Smart Healthcare Applications over 5G Networks: A Systematic Review

Abstract

Provisioning of health services such as care, monitoring, and remote surgery is being improved thanks to fifth-generation cellular technology (5G). As 5G expands globally, more smart healthcare applications have been developed due to its extensive eMBB (Enhanced Mobile Broadband) and URLLC (Ultra-Reliable Low Latency Communications) features that can be used to generate healthcare systems that allow minimizing the face-to-face assistance of patients at hospital centers. This powerful network provides high transmission speeds, ultra-low latency, and a network capacity greater than that of 4G. Fifth-generation cellular technology is expected to be a means to provide excellent quality of medical care, through its technological provision to the use of IoMT (Internet of Medical Things) devices. Due to the numerous contributions in research on this topic, it is necessary to develop a review that provides an orderly perspective on research trends and niches for researchers to use as a starting point for their work. In this context, this article presents a systematic review based on PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analyses), with article selection based on inclusion and exclusion criteria that avoid bias. This research was based on research questions that were answered from the included works. These questions focus on technical characteristics, health benefits, and security protocols necessary for the development of smart healthcare applications. We have identified that a high percentage of existing works in the literature are proposals (56.81%, n = 25) and theoretical studies (22.73%, n = 10); few implementations (15.91%, n = 7) and prototypes (4.55%, n = 2) exist, due to the limited global deployment of 5G. However, the panorama looks promising based on proposals and future work that these technological systems allow, all based on improving healthcare for people.

1. Introduction

Technological evolution has not only allowed intelligent systems’ development in the manufacturing, transportation, and automation industries, but also in the medical industry, leading to an improvement in people’s quality of life. On this specific topic, smart devices’ use originated the concept of smart healthcare—healthcare based on technological support.

The main applications of smart healthcare are remote monitoring, telemedicine, telesurgery, and mobile medical assistance [1]. The deployment of smart healthcare applications in recent years has been developed based on the use of 4G [1,2] technology, while also using resources from previous generations. However, these have not been able to meet real needs that the smart health field demands, due to technological restrictions such as: limited bandwidth, slow transfer speed, and high latency [2].

Although 4G has been used in Internet of Medical Things (IoMT) applications for connecting various physical devices, such as cell phones, physiological monitors, body sensors, and wearable devices [3], it faces numerous challenges when the number of device connections increases. Therefore, to meet the demands of IoMT, a more evolved network than 4G is required in terms of speed, latency, support for simultaneously connected devices, and data storage capacity [2,3].

To solve this problem, the 5G network presents itself as an excellent alternative thanks to its greater performance compared to the 4G network. With the use of 5G, the current limitations will be fully eliminated. High transmission speed, ultra-low latency, enhanced mobile broadband, and a network capacity superior to that of 4G will enable innovation in healthcare [2]. For example, with ultra-low-latency communication, an ambulance can support remote and diagnostic operations [4]. In addition, smart healthcare is expected to grow significantly by 12.5% by 2025 [5].

Fifth-generation technology in smart healthcare can not only be used for healthcare for sick people, but also facilitates preventive care [6]. At the same time, it has the ability to transfer and manage medical data in bulk, interactive video and real-time remote device control [7]. These data can be archived or transmitted between patients and medical professionals, so the technology can meet the demands of multiple consultations.

Likewise, it will be possible to have immediate access from any place and at any moment to the patients’ personalized medical data through digital networks [3]. For example, large X-ray (XR) image files, such as magnetic resonances, can be sent by patients to specialists for review [1,6].

As 5G deployment advances, new studies are emerging in the medicine field, showing an increase in clinical cases covered by this technology. Through the application of telesurgery, teleconsultation, sanitary assistance, diagnosis and preventive care, platforms, systems, and apps have been created to facilitate healthcare. The existence of growth in the development of technological solutions also increases the existing scientific literature in this regard. For this reason, a systematic review analysis is required that allows interested developers and researchers to have a useful and understandable resource where they can learn about the progress of the solutions generated and the new challenges that must be faced.

In this context, several authors have made reviews of the states of the art that provide multiple perspectives on the advancement of smart healthcare applications developed on the 5G network. The most significant works and the main differences between them and the work presented in this article are mentioned below.

• Latif et al. [8], in 2017, presented the challenges and difficulties that could hinder 5G development in the health sector. The authors mentioned the fact that 5G, in conjunction with IoT devices, big data, artificial intelligence (AI), and machine learning (ML), has the potential to transform health systems globally. They also performed a comparison of latency and performance between 2G, 3G, 4G, and 5G technologies. They emphasized that coverage and reliability are key to enabling electronic health applications, such as home health monitoring.
• In 2019, Selem et al. [9] reported the state of the art in e-health applications that use 5G, in which various medical scenarios, such as robotic surgery and remote monitoring, are presented, highlighting the benefits of implementing 5G. In addition, they indicate the use of a wireless body sensor network (WBSN) in real-time monitoring applications, data transmission, emergency situations, and patient location. The emphasized that WBSN will be the backbone for the development of health applications with 5G.
• In 2020, Contreras et al. [10] contributed through a study on the use of telemedicine in the context of the pandemic caused by COVID-19. Through an analysis, they indicated that telemedicine can be enhanced later, after the health system was violated as a result of the pandemic. This review shows a descriptive approach and not a systematic analysis, since the selection of previously published articles was not performed. In the same year, Zhang Chunming et al. [11] also highlighted 5G use in smart healthcare in the control and management of the pandemic caused by COVID-19. The implementation of one of the methods consists of a teleconsultation system to attend two critical cases. The “West China Hospital” of Sichuan University was the developer of this implementation. Kim et al. [12] discussed the use of 5G technology in pre-hospital emergencies. They mentioned that it can support paramedics when assisting patients in emergency. Mainly, they highlighted the use of a remote connection to provide real-time information to external doctors and hospitals. In addition, they emphasized several challenges to which it is exposed, such as security, privacy, patient information sharing, and lack of access in rural areas. Siriwardhana et al. [13] presented the opportunities and challenges of 5G use in the health area. The solutions it can provide for the fight against the COVID-19 pandemic are detailed. The authors emphasize 5G use in telemedicine systems, telesurgery, remote monitoring, and drug delivery. Furthermore, they describe the technical requirements expected for different telehealth applications, such as telemedicine, telenursing, telesurgery, and telepharmacy. Finally, Ahad et al. [14], through a literature review, highlighted the solutions that 5G can provide in the medicine area. The authors presented a categorization of smart healthcare based on communication technologies, services, applications, and requirements that are key to the development of these systems.
• In 2021, Penn et al. [15] provided a narrative review of 5G opportunities in neurosurgery, in which they refer to the enhanced features of 5G such as ultra-reliable low-latency communication (URLLC), massive machine-type communication (mMTC), and enhanced mobile broadband (eMBB). They mentioned that these features could promote the development of neurosurgery by eliminating geographical barriers. They consider that 5G facilitates neurosurgeons’ real-time remote connection in pre-hospital emergencies. In addition, they revealed that in 2019, the China Daily newspaper announced a first neurosurgical case developed with 5G.

In this article, a multidisciplinary approach to the application of 5G technology in the medicine field is provided, through which, wireless healthcare services can be offered, taking into account various technical aspects. Our analysis provides a more complete perspective that includes different variables, such as technical aspects, application areas, types of devices, communication protocols, and security. The correlations between these variables is analyzed, showing those that show greater development and those that have not yet received much attention and could become new lines of research.

The paper is organized as follows. In Section 2, we present the methodology based on the PRISMA method, with the formulation of three research questions that were the basis for the entire study. In addition, the process followed to search for articles and filter them by applying inclusion and exclusion criteria is detailed. In Section 3, we answer the questions raised based on an analysis of the selected articles, while describing and discussing the main results. Finally, in Section 4, the conclusions of our work are presented.

2. Method

To develop this systematic literature review about 5G technology applications in smart healthcare, information discrimination was carried out following a systematic process. Through the PRISMA method [16], the research questions (RQ) that guide our study are raised.

