions and the value they create for the healthcare sector. Within AI, two critical concepts, machine learning and deep learning, are also discussed. Machine learning, AI, deep learning, and business intelligence have been addressed in numerous studies under various topics. However, there has been no study in the literature that comprehensively evaluates these technologies collectively. Additionally, no research has been found that discusses these topics specifically within the field of health sciences. This study aims to fill this gap. In light of the significant investments many countries have made in AI in recent years, this study also explores the potential economic benefits Turkey could achieve through AI applications. It presents solutions and example applications for policymakers and policy implementers regarding future healthcare policies.
EXTENDED ABSTRACT
Current technological advancements, innovations, and the rapid developments in artificial intelligence provide tremendous opportunities for information and communication technologies to simulate human cognitive activities. This situation deeply influences the healthcare sector, the methods of delivering healthcare services, and developments in many areas related to health sciences. In their study, Boddu et al. (2022) emphasized that machine learning, artificial intelligence, and automation processes have created rapid changes in the pharmaceutical industry and noted that the use of different analytical techniques has helped reduce global mortality rates, and the use of artificial intelligence tools has aided in identifying critical diseases. As seen, integrating artificial intelligence into healthcare has excellent potential to improve disease diagnosis, treatment selection, and clinical laboratory tests. Our study is a review-type research. Artificial intelligence is a powerful and groundbreaking field of computer science with the potential to fundamentally change medical practices and the delivery of healthcare services (Bajwa et al., 2021). Artificial intelligence/machine learning developments present significant potential to improve healthcare services (Nallamothu & Cuthrell, 2023). The potential use of artificial intelligence and business intelligence in the healthcare sector looks promising in the future.
Conceptually, the Decision-Making Process
Decision-making is not only one of the most common types of problems encountered in our daily and professional lives, but it also represents the fundamental processes in solving more complex and poorly structured problems. Decisionmaking is a critical component in more complex problems such as diagnosis, negotiation, design, situational assessment, and command and control (Means et al., 1993). Essentially, decision-making involves selecting one or more beneficial or satisfactory options from a broader set of alternatives.
Conceptually, Machine Learning and Deep Learning
Typically, machine learning refers to a system that identifies data patterns from input and trains a prediction model, then uses such a model to make useful predictions from new, previously unseen data. Machine learning is widely used in other artificial intelligence technologies such as natural language processing, speech technology, and robotics (Chen & Decary, 2020). Machine learning refers to basic algorithms applied using numerous similarly structured tasks to identify patterns. Deep learning, on the other hand, is a set of approaches used in machine learning that employs feature/representation learning techniques without specific algorithms for particular tasks (Goodfellow et al., 2016).
Conceptually, Artificial Intelligence
Artificial intelligence, through data-driven analysis, attempts to reflect human cognition, and it may also reflect the biases present in our collective conscience (Thomasian et al., 2021). Artificial intelligence minimizes human errors while providing enhanced accuracy, reduced costs, and time savings (Alowais et al., 2023). Artificial intelligence typically refers to computational models that automate tasks performed by humans, and this umbrella term includes machine learning algorithms (Esteva et al., 2019).
Conceptually, Business Intelligence
Technology Business intelligence systems can be expressed as a significant part of the portfolios of many organizations within information and communication technologies. Although the evolutionary nature of other decision support technologies has been stated, there is limited research examining the evolutionary nature of business intelligence systems (Safwan et al., 2016). This is particularly evident when the Turkish literature is reviewed. The standard of a business intelligence system integrates data from an organization’s internal information systems and combines it with data from specific environments, such as statistics, financial and investment portals, and various databases. These systems are designed to provide adequate and reliable up-to-date information about different aspects of organizational activities (Olszak & Ziemba, 2007). This is part of the decision-making process.
