In the past year, we've become familiar with the factors that can make a person more vulnerable to COVID-19 infection. The elderly are more at risk, as are those who smoke and are already dealing with other diseases, such as cancer and Type 2 diabetes.
At a deeper level, though, there are dozens of other factors that may come into play and influence a person's susceptibility to disease. A recent analysis of hospitalized COVID-19 patients in 14 states found that among patients ages 50-64 that obesity was the most prevalent underlying medical condition. Similarly, there's growing evidence to suggest that vitamin D deficiency contributes to COVID-19 infection.
The emerging field of network medicine, powered by this type of digital analysis of large data sets, sheds light on the interplay between microbial virulence and the ability of a person's immune system to defend against diseases such as COVID-19. Network medicine allows researchers and physicians to look beyond the traditional root causes of disease and take a more holistic approach to identify agents that can influence a person's susceptibility to disease.
In an article that I co-authored with Paul Cerrato and Adam Perlman, M.D., MPH, for Mayo Clinic Proceedings: Innovations, Quality and Outcomes, we describe how the analytic power of supercomputers and the emergence of big data sets has given researchers new insights into the causal relationships that influence susceptibility to disease. This technology dramatically improves our ability to assess the relative strengths and weaknesses of factors as contributing agents.
Some of these agents are not surprising — nutritional status, for one, and environmental factors. Others may be harder to assess, like sleep habits, exercise, physical and psychosocial stressors, obesity, protein-calorie malnutrition and emotional resilience. Genetic variations such as single-nucleotide polymorphisms also are examined as possible agents affecting a person's vulnerability to disease.
With possible factors identified, deep learning algorithms can assess each's likely strengths and weaknesses as contributing factors to disease and help identify therapeutic options.
Using machine learning-enhanced algorithms to analyze risk factors and their interactions can help determine which ones can predict a person's risk of COVID-19 infection or the prognosis for someone who already has tested positive.
At a time when we're all looking for reasons for hope and encouragement — and the national rollout of a COVID-19 vaccine is a big one — it's good to remember that our capabilities to gain essential insights from AI, network medicine and deep learning algorithms are ever-growing and that we have the potential not only to resolve this pandemic more quickly but to completely redesign how we respond to pandemics in the future.