• RQ1: What are the healthcare applications that can be developed with the 5G implementation in smart health systems?
  Through this question, we intended to know the different health systems developed with the use of 5G technology, its benefits, and marked trends in terms of the deployment of new supports.
• RQ2: What is the type of technology used, and what technical requirements does an intelligent health system demand through the use of 5G?
  This question was aimed at identifying the devices, tools, and technical characteristics that enable the development and operation of smart healthcare systems. It also aims to detect the technical characteristics that developers consider essential in the generation of applications.
• RQ3: What are the security mechanisms or protocols required by smart health systems based on 5G?
  This question was aimed at finding out what security measures are used in smart health systems based on 5G, related to security architectures and encryption protocols in the information transmission.

After raising the research questions, the keywords that allow us to structure the research string (RS) were established: “Mobile network 5G”, “smart healthcare”, “IoMT”, “telemedicine”, and “eHealth”. The research string (RS) was defined as follows:

RS = (Mobile network 5G OR 5G) AND (smart healthcare OR telemedicine OR eHealth OR IoMT) AND (applications OR apps OR devices)

Once the RS was structured, the document search stage began. The databases that were consulted were: IEEE, Science Direct, Scopus, Web of Science, Springer, Wiley Online Library, Gale Power Search, and Association of Computing Machinery (ACM). These databases were selected based on the multidisciplinary nature of their publications and previous work. An attempt has been made to include all those databases used by other authors in previous work. As the number of articles obtained was adequate, no additional technique was applied to expand the results.

In order to filter the scientific articles, the systematic process established in the PRISMA method was applied. As shown in Figure 1. We followed three stages, each with inclusion and exclusion criteria. Before starting with the identification stage, a filter was applied in the databases, considering that 5G technology appeared approximately in 2013 [17].

• In stage 1 (identification), the metadata were extracted from the retrieved articles, and the inclusion and exclusion criteria were used, starting with the publication year. As shown in Figure 2, since 2016 there has been an increase in the number of publications produced. With this background, we decided to review the documents published from 2016 to 2021. Within this framework, we also eliminated systematic reviews and state-of-the-art studies that were not based on a technological foundation and showed previous analyses that were inappropriate to the objective of this research.
• In stage 2 (screening), we reviewed the articles according to title, abstract, and keywords. Articles unrelated to our topic and brief, non-relevant articles were removed. Articles that could not be accessed were also discarded.
• In stage 3 (eligibility), we reviewed each article in depth, including for review those that clearly expressed and detailed their applications.

Table 1. Summary of articles included in the review.

Ref.	Author	Source	Year	Country
[17]	Mattos and Gondim	IEEE	2016	Brasil
[18]	Jusak et al.	IEEE	2016	Indonesia
[19]	Philip and Rehman	Springer	2016	United Kingdom
[20]	Nasri and Mtibaa	Web of Science	2017	Tunisia
[21]	Lloret et al.	Web of Science	2017	Spain
[22]	Nasri and Mtibaa	IEEE	2018	Saudi Arabia
[23]	Guo	IEEE	2018	China
[24]	Fang and Ye	IEEE	2018	USA
[25]	Dhaya et al.	Springer	2018	India
[26]	Mohanta et al.	IEEE	2019	India
[27]	Zhang et al.	IEEE	2019	Japan
[28]	Mamun et al.	IEEE	2019	Bangladesh
[29]	Gupta et al.	IEEE	2019	India
[30]	Imran et al.	IEEE	2019	India
[31]	Usman et al.	Scopus	2019	United Kingdom
[32]	Haider et al.	Web of Science	2019	China
[33]	Morosi et al.	Springer	2019	Italy
[34]	Padmashree and Nayak	IEEE	2020	India
[35]	Tsoulchas et al.	IEEE	2020	Greece
[36]	Meshram and Patil	Science Direct	2020	India
[37]	Kumar et al.	Scopus	2020	India
[38]	Zheng et al.	Web of Science	2020	China
[39]	Ye	Web of Science	2020	USA
[40]	Khorashadizadeh et al.	Springer	2020	Malaysia
[41]	Chen et al.	IEEE	2021	China
[42]	Angelucci et al.	IEEE	2021	USA
[43]	Zhai et al.	IEEE	2021	China
[44]	Zhan	Science Direct	2021	China
[45]	Pustokin et al.	Science Direct	2021	Spain
[46]	Braeken and Liyanage	Scopus	2021	Finland
[47]	Zhang et al.	Scopus	2021	China
[48]	Zhang et al.	Scopus	2021	China
[49]	Lin and Hsu	Web of Science	2021	Taiwán
[50]	Hameed et al.	Web of Science	2021	Pakistan
[51]	Chen et al.	Web of Science	2021	China
[52]	Guo et al.	Web of Science	2021	China
[53]	Lin et al.	Web of Science	2021	Taiwan
[54]	Minahil et al.	Web of Science	2021	Pakistan
[55]	Tang et al.	Web of Science	2021	China
[56]	Ying and Yu	Springer	2021	China
[57]	Mishra et al.	Springer	2021	India
[58]	Wang et al.	Wiley Online Library	2021	China
[59]	Duan et al.	Gale Power Search	2021	China
[60]	Talukder and Haas	ACM Digital Library	2021	India

shows a summary of the basic characteristics of the studies finally included. These are: authors, source, publication year, and country in which each study was conducted. Regarding the type of document, 65.91% (n = 29) are journal articles, 20.45% (n = 9) are conference papers, and finally, 13.64% (n = 6) are chapters and sections of books.

Regarding the origins of the studies, the country with the highest number of relevant works was China, with 31.80% (n = 14) of the total. This was followed by a considerable contribution from India, 20.50% (n = 9). The US was in third place, with 6.70% (n = 3). Spain, the United Kingdom, Taiwan, and Pakistan were in fourth place, each with a contribution of 4.50% (n = 2). Other countries, such as Brazil, Indonesia, Tunisia, Saudi Arabia, Japan, Bangladesh, Italy, Greece, and Finland, presented individual contributions of 2.3% (n = 1)—that is, only one article each. It should be noted that the countries mentioned above are mostly developed countries. In addition, several of them have 5G networks. On the other hand, from the South American continent, we found only one study (Brazil), which implies that most countries in this region have not yet implemented a 5G network, so research on this topic is limited.

Using the VoSviewer tool, Figure 3 was developed, which shows a map of a bibliometric network generated among the countries that have made the most contributions to the smart healthcare topic regarding 5G networks. It also shows networks formed with other countries that are also developing related research. Some of these countries are not part of our review due to inclusion and exclusion criteria we used, but the map shows that there are related contributions based on papers’ metadata. This figure provides an overview of the countries where there is most interest in this topic.

3. Results and Discussion

For the characterization of the studies covered in this review, from the technical and application points of view, the following variables were considered:

• Application area.
• Healthcare benefits.
• IoMT devices.
• Communication protocols.
• Networking technologies.
• Network technical requirements.
• Security mechanisms or protocols.

The objective of the study is to identify the main characteristics from the technical and application points of view focused on healthcare, which are used in smart healthcare systems. In order to show the results graphically, pie diagrams were used that allow a simple and quick interpretation of the works’ distribution, considering the variables proposed.

When searching for papers, we identified great diversity in the publication types obtained. Since 5G network deployment is a relatively new topic, we found several papers that make proposals or describe prototypes under development. There are also implemented systems (although few) and theoretical studies with perspectives that deserve to be analyzed. Figure 4 shows the percentage of studies classified according to the type of work presented in each of the articles: 56.81% (n = 25) of the studies are proposals—that is, unimplemented work, thereby containing almost no technical and safety features. However, they show theoretical support and propose detailed models, which contribute to the identification of analysis variables. On the other hand, 22.73% (n = 10) of the articles report on real-world implementations. In general, these documents focus on the benefits, merits, and advantages of the systems, without neglecting technical characteristics and safety requirements.

