Home / Resources / News and Trends / ISACA Now Blog / 2019 / AI and Healthcare A Life-Saving Combination


AI and Healthcare: A Life-Saving Combination

Author: Ryan Abdel-Megeid
Date Published: 3, December 2019

Artificial intelligence (AI) and machine learning are common terms in the world of emerging technology. Although still sounding futuristic to some people, AI is already being deployed everywhere from fantasy football weekly recap emails, to retail environments, to advanced, state-sponsored surveillance systems. In ISACA’s Next Decade of Tech: Envisioning the 2020s research, a recent survey of more than 5,000 global technology professionals, 38% of respondents expect AI and machine learning to be the most important enterprise technology of the next decade – more than cloud platforms (22%), big data (16%) and even blockchain (8%). Ballooning costs, labor shortages, poor service quality, strong public interest, and recent market shifts forcing the enhanced availability of electronic records are strong indicators that few industries will experience the impact of AI more than healthcare.

Taking a step back, the healthcare field has an essential yet polarizing role in today’s society – residing precariously at the inflection point of people’s health and a multi-trillion dollar for-profit industry. As introduced In William Kissick’s book, Medicine’s Dilemmas: Infinite Needs Versus Finite Resources, the “Iron Triangle of Health Care” is a simple but effective depiction of the resulting balancing act:

Figure 1

The three interlocking factors of the Iron Triangle – access, cost and quality – demonstrate an industry forced to make seemingly impossible trade-offs between the ability to provide quality care to everyone that needs it while also attempting to contain skyrocketing costs. Where do AI and machine learning fit into this seemingly hopeless triangle of despair? They don’t. Instead, they have the potential to be a disruptive force powerful enough to break the traditional model and improve all three factors at once – albeit not without consequences.

A telling illustration of this phenomenon is the partnership between IBM Watson and University of North Carolina Lineberger Comprehensive Cancer Center, as reported in a 2016 episode of 60 Minutes on CBS. In a pilot designed to test how AI could be deployed in a clinical oncology environment, IBM “taught” Watson to read medical literature in about a week, and a week later it read 25 million published medical papers and had the ability to continuously scan the web for the latest medical research. Traditionally, analyzing a patient’s individual genetic mutations and other relevant information against the vast population of medical literature and open clinical trials could take days or weeks. The analysis performed was highly manual, and relied on doctors’ ability to stay current on clinical trials happening around the world. Armed with this vast body of knowledge, doctors fed Watson actual cases from cancer patients whose treatment had exhausted all options known to the panel of experts at the time, and in over 90% of the cases, Watson identified the same experimental treatment options as the panel of experts. However, even more striking, in roughly 30% of the cases, Watson was able to identify a potential treatment not previously considered by the panel.

The results of this limited trial are by no means a silver bullet, but the outcome is especially promising in the context of the Iron Triangle because it demonstrates how AI can be used to improve the quality of care (Quality) for more patients (Access), with fewer doctors and in less time (Cost). While AI cannot offer a comforting presence and doesn’t have the capacity for genuine empathy – both considered important factors leading to positive medical outcomes – as AI penetrates the time-intensive world of health care administration (claims, billing, fraud detection, etc.), it should allow doctors to spend far more of their time in patient-facing roles.

The UNC/Watson pilot is one demonstration of the countless potential use cases for AI in healthcare. Drug development, virtual chatbots, dictation support for physicians and complex data analysis capable of predicting likelihood of hospital re-admittance are among the other uses currently being piloted and implemented across the world.

Given the acute focus on healthcare costs and outcomes in today’s political climate, and a lack of concrete solutions with unified support, AI is poised to take center stage in the next decade. Without question, there are legitimate concerns around the accuracy of AI-driven results, patient privacy and the potential for companies to misuse the vast troves of newly available data. However, as information technology professionals focused on compliance, risk management, and security, our role in this impending shift will prove critical in ensuring AI is deployed securely, data is used appropriately, and the results delivered are accurate and actionable.

ISACA Now By Year

Check Mark

Check Mark

Check Mark

Check Mark

Check Mark