Edge-of-Things in Personalized Healthcare Support Systems

By Michael Sheng

Edge-of-Things in Personalized Healthcare Support Systems discusses and explores state-of-the-art technology developments in storage and sharing of personal healthcare records in a secure manner that is globally distributed to incorporate best healthcare practices. The book presents research into the identification of specialization and expertise among healthcare professionals, the sharing of records over the cloud, access controls and rights of shared documents, document privacy, as well as edge computing techniques which help to identify causes and develop treatments for human disease. The book aims to advance personal healthcare, medical diagnosis, and treatment by applying IoT, cloud, and edge computing technologies in association with effective data analytics.

• Provides an in-depth analysis of how to model and design applications for state-of-the-art healthcare systems
• Discusses and explores the social impact of the intertwined use of emerging IT technologies for healthcare
• Covers system design and software building principles for healthcare using IoT, cloud, and edge computing technologies with the support of effective and efficient data analytics strategies
• Explores the latest algorithms using machine and deep learning in the areas of cloud, edge computing, IoT, and healthcare analytics

• Language: English
• Publisher: Academic Press
• Release date: Jun 19, 2022
• ISBN: 9780323907088

Book preview

Chapter 1

Exploring the dichotomy on opportunities and challenges of smart technologies in healthcare systems

S. Prabavathy and I. Ravi Prakash Reddy,    Department of Information Technology, G. Narayanamma Institute of Technology and Science (For Women), Hyderabad, India

Abstract

In the information era, healthcare systems use the advancement of digital revolution. The need for digitization in healthcare system are leveraging medical data, improving patient-care service and reducing the cost of medical service. The key revolutionizing technologies in healthcare are: Internet of Things (IoT), Artificial Intelligence, and Cloud computing. These technologies not only provide digitization but also made multidimensional change in the healthcare systems. It changes the medical model, that is, from disease-centered care to patient-centered and changes in treatment concept, that is, the focus from disease treatment to focus towards the preventive healthcare. In addition, there was also change in information collection and management including clinical data to regional medical data and general management to personalized management. These multilevel changes improved the efficiency of healthcare system and supports people to lead a healthier life.

Keywords

Internet of Things; cloud computing; edge computing; artificial intelligence; healthcare system

1.1 Introduction

In the information era, healthcare systems use the advancement of digital revolution. The need for digitization in healthcare system are leveraging medical data, improving patient-care service and reducing the cost of medical service (Velthoven et al., 2019). The key revolutionizing technologies in healthcare are: Internet of Things (IoT), Artificial Intelligence, and Cloud computing. These technologies not only provide digitization but also made multidimensional change in the healthcare systems. It changes the medical model, that is, from disease-centered care to patient-centered and changes in treatment concept, that is, the focus from disease treatment to focus towards the preventive healthcare (Jayaratne et al., 2019). In addition, there was also change in information collection and management including clinical data to regional medical data and general management to personalized management (Hassanalieragh et al., 2015). These multilevel changes improved the efficiency of healthcare system and supports people to lead a healthier life.

The vision of IoT is to deploy intelligent sensor and using Internet Protocol for communication to provide pervasive and ubiquitous environment and experience (Prabavathy et al., 2018a,b). IoT has a potential impact in healthcare system. In smart healthcare, it improves the convenience of physicians and patients by using real-time monitoring with efficient information processing. It enables an effective treatment for the patients using data collected through remote monitoring of IoT devices used by patients (Ungurean & Brezulianu, 2017). A variety of body sensor devices have been used for monitoring patients, providing physical activity consciousness and maintaining individual fitness.