Regarding the implementations, which comprise 15.91% (n = 7) of the studies, the results are detailed for developed and functional systems which address issues related to security and privacy mechanisms, although not in their entirety.

Finally, 4.55% (n = 2) of papers describe prototypes’ development, that is, models that serve as apparent and concrete representations of operational systems. These works focus on most of the variables analyzed, giving details of technical requirements, proposed applications, and safety considerations.

The following sections review the results obtained in response to the research questions initially posed and that guided the research.

3.1. RQ1: What Are the Healthcare Applications That Can Be Developed with the Implementation of 5G in Smart Health Systems?

The medical system has dynamically evolved, and this has made smart health a reality. The development of these systems lies in the possibility of improving people’s life quality. Figure 5 shows a classification generated based on the application developed in each work. We have identified three specific application areas: remote care, data transmission, and continuous real-time monitoring. In addition, in

Table 2. Summary of articles according to the application area.

Application Area	Article
Remote care: surgery and diagnosis	[26,27,29,30,34,36,37,38,39,40,41,45,47,51,53,54,58,59,60]
Real-time continuous monitoring	[17,18,20,21,22,23,24,25,28,32,33,35,42,44,46,48,49,50,52,55,56,57]
Data transmission: smart ambulance	[19,31,43]

, articles that belong to each classification are listed.

3.1.1. Remote Care: Surgery and Diagnosis

Through remote care, healthcare specialists provide medical services remotely via the Internet. [39]. This enables the sharing of physiological data collected by sensors and IoMT devices, which allow the provision of diagnostic, surgical, and patient treatment services remotely [2]. Figure 6 shows a graph that generally describes the appearance of a remote surgery performed by a teleoperated robot. This image serves as a reference for the application area under discussion. Work focused on remote medical care is presented below.

Zhang et al. [27] presented the implementation of a remote clinic and ambulance system at Zhengzhou University Hospital. On the one hand, in the remote clinic, the diagnosis and treatment system is used to diagnose patients without requiring them to travel to the hospital. In the tests carried out, the HD video consultation was successfully performed. On the other hand, in remote surgery, the system includes a control center where the specialist is located and a work center where the surgical robot is located. Although the system is mentioned, no surgeries have been reported yet, due to the unsolved operational risk.

Zheng et al. [38] illustrated the implementation of a remote laparoscopic surgery system. This system uses a surgical robot called “Micro Hand” manipulated remotely by a surgeon. The surgery was performed on a pig that was 3000 km away from the surgeon, and there were no complications during the process.

Chen et al. [51] reported the development of a system that performs retinal photocoagulation in patients with diabetic retinopathy (DR). The system includes a teleophthalmology platform, a laser device, a platform for remote computer control (TeamViewer GmbH), and a platform for teleconsultation through videoconference (Cisco Systems Inc.). This is the first remotely controlled retinal photocoagulation procedure for which a report is available.

Wang et al. [58], through the use of a remote ultrasound system, developed an application for the diagnosis of COVID-19. The system examines the heart, lungs, blood vessels, and inferior vena cava. It is composed of two environments: The first, where the doctor is located, consists of a workstation, a console, image monitors, and a transducer. The second environment is where the patient is located and is composed of an ultrasound (US) system and a 6-degrees-of-freedom robotic arm equipped with transducers and a force sensor. The ultrasound was performed on a patient who was in Zhongnan Hospital 700 km away from the sonographer.

Duan et al. [59] proposed a teleultrasound diagnostic system for an intensive care unit (ICU). The system is equipped with a ultrasound assisted by a robotic arm, MGIUS-R3, on the patient’s side, and a control unit on the physician’s side. In addition, audio, video, and main-frame systems are located on both the surgeon’s side and patient’s side. The ultrasound was used to perform liver, pancreas, spleen, gallbladder, and kidney examinations of 32 patients. One patient was excluded due to interference caused by intestinal gases and poor image quality. Regarding the rest of the patients, there were no problems during the examination, and the quality of the images was high.

Zhang et al. [47] developed a diagnostic system for postpartum hemorrhage (PPH). The system uses machine learning algorithms such as a support vector machine (SVM) and an artificial neural network (ANN) for classification and diagnostic prediction. The model collected 3842 vaginal delivery cases performed in 2017 at the Beijing Hospital of Gynecology and Obstetrics. This dataset enabled the development of the machine learning prediction model, collecting 23 PPH-relevant features from each patient.

Gupta et al. [29] presented a proposal for the telesurgery architecture. It consists of three domains: master, slave, and network. The system consists of a surgical arm, which is located in the slave domain; in the master domain, the surgeon is in charge of remotely manipulating the robotic arm, through a human system interface (HSI); and the network domain is used as the communication channel.

Meshram and Patil [36] proposed a tactile robotic telesurgery architecture. This system consists of a master–slave console. The master console houses the surgeon, who through commands, remotely controls the robotic arm located in the operating room (slave console). In addition, this architecture has two links, a direct link and a feedback link. The direct link is used to control the movement of the robotic arm, and the feedback link allows monitoring the patient’s vital signs.

Pustokin et al. [45] presented a model for the diagnosis of diseases based on the deep belief network (DBN) using big data and deep learning (DL) techniques. This model consists of an acquisition stage in which physiological data are collected through body sensors and wearable devices. The generated data are stored in the cloud and managed through a Hadoop echo system. Subsequently, the information was analyzed and classified by DBN, allowing the diagnosis of diseases to be made based on patients’ medical data.

Padmashree and Nayak [34] conducted an analysis of remote-surgery systems, detailing that it reduces the time that the surgeon spends to enter an operating room in a traditional surgery. In addition, it eliminates the need for the doctor’s presence in the operating room and allows more experienced surgeons to provide support to less experienced ones.

Finally, Ye [39] is one of the authors who emphasized the use of remote systems for COVID-19 pandemic implications. He mentioned that through technologically supported systems, remote care and prevention services can be provided. He highlighted the use of AI as a prediction and detection system during the spread of the virus. These systems use X-ray images, and their use in learning algorithms helps specialists when making decisions.

3.1.2. Real-Time Continuous Monitoring

As shown in Figure 7, continuous monitoring consists of periodic and real-time transmission of physiological parameters and vital signs of patients through the use of portable devices (body sensors and wearable devices) [18]. This allows the transmission of the collected parameters to be assessed and analyzed by a doctor [28] or an expert system. In addition, generally, these continuous monitoring systems are based on a two-hop architecture, where the parameters collected from the sensors are sent to a gateway using short-range technologies (Bluetooth, Zigbee, NFC, Wifi) and then to a data system using cellular technology or other access network [42].

In the context of the pandemic caused by COVID-19, Zhang et al. [48] presented a system for diagnosing patients infected with the virus. It began operation on 29 January 2020 in Wuhan city. The main system’s objective is to reduce physical contact between doctors and infected patients to mitigate the spread of the virus. The system includes the use of body sensors to collect vital signs such as blood pressure, oxygen saturation, heart rate, and temperature, which are essential in monitoring the disease. The monitoring was carried out by specialists in respiratory pulmonology. In addition, 161 infected individuals were followed up, of which 38 cases had a severe diagnosis. Although the number of infection cases increased, severe cases and deaths were controlled.

Talukder and Haas [60] develop an intelligent care system called Homing. This system consists of a solution to measure the patient’s vital parameters, through the use of devices such as wearables, IoT, smartphones, and tablets. In addition, it integrates an app for smartphones in which the vital parameters’ information is stored, and these are updated in real time to be monitored by a doctor.

Jusak et al. [18] created a prototype for monitoring cardiac activity in the form of a phonocardiograph signal (PCG). It consists of an agent that collects and transmits data (heart sound sensor) and a manager (Raspberry Pi). It is a multiprotocol device to back up the collected data. In addition to capturing and monitoring heart sound activities, the prototype has a web server so clinicians can access data remotely 24 h a day.