The Difference Between Artificial Intelligence and Traditional Statistics
Statistics has been the standard method for medical research to demonstrate the benefits of new treatments, predict prognoses, identify risk factors, and reveal disease mechanisms. Interestingly, there are significant overlaps in techniques and methodologies between traditional statistics and artificial intelligence. The fundamental difference can be traced to their philosophies. Statistics is a science that predicts and explains data, while artificial intelligence or machine learning aims to derive practical predictions from available data. Statistical analysis typically provides quantitative results, such as probability, confidence intervals, test results (p-value), or the accuracy of a particular model. In contrast, artificial intelligence generally produces more practical results, such as predictions, classifications, recommendations, or forecasts.
The Importance of Decision Support Tools within the Healthcare System
AI-powered decision support systems assist healthcare providers by offering real-time suggestions to aid diagnostic and treatment decisions. In emergency departments, faster clinical data interpretation is crucial for classifying the severity of the condition and the need for urgent intervention. The risk of misdiagnosing patients is one of the most critical issues affecting medical practitioners and healthcare systems (Alowais et al., 2023). On the other hand, many services in the healthcare sector are provided simultaneously. In such cases, healthcare managers may need to analyze data and make decisions through a single information system or multiple information systems, depending on the situation. Here, business intelligence technology emerges as a vital strategic tool for healthcare sector managers. Healthcare institution managers need effective communication and interaction with various data sources between doctors, patients, and administration in decision-making processes, and business intelligence can assist as a decision support tool.
Applications of Artificial Intelligence, Machine Learning, Deep Learning, and Business Intelligence in the Healthcare System
Artificial intelligence is rapidly becoming the cornerstone of modern healthcare services by offering unprecedented capabilities in diagnosis, treatment planning, patient care, and healthcare management (Alzamily et al., 2024). Healthcare services are rapidly expanding to include remote and mobile delivery modes, making it timely and crucial to integrate artificial intelligence technologies to assist in diagnosis, treatment, and prevention. AI is primarily used to assist and automate existing healthcare services, particularly in diagnostic and pharmaceutical fields (Palaniappan et al., 2024). Another example of AI’s effective use in preventive care includes healthcare services that incorporate AI for personalized nutrition aimed at managing chronic diseases long before they emerge. Additionally, patients, particularly those with chronic conditions such as diabetes, hypertension, and mental health issues, may be able to manage their medical conditions, especially chronic illnesses, with the assistance of AI (Palaniappan et al., 2024). As seen, many applications, supported by artificial intelligence, are finding their place within the healthcare sector. Furthermore, business intelligence is recognized as a popular tool in business management and decision support systems. Business intelligence helps transform raw data into intelligent information. There are various business intelligence tools such as extraction, transformation, and loading, data warehousing, online analytical processing (OLAP), and dashboards. Business intelligence tools are typically used in public health sectors for financial and administrative purposes (Jinpon et al., 2011).
Applications of Artificial Intelligence, Machine Learning, Deep Learning, and Business Intelligence in Clinical Processes
AI is primarily used in diagnostics to read images and support clinicians in the decision-making process (Palaniappan et al., 2024). Artificial intelligence is rapidly developing in healthcare due to its potential to unlock the power of big data, support evidence-based clinical decision-making, and gain insights for value-based care (Chen & Decary, 2020). Developing intelligent decision support systems for medical fields to improve clinical activities is not a new area of research. In recent decades, many studies have identified different challenges in developing intelligent solutions for various purposes (Miah, 2018). Additionally, AI in the healthcare sector has become a transformative force that offers exciting opportunities to improve clinical conditions and patient care (Chauhan & Degan, 2024). AI also has the potential to assist medical professionals in diagnosing and treating patients (Dave & Patel, 2023). Moreover, business intelligence stands out as another key technology, especially for managing administrative processes. It can act as part of business analytics and can be used to detect healthcare fraud, including false billing and prescription drug abuse (Ramalingam et al., 2024). Business intelligence can also be used, especially for middle and upper-level managers, to coordinate data across different systems and activate decision-making processes. With data-centered management, institution managers can lead healthcare organizations more effectively.