The real-time data from the IoT medical devices need to be stored, managed, and transported with security and privacy. This need is satisfied by cloud computing which provides computing services and analytics over the internet (Pattnaik et al., 2017). Integrating IoT and cloud computing in healthcare allows accessing of shared medical data in common infrastructure transparently by providing on-demand services through the network. Healthcare system needs availability of critical information and processing of real-time data with low latency. Edge computing addresses this need by storing and processing critical data near to the devices where it is collected. Edge computing is highly valuable in healthcare where data is needed immediately and delay for processing critical data on cloud is not efficient. The large amount of data collected from the IoT medical sensors, which are stored on the cloud, need to be analyzed for providing insights that is equivalent to human level of accuracy. This is possible by integrating artificial intelligence (AI) along with IoT and cloud computing in healthcare systems. The goal of AI is to mimic the cognitive functions of human beings (Martınez-Miranda & Aldea, 2005). AI is receiving greater interest in healthcare system because of its capability in analyzing large volume of medical data. The massive progress in AI techniques such as deep learning, robotics, computer vision, and natural language processing has brought a paradigm shift in the medical field (Bartoletti, 2019). Fig. 1.1 describes the integration of smart technologies in healthcare sector.

Figure 1.1 Smart technologies integrated healthcare.

This chapter provides recent breakthrough in integrating IoT, AI, and cloud computing in healthcare systems along with applications and challenges for further progress in healthcare systems. The main objectives of this chapter are

• motivation for integration of IoT, AI Cloud Computing, and Edge computing with healthcare systems

• challenges in integration of IoT, AI Cloud Computing, and Edge computing in healthcare applications

• future research directions in healthcare systems.

The remainder of this chapter is organized as follows—Section 1.2 presents the opportunities and risk in integrating IoT in healthcare systems. In section 1.3 the applications and challenges in integrating Cloud computing in healthcare system are elaborated. Section 1.4 provides the opportunities and challenges in integrating artificial intelligence in healthcare system. Section 1.5 presents the challenges and opportunities in integrating edge computing in healthcare system. Section 1.6 presents the forthcoming research technologies and future trends in healthcare system. Finally, we conclude the chapter in Section 1.7.

1.2 Internet of things in healthcare system

IoT is an emerging field of research, and its use in healthcare area is still in its early stage (Shah et al., 2019). In this section, IoT technology is studied and its applicability for healthcare system is analyzed. Several researches towards building healthcare IoT systems are reviewed to analyze the opportunities and risks in involving IoT in healthcare systems.

1.2.1 Internet of things

IoT is a virtual network that interacts with real world, using wireless sensor network and Internet as its core technology (Prabavathy et al., 2018a, 2018b). The goal of IoT is to use intelligent sensors and actuators for enabling pervasive environment with ubiquitous experience. It allows automation in a large range of industries with collection and processing voluminous data. Fig. 1.2 shows the layered architecture of IoT. Recently, IoT technology is penetrating into commercial application such as smart home (Stojkoska & Trivodaliev, 2017), agriculture (Ruan et al., 2019), smart parking (Khanna & Anand, 2016), smart grid (Al-Turjman & Abujubbeh, 2019), etc. Healthcare system is different from the aforementioned fields because the outcome of IoT-based healthcare system should be plausible and reliable. IoT technology in other commercial areas have proven that real-time remote monitoring with detection, collection, and reporting of data are achievable (Butler et al., 2019; Saravanan et al., 2018; Yu et al., 2005). This capability of IoT can be expanded and used in healthcare systems for monitoring patient health and reporting it to healthcare related people such as caretakers, doctors, or emergency services.

Figure 1.2 Internet of things layered architecture.

1.2.2 Opportunities of internet of things in healthcare

IoT provides a promising solution to digitize healthcare systems, and many researches are being performed in IoT healthcare system (Farahani et al., 2020; Negash et al., 2018; Strielkina et al., 2017). Healthcare is a vast ecosystem hence it facilitates numerous IoT applications in patient monitoring (Archip et al., 2016), smart pills (Goffredo et al., 2015), robots in disease treatment (Cianchetti & Menciassi, 2017), wearable sensors (Rodgers et al., 2014) etc. The data collected by the IoT devices in healthcare system provide detailed contextual analysis and clear insights about the disease and patients’ health. IoT not only helps people lead healthier life but also supports doctors to improve treatment and researches.