Nasri and Mtibaa [20] designed a prototype for monitoring patients in general, through the measurement of physiological signals such as body temperature, oxygen saturation level, and heart rate. It is considered as a preventive system because it alerts one variations of these physiological signals and changes in environmental parameters. For this purpose, a WBSN is used to obtain the patients’ vital parameters. The system has a platform where users can make queries and doctors can provide follow-up. The monitoring system proposed by Lloret et al. [21] consists of an architecture that allows remote monitoring of chronic patients suffering from heart and lung diseases, cancer, diabetes, bone, mental disorders, and genetic disorders. The architecture consists of body sensors that collect physiological variables from the body and sends the collected information to a mobile device via Bluetooth. The information is stored in a database, and using AI with data collected from hospitals, it decides whether the parameters collected by the sensors are typical data or data with anomalies. In addition, the system can send alarms if it detects anomalous data.

Hossain and Muhammad [22] proposed an intelligent health system capable of detecting emotions in the elderly, children, and people with mental disorders, using image and voice signals. To do this, the system has cognitive sensors that capture these signals, which are then processed to detect if the patient expresses emotions related to pain. It uses a local binary pattern (LBP) and an interleaved derivative pattern (IDP) for feature extraction from images. If it detects the person is in pain, the system sends a notification to related entities, such as hospitals, health centers, general practitioners, and caregivers, so that they can be directed to the patient for immediate assistance.

Guo [23] presented a proposal for a miniature electrochemical “Dongle” (USB flash device comprising a reader and a sensor). This system allows diabetic patients’ monitoring through the glucose and uric acid measurement. It includes a smartphone as an electrochemical analyzer, which powers the “Dongle” with an OTG cable; and a medical service platform for consulting, monitoring, and storing biochemical parameters, so that the doctor can monitor the patient’s health status.

The intelligent health monitoring system presented by Zhang et al. [27] was used to monitor a patient’s vital signs and venous transfusion. The doctor in charge of performing the procedure was miles away from the individual. In addition, the system notifies the nursing service once the transfusion is completed for the removal of the needles. In case of emergency, it has a button to communicate with the nursing department. The system has body sensors and an intravenous injector.

Autilife is a proposal submitted by Mamun et al. [28] to monitor children in autism centers. Emergency situations such as epilepsy, heart attacks, strokes, hysteria, and anxiety are situations that can occur in an autism center. The system focuses on improving their life quality, through the constant monitoring of physiological parameters such as blood pressure, body temperature, heart rate, speech signals, and body movements. The system includes WBSN in various accessories, such as a watch, belt, shoes, shirt and glasses, for the collection of parameters that are sent to a database through the cellular network. The SVM is used for classification of collected data and signals. The classification contains two classes: normal-health condition and endangered-health condition. If one’s health is in danger, the system sends alarms to a nearby ambulance or hospital and to the autism center caregiver.

The intelligent patient-monitoring system proposed by Haider et al. [32] allows one to detect postsurgical falls and body movements, such as fast and slow walking, squatting, lying down, and sitting. The system’s objective is to protect the patients’ health by preventing falls that could lead to injuries and even death. The patients who benefit most from the system are the elderly, due to their lack of mobility and balance. With this resource, doctors can monitor patients 24 h a day. ML algorithms such as SVM, Naive Bayes, and decision tree were used for classification of six body movements, including falling. For training, the system uses repetitive body movement signal information from eight different individuals of different ages, through the use of low-cost body sensors. The results showed that SVM performed better in classifying body movements.

Morosi et al. [33] proposed a telemonitoring and teleassistance system for monitoring patients suffering from diabetes. Body weight and blood glucose measurements can be taken through a glucose test kit that incorporates an NFC (near-field communication) glucometer, a lancing device, and a scale. By using a mobile app, health parameters can be recorded and monitored at any time. The mobile app can interact directly with the scale via a Bluetooth connection and with the glucometer via an NFC connection. The data parameters acquired on the mobile phone through the app are sent to a cloud platform via the cellular network. Medical staff can access the platform and monitor the patient’s parameters in real time.

Focused on people with diabetes, Tsoulchas et al. [35] proposed a system that tracks the user’s glucose level. To capture the measurements, an smart device (Dexcom G6) is used, which transmits the data to a smartphone via Bluetooth, and stores the values and generates an alert in case the measured values are not adequate.

The home telemedicine system presented by Angelucci et al. [42] consists of a mechanism for monitoring the patients’ respiratory status with chronic respiratory diseases. The system is equipped with three devices: an Airgo band, which provides a signal similar to the thoracic circumference; a uHoo environmental sensor, which measures parameters such as temperature, humidity, and pressure; and a SAT-30 oximeter, to obtain parameters such as heart rate and blood oxygen saturation. The band is connected to an iPad and the parameters are collected through an iOS app to be analyzed by doctors.

The neural network system suggested by Hameed et al. [50] was developed for disease monitoring. It uses body sensors such as a pulse sensor, blood-pressure sensor, temperature sensor (MAX30205), and oximeter sensor (MAX30100). The medical parameters (blood pressure, body temperature, heart rate, and oxygen saturation) collected by the sensors are stored on a cloud server. An ANN is used to analyze and predict whether the person’s parameters are normal values or threshold values indicative of disease. The ANN training phase is performed using data from patients’ clinical records and previous data from others.

Guo et al. [52] is another author who proposed a system for quantitative detection of people infected with COVID-19. Unlike the system presented by Zhang et al. [48], this system uses an ultra-sensitive fluorescence sensor that performs the measurement of COVID-19 peak protein and nucleocapsid protein through nasopharyngeal swab samples. The information captured by the sensor is sent to a mobile device via short-range communication technology and then to a database via the cellular network. In addition, monitoring is done through the use of a mobile app that integrates DL and fuzzy logic algorithms.

Tang et al. [55] also presented a system for biomedical signals’ remote monitoring in patients infected with COVID-19. The authors mentioned that a monitoring experiment was carried out on 2000 individuals of different ages for 24 weeks to obtain authentic data of biomedical signals (body temperature and heart rate). The data collection is performed using wearable IoT devices that transfer data to a cell phone via short-range communication (Bluetooth, Wifi), and the data are sent in real time to a server in the cloud via the cellular network. For the training, 8,064,000 biomedical data were collected; 613 individuals presented health problems (according to the doctor) during the experiment process.

Zhan et al. [44] proposed the use of body sensors and wearable devices for health monitoring in athletes by collecting vital signs such as oxygen saturation level, heart rate, body temperature, and blood pressure. The information collected by the devices can be transmitted to a database where the doctor can monitor and analyze their physical health status and their performance related to the sport they practice.

Mohanta et al. [26] proposed a system for smart health management by using AI, wearable devices, and body sensors (electrocardiogram sensor, blood pressure sensor, temperature sensor, electromyography sensor, blood oxygen sensor) for information gathering and disease monitoring from the extracted sensory parameters. The system uses a cloud platform for the medical records storage collected from the sensors. In addition, with the help of AI applied to RX images, electronic health reports (ERHs) are generated, which helps specialists in the analysis and diagnosis process.

Dhaya et al. [25] exposed the merits and new smart health innovations in healthcare. They mentioned that several systems have been developed, such as glucose monitoring programs for patients with high blood sugar levels, through the use of smart meter sensors; a system for monitoring patients suffering from asthma with the use of IoT technology; a system for Parkinson’s monitoring through ingestible sensors; and a system for monitoring patients during the burning of calories, through the use of the “Hike Run Beta” app that tracks the amount of steps an individual takes.

Dhaya et al. [25] and Ying and Yu [56] highlighted that by integrating technologies such as IoT, big data, AI, and wearable devices, an intelligent health monitoring system can be implemented which would support the collection of physiological parameters such as blood pressure, heart rate, and body temperature. In addition, the collected parameters can be stored in a cloud platformm.

