Viewpoint: AI should be used not just to predict disease, but to understand the underlying causes

Most artificial intelligence tools developed for use in healthcare purport to predict and detect disease with superhuman speed and accuracy, but the technology's true life-saving potential lies in discovering why those diseases occur at all.

In an op-ed for the Harvard Business Review, two professors from the Boston-based Harvard T.H. Chan School of Public Health wrote that while the predictive capabilities of some AI offerings have certainly saved lives, those systems simply "take what we already do and improve it by shifting the task from a human … to an algorithm."

To maximize the capabilities of AI, then, healthcare organizations should turn to causal algorithms, which can "infer causal relationships from observational data, telling us how different factors interact with each other and which one is causing what" — tasks that, for humans to perform, require costly, time-consuming clinical trials that are often not wholly representative and may not be feasible at all.

In the article, Sema Sgaier, PhD, and Francesca Dominici, PhD, listed three areas in which causal AI algorithms show the most promise:

  • Discovering mechanisms of disease
  • Optimizing treatment
  • Addressing social determinants of health

"Asking 'why?' through causal AI offers new ways of unraveling complexity to act on the most important factors that cause disease by leveraging the power of data," they concluded. "Creating algorithms that can perform this task is harder than creating ones that can make predictions. But it's not only a worthwhile endeavor, it's a necessary one."

More articles on AI:
Why AI is being used more in clinical workflows, quality reports
To get executives on board with new tech, put ROI before AI
Moving beyond the hype: How hospitals are applying AI to daily functions

© Copyright ASC COMMUNICATIONS 2019. Interested in LINKING to or REPRINTING this content? View our policies by clicking here.

 

Top 40 Articles from the Past 6 Months