Remote patient monitoring is the most important application for healthcare system because it enables real-time personalized healthcare. It also helps them reduce cost and time (Kumar, 2017). The patients can be monitored using IoT wearable medical devices such as heart rate monitoring cuffs, sleep analyzer, blood pressure measuring bands, fitness bands, glucometer, etc., which reduces the burden of medical professionals (Khan et al., 2016). It is very much suitable for tracking the health conditions of elderly patients to detect the risk in advance.

The data from the IoT medical devices has many applications. Fig. 1.3 gives the benefits of IoT in healthcare. The doctors can use patient data from smart medical devices for faster treatment. The doctor can alter the dosage of medicines accurately using those data (Laranjo et al., 2012). It also helps in analyzing risk and benefit of particular drug. The data is available anywhere and anytime in the hospital IT infrastructure helps the physicians to make reliable decisions and services. These data also helps government to predict the viral spread, track patient journey for medical research and resource allocation.

Figure 1.3 Healthcare with internet of things benefits.

The success of remote health monitoring have initiated more researches in integrating IoT in various healthcare applications. There are several works where IoT healthcare systems have been built for specific purposes, including rehabilitation (Celesti et al., 2020), diabetes management (Ara & Ara, 2017), and assisted ambient living for elderly persons (Abdelgawad et al., 2016). Though these systems are designed for many different purposes, they are each strongly related through their use of similar enabling technologies.

1.2.3 Risks of integrating internet of things in healthcare

IoT is incredible to revolutionize the healthcare systems with its technology. But there are some challenges in integrating it within healthcare systems. Two of the most important risks in integrating IoT are data security and privacy issue. IoT medical devices monitor and detect patients’ real-time data, which are vulnerable to multiple data security and privacy threats (Chacko & Hayajneh, 2018). As the IoT medical devices continuously detect and share personal data of patient, it is essential to ensure security and privacy. The impact of attack on these devices may lead to disaster or loss of life (Devi & Muthuselvi, 2016). Technologies for IoT healthcare are in the developing stage and are subjected to increased technical implications. Hence these new technologies of IoT healthcare have new challenges in ensuring security and privacy. The cybercriminals can compromise the patients’ IoT medical device and use that personal health data for fraudulent health insurance claims and sometimes for buying and selling medical drugs illegally using fake IDs.

The next major challenge is integration of multiple IoT devices and protocols in healthcare systems. Most of the IoT devices are in developing stage and manufacturers do not follow standardization in communication protocols and device development (Knickerbocker et al., 2018). Hence there can be interoperability issues when multiple devices are involved in healthcare application. The various disciplines captured by healthcare IoT devices are regulated by a diverse group of regulatory agencies (Huycke & All, 2000). This creates complexity in healthcare system wherein medical standards are subjected to strict regulations. The lack of interoperability among the heterogeneous device platforms and standards that exist for authentication is an important vulnerability that leads to data privacy threats, compliance regulation issues, and backward compatibility with legacy systems. The IoT medical devices must be compatible with many transmission formats and protocols for authentication and encryption.

IoT integrated healthcare system involves data management challenges. The real-time monitoring data from body sensors will be voluminous and it is in various data formats. The healthcare system should be developed using data-driven learning techniques to handle these voluminous and variety of data formats (Reda et al., 2018). In healthcare system, data are collected from various sources, therefore proper authentication is needed to assure that healthcare data are submitted from registered clinics, hospitals, and medical institutions. Overloading of data can affect the accuracy in decision-making during the treatment. IoT-based healthcare are not always easy to use by medical professionals and physicians involved in it. The vast number of features involved in it could sometimes make healthcare system complicated which in turn discourage healthcare workers in learning.