For remote monitoring of elderly people, Mishra et al. [57] analyzed the use of wearable devices for the acquisition of physiological parameters, such as oxygen saturation level, blood pressure, body temperature, rhythm heart rate, and respiratory rate. In addition, a ML system generates reports of the physiological parameters collected from the person, and after a period of time they are sent to the specialist to provide a follow-up and diagnosis.

3.1.3. Data Transmission: Smart Ambulances

The information transmission (video, images, and vital signs) in real time in ambulances is of vital importance in emergencies and in cases of rescuing people at risk when their health is compromised [43]. During the pre-hospital care performed by paramedics, they can be guided or supervised by a doctor remotely, in order to provide treatment and monitor the patient’s current condition [31]. In addition, one of this system’s benefit is that unnecessary hospitalizations can be prevented because doctors provide appropriate care remotely, and during the process, they can decide whether or not patients’ hospitalization is necessary. Therefore, this system type requires connectivity and high-speed data transmission to ensure uninterrupted communication [19]. In this context, in this section, the articles related to the application area in the field of information transmission in ambulances are summarized. In addition, an explanatory schematic diagram of how this type of application works is shown in Figure 8.

The ambulance and the remote clinic presented by Zhang et al. [27] act as a mobile diagnostic and treatment system, enabling real-time patient information sharing with doctors and specialists. The system can share patients’ vital signs with the use of body sensors and high-fidelity cameras for image and video transmission. In addition, during the patient’s transfer to the hospital, paramedics can remotely look up the patient’s medical history, upload vital signs, and communicate with the specialist to minimize the time spent during the rescue.

Usman et al. [31] presented an ambulance proposal for emergency situations. The system includes US video transmission, ambulance video transmission, and transmission of vital signs, such as oxygen saturation level, heart rate, body temperature, blood pressure, respiratory rate, electroencephalography, and electrocardiography. During the video transmission, the doctor can observe the patient and the paramedics onboard the ambulance, in order to provide support to the personnel who are attending the emergency. In this sense, doctors can remotely examine the US performed by an ultrasound system equipped in the ambulance, to generate a diagnosis or referral to an appropriate hospital department for further treatment. In addition, the authors mentioned that the video and vital sign transmissions are compressed for transfer to the hospital center because they generate a large amount of information, thereby avoiding unnecessary data consumption and minimizing network resources and costs.

The smart ambulance architecture proposed by Zhai et al. [43] is equipped with a multi-monitoring system composed of a vital-signs-acquisition device, a medical workstation, and an HD camera. This equipment enables audio and video transmission from the ambulance’s central command to the hospital, which facilitates doctors’ interaction with patients in real time. At the same time, they emphasize that the use of virtual reality (VR) glasses is the main advantage in remote video transmission, since it facilitates patient care by making the doctor feel immersed during the transfer process. In addition, by sending data to the hospital center, the patient registration process is simplified, improving the efficiency of the procedure that follows. Finally, the authors performed a simulation of the video transmission in 5G and determined that the uplink and downlink speed values are higher compared to those obtained through the simulation performed with the 4G network.

Healthcare Benefits

After analyzing each of the clinical applications proposed by the works included in the review, we were able to identify that each application presents different healthcare benefits, categorized into three large groups: constant and immediate attention with intelligent systems, specialized medical care in remote areas, and immediate assistance in emergencies.

In Figure 9, it can be seen that the majority of clinical applications provide benefits related to constant and immediate care through intelligent systems (52.27%, n = 23) that allow specialized healthcare to be provided through remote monitoring and tracking.

Table 3. Summary of articles according to the healthcare benefits.

Healthcare Benefits	Article
Constant attention and inmediate with intelligent systems	[17,18,20,21,22,23,24,25,26,28,32,33,35,42,44,48,49,50,52,55,56,57,60]
Specialized medical care in remote areas	[27,29,30,34,36,37,38,39,40,41,45,46,47,53,54,56,58,59]
Immediate emergency assistance	[19,31,43]

shows papers included in each classification.

Remote diagnostic and surgical systems make it possible to minimize patient attendance at hospital centers through the use of digital platforms. This facilitates safe and specialized care, including major surgical interventions, through precise and non-invasive procedures. It is observed that the applications related to this description can be included within the category of specialized medical care in remote locations, and it comprises 40.91% (n = 18) of the total.

Finally, with a much lower percentage than the previous types (6.82%, n = 3), the applications offer immediate assistance in emergency situations. In this case, the benefit consists of providing emergency treatments and first aid to safeguard the patient’s life until he or she can be hospitalized.

3.2. RQ2: What Is the Type of Technology Used, and What Technical Requirements Does an Intelligent Health System Demand through the Use of 5G?

To answer this question, we have generated several subsections that address different specific topics related to the technology type used and technical requirements. In terms of the technology used, Section 3.2.1 analyzes the articles and presents a classification according to the type of device used in development. Then, in Section 3.2.2 another classification is presented based on the communication protocols used by the systems. Along with the communication protocols, the analysis of short- and long-range network technologies is proposed and is analyzed in Section 3.2.3, providing a broader vision of the diversity of technological solutions. Finally, in Section 3.2.4, the technical requirements proposed by the authors for the correct operation of the systems are analyzed.

3.2.1. IoMT Devices

The medical applications presented in Section 3.1 are associated with the use of IoMT devices that allow one to provide patients with services such as monitoring, telemedicine, and even telesurgery.

The use of body sensors, wearable devices, software, and communication interfaces, make up the well-known IoMT [4]. It is a technology that allows managing health by reducing costs and improving productivity, quality of service, and the patient’s experience.

In order to know the IoMT technologies that are behind the presented systems, an analysis and classification of each one was generated, thereby establishing five groups: body sensors, artificial intelligence, teleoperated robots, ultra high definition cameras, and wearable devices. The percentages of studies related to the types of devices that smart healthcare systems manage are shown in Figure 10.

In general, the classification is similar to that presented by Haghi Kashani et al. [61], although their analysis emphasizes the use of body sensors and wearable devices. Our analysis focuses on the type of technology—sensors, devices, or algorithms. The following is a summary of the main ideas of the types of technology classified in each category.

Figure 10 shows that body sensors were used in 40.91% of the studies. This is related to the fact that in most continuous monitoring systems, the use of a WBSN [19] is beneficial for the collection of physiological signals. These IoMT devices are oriented to the performance of a specific activity, unlike wearable devices that can perform various tasks. For example, a smartwatch shows the time and at the same time can measure the oxygen saturation level thanks to an SpO2 sensor, which is integrated internally. In addition, the sensors not only collect physiological signals from the body, but also capture voice and image signals [22].

The integration of AI is useful in diagnosing, predicting, and monitoring patients remotely. Generally, this technology is used in continuous monitoring, so that systems estimate results without the need for human intervention. The use of this technology makes an intelligent system and improves the ability to provide sanitary attention, so that it is within people’s reach. Systems that integrate AI mostly use ML algorithms for labeling and classification of data [28] and medical images [39]. This is essential due to the limitations of the algorithms. In this aspect, our research reflects a contribution of 25.00% (n = 11) of systems using AI for the development of optimal sanitary systems.

Another type of IoMT device used in the systems presented in Section 3.1 is the teleoperated robot, to which, 20.45% (n = 9) of the research is related. They are generally employed in remote surgery and diagnostic systems for performing surgical operations and US examinations thanks to their utility and efficiency. In other words, the integration of such robots in telesurgery is considered a broad field [30] in intelligent systems for healthcare. Generally, these robots are manipulated by a surgeon remotely, making use of a control panel [29] or a console [36]. One of the teleoperated robots presented by Gupta et al. [29] contains two arms or manipulators (right and left) that control equipment, and a main manipulator that controls an endoscope. “Microhand”, presented by Zheng et al. [38], is a teleoperated robot consisting of a console for the surgeon, which includes an image and control panel, two manipulators (right and left), and pedals for motion control. It also has a system for the patient, which includes a rotating head, a passive arm, and three slave arms. Similarly, a teleoperated robot used for diagnostic ultrasound of COVID-19 [58] consists of two systems: on the doctor’s side, image monitors, a console, and a transducer to control the robotic arm; and on the patient’s side, it has an US device, a 6-degree-of-freedom robotic arm equipped with transducers, and a force sensor. Lastly, MGIUS-R3 [59] is a robotic arm that is also equipped with an US device on the doctor’s site. Additionally, the audio, video, and mainframe systems are located on both surgeon and patient sites.