Applications of Artificial Intelligence, Machine Learning, Deep Learning, and Business Intelligence in Pharmaceutical Processes
Technological innovations or the implementation of artificial intelligence, machine learning, and automation processes have led to significant changes in the pharmaceutical sector (Boddu et al., 2022). Artificial intelligence technology is useful for healthcare support and pharmaceutical assistance. An initiative named Molly, which features a pleasant voice and a friendly face, aims to help people monitor their conditions and treatments. It supports patients with chronic conditions (Manikiran & Prasanthi, 2019). In drug research, AI is used for drug discovery and the prediction of chemical and pharmaceutical properties (Lu et al., 2017). For example, the drug synthesis process within the research and development cycle can be shortened by using machine learning models to automate chemical experiments (Ahneman et al., 2018). As seen, the pharmaceutical sector is one of the areas in the healthcare industry where artificial intelligence and machine learning are extensively applied. AI-based solutions can be developed to address many problems in the pharmaceutical industry.
Applications of Artificial Intelligence, Machine Learning, Deep Learning, and Business Intelligence in Medical Education and Research Processes
The integration of artificial intelligence into dental education has shown promising results in enhancing the learning experience and improving patient care. Dental students can benefit from virtual simulations, allowing them to perform complex procedures such as fillings and root canals without putting real patients at risk (Dave & Patel, 2023). AI technologies can personalize learning experiences, process large amounts of data to access the latest learning resources, and provide interactive training through simulations (Masters, 2019). AI algorithms can analyze student data to provide personalized learning experiences, assist in grading assignments, and create intelligent tutoring systems (Dave & Patel, 2023). In addition to this, business intelligence technology emerges as another key concept for organizing research data and analyzing the scientific productivity of institutions, as well as managing educational processes. Through this technology, the scientific publication and publishing activities of higher education institutions can be assessed on a national scale (Damar et al., 2023), research collaborations between countries can be evaluated (Damar et al., 2022), scientific collaboration between two different countries can be examined (Damar, 2022a), university business processes can be coordinated (Gökşen et al., 2016), and it can be used as a data analysis and visualization tool in scientific research (Celik et al., 2023).
Ethical and Legal Discussions Regarding Artificial Intelligence, Machine Learning, Deep Learning, and Business Intelligence Technologies in the Healthcare Sector
Ka & Khokhlov (2024) stated that the integration of AI into healthcare offers unprecedented opportunities to improve patient care and outcomes. Redrup et al. (2023) highlighted, as an example, automation bias, where AI trained on data that does not represent minority ethnic groups can lead to discrimination and inequality. Rosemann and Zhang (2022) identified various challenges arising from the integration of AI and big data analytics into medical and healthcare settings. Additionally, as business intelligence involves aggregating data from many different systems, it is important to consider the interaction with various legal regulations and legislation related to data protection, privacy, and intellectual property rights.
Discussions on the Sustainability of Business Intelligence and Artificial Intelligence
While artificial intelligence has the potential to revolutionize clinical applications, there are challenges that must be addressed before it can reach its full potential. Among these challenges is the lack of high-quality medical data, which can lead to incorrect results. Data privacy, usability, and security are also potential limitations to the implementation of AI in clinical practice (Alowais et al., 2021). Healthcare organizations need to overcome several challenges to ensure the success of AI (Chen & Decary, 2020). Furthermore, business intelligence technologies generally feed on data from various systems to meet management needs. Therefore, especially when working with large datasets or big data, it is critical to ensure that the interactions within systems are thoroughly reviewed and adjustments within subsystems are made accordingly.
Conclusion and Recommendations
This study provides a comprehensive evaluation of business intelligence and artificial intelligence, with a focus on the healthcare sector. The medical field has become too complex and interrelated with multiple disciplines to be left solely to healthcare professionals. Therefore, as mentioned earlier, fields such as artificial intelligence, data management and analytics, robotics, and autonomous systems will continue to gain prominence in the healthcare sector.