1.3 Cloud computing in healthcare system

Cloud computing is not a new term in healthcare systems. In the last few years, the impact of cloud computing technology is increasing at faster rate in healthcare industry. In this section, cloud computing technology and its influence on healthcare has been analyzed. Recent researches in adopting cloud technology in healthcare systems are reviewed to understand the opportunities and challenges in taking up cloud based healthcare systems.

1.3.1 Cloud computing

Cloud computing has modified the basic building blocks of computing and revolutionized the ability of computing in wide variety of applications. Cloud computing facilitates a shared pool of computing resources containing networks, servers, storage applications, and service (Velte et al., 2009). These computing resources can be configurable and used with minimal effort. Based on demand of the customers the computing resources are allocated and released from the shared resources. This guarantees efficient and optimized resource allocation. The users need not invest on IT infrastructure and they can pay based on their utilization, hence it is cost effective too. Cloud computing facilitates customers with multiple technologies such as virtualization (Lombardi & Di Pietro, 2011) and multitenancy (AlJahdali et al., 2014). Using the web applications, the shared computing resources are managed and customers use the shared physical resources through virtual environment.

Based on a deployment model, the cloud model is classified as public, private, hybrid, and community cloud (Savu, 2011). Public cloud holds vast amount of resources and has high scalability which can be accessed by multiple organizations. Private cloud is designated to a single organization with high security. Hybrid cloud is a combination of private and public cloud and allows seamless interaction between the platforms. Community cloud is a multitenant platform provided to multiple organizations for sharing the same applications. Fig. 1.4 describes cloud deployment models. Based on the service provided, cloud models are classified as Infrastructure-as-a-Service (IaaS), Platform-as-a-Service (PaaS), and Software-as-a-Service (SaaS) (Weinhardt et al., 2009). Fig. 1.5 shows the various Cloud Service Models. Further these models can be specific to Storage, Database, Information, Process, Application, Integration, Security, Management, and Testing-as-a-service

Figure 1.4 Cloud deployment models.

Figure 1.5 Cloud service models.

1.3.2 Opportunities of Cloud computing in healthcare

The generation, consumption, storage, and sharing of data in healthcare system has increased tremendously. The digitization of data in healthcare industry has made massive shift in the data management processes (Fabian et al., 2015). The adoption of cloud computing technology has revolutionized the data management in healthcare system. Fig. 1.6 gives the important features of cloud computing. The extensive cloud adoption in healthcare system have gained benefits not only in data storing but also in other areas such as workflow optimizations, cost benefits, and personalized care.

Figure 1.6 Features of cloud computing.

The cloud adoption in healthcare system provides significant cost benefits for patients as well as hospital management. High capital investments are avoided and operating cost alone is considered because the required computing resources are obtained on demand and released in cloud adopted healthcare system (Sultan, 2014). Since the computing and data management resources are maintained by cloud service providers, the related cost for IT staff resources in healthcare management is reduced.

Cloud services offers interoperability for data integration at every point of data management. The cloud adoption in healthcare provides access to patient data from multiple sources and share among the healthcare stakeholders to provide efficient treatment. The insights gained from multiple sources of patient’s data facilitates efficient planning and healthcare service. The geographic limitations are avoided and provides seamless data transfer among various stakeholders such as pharmaceuticals, labs, and insurance companies, owing to interoperability of cloud service (Lupşe et al., 2012). Fig. 1.7 shows the cloud integrated healthcare system. Thus cloud service facilitates the data transformation from volume to value in healthcare.

Figure 1.7 Cloud integrated healthcare system.

Scalability of cloud services supports the healthcare system abundantly by providing on-demand resources scaling (Ahn et al., 2013). In recent times large number of medical sensors are used in healthcare industry. These sensors are sampled at high frequency which leads to vast amount of data. Cloud services supports heterogeneous data of varying size by providing high storage and computing capabilities. Cloud adoption in healthcare system provides remote accessibility which aids numerous healthcare functionalities such as high personal care schemes, telemedicine, and posthospitalization monitoring schemes. Telehealth application involving cloud computing enables convenient sharing of data for patients during the treatment as well as posttreatment (Wang et al., 2017). Cloud services offer wide variety of capabilities for healthcare management staff to enable improved ways of working and provide new services to patients. Cloud adoption in healthcare supports faster upgrading of services with zero service interruption and minimal cost.