Continuing with IoMT devices, it is observed that the use of ultra-high-definition cameras was involved in 6.82% (n = 3) of the systems. These are usually equipped inside ambulances. The main objective of these devices is to perform image and video transmissions in real time, in order to provide patient information to a hospital facility and a physician so that remote care and guidance to paramedics can be provided during the process of transferring the patient to the hospital. In the ambulance system presented by Usman et al. [31], through the process of transferring a patient to the hospital center, an ultrasound scan was performed with the use of US. The process was monitored by a remote doctor, who moved the camera to perform the whole ultrasound process.

Wearable devices were involved in 6.82% of studies (n = 3), as were the ultra-high-definition cameras. These types of devices, and body sensors, are used to collect biomedical signals (oxygen saturation level, blood pressure, body temperature, heart rate, and respiratory rate). Their use is involved in surveillance, continuous monitoring, and remote care systems. In addition, these types of sensors not only fulfill the function of signal collection, but are also devices developed to fulfill other types of tasks.

3.2.2. Communication Protocols

The IoMT devices presented in Section 3.2.1 use various communication protocols. Among these are device-to-device communication (D2D), machine-to-machine communication (M2M), and human-to-machine communication (H2M). M2M allows communication without the need for human intervention, unlike H2M communication, in which a human–machine interaction is required. D2D is an optimal alternative for device-to-device communication in situations where cellular communication technology is not available. Figure 11 shows the classification based on the types of communication used by these systems.

The classification of the type of communication protocol used in smart health systems is based on that presented by Ahad et al. [14], which focuses on the technologies and techniques required in a smart healthcare network.

Figure 11 shows that the majority of smart healthcare systems (59.09%, n = 26) use device-to-device (D2D) communication. This is due to the fact that large parts of the systems and architectures dedicated to intelligent healthcare are made up of wireless devices (body sensors and wearable devices) that communicate with each other and share information.

Note that 27.28% (n = 12) of the systems use human–machine communication (H2M). This technology is used in robotic telesurgery systems, in which there is interaction between a surgeon and a robotic arm, which through a human system interface (HSI), can remotely manipulate the surgical robot.

Only 11.36% (n = 5) of the studies presented made use of D2D and M2M communication. Some studies propose systems that require these two types of communication. According to [25], D2D and M2M communications in conjunction with intelligent medical devices generate significant improvements in various health scenarios.

Like D2D, M2M communication in IoMT devices plays an important role in the development of smart healthcare systems, as it enables multiple interconnections with smart devices [17]. However, only 2.27% (n = 1) of the works presented by the authors exclusively used this type of communication.

3.2.3. Networking Technologies

Clinical applications based on the use of body sensors and wearable devices, use short- and long-range network technologies to communicate and send information. Among the short-range technologies are WiFi, Bluetooth, NFC, and Zigbee [53]. For long-range communication, the cellular network is used.

Figure 12 shows that 43.18% (n = 19) of smart healthcare systems use Bluetooth and WiFi for communication between devices. This corresponds to the fact that a large portion of the devices are integrated with this type technology, due to its low cost and energy consumption [20]. For this reason, there is extensive development of smart health systems based on these transmission technologies [39] for sending and receiving medical data collected from patients through body sensors or wearable devices. That said, the system presented in [42] for monitoring patients with respiratory diseases uses a portable sensor (Airgo band), which sends vital signs data to an iPad device via Bluetooth, and via Wi-Fi, the data of oxygen saturation (SAT-30 oximeter) and environmental parameters (uHoo environmental sensor). Similarly, in [50,55], using these technologies, medical devices and sensors transmitted user information to a mobile device.

Likewise, with a proportion of 29.55% (n = 13), telesurgery and ambulance systems transmit patient data through the cellular network. In other words, these systems do not use short-range communication technologies (Bluetooth, NFC, Zigbee, and WiFi) because these generate a fairly large volume of data, so the data transmitted directly.

Zigbee is another of the communication protocols that is used in conjunction with WiFi and Bluetooth. It is a robust protocol, it is low cost, and its power consumption is relatively low [20]. However, only 11.36% (n = 5) of the presented studies used these three communication protocols in a hybrid way. According to the above, one of the systems that uses these three types of communication is the Nasri and Mtibaa [20] monitoring system. The system has an app for Android devices in which it receives physiological data captured by WBSN, which are later sent to a platform in the cloud.

The same thing happens with other systems that use Bluetooth and NFC to share users’ data. The use of these technologies is reflected in the system for monitoring diabetic patients proposed by Morosi et al. [33], which makes use of a scale via a Bluetooth connection and a glucometer through an NFC connection for the acquisition of parameters such as body weight and blood glucose. The use of this hybrid protocol represents 2.27% (n = 1) of the total number of studies incorporating these two short-range network technologies.

Finally, it is noteworthy that only 2.27% (n = 1) of the systems use WiFi technology for device communication as the only means of transmitting medical data, considering that it is one of the most used wireless technologies worldwide; however, it is the least used in in IoMT applications because it has high power consumption.

3.2.4. Network Technical Requirements

When it comes to medical scenarios, network requirements turn out to be higher and more critical than other types of scenarios that also require massive data transmission, such as agriculture and road traffic management [34]. Uninterrupted communication is needed to connect patients, doctors, and emergency services, generating diverse requirements, which result in higher performance from a technical point of view.

Table 4. Summary of technical requirements.

Ref.	Capacity	Latency
[17]	Peak data rate 10 Gbps	≤1 ms
[19]	Peak data rate 10 Gbps	≤1 ms
[21]	Bandwidth 1000 MbpsMbps	1–5 ms
[25]	Bandwidth 1 Gbps	-
[26]	Data rate 137 MbpsMbps-1.6 Gbps	-
[27]	Downlink peak speed 1 Gb/s, Uplink peak speed 100 MbpsMb/s	7.6 ms
[29]	-	<1 ms
[30]	-	100 ms, 10 ms y 1 ms
[31]	Data rate 10 Gbps	1–4 ms
[33]	200 MbpsMbps	-
[34]	Bandwidth > 1 Gbps average speed 200–400 MbpsMbps	<10 ms
[36]	Ultra high data rate > 1 Gbps	<5 ms
[37]	Data rate of 10 Gbps	<1 ms
[38]	Bandwidth Gb/s	264 ms
[39]	Peak data rate 10 Gb/s	1 ms
[40]	Data rate > 1 Gb/ss	≤ 1 ms
[43]	Data rate 10–20 Gbps	12.88 ms
[48]	Uplink transmission speed 100 MbpsMbps	10 ms
[51]	Upload speed 88.45 MbpsMb/s and Download speed 853.63 MbpsMb/s	20 ms
[57]	20 Gbps for downloading and 10 Gbpsfor uplink	1 ms
[58]	Transmission rates 930/132 MbpsMbps audio/video	1 ms
[59]	Download speed 580 MbpsMbps upload speed 92 MbpsMbps	-
[60]	Download speed 20 Gbps	1 ms

shows the classification of the articles according to the technical requirements presented in each of the studies.

As seen in Table 4, the authors pay much attention to system latency. This is because it is one of the key requirements for end-to-end communication in mobile health [17], since it involves the time needed for the acquisition, emission, reception, preprocessing, and display of the data at the endpoint. In addition, it is vital in the transmission of high-definition information, such as the transmission of medical multimedia content in ambulance scenarios [31]. Similarly, in remote-surgery systems, ultra-low latency allows the surgeon to have haptic feedback [30]; that is, despite being located miles from the patient, the specialist can feel the person’s tissues through the robotic arm just as in conventional surgeries. On the other hand, with ultra-low latency, machine-machine (M2M) and device-device (D2D) communications, and even human–machine interactions (H2M) [29] can be carried out, making possible the development of telesurgery, teleassistance, remote monitoring, and surveillance with drones.

Likewise, capacity is another requirement shown by a large number of studies, which address issues such as system bandwidth, which should guarantee quality of service (QoS) to patients [21,25]. To a lesser extent, data speed is also considered, since it can contribute to the fast and reliable transmission of medical files and multimedia information successfully [26,31].

It should be noted that all the works initially identified as implementations detail the specifications and technical details of their systems. However, the proposals, prototypes, and studies under development do not always expose these characteristics. This is because works of this nature are not being implemented due to the fact that the 5G network is not yet available in most countries. Nevertheless, it is possible to perform an analysis because they have theoretical foundations and relevant information related to the variables of analysis.

3.3. RQ3: What Are the Security Mechanisms or Protocols Required by Smart Health Systems Based on 5G?

By using IoMT devices in smart healthcare, patient information and medical systems are at risk of cyber attacks. Through them, patient privacy is compromised, and hackers can make use of user data, violating the confidentiality of the information with perverse intentions.

The security and privacy of the information contained in the systems are paramount and must be safeguarded. However, the results obtained in Figure 13 show that 47.73% (n = 21) of the studies do not mention or consider as a priority the implementation of security methods or protocols in their work.

Based on the analysis of the works, three systems related to the security of information management, both in transmission and in storage and access, were identified. These systems are authentication, encryption, and coding, which are detailed below.

3.3.1. Authentication

This protocol is established in three ways: between the user and the service provider, between two devices, and between a device and the network [24].

Figure 13 shows that authentication (31.82%, n = 14) is the security protocol that attracted most the attention from the authors. Fang and Ye [24] proposed an identity management (IdM) framework for mutual authentication, identity protection, and trusted security. This system was considered for both wireless access and cloud access. Additionally, it is mentioned that through the use of AI [26,48], an authentication system for users can be achieved, which ensures the security of updating patient information periodically via portable devices through scripts [34]. Additionally, [41] presented Zero-Trust, which is a security architecture that constantly evaluates network vulnerabilities based on traffic detection, logging, and analysis. In this, way it reduces malicious access and continuously monitors client behavior.

Another type of security architecture that uses authentication is EAP-TLS (Extensible Authentication Protocol—Transport Layer Security) [46]. This architecture is used in private networks and in systems that make use of IoT devices, for the authentication of the user equipment (UE) and the virtual network. It consists of three main actors: user, authenticator, and server. It is focused on the use of symmetric keys and has the ability to manage user accessibility to virtual systems through user and system authentication.

To ensure the protection of medical data in telemedicine systems, Lin and Hsu [49] proposed an architecture for federated anonymous identity management (FAIDM). The presented scheme has three nodes: a restricted node, a gateway node, and a remote server node. The restricted node is in the layer of devices (body sensors and wearable devices) that collect physiological data. The gateway node is in the communication layer, and the remote server node is in the system architecture layer. This security scheme provides a key to the user for secure transmission of information. Validation is performed by mutual authentication. Additionally, it provides the devices with an anonymous identity to reduce the risk of patient data leakage.

UCSSO (User Controlled Single Sign-On Authentication) is another authentication system proposed by Lin et al. [53], which through authenticated session keys can establish a reliable communication channel between patients and services. In addition, the user can use a smartcard to access the server securely. The proposed system consists of four stages (initialization stage, registration stage, authenticated password exchange stage, and offline key change stage) that guarantee the privacy of the user’s medical data.

The protocol proposed by Minahil et al. [54] consists of three stages: registration, login, and authentication. In the registration stage, the user and the server are connected through the Internet. The server registers and provides the user with a smart card. In the login stage, the user uses a smart card reader to communicate with the server. Additionally, in the authentication stage, if the user is legal, he can make use of the services provided by the server.

3.3.2. Encryption

Note that 18.18% of the studies were based on encryption as a security protocol and focused on the use of encryption algorithms. That is, algorithms were used to encrypt the sending of information. In fact, in an implemented laparoscopic surgery system, Zheng et al. mentioned that the security was provided by a VPN (virtual private network) encryption system [38]. However, the authors did not define the mechanism. In [44,52,59] the authors briefly mentioned that their systems use data encryption, without detailing the methodologies or protocols implemented.

In the same way, Hameed et al. [50] proposed an intelligent system with neural networks that stores information in a distributed database (blockchain), which uses an encryption mechanism to protect against cyber attacks. It consists of a blocks chain that contains information, in which each block is connected to the next in a mathematical way, and if an attempt is made to modify the information of any block, the chain is broken; the rest of the chain and the information contained remain unchanged. It allows digital information to be stored in a secure and fragmented manner, which allows providers to generate safer and more complete patient history storage; that is, the information is protected in such a way that it cannot be modified or deleted.

3.3.3. Coding

Finally, encryption is another of the mechanisms used to guarantee and protect the confidentiality of smart health systems from attacks that seek to steal information for impersonation or invasion of privacy. Only 2.27% (n = 1) of the clinical applications used this mechanism.

Gupta et al. [29] employed physical network layer (PCN) encryption in telesurgery systems, because the use of cryptographic algorithms requires more resources due to their complexity and could reduce system efficiency. This technique is applied to protect wireless network information and to achieve high-speed, low-latency communications. It uses some PCN techniques, such as polar coding, low-density-parity check coding, and lattice coding. It uses the properties of wireless channels (noise, fading, and interference) to deteriorate a signal received by malicious users who try make use of the system. It allows secure, keyless transmission through the design and signal processing mechanisms. Unlike cryptography, this technique is not based on data encryption and decryption, so it can realize flexible security configurations for processing.

3.4. Correlation Analysis

In order to generate a correlation analysis between the different categorizations of the selected papers, tables were constructed to analyze each of the variables in pairs. This correlation analysis was based on the methodology developed by Guerrero-Vásquez et al. [62]. With the resulting tables, it is possible to identify trends to strong interrelationships that have been widely addressed, or on the contrary, possible research niches waiting for contributions.

The tables are shown in Appendix A. The values are based on a simple counts of the number of items crossing the specific subcategories. The dimensions analyzed are the study variables detailed:

• Paper type: theoretical study; proposal; implementation; prototype.
• Application area: remote care: surgery and diagnosis; real-time continuous monitoring; data transmission: smart ambulances.
• Healthcare benefits: constant and immediate attention with intelligent systems; specialized medical care in remote areas; immediate emergency assistance.
• IoMT devices: body sensors; teleoperated robots; artificial intelligence; wearable devices; ultra high definition cameras.
• Communication protocols: D2D; M2M; H2M; hybrid: D2D–M2M.
• Networking technologies: WiFi; Bluetooth; cellular network; hybrid: Bluetooth–NFC; hybrid: Bluetooth–ZigBee; hybrid: Bluetooth–WiFi; hybrid: WiFi–ZigBee–Bluetooth.
• Security protocol: authentication; encryption; coding; not defined.

With the intention of establishing a relationship between the different categorizations made of the 44 selected papers, an analysis based on the standard deviation of the count was added to the tables presented in Appendix A.

Figure 14 shows the values of the standard deviation of the counts of each table presented in Appendix A. On the one hand, the higher values represent less uniform combinations of variables, i.e., with marked tendencies toward certain point combinations. On the other hand, lower values represent a more uniform distribution around the mean, without considerable variations representing drastic trends.

The highest result is observed in the crossing of the variables “Healthcare benefits” and “Application area”, with a value of 8.25. To understand this value, we can observe that between the subclassifications “constant and immediate attention with intelligent systems” (healthcare benefits) and “real-time continuous monitoring” (application area), 23 studies cross over. Such a value is expected, since there is a logical relationship between the subcategories, where the application has a specific benefit.

Leaving aside the values that imply an explicit logical relationship, high results, such as those for the relationship between “Application area” and “Communication protocol” (SD = 5.33), merit a specific analysis. In this case, there is a tendency to prefer a hybrid communication protocol: Bluetooth–WiFi, for real-time constant monitoring applications (n = 12). Remote care applications mostly used the cellular network (n = 9), and the next most, the hybrid Bluetooth–WiFi protocol (n = 7). Such results provide a suggestion of the trends on which researchers have concentrated their efforts and can be interpreted as routes for new developments. Another way of interpreting the heat maps of the crossroads of papers is to consider those areas without overlaps as possible research niches that are waiting for contributions. Analyzing the same example above, it can be seen that the applications oriented toward information transmission use only the cellular network, without considering the possibility of inserting other long-range communication technologies, such as LoRa.

With this in mind, the heat maps presented in Appendix A can be analyzed individually for trends or gaps, and in both cases, interpreted as research opportunities. Although correlation analysis shows a summary of study variables characteristics, each heat map allows a more in-depth analysis. The first heat map (networking technologies vs. communication protocol) shows that the D2D communication protocol is present in all technologies; there is a clear trend towards a hybrid Bluetooth–WiFi network. Apparently, these network technologies are the favorites when developing smart-healthcare-related apps. The H2M protocol uses only the cellular network, leaving aside all other technologies. This trend is interesting, because one would expect more interactions using WiFi or Bluetooth. This can be explained by knowing that apps with H2M protocols are developed for remote sites whose only access technology is the cellular network. According to the heat map, the M2M protocol is the least used, and applications developed with it use complex systems of network technologies that combine Bluetooth, WiFi, and ZigBee. Despite the importance of the M2M protocol, as it allows the interconnection of devices without the need for human actors, it is apparently not of interest to researchers and developers. Similarly, few applications exist that combine M2M and D2D protocols (only 11.36%, n = 5), and in all cases, they combine Bluetooth and WiFi technology (two include ZigBee technology). This trend is striking, because smart city proposals tend to strengthen the interconnection between devices and machines, but our study reveals that smart healthcare applications do not prioritize this aspect. In the same way, the other heat maps can be analyzed and discussed, finding interesting opportunities for research and questioning current developments.

4. Conclusions

The results obtained by the authors show that 5G in smart healthcare is still under development in the stages of proposal and feasibility analysis. This is mainly because the technology is not implemented globally. Barely 15.91% (n = 7) of the studies reviewed involved of smart healthcare systems that were implemented and operational. However, after reviewing other related works, we saw a tendency to generate contributions related to research on 5G in smart healthcare, in theoretical proposals in the last several years [63,64,65], yet large-scale implementations and deployments have not gained momentum. In view of this, it is worth questioning whether the analyzed contributions really intend to achieve smart cities with real and functional smart healthcare applications or are simply marketing based on rhetoric.

Considering the application areas, these results indicate that information transmission is the area least addressed by researchers. This is due to the technical requirements involving uninterrupted communication with ultra-low latency and high speed aspects. Additionally, in comparison with previous works, in [8], it is mentioned that telemedicine, or remote care, is one of the most prominent applications which can be enabled with 5G. This is evidenced by the classification in Figure 5, which shows that 38.63% (n = 17) of systems are based on remote care.

In addition, according to “digital health trends 2022” study Cortes [66], digital health will continue being related to the pandemic caused by COVID-19, in order to assist patients in monitoring and managing chronic diseases. Undoubtedly, this relates to continuous and real-time monitoring, which in our research represents 54.55% (n = 24) of systems.

Based on IoMT technology, it is observed that body sensors (40.91%, n = 18) are the most used by different medical applications, for the transmission of physiological signals. Although wearable devices are also used for the same purpose, they only represent 6.82% (n = 3) of the research. This is explained by the fact that sensors fulfill a specific task, unlike wearable devices, which are devices dedicated to several tasks.

With the deployment of the 5G network, new remote-surgery systems are being developed which employ technologies such as teleoperated robots to operate on patients, even kilometers away from the surgeons. Nine (20.45%) of the studies presented used this technology, mainly in telesurgery systems. However, not all teleoperated robots were used in these clinical scenarios, as they are also used for other types of applications, such as robot-assisted remote ultrasound systems.

We note that another key technology in the development of intelligent healthcare systems is AI, which, through machine learning algorithms, enables the diagnosis, monitoring, and prevention of diseases. Undoubtedly, this coincides with what was mentioned in a report published by the WHO, in which they highlight that AI has the capacity to improve the provision of medical services worldwide. In addition, they remind the readers that in the implementation of AI systems, ethical principles and people’s rights must be respected, since, as with other technologies, they could be misused [67].

We observed that the communication protocol most used by the different systems is D2D communication (59.09%, n = 26), since this technology allows direct connection of medical devices with or without network assistance [17] and is therefore widely used. In addition, it was evident that systems using teleoperated robots employ H2M communication, although in a lower percentage (27.28%, n = 12), compared to D2D. Mainly, this is due to the fact that the 5G network is not yet available in all territories, so the contributions with this type of H2M interaction are still limited. It is expected that in the future when 5G is fully implemented, H2M communication will increase, enabling new remote-surgery systems, which is the field where this communication is required.

Additionally, it could be seen that the body sensors and wearable devices used in the different smart healthcare systems use short- and long-range communication technologies to share patients’ medical information. As was observed in Figure 12, 40.91% (n = 18) of smart healthcare systems used Bluetooth and Wifi as communication technologies. However, Bluetooth is also employed in other systems, along with short-range technologies such as NFC and Zigbee. In other words, it is widely used in the vast majority of the works presented. This is because Bluetooth is compatible with all types of body sensors and wearable devices, such as oximeters, glucose monitors, and smartwatches. This makes it adaptable to most of the medical applications presented. Still, it was observed that the cellular network plays an important role in the transmission and communication of intelligent health systems, especially in those that generate large amounts of medical data, such as in the case of telesurgery and ambulances, where reliable and uninterrupted communication is required for direct communication during the transmission of medical information.

Regarding the technical requirements, it was observed that the presented smart health systems require improved mobile broadband, and highly reliable and low latency communications. However, as shown in Table 4, most of the authors focused mainly on system latency, since this parameter is of vital importance for all medical applications (monitoring and remote care), but mainly for telesurgery scenarios, where the patient’s life is at risk. Uninterrupted communication is necessary to safeguard their safety. A delay or disconnection in this type of intervention could be fatal. Along with latency, parameters such as data rate, bandwidth, and uplink and downlink speed are fundamental to smart healthcare systems, as they allow the transmission of medical data in a fast, accurate, and reliable way. However, not all the works identified detail the technical specifications of their systems.

When analyzing the security systems or protocols employed by smart healthcare systems, it was observed that most studies do not address this issue. This feature should be considered by researchers or developers, as the privacy protection of patient information and medical systems must be safeguarded. In addition, an information leak or a cyber attack could put human lives at risk.

The reviews and states of the art studies presented above mostly do not indicate the technical requirements of the systems. They do mention security for cyber protection, but most do not present any mechanism or architecture. However, in our work, we have addressed the technologies, technical requirements, and security that are part of a smart healthcare system with 5G, intending to provide the reader with a broader and more complete picture.

Today, smart healthcare is key to safeguarding people’s lives, especially in the current situation of the healthcare system, which has been violated worldwide by the presence of the COVID-19 pandemic. This is supported by the study published by WHO on 31 August 2021 [67], which presents a compendium of technologies for the fight against COVID-19 in order to support countries in the improvement of health services. The need to migrate to intelligent health systems to safeguard patients’ health and improve access to medical services is notable.

Future work will include the development of a SWOT analysis to provide a more complete perspective on existing technological supports. In addition, based on the correlation analysis, it is possible to make proposals for the development of technological supports in lines of research that are not currently being studied. Each of the empty spaces within the analysis represents an opportunity to propose alternatives and to be pioneers in this area.