The emerging advances in technology promise to help man solve some of the most complex health challenges, and Artificial Intelligence (AI) seeks to play a leading role.
AI, a wide-ranging branch of modern technology concerned with building smart solutions, is currently being used to perform health tasks that typically require human intelligence, in a better, faster and more accurate manner.
The ideal characteristic of AI is its ability to rationalize and take actions that have the best chance of achieving a specific goal, using data.
As the Fourth Industrial Revolution unfolds across the world, disrupting every sector, medical practitioners and research centres are increasingly relying on the technology to help control cardiovascular diseases.
“AI does things that ordinary human beings cannot store in their long term memory. It can monitor blood flow and predict adverse effects accurately,” says Prof Bitange Ndemo, chairperson of the Blockchain and AI taskforce.
Rheumatic, congenital, coronary, cerebrovascular, peripheral arterial heart diseases and deep vein thrombosis are some of the maladies that doctors world over are optimistic will be controlled using AI in the next few years.
The development of Machine Learning and Deep Learning – subtypes of AI – has exhibited AI’s application in controlling cardiovascular diseases, with studies demonstrating that it can predict strokes and heart attacks.
“AI is the technology that can be applied anywhere. You build the AI technology using Machine Learning where you teach a machine to remember tasks from where it can act on its own when it notices a deviation from the norm,” notes Ndemo, who is also a professor at the University of Nairobi’s Business School.
Two ways of control
According to Dr Beatrice Kariuki, an early adopter of AI and data science at local data firm Predictive Analytics, there are two main areas where AI in health can impact the control of heart diseases.
“We have ‘test and treat’ and ‘predict and prevent’. Test and treat interventions apply to persons who already have a disease or condition. The main goal of these is cure or managing symptoms where disease conditions are not treatable. Patients who have suffered a heart attack or stroke fall in this category,” explains the medical doctor.
In ‘predict and prevent’, she expounds, the use of AI would be most beneficial in the early detection of the disease, giving medical teams more time to proactively react and personalize treatment for better recovery outcomes.
Both a heart attack and a stroke are medical emergencies requiring the right care at the right time to avoid death and minimize disability. Right care at the right time is resource intensive but it is not always available.
“The use of AI to classify persons by risk levels is helpful. High-risk individuals are offered preventive programs to prevent cardiovascular arrests. Thus, AI helps to make informed decisions which are critical to millions of people suffering from heart diseases in the world,” she accentuates.
Opportunities also exist for the use of AI at a population level, for example, in identifying location-specific risk factors contributing to control diseases of the cardiovascular system.
Generally, AI applications targeting ‘test and treat’ address acute conditions while those applications targeting ‘predict and prevent’ address chronic conditions.
Speedy and accurate diagnosis of acute conditions is desired and applications that use algorithms and machine learning for diagnosis are in various stages of testing.
“It is envisioned that these will be used and replace clinicians for diagnosis. It has been possible to use AI to hasten diagnostics for cardiac diseases as most diagnostic tests are already in digital formats such as radiological imaging ECG, X-rays and Magnetic Resonance Imaging (MRI) CT scans,” says Dr Kariuki.
Genetic, behavioral and environmental risk factors are the cornerstone of ‘predict and prevent’ interventions in chronic heart diseases and conditions.
Behavioral preventive interventions, she says, are more widespread due to ease of use and acceptability to the population.
Common AI applications use a wearable device to track select physical and behavioral indicators like vital signs over a period of time, establishing a base pattern for an individual.
Real-time AI monitoring of changes in patterns are used to flag potential emergencies and worsening of diseases.
Wearables also make the tracking of health behaviors like exercise, sleep and nutrition easy and fun to track and empower and individuals to decide how they would like to respond according to the observed behavior.
New findings
Findings of a new study published on February 14, 2020 by the Circulation journal reveal AI’s unmatched potential in this field of medicine.
Led by Dr Kristopher Knott, a research fellow at the British Heart Foundation, the team conducted a study on cardiovascular magnetic resonance imaging (CMR) and AI, involving more than 1,000 participants, at the St Bartholomew’s Hospital and the Royal Free Hospital in London.
“CMR is a scan that measures blood flow to the heart by computing the difference in heart muscle strength. A stronger flow of blood means there is a smaller possibility of heart vessels getting blocked,” says Dr Knott.
Study participants were people who needed a CMR scan because they had been diagnosed with a heart condition or because doctors suspected they were exposed to a stroke or heart attack.
The model was then used to predict which of the patients were more likely to suffer from the pestilence. Researchers then followed up the patients for 20 months to measure the outcomes.
The AI-generated results were then compared with those given by doctors who had manually looked at the scans.
When patient outcomes for both sets were compared, the blood flow measurements calculated by the AI models were found to be accurate predictions of future heart disease occurrences.
According to Dr Knott, the process of reading the scans by humans takes a lot of time and is more dependent on what the medic will see as opposed to solid scientific fact.
This exposes the results or diagnosis to glaring errors if the doctor is exhausted or has poor eyesight.
“The predictive power and reliability of the AI was impressive and easy to implement within a patient’s routine care. The calculations were happening as the patients were being scanned, and the results were immediately delivered to doctors.
“As poor blood flow is treatable, these better predictions ultimately lead to better patient care, as well as giving us new insights into how the heart works,” says Dr Knott.
Scientists found out that the risk of death from a heart risk nearly doubled for every decrease in blood flow to the heart by 1 millilitre per gram per minute or 1ml/g/min.
At the same time, the risk of having a heart attack, stroke or other heart infirmities more than doubled.
Dr Knott said the findings were proof that AI can be used to spot potential cardiovascular attacks in patients, therefore enabling doctors to take preventive measures that can save their lives.
“Rather than a qualitative view of blood flow to the heart muscle, we get a quantitative number,” he says.
“And from that number, we have shown that we can predict which people are at higher risk of adverse events,” he adds.
Involving AI in the analysis of blood flow as part of standard tests for heart disease patients could prove to be a critical milestone in efforts to reduce the incidence of heart afflictions.
Is Kenya ready?
However, the globe is moving at different speeds in the adoption of AI in health, as these advances are only being witnessed in developed economies, leaving third-world nations stuck in traditional methods of diagnosis and treatment that are slow and inaccurate, giving room for a higher death frequency.
Medical practitioners and health ministries in developing economies have been dragging their feet in adopting the technology that is already moulding the future of medicine.
“Although developing populations are already using AI in their smartphones, they need to develop more awareness and capacity to deploy it in healthcare, which is a key aspect of economic progress,” says Prof Ndemo.
In Kenya and Africa, the adoption of AI in medicare has been minimal since the technology requires a health information governance and leadership framework, infrastructure, policies and laws to provide guidance to any interventions.
“Health data is very personal and confidentiality standards that exist internationally and locally must be met. This makes health AI projects more costly than non-health projects,” observes Dr Kariuki.
The adoption of AI in health, experts hope, will be shaped greatly by the new Data Protection Law that was assented last November.
Governing documents that provide common blueprint for health information enterprise architecture for Kenya, data ownership and data sharing within the health sector will be foundational regulatory and infrastructural pillars required to create the conducive environment for AI to flourish in Kenya.
Otherwise, as in the case of most wearables, the data generated is prepackaged and availed to the individual through a dashboard but the big data is owned by the device maker.
“Kenya and Africa need a clear commitment to transition from consumers of AI to developers. Local health challenges need AI interventions that are locally developed from our data,” Dr Kairuki advises.