Cloud services can integrate with the emerging IoT devices which enables rapid development in the innovation of new pervasive and ubiquitous services in healthcare environment. Cloud technology is based on internet which implements standard protocols hence connects to healthcare applications without much technical complications. The medical mistakes can be avoided by using cloud services with cognitive capabilities such as intelligent business process management suites and case management frameworks. The latest analytical techniques in cloud service enable improved knowledge on disease diagnosis, treatment techniques, and patient care using the vast amount of data in the cloud (Iyengar et al., 2018).

1.3.3 Risks of integrating cloud computing in healthcare

Cloud adoption in healthcare is no doubt exciting, but it has various potential problems that need to be addressed. Security is a major concern in cloud based healthcare system because it stores and processes sensitive data held by healthcare systems (Khattak et al., 2015). The most common security problems are data breach and data loss, account hijacking, denial of service, and insecure interfaces. The patient data should be accessed with effective authorization else the attackers will compromise the system to access the patient data. There are possibilities for ransomware attack in which the attacker deliberately corrupts or deletes the patient data which may lead to data loss (Spence et al., 2018). Sometimes attacker may also perform identity theft to access the account of any stakeholders involved in the healthcare system. Adversaries seek to hijack the account of healthcare system stakeholders by stealing their security credentials and then eavesdropping on their activities and transactions (Flynn et al., 2020).

As the number of connected devices are growing in healthcare system, they are more prone to distributed denial of service attacks (DDoS) (Latif et al., 2014). During a DDoS attack, the attacker will flood healthcare IT systems by using a myriad of connections to overwhelm the system. The attacker use bots to generate numerous attacks from the connected devices, and it is hard to block. In healthcare, DDoS attackers can shut the cloud access for the stored data and services, which could prevent healthcare professionals from accessing critical patient information.

The security of cloud adopted healthcare mainly depends on the security of interfaces and application program interfaces (APIs) through which the stakeholders of healthcare system interact and manage the healthcare cloud infrastructure (AbuKhousa et al., 2012). The configuration of cloud infrastructure must be shipped with security else these configurations for API can be altered and the entire healthcare cloud infrastructure can be compromised by attackers.

The cloud providers offer services by combining them from subcontractors. Multiple subcontractors may be involved in cloud adopted healthcare system because of its complexity. Sometimes, service provided by these subcontractors will not be under the control of healthcare professionals hence there exist lack of influence and they alter the service during course of contract. This may lead to inefficiency in providing the health service.

The healthcare professionals do not have complete information about processing methods of cloud service providers (Al-Issa et al., 2019). This will affect the healthcare service during the maintenance phase. Sometimes there exist inadequacy of tools for managing and accessing cloud adopted healthcare data due to the dynamic and voluminous growth of healthcare data. Patients’ personal data privacy is very hard to maintain in cloud adopted healthcare system because the cloud service provider has the physical control over the data.

1.4 Edge computing in healthcare system

The critical information storing and processing is one of the major requirement of digital healthcare system which can achieved efficiently by edge computing. In this section, edge computing technology is reviewed. Several researches towards involving edge computing in healthcare systems are reviewed to analyze the opportunities and challenges in integrating edge computing in healthcare systems (Chen et al., 2018; Ray et al., 2019).

1.4.1 Edge computing

In edge computing data is processed closer to location where it was generated such as devices or sensors. It reduces the amount of data transfer between the devices to cloud. The network load is reduced by localizing data storage and processing (Shi et al., 2016). It also reduces the data processing delay since it is performed near to location where it was generated instead of transferring to cloud back and forth. The localized data storage in edge computing allows the data to be accessed even when network is offline. The data are stored closer, hence it reduces data transmission cost.

Edge computing model allows even complex image processing algorithms to run locally on edge system having increased processing power. The real-time applications involving quick processing and response such as autonomous vehicles, and augmented reality, can be implemented using edge computing. The edge devices include simple sensor, smart phone, security camera, laptop, and gateways. Fig. 1.8 provides the basic architecture of edge computing.

Figure 1.8 Edge computing architecture.

1.4.2 Opportunities of edge computing in healthcare system

Edge computing facilitates healthcare sector to identify and analyses data as it is produced from IoT devices for quick understanding and action. This moves the healthcare industry towards improved outcome for healthcare stakeholders. Edge computing allows to store and process critical information task on the edge devices which helps the healthcare industry to perform fast real-time medical data analysis and quickly respond to emergencies (Oueida et al., 2018). The noncritical medical data can be stored and processed in cloud environment.

Cloud computing provides services to store and compute data, still most of the healthcare stakeholders prefer to handle their data in on-premises data centers. Storing data in on-premises data centers provides more control over the data in terms of security and compliance. This also reduces the risk of downtime when compared cloud based systems (Shi et al., 2016). It allows autonomous control and security policies over the medical data. These security policies ensure that only authorized access of data and services are provided. Data is stored on-premises, hence the potential security breaches can be identified and prevented earlier, and thus it enables on-site security in high-risk environment.

Healthcare industry involves media-rich contents such as X-ray, and scan videos. There could be increase in delay, when these data are stored and accessed using conventional network. This issue is solved edge computing by storing and retrieving the data locally in on-premises data centers (Cha et al., 2018). Thus it provides fast access to critical media-rich data which supports on-time diagnosis and treatment. Fig. 1.9 shows the integration of Edge computing in healthcare sector.

Figure 1.9 Edge computing in healthcare.

Interconnectivity is a major problem in healthcare infrastructure. Healthcare systems was suffering with recordkeeping and incompatible medical equipment and systems. Edge computing along with IoT devices solved this problem by providing communications quickly across the devices and rapid access and processing of data (Hassan et al., 2018). In addition to it, healthcare industry faces other challenges such as limited availability of skilled staff, data management and administration problems, and complex computing system and network which becomes life threatening for patients. Edge computing handles these issues by providing support to staff, bringing structure to the unstructured data collected from the IoT devices for efficient data management.

1.4.3 Challenges in integrating edge computing in healthcare system

Edge computing will revolutionize the healthcare systems with its technology, but still there are some challenges in integrating it with healthcare systems. The major challenge in integrating edge computing with healthcare system is interoperability (Sigwele et al., 2018). Healthcare infrastructure involves heterogeneous edge devices and systems from different vendors and using different communication protocols. Interoperability between these edge devices are required for implementing edge computing in healthcare system.

Edge computing architecture involves computing among heterogeneous devices with different computational power and storage (Cicirelli et al., 2017). Therefore load distribution among these heterogeneous devices with efficient resource allocation and usage is one of major challenge to process media-rich data of healthcare system.

Edge computing involves distributed process which enlarges the scope of the attack surface (Xiao et al., 2019). The healthcare edge devices are made smarter which will increases the vulnerability of the device. Not all edge devices are filled with resources hence implementing complex security algorithms are not feasible with those devices. The distributed dynamic architecture of the edge devices in healthcare environment increases the dimension of the attack surface.

Edge computing is in the development stage with nonstandard protocols and technologies. It has lot of implications in implementing in critical infrastructure like healthcare systems.

1.5 Artificial intelligence in healthcare system

AI is developed to mimic human cognitive capabilities (Müller & Bostrom, 2016). It is changing the medical field drastically using the digitized medical data and various machine learning algorithms. It assist the medical professionals in disease diagnosis and treatment efficiently. The voluminous healthcare data and vast growth in big data analytics using AI algorithms has made paradigm shift in healthcare system (Jiang et al., 2017). It is necessary to study the potential capabilities of AI to transform healthcare system and the risks involved in applying AI for healthcare system.

