You’ve probably already heard a lot about how AI will change healthcare. Me too. Still, given its potential, I’m always interested in hearing more, and the following article struck me as offering some worthwhile ideas.
The article, which was written by Humberto Alexander Lee of Tesser Health, looks at ways in which AI tools can reduce data complexity and detect patterns which would be difficult or even impossible for humans to detect.
His list of AI’s transformative powers includes the following:
- Identifying diseases and providing diagnoses
AI algorithms can predict when people are likely to develop heart disease far more accurately than humans. For example, at Google healthcare technology subsidiary Verily, scientists created an algorithm that can predict heart disease by looking at the back of a person’s eyes and pinpoint early signs of specific heart conditions.
- Crowdsourcing treatment options and monitoring drug response
As wearable devices and mobile applications mature, and data interoperability improves thanks to standards such as FHIR, data scientists and clinicians are beginning to generate new insights using machine learning. This is leading to customizable treatments that can provide better results than existing approaches.
- Monitoring health epidemics
While performing such a task would be virtually impossible for humans, AI and AI-related technologies can sift through staggering pools of data, including government intelligence and millions of social media posts, and combine them with ecological, biogeographical and public health information, to track epidemics. In some cases, this process will predict health threats before they blossom.
- Virtual assistance helping patients and physicians communicate clearly
AI technology can improve communication between patients and physicians, including by creating software that simplifies patient communication, in part by transforming complex medical terminology into digestible information. This helps patients and physicians engage in a meaningful two-way conversation using mobile devices and portals.
- Developing better care management by improving clinical documentation
Machine learning technology can improve documentation, including user-written patient notes, by analyzing millions of rows of data and letting doctors know if any data is missing or clarification is needed on any procedures. Also, Deep Neural Network algorithms can sift through information in written clinical documentation. These processes can improve outcomes by identifying patterns almost invisible to human eyes.
Lee is so bullish on AI that he believes we can do even more than he has described in his piece. And generally speaking, it’s hard to disagree with him that there’s a great deal of untapped potential here.
That being said, Lee cautions that there are pitfalls we should be aware of when we implement AI. What risks do you see in widespread AI implementation in healthcare?