What Will AI’s Biggest Contribution to Healthcare Be?
Artificial intelligence is already improving many aspects of our lives, including how we drive, how we socialize, and what we buy. It also has the potential to transform healthcare in variety of ways, but the biggest impact might surprise you.
It would be overly simplistic to say there’s only one way that AI and related technologies, like deep learning and streaming analytics, will be used in healthcare. Obviously, a technology as powerful as AI will be used in multiple ways in healthcare, a huge industry that accounts for tens of trillions of dollars in annual spending around the world, and touches nearly every one of us.
Diversity has been the rule with AI and big data analytics in healthcare up to this point. Readers of Datanami will remember how:
- Hospitals affiliated with the group purchasing company (GPO) Vizient are using big data analytics, including data lakes and machine learning, to spot patterns in lab data to improve care;
- Closer monitoring of intensive care unit (ICU) telemetry data can give doctors and nurses a heