Overview

According to the Global Burden of Diseases, Injuries, and Risk Factors Study (GBD) 2019, cardiovascular diseases (CVDs) remain the leading cause of death globally, with the total number of CVD cases increasing from 271 million in 1990 to 523 million in 2019, and deaths rising from 12.1 million to 18.6 million during the same period.¹ 

The rise of cardiovascular diseases as a leading risk factor for mortality comes in lieu of an epidemiological transition as a result of the constant progressions within the scientific field and healthcare services; this means that the previously leading cause of death, communicable diseases, are adequately managed by recent scientific and medical advances, such as the discovery of antibacterial and antiviral medication. Thus, this has led to the dominance of chronic, non-communicable diseases such as cardiovascular diseases. 

Cardiovascular diseases are both complex and multifactorial in nature – meaning that early detection and prevention are vital in managing the condition appropriately. This can be achieved through artificial intelligence by leveraging predictive analytics and facilitating early intervention. 

AI and predictive analysis

Artificial Intelligence (AI) is the simulation of human intelligence processes by machines. Some of these processes include natural language processing (NLP) and predictive analysis. 

Natural language processing is a subset of AI. It refers to the interaction between the computer and the human language, and how computers interpret human languages in a meaningful way. NLP tools achieve many tasks to enable computers to understand human languages: text processing, linguistic analysis, sentiment analysis and question answering.² NLP tools are useful in medicine as they allow computers to extract useful clinical information from patient records and notes, which can be utilised for secondary analysis.³

What this can also do is train an AI model using current patients using previous patients’ clinical notes and records to predict future cardiovascular events, such as heart attacks and strokes: this is what we call predictive analysis. 

An example of this is the Framingham Heart Study: this is a long-term cohort study that demonstrated the use of AI models to predict cardiovascular risk in an individual patient with remarkable accuracy. This is achieved by analysing data points like blood pressure, cholesterol levels, smoking status and age.⁴ The prediction from these models allows clinicians and other healthcare providers to formulate a personalised prevention strategy, thus improving patient outcomes.

AI and early intervention

Early intervention is vital to avoid the fatal impact of cardiovascular diseases. This is enabled by continuous monitoring, personalised treatment plans and well-timed medical responses.

Wearable devices

AI can allow for continuous monitoring of the patient through wearable technology.⁵ Examples include smartwatches and fitness trackers that have become more and more sophisticated over the years. They are made with sensors that enable the remote monitoring of many factors including:

• Heart rate
• Blood pressure
• Oxygen saturation
• Physical activity
• Sleep patterns
• Electrocardiograms (ECGs)

These devices produce a constant generation of data that AI tools analyse in real-time to detect any abnormalities in the patient’s physical well-being. An example of this would be AI-powered wearables that can identify irregular heart rhythms that indicate atrial fibrillation, a common risk factor for strokes. Previous methods may miss episodes of intermittent atrial fibrillation, but continuous, real-time monitoring with artificial intelligence can detect these episodes as they occur, allowing patients to seek medical attention in a timely manner before a major cardiovascular event like stroke occurs.⁶

Personalised medicine

Artificial intelligence has the ability to analyse vast, complex datasets, also known as Big Data. This enables the development of personalised medicine in which medical care is provided in a way that is completely customised to the individual characteristics of the patient. Artificial intelligence allows for the integration of genetic information, lifestyle factors and clinical data, which allows us to predict how individual patients respond to various treatments.

In terms of cardiovascular health, personalised medicine can be aided by artificial intelligence in two ways: pharmacogenomics and treatment optimisation. 

Pharmacogenomics refers to artificial intelligence’s analysis of genetic data to determine how a patient will metabolise and respond to specific medications, thereby allowing the clinician to prescribe the most effective drug with the least side effects to the patient. For example, patients with certain genetic markers or of a certain age range might respond to specific blood pressure medication or cholesterol-lowering drugs.⁷ 

Treatment optimization refers to the optimal combination and dosage of medications for individual patients. AI models can analyse data on how different patients respond to treatment and suggest personalised drug regimens that maximise efficacy and minimise side effects.

Clinical decision support systems

Clinical decision support systems refer to the use of artificial intelligence to assist healthcare providers in making informed decisions. These systems analyse patient data and using current guidelines, they provide evidence-based recommendations for diagnosis, treatment and management. Clinical decision support systems can help clinicians to not only identify high-risk patients but also recommend appropriate diagnostic tests and suggest optimal treatment plans.⁸

For instance, a clinical decision support system might analyse a patient’s electronic health record and identify a pattern consistent with early-stage coronary artery disease. To investigate this further, the patient may recommend further diagnostic testing, such as a stress test or coronary angiography, and suggest lifestyle modifications or medications to prevent disease progression. This system enhances the quality of care by supporting clinicians with data-driven insights, thus improving patient outcomes.

Challenges and ethical considerations

Having outlined the immense potential of artificial intelligence in cardiovascular health, the challenges of this field should also be addressed.

Data quality

The predictions produced by an AI model are heavily dependent on the quality of data that the model was trained on. Incomplete, inaccurate, or biased data can lead to the model predicting incorrectly, thereby providing suboptimal treatment recommendations.

Privacy and security

Artificial intelligence is used to handle sensitive patient data, raising concerns about privacy and security, and ensuring that patient data is protected. This can be achieved through regulations such as the Data Protection Act (DPA) in the United Kingdom which provides guidelines for data protection.

Ethical considerations

AI models are trained on existing datasets in healthcare, which means that the trained AI models can inadvertently perpetuate existing biases in healthcare if the training data reflect historical biases in medical treatments. Ensuring that AI systems are fair and unbiased is vital in order to use these technologies with trust and confidence.

Another matter to consider is accountability in terms of the clinical decision-making process. If the clinical decision support system provides a recommendation that ultimately leads to an adverse outcome, is it the fault of the clinician or the AI system? Guidelines and regulations regarding this must be outlined clearly in order for artificial intelligence to enhance and not undermine patient care.

Summary

Artificial intelligence can be integrated with genomic data in order to provide precise risk predictions and treatment recommendations, and with the decreasing cost of genomic sequencing, incorporating genetic information into routine clinical practice will become increasingly prevalent. AI and cloud computing will advance in the future and will pave the way for real-time analysis of patient data, which is particularly valuable for the early signs of deterioration in those with chronic cardiovascular conditions, invoking prompt intervention and potentially preventing hospitalisations.

AI can also allow for the improvement in accuracy and efficiency of image analysis when it comes to cardiovascular imaging like ECGs, CT scans and MRIs, and detecting cardiac abnormalities. This will reduce the burden on radiologists and improve diagnostic accuracy.

Artificial intelligence can enhance the detection, prevention and management of cardiovascular diseases, but this must be done with caution due to concerns relating to data quality, privacy, security and ethical considerations. As technology advances, artificial intelligence will grow to hold a prominent role in improving cardiovascular health outcomes, thus transforming healthcare delivery.

AI has a huge scope for the predictive analysis of cardiovascular health through the use of wearable devices, personalised medicine and clinical decision support systems. In the future, the role of artificial intelligence will only grow with technological advancements.