.png){kind=small}
Artificial intelligence (AI) and Generative AI are on track to adding trillions of dollars to the global economy through an increase in worker productivity by 0.1% to 0.6% into 2040 according to statistics by McKinsey. In particular, we see the impact of AI on global healthcare, offering solutions from advanced diagnostics to the management of chronic conditions. 38% of Americans think AI usage in healthcare would improve patient outcomes according to a Pew Research Center Survey. Here are six ways that AI is reshaping global healthcare systems, and reducing pressure on them due to supply chain and climate chain problems:
At the University of Oxford, researchers are demonstrating how AI can improve CT scans and their accuracy in detecting artery blockages and narrowing of arteries. AI’s analysis of CT scans outperforms traditional scans, providing a more thorough analysis and greater diagnostic accuracy. It is the case that regular CT scans struggle to pick up on some issues. As a result of AI’s involvement in cardiac imaging, there is a significant improvement in the prescription of appropriate medications and an increase in life-saving treatment earlier.
The AI tool named “Sybil” created by MIT represents a significant advancement in the fight against lung cancer. From a single low-dose chest scan, Sybil can predict the risk of lung cancer 1-6 years in advance. Its precision surpasses that of human radiologists checking images with the naked eye, and it can identify subtle abnormalities. Sybil has been trained by scientists at MIT using hundreds of CT scans. AstraZeneca and Qure.ai have their own lung cancer detection tool which is 17% more accurate than human radiologists. South Korea’s Lunit is similarly contributing with AI that pinpoints early-stage cancers and predicts patient responses to treatments, an essential step for personalized healthcare.
AI is speeding up drug discovery. Insilico Medicine’s approach to drug discovery using Generative AI shows how AI is revolutionizing the drug development process for rapid development with reduced risk.
Picture this: A protein target that has been associated with a variety of diseases by scientists is attacked using rapidly generated tiny drug molecules that contain properties to target these proteins. Generative AI is then used to predict the outcome of targeting the protein with specific drugs in clinical trials.
The failure rate for new treatments is costly, time-consuming, and comes with a high level of risk; AI works to reduce the degree of loss concerning all of these factors. Traditional approaches to drug discovery have a whopping 90% failure rate with a substantive cost of $2.6 billion if we start with a pre-discovered protein target through to clinical trials. This does not even include the decades of research it can take to find the initial protein target. With the advent of AI, the research required to find the protein target can be cut to approximately plus or minus 8 years.
“Rather than hunting for molecules with particular effects, Insilico’s AI can design new molecules with the required properties from scratch. One recent example is a potential treatment for fibrosis, a lifelong condition in which scar tissue builds up inside the body. Insilico uses AI to find a safe molecule chemical that targets fibrosis. This drug has undergone Phase I trials, and is now in Phase II. The probability of failure for pharmaceutical development is extremely risky, it’s a molecular casino. But now we’ve managed to demonstrate something that has achieved 1% probability. We took a new novel target that has never been in human clinical trials before, generated novel molecules for it, and took it into human clinical trials and reached Phase II. It is a very low probability event for one company to do internally, especially for an AI company.”
-- Alex Zhavoronkov, PhD, founder and CEO of Insilico Medicine
The application of Generative AI on large data sets is streamlining the tedious and error-prone documentation processes within healthcare. This AI technology is currently being tested in various aspects of healthcare documentation, such as creating detailed visit summaries and aggregating pertinent research to support patient care decisions. Automating these administrative tasks can significantly reduce the workload of healthcare professionals, allowing them more time to focus on patient care.
An app, called “PainScale”, is being used to determine patterns from patient pain input data – giving insight into the environmental factors such as diet and physical activity which are improving or worsening a person’s pain levels. A tool called “Elipsis” can detect an increase in pain levels by analyzing changes in voice patterns caused by a patient’s fluctuating mental state, which are interrelated.
The anonymous AI chatbot called “Wysa”, accessible through the app, online, and on WhatsApp, provides mental health support by communicating directly with users through natural language processing and providing evidence-based therapeutic methodologies created by specialist therapists, such as breathing techniques for relaxation and prompts for recentering and reframing deleterious thoughts. It was created to reduce barriers to mental health support.
“Wysa started with my own struggle with mental health. As I was struggling, I began to realize that even people like me who are privileged have so many access barriers to mental health and how bad the situation could be for everybody else. And once you see the issue, it’s very hard to unsee it. We started out by trying to create an interface that could use natural language processing and be a place where people could talk about their issues without worrying about being seen by anybody else but still feeling heard. We began to realize that it’s reaching so many people who otherwise have no access to mental health support and it’s really helping them.”
-- Jo Aggarwal, CEO of Wysa
Wysa has reached millions of users worldwide and continues to expand its language offerings to extend its reach and impact.