1.5.1 Artificial intelligence

The goal of AI is to emulate the cognitive capabilities of human intelligence such as ability to perceive, reasoning, planning, decision-making, learning and understanding (Bogue, 2014). The recent advancements in AI is leading it as potential contributor of fourth industrial revolution. Fig. 1.10 shows the important components of AI. Initially AI had some limitations due to limited computation capability, data availability, and storage but now the cloud and IoT has powered AI with high computing resources and voluminous data. The developments in AI have generated algorithms which can learn and predict from voluminous complex raw data. Presently, decision-making using AI involves huge volume of data and high computation so the errors are reduced and decision are made faster with higher accuracy. The AI algorithms can efficiently detect the unforeseen relationship and complex nonlinear interactions in the voluminous data which cannot be done with standard statistics (Hofmann et al., 2017). AI algorithms have been implemented successfully in many application areas such as agriculture, business, healthcare, and science.

Figure 1.10 Components of artificial intelligence.

1.5.2 Opportunities of artificial intelligence in healthcare

AI in healthcare uses huge amount of data for analysis and interpretation to assist medical professional for making accurate decisions faster (Bennett & Hauser, 2013). Various pattern recognition AI algorithms help the doctors to handle complicated health conditions more efficiently. AI enables early detection of diseases using predictive algorithms. The health conditions of patients are diagnosed and decisions are made faster because time is a life-altering parameter for patients.

AI supports information management in healthcare systems. It helps healthcare stakeholder for efficiently managing the information. The AI enabled telemedicine helps both doctor and patients to save time and money with better treatment (Lysaght et al., 2019). Thus it takes the strain off of healthcare professionals in information management and improves the comfort of patients. AI also helps in improving the quality of Electronic Health Record (EHR) without human error (Koppel & Lehmann, 2015). Using deep learning combined with speech recognition technology, doctor interactions with patients is recorded along with clinical diagnosis and treatments provided are documented more accurately.

Presently AI has the ability to analyses voluminous patient data to detect the various treatment options using the natural language processing algorithms to provide personalized medication. Drug discovery is one of the greatest opportunity of healthcare where AI can be used efficiently to reduce long and expensive process of drug discovery (Smalley, 2017). Various AI algorithms are used in discovering new drugs with reduced risk in developing and testing. AI technology helps researchers to identify suitable patients to involve in the experiments and assists in monitoring medical response of patient more accurately.

The hospital operations such as managing emergency rooms, handling in-patient wards and scheduling doctor appointments are automated using AI algorithms. AI enabled Chatbots and virtual assistants are provided through telehealth (Sharmin et al., 2006). It also has preliminary diagnosis also provided through machine learning algorithms which helps patients to save time and money. The high risk patients are efficiently treated using AI algorithms by analyzing large dataset from various sources using personalized drugs dosage.

In addition to the above, the other applications of AI include, surgical robots provides surgeons support in surgical procedures (Palep, 2009), the virtual nursing assistants help patients and care providers in communication 24/7. AI also helps the health insurance companies to detect the false claims through the connected healthcare system. Fig. 1.11 gives the list of applications in integrating AI within healthcare sector.

Figure 1.11 Applications of artificial intelligence in healthcare.

1.5.3 Risks in integrating artificial intelligence in healthcare

AI in healthcare has more development in technologies but still it has some limitations. The main challenge in integrating AI with healthcare system is lack of standard training datasets (Iliashenko et al., 2019). Many AI algorithms need training dataset to train itself for detection or prediction mechanism. The accuracy of result in AI algorithms depend on training data used to train the said algorithms. The training dataset may contain bias which may lead to incorrect decision-making. The genetic and behavioral data are required for some diseases which are hard to collect. If some information is underrepresented in training data then AI algorithms will not provide accurate results. AI algorithm-based decisions may be accurate but it may not be always optimal. Collecting complete patient data for decision-making is a daunting