A Review of the Scope, Future, and Effectiveness of Using Artificial Intelligence in Cardiac Rehabilitation: A Call to Action for the Kingdom of Saudi Arabia

ABSTRACT

The burden of cardiovascular disease (CVD) is still rising after decades of growth. CVD is the leading cause of death in the Kingdom of Saudi Arabia (KSA), accounting for more than 45% of all CVD-related deaths. The development of cardiac rehabilitation (CR) programs is significantly aided by artificial intelligence (AI) approaches, including machine learning and deep learning models. However, the KSA has a limited supply of CR, and AI methods are unavailable. This review aims to assess contemporary research on AI approaches’ application, potential, and efficacy in CR as a call to action for harnessing it in the KSA. Using the keywords artificial intelligence, AI, machine learning, deep learning models, cardiac rehabilitation, and Saudi Arabia, electronic databases of PubMed, CINHAL, Web of Science, PEDro, and SCOPUS were searched to find relevant articles. Evidence from the literature supports the idea that using AI techniques in CR can improve the ability to effectively diagnose more patients in areas without doctors in Saudi Arabia. Using AI in CR has constrained CR resources, which will lessen the need for outsourcing and enhance healthcare. Patients can receive an accurate diagnosis online thanks to machine learning algorithms and the expanding capabilities of AI.

Introduction

Cardiovascular diseases (CVDs) are the most significant cause of death worldwide and a major contributor to disability (Roth et al. 2020). The prevalence of total CVD doubled to 523 million in 2019 from 271 million in 1990. The CVD deaths steadily increased from 12.1 million to approximately 18.6 million in the same period. Years with disability doubled from 17.7 million to 34.4 million over that period. Moreover, the CVD burden continues its decades-long rise for almost all countries outside high-income countries (Roth et al. 2020). The age-standardized CVD rate has risen alarmingly, whereas it was declining in some high-income countries, particularly in the Kingdom of Saudi Arabia (KSA) (Roth et al. 2020).

KSA is the largest country in the Arabian Peninsula with an area of 2,150,000 km² (Bindawas and Vennu 2016), a population of more than 35 million (GASAT 2022), and a rapidly adapting increasingly urbanized lifestyle in the country (Soofi and Youssef 2015). Parallel to the rapid urbanization and developing economies in KSA, more than 45% of all deaths account for CVDs (Aljefree and Ahmed 2015), which is incredibly high among women (Alshaikh et al. 2016). In addition, the world’s biggest killer of ischemic heart disease (IHD) is contributing to a mortality of 7–14 for both sexes per 100,000 populations per year in the country (Huang et al. 2021). There is a call to action for secondary prevention in KSA.

Cardiac rehabilitation (CR) is an effective secondary prevention strategy for CVDs related mortality and morbidity cost-effectively (Balady et al. 2007; McMahon, Ades, and Thompson 2017; Wong et al. 2012). Shreds of evidence from randomized controlled trials and meta-analyses suggest that CR is cost-effective (Edwards et al. 2017; Shields et al. 2018), especially with exercise decreasing morbidity and mortality compared to usual care. This significantly reduces risk factors, improves clinically meaningful health-related quality of life (McGregor et al. 2020), and promotes a healthy lifestyle in patients with cardiovascular disease (Anderson and Taylor 2014; Edwards et al. 2017; Xia et al. 2018). However, home-based training appears more cost-effective, with higher satisfaction than center-based training for patients with low-to-moderate cardiac risk (Bravo-Escobar et al. 2017; Edwards et al. 2017). CR for low-risk patients can deliver equivalent benefits, if not more than medical therapies (Naci and Ioannidis 2015). The results of the recent meta-analysis indicated that technology-assisted cardiac rehabilitation is a potential alternative to center-based CR to remove CR barriers during the prolonged pandemic (Chong et al. 2021). In addition, a recent systematic review suggested that harnessing artificial intelligence (AI) in CR holds excellent potential for early detection of cardiac events, allowing for home-based monitoring and improved clinician decision-making (Sotirakos et al. 2022).

AI techniques, such as machine learning (ML), deep learning (DL), and cognitive computing, are on their way to having an impact on practically every aspect of human conditions, and cardiology is no exception (Mathur et al. 2020). AI techniques can radically change how cardiology practices, especially imaging, interpret data and make clinical decisions (Romiti et al. 2020). AI techniques can also improve medical knowledge due to the increase in the volume and complexity of the data, unlocking clinically relevant information (Murdoch and Detsky 2013). Likewise, AI techniques are becoming pivotal to creating a pervasive healthcare service through which patients with CVDs can receive CR at their homes, reducing hospitalizations and improving their quality of life (Sotirakos et al. 2022). Thus, harnessing AI techniques in CR can play a critical role in the early detection and diagnosis of CVDs, along with outcome prediction and prognosis evaluation (Sotirakos et al. 2022). Moreover, AI techniques are required to interpret massive datasets, such as quantitative, qualitative, and transactional data (Murdoch and Detsky 2013). As a result, physicians can make better clinical decisions by detecting clinically relevant information that can be found in the massive amount of data (Jiang et al. 2017; Visco et al. 2021).

Despite these benefits, a recent review of the literature suggested that tremendous gaps exist in how many patients can access CR worldwide (Turk-Adawi, Sarrafzadegan, and Grace 2014). According to that review, CR services are available in less than 40% of countries worldwide. More centrally, the characteristics of CR programs, the components delivered, their capacity, and the degree of patient access are known to be insufficient, including in Latin America and the Caribbean (LAC) countries (≥1 CR program in 24 LAC countries) (Chacin-Suarez et al. 2021). Given this uncertainty, the previous global survey ascertained CR availability, volumes and their drivers, and density (Turk-Adawi et al. 2019). According to that survey, CR is available in 54.7% (111 out of 203) countries globally (Turk-Adawi et al. 2019). The survey also indicated the capacity of 5,753 programs serving 1,655,083 patients/year is grossly insufficient because of an estimated 20,279,651 incident IHD cases globally per year. The average of 1,677 programs in 12 Eastern Mediterranean countries served a median of 120 (Q25–Q75 = 93–227) patients per year each. In KSA, only 1 CR program helps a median of 90 patients per year each, as shown in that global survey (Turk-Adawi et al. 2019).

A recent review indicated four established CR programs in KSA (Rashed et al. 2020). According to that review, the Physical Therapy Department in the King Faisal Specialist Hospital and Research Centre in Riyadh implemented the first CR program. Another CR service for cardiac patients was implemented at Prince Sultan Cardiac Center in Al-Ahasa. However, there is no specific information about the characteristics and duration of the program. In addition, King Khalid University Hospital (KKUH) and King Abdul-Aziz University Hospital (KAUH) have CR facilities. Still, there are several obstacles regarding patient safety standards, limited capacity, and staff qualifications (Rashed et al. 2020). Given the enormous burden of CVDs (Aljefree and Ahmed 2015; Huang et al. 2021) and the vastly insufficient availability of CR in KSA (Turk-Adawi et al. 2019), there have been no comprehensive studies that have explored AI techniques in CR delivery (Sotirakos et al. 2022). Thus, this article aims to review the recent literature on the scope, future, and effectiveness of AI techniques in CR as a call to action for suggesting harness of it in KSA.

Method of Literature Collection

It is customary to make the paper selection criteria for a literature review clear to ensure representativeness and confidence in the information source. The method for choosing the papers is explained as follows:

The three main components of the search scope are AI, healthcare, and CR, covering the spectrum, future, and efficacy of applying AI in CR. The collection scope is based on the primary content of this work. An incomplete reverse snowball search is conducted on the article’s reference list identified by an early keyword search to uncover more pertinent papers that the keyword might not have acquired.

The categories dictate the keywords for each part. For instance, “machine learning” OR “deep learning” AND “healthcare” are applied after fusing the keywords as AI can be divided into ML and DL according to the type of extracted characteristics. The terms “artificial intelligence,” “machine learning,” “deep learning,” and “cardiac rehabilitation” are combined to create the keywords for CR. The CR in the KSA uses the same searching procedure.

The literature that was primarily found comes from journals. The most significant pertinent works have been considered, including research articles, randomized studies, systematic reviews, and meta-analyses published in English. Most of the reviewed articles compiled in this report came from the electronic databases of PubMed, Web of Science, CINHAL Plus, and PEDro. In addition, external web archives like ResearchGate, SCOPUS, and Google Scholar are utilized.

The title and abstract of the papers that were discovered during the search were used to identify and categorize them. I studied the full texts of the relevant articles to ensure they met the criteria for high relevance and high-quality material, and I then considered them if I thought they were appropriate. As a result, collecting existing articles could serve as a good foundation for this paper’s review effort.

Cardiac Rehabilitation

The World Health Organization (WHO) is defined CR as “The sum of activity and interventions required to ensure the best possible physical, mental, and social conditions so that patients with chronic or post-acute CVDs may, by their efforts, preserve or resume their proper place in society and lead an active life” (WHO 1993). CR, also known as cardiac rehab, is a customized inpatient and outpatient program comprising exercises and education (Kalix 1988). CR programs are patient-centered and delivered by a multidisciplinary team (such as a cardiologist, nurse, physiotherapist, dietician, and psychologist) to improve patients’ quality of life, psychosocial well-being, and functional capacity (Francis et al. 2019). They are cost-effective (Shields et al. 2018). CR is integral in providing comprehensive care to patients suffering from CVDs. Therefore, it plays a critical role in the secondary prevention of coronary heart disease (Balady et al. 2007; McMahon, Ades, and Thompson 2017). CR contains exercise programs, emotional support, and education regarding lifestyle changes to reduce CVDs risks, and lifestyle changes include eating heart-healthy food, maintaining a normal weight, and quitting smoking (Balady et al. 2007; Leon et al. 2005). All cardiac rehabilitation/secondary prevention programs, according to the American Heart Association (AHA) and the American Association of Cardiovascular and Pulmonary Rehabilitation (AACVPR), should include specific core components aimed at reducing cardiovascular risk, fostering healthy behaviors, and compliance with these behaviors, reducing disability and encouraging patients with cardiovascular disease to live an active lifestyle (Suaya et al. 2007).

Phases in Cardiac Rehabilitation

The CR consists of three phases (Suaya et al. 2007). Phase I is an inpatient program that occurs when the patient is hemodynamically stable in a safe, supervised setting within 48 hours of a cardiovascular event (T. J. Wang et al. 2020). In this phase, the patient performs passive and active movements to maintain muscle tone and reduce sedentary risks and unfavorable complications.

Phase II is an outpatient program that recommends patients begin within 2–7 days following a percutaneous intervention, 4–6 weeks after cardiac surgery, or the intervening 6–8 weeks after discharge from the hospital (Shajrawi et al. 2020; Zhang et al. 2018). Patients must generally obtain a physician’s referral to participate in this program because of fears of overexertion or a recurrence of heart issues (Shajrawi et al. 2020; Supervia et al. 2019). Typically, participation in this program begins with an intake evaluation of cardiac risk factors, such as lipid measures, blood pressure, body composition, depression/anxiety, and tobacco use (Grace et al. 2016). In addition, an exercise stress test is performed to determine if exercise is safe to allow for the customized exercise program (Supervia et al. 2019). Risk factors are addressed, and patients’ goals are established to achieve their targets through a cardiac-trained Registered Nurse, Physiotherapist, or exercise physiologist. The patient’s heart rate and blood pressure may be monitored during exercise to check the activity intensity (Supervia et al. 2019). Overall, the duration of CR varies from program to program and can range from six weeks to several years. Usually, the better well-established median of 24 sessions is offered worldwide (Chaves et al. 2020; Santiago de Araujo Pio et al. 2017).

Phase III is a long-term maintenance program available to interested patients after finishing phase II of CR (Chowdhury et al. 2021). In this program, benefits are optimized with long-term adherence. However, this program is cost-effective as patients generally have to pay out-of-pocket. CR programs are recommended by the American Heart Association/American College of Cardiology (Smith et al. 2011) and the European Society of Cardiology (Piepoli et al. 2016), among other associations (Guha et al. 2017) based on their role in proving CVDs outcomes (Taylor, Dalal, and McDonagh 2022).

However, CR is significantly under-used globally with widely varying rates because of causes by the following multi-level factors (Grace, Kotseva, and Whooley 2021; Santiago de Araujo Pio et al. 2020). The health system level (lack of available programs) (Turk-Adawi et al. 2019), the provider level (low referral rates by physicians) (Ghisi et al. 2013), and the patient-level factors, such as transportation, distance, cost, competing responsibilities, lack of awareness and other health conditions are responsible (Shanmugasegaram et al. 2012). Other factors, such as women (Samayoa et al. 2014), ethnocultural minorities (Midence et al. 2014), older patients (Grace et al. 2009), those of lower socioeconomic status with comorbidities, and those living in rural areas (Leung et al. 2010) are less likely to access CR even the fact that these patients often need it most (Ruano-Ravina et al. 2016). To mitigate these barriers to CR use (Santiago de Araujo Pio et al. 2019), it is crucial for treating cardiac patients using the phone (Falter et al. 2019; Parak et al. 2021) and other technologies to institute automatic, systematic, or electronic referrals to CR (Grace et al. 2011). Moving to a generally safe home-based CR (Thomas et al. 2019), which depends on remote monitoring and coaching tactics, has now been seen to be a viable option for boosting participation rates (Sotirakos et al. 2021). This was one of the most effective strategies used during the COVID-19 pandemic too wherein all these rehabilitation programs were seen to be carried out remotely because of not only the risk of infection but also the overburdened healthcare system. These issues have opened new doors for harnessing AI techniques (Mathur et al. 2020; Romiti et al. 2020) to penetrate the traditional CR programs and expand the spectrum of interventions that can be done by clinicians by overcoming the challenges faced by not only the patients but also the clinicians.

Artificial Intelligence in Healthcare

Artificial intelligence (AI) techniques are on their way to impacting practically every aspect of human conditions (Mathur et al. 2020). AI refers to the use of technology that can make decisions based on data (Duan, Edwards, and Dwivedi 2019). In clinical practice, AI unveils a new realm of possibilities (Khalsa et al. 2021). By evaluating patterns in patient data to predict outcomes and guide treatment, AI can be efficiently utilized in healthcare to provide constructive data to clinicians (Miller 1994). AI has spread its roots in almost every area of medicine (Patel et al. 2009; Schork 2019). The first application of AI occurred in 1976 when it was used to successfully identify severe abdominal pain via computer analysis (Gunn 1976). In the real world, medical practice is changing hand in hand with the development of new AI systems, and problems from different areas have been successfully solved using AI algorithms (Davenport and Kalakota 2019). In terms of illness discovery and targeted treatment, the application of AI approaches in setting up or creating precision medicine is critical (Schork 2019). Moreover, with technology, clinical personnel can deliver a very efficient healthcare service (Visco et al. 2021).

AI algorithms showed promising results in clinical (Bhinder et al. 2021; Blease et al. 2019; Cao et al. 2021; Hibler, Qi, and Rossi 2016; Tran et al. 2021); systems (Christopoulou et al. 2020; Liang et al. 2019; Pouke and Hakkila 2013), and industry settings (Pisarchik, Maksimenko, and Hramov 2019). For example, in clinical settings, AI algorithms are an initial triage tool for accurately diagnosing and risk-stratifying patients with coronary artery disease (Wang et al. 2021). The accuracy of machine learning models on clinician diagnostic ability has been compared directly in a few studies (Stewart et al. 2021). Patient mortality, medication effects, and adverse events following treatment for acute coronary syndrome have been predicted using other algorithms (Wang et al. 2021). The various settings of AI models potentially enabled earlier detection of cardiac events occurring outside the hospital (Sotirakos et al. 2022). The utility of AI in classifying heart sounds and diagnosing valvar disease is another growing area of research (Chen et al. 2021). However, the challenges of AI in limited data on social determinants of health as they pertain to cardiovascular disease (Zhao et al. 2021). Despite challenges, AI allows for a good experience by resourcing files to find the best treatment for patients in areas where healthcare is scarce (Guo and Li 2018). The ability of AI to adjust to a level of individualized care that is nearly non-existent in developing countries also allows the patient to have their treatment modified based on what works for them (Guo and Li 2018).

Artificial Intelligence in Cardiac Rehabilitation

Most CVDs are multi-system clinical syndromes with no single quantitative indicator representing the complexity of the data needed to determine disease status (De Cannière et al. 2020). As a result, to adequately portray the physiological condition of CVD patients, a multi-parametric approach is required. Sensory inputs by a system integrate measurements from a heart rate monitor, an inertial measurement unit (IMU) (reporting the treadmill inclination), a Laser Range Finder-LRF (to estimate gait parameters), and periodic results from the Borg scale, as well as a user interface and an autonomous humanoid social robot platform-NAO (pronounced as NOW) (SoftBank Robotics Europe, France). Such a system was designed by (Lara et al. 2017)to present the three primary metrics considered in CR.

First, gait spatiotemporal parameters measurements from an LRF are used to estimate the patient’s cadence, step length, and speed of the patient. This node estimates the parameters of the kinematics of lower limbs and performs filtering of the oscillatory components contained in the user movement intention. Second, regarding cardiopulmonary parameters, a heart rate monitor (Zephyr HxM BT) is located on the user’s chest and reports a wireless and continuous heart rate measurement using Bluetooth communication. Third, physical activity difficulty parameters are used to evaluate the biological activity difficulty: one being the inclination of the treadmill and the other the patient’s reported problem with the exercise. A tactile computer monitor (i.e., a tablet) uses a graphical user interface to measure the patient’s fatigue using the Borg scale. In a CR supported by a social robot, an NAO robot periodically asks the patient to report a value on the Borg scale entered on the tablet. Novel methods to help patients with cardiac problems outside of a hospital environment have been proposed using mathematical algorithms and mobile applications (Lara et al. 2017; Medina Quero et al. 2017).

AI for Quitting Tobacco Initiative

Lifestyle modifications play an essential role in CR, and quitting tobacco is crucial for many individuals (Balady et al. 2007). Amazon Web Services and Google Cloud have supported San Francisco and New Zealand-based Digital People company Soul Machines to create Florence. Florence is a robot that develops a tobacco quit plan that is outlined as follows (Kaplan et al. 2001):

• The patient is required to set a quit date. They must set up a quit date as soon as possible. The reason is that giving oneself a short period to quit will keep one focused and motivated to achieve a goal.

• The patient is advised to involve their friends, family, and coworkers since it is essential to share their goal of quitting with those they frequently interact.

• The patient is required to anticipate challenges to the upcoming quit attempt. The first few weeks especially are critical and the hardest to overcome due to the potential nicotine withdrawal symptoms and the obstacles usually presented while breaking any habit.

• The patient is asked to remove tobacco products from their environment to avoid distractions.

Mobilizations and Walking Monitors Using Smartwatches

The program in CR is embedded in a wrist-worn device with a heart rate sensor, which provides real-time physical activity monitoring during sessions safely and effectively. Moreover, wearable devices have been demonstrated to favor strategies for changes to healthy habits and the promotion of healthful physical activity (Case et al. 2015). For this, a linguistic approach based on fuzzy logic was proposed to model the cardiac rehabilitation protocol and the expert knowledge from the cardiac rehabilitation team. Fuzzy logic has not only provided successful results in developing intelligent systems from sensor data streams but has also been described as an effective modeling tool in CR (Peláez-Aguilera et al. 2019). The following steps can effectively achieve this:

• It integrates a high-quality protocol based on clinical guidelines for monitoring a patient’s heart rate in a personalized way that will suit the patient’s needs, like diet specifications, exercise type, and duration, and titrating the dose of medications regularly.

• A wearable wrist-worn device with a heart rate sensor provides real-time monitoring during sessions.

• A wearable-mobile-cloud platform is being used to collect and synchronize data between patients and the CR team.

Decision Support Systems (DSS)

DSS are computational applications that improve the abilities of an individual or group of individuals to make decisions (Main et al. 2010). In recent years, its use in clinical practice is gaining acknowledgment of its benefits in terms of reducing diagnostic errors (Linder et al. 2009), increasing conformity with clinical guidelines (Roshanov et al. 2013), and potentially increasing healthcare quality (Roshanov et al. 2013). There are different types of DSS, but the most suitable for clinical DSS is the knowledge-driven, which uses AI and statistical inference in codified knowledge to assist in specialized problem solutions, suggesting or recommending actions. Rules, clinical standards, or probabilistic networks can all be used to represent knowledge in a knowledge-driven clinical DSS (Kong, Xu, and Yang 2008).

Virtual Cardiac Rehabilitation (VCR)

A home-based CR program with guided exercise usually takes place post-acute CR. In the last two years, telerehabilitation has been implemented worldwide, and VCR was implemented to overcome the challenges created by the COVID-19 pandemic. An application that offers home-based CR through videos and monitors has been adapted as a solution to address this situation (Pérez-Robledo et al. 2021).

Sensor Derived Biomarkers

A reasonable interpretation of physical fitness is required to track a patient’s progress during a CR program. Various wearable sensors that can track multiple physiological signals reflective of physical fitness are available today, and their implementation within a cardiac population has been investigated extensively (Gevaert et al. 2020). However, for clinical acceptance, the easy interpretability of the AI models is crucial (De Canniere et al. 2020). The sensors can visualize the relationship between sensor-derived biomarkers and functional capacity, which monitors the clinical progression of patients throughout the CR program. These findings pave the way for the use of AI approaches that can be interpreted by healthcare professionals and help translate CR to a home environment (De Canniere et al. 2020). Studies have shown that the combination of sensor-derived biomarkers, i.e., during a standardized 6-Minute Walk Test (6MWT), chronotropic reaction, and effort can be used as a surrogate for functional capacity in a CR (De Canniere et al. 2020).

Future of Artificial Intelligence in Cardiac Rehabilitation

Covid-19 has thrown light on people’s capacity in terms of what they can do while they are stuck at home. The whole world has learned to function virtually and effectively as a consequence. There is no single problem in this world that does not have a solution. This being said, the day is not far from when we incorporate AI in our daily lives and make CR more simple with algorithms, devices, and virtual applications that can safely and accurately monitor safety, adherence, and compliance and ensure a successful treatment (Barrett et al. 2019; Gensini et al. 2017). This would be a significant path-breaking step in medicine and will result in a reduction in mortality and morbidity due to CVDs which are the most common cause of death worldwide, specifically in KSA. The AI techniques in CR provide various tools to enhance and extend the effectiveness of CR teams and programs (Johnson et al. 2018). Thus, more comprehensive and interdisciplinary courses ought to come, and innovative teams will bring out solutions with creative and brainstorming sessions in the future.

Because AI-based technologies assist clinicians in daily medical activities, enhancing illness identification and treatment, personalized patient care via telemonitoring may soon improve clinical practice (Visco et al. 2021). Indeed, AI may obviate much of the tedium of modern-day clinical practice, such as interacting with Electronic Health Records and billing, which will soon be intelligently automated to a much greater extent (Steinhubl and Topol 2015). Although personnel with specialized training often perform machine learning, these methods will become increasingly easy and commoditized in the future.

The expert knowledge of pathophysiology and clinical presentation that clinicians acquire over their training and career will remain vital. Therefore, clinicians should lead in deciding where to apply and how to interpret these models (Konstam et al. 2017). AI will help deliver improved patient care because clinicians can analyze more data in greater depth than ever. Reinforcement learning algorithms will be used as companion clinicians’ aids, inconspicuously supporting clinicians while expediting clinical treatment (Konstam et al. 2017; Steinhubl and Topol 2015). Advances in unsupervised machine learning will enable far greater characterization of patients’ disorders and ultimately lead to better treatment selection and improved outcomes (Shameer et al. 2018). Future physicians will be able to monitor, analyze, and respond to new streams of biological data collected remotely and automatically because of the spread of AI (Shameer et al. 2018).

A recent network meta-analysis of patients with heart failure, stroke, hypertension, and diabetes was performed by comparing the machine learning and deep learning models for CVDs prediction (Baashar et al. 2022). A piece of evidence from the network meta-analysis indicated that the deep learning algorithms predicted well in heart failure with an area under the curve of 0.843 [CI = 0.840–0.845], while the machine learning algorithm achieved an average accuracy of 91.1% in predicting heart failure. These findings suggest that the prediction of and knowledge about heart failure is effective with deep learning models, but these models in the field of CVDs are unknown. A very recent systematic review evaluated the efficacy of AI in cardiac rehabilitation using the current body of research in the literature (Sotirakos et al. 2022). Evidence from this review showed that AI-driven CR programs hold immense potential for patients and clinicians alike. Mainly, good clinical outcomes were observed with the telehealth delivery of CR, as shown previously (Jin et al. 2019). The evidence also showed that harnessing AI tools in existing telehealth cardiac rehabilitation programs could provide a personalized and adaptive rehabilitation program that responds to individual clinical needs (Falter, Scherrenberg, and Dendale 2020). Findings indicated that AI in CR improved communication between patient and clinician by increasing patient-clinician interaction and making cardiac rehabilitation more accessible to those in geographically or socially disadvantaged areas or during a pandemic (Widmer et al. 2017). Additionally, the review suggested that clinicians may benefit from AI-based tools by using clinical decision-making support tools or accessing a patient’s smartwatch data (Diciolla et al. 2015). Such a system may serve helpful in the early detection, diagnosis, and treatment of cardiac arrhythmias or in tailoring a rehabilitation program of care (Omura et al. 2017).

To summarize, the research indicates that CR programs are patient-centered, affordable, and provided by a multidisciplinary team. It is essential for the secondary prevention of coronary heart disease. It can help with a specialized program that includes workouts and education to enhance patients’ functional ability, quality of life, and psychosocial well-being. However, evidence from the literature revealed that due to the multi-level variables discussed above, CR is severely underutilized internationally and in the KSA, with rates that vary greatly. The evidence implies that using AI technology to lessen these barriers to CR use is essential for treating cardiac patients (Santiago de Araujo Pio et al. 2019). Given how AI techniques practically affect every area of human conditions (Mathur et al. 2020), a request is that the KSA utilize AI methods for identifying trends in patient data to forecast outcomes and direct care, especially in CR. Using algorithms, gadgets, and virtual applications that can reliably and safely monitor safety, adherence, and compliance to ensure a successful treatment, according to lessons learned from the COVID-19 epidemic, is another way that integrating AI in CR may be made simpler (Barrett et al. 2019; Gensini et al. 2017).

Artificial Intelligence Techniques in the Kingdom of Saudi Arabia

The KSA has taken severe and broad steps to harness AI techniques in various fields, especially after announcing a vision for 2030, in which the transition to digitalization is one of its biggest goals. Therefore, it considers AI an important strategic option to achieve this goal and its ambition to hold a highly distinguished position among the world’s nations. Moreover, the recent paper showed KSA’s efforts in developing AI techniques in different fields, especially in the education and healthcare sectors (Al-Jehani et al. 2021). The report showed that KSA has contributed by AI techniques to facilitating the performance of indispensable daily work, such as holding E-Learning, performing remote work, and even serving people who need medical information about Coronavirus. In addition, the paper showed the use of robots, innovative pharmacies, and some applications, such as TABAOD and TAWAKKALNA. Similar to how they help in daily medical tasks, AI approaches in CR across the KSA improve sickness recognition and treatment while customizing patient care. The monotony of contemporary clinical practice, such as communicating with electronic health records and billing, may also be much reduced by AI in the future. These tasks will soon be intelligently automated to a far more considerable degree.

Recommendations

The flowchart of recommendations is shown in Figure 1. In KSA, the most advanced AI industry techniques were invested in 2017 (Al-Jehani et al. 2021). However, this nation can consider a call to action for harnessing AI techniques in CR because more than 45% of all deaths account for CVDs (Aljefree and Ahmed 2015) parallel to the rapid urbanization and developing economies, exceptionally high among women (Alshaikh et al. 2016). In addition, the world’s biggest killer of IHD is contributing to a mortality of 7–14 for both sexes per 100,000 populations per year in this country (Huang et al. 2021). Thus, there has to be a call to action for secondary prevention in KSA using AI in CR, which is an effective secondary prevention strategy for CVDs related mortality and morbidity cost-effectively (Balady et al. 2007; McMahon, Ades, and Thompson 2017; Wong et al. 2012). Moreover, AI techniques can radically change the way cardiology practices (Romiti et al. 2020). Especially imaging, interpreting data, unlocking clinically relevant information (Murdoch and Detsky 2013), making clinical decisions, reducing hospitalizations, and improving quality of life (Sotirakos et al. 2022) aligned with the aspirations of Saudi Arabia’s vision 2030 (Sotirakos et al. 2022). Therefore, the capacity of AI to alter course is essential since it enables patients in the KSA to have their therapy updated based on what is most effective for them. Additionally, patients can receive an accurate diagnosis online thanks to sophisticated ML algorithms and the expanding capabilities of AI.

Figure 1. The flowchart of recommendations.

Conclusion

AI has roots in practically every area of medicine, including in CR. The algorithms used by AI have produced encouraging outcomes in clinical, system, and industry settings to give helpful information to physicians. The ability of various wearable sensors to measure multiple physiological signals indicative of physical fitness and their application within a cardiac population has been thoroughly studied. Artificial intelligence is predicted to simplify algorithms, hardware, and software in CR in the future so that it can accurately and safely monitor patient safety, adherence, and compliance. These results pave the path for applying AI techniques that support the translation of CR to the home environment and can be understood by healthcare practitioners. Therefore, in line with the KSA’s 2030 vision, this review issues a call to action for using AI approaches in CR. Because CVDs are the leading cause of death in KSA, using AI techniques in CR can increase the effectiveness of diagnosing more patients in KSA without doctors.

Acknowledgements

The author extends his appreciation for technical support to Dr. Vishal Vennu.

Disclosure Statement

No potential conflict of interest was reported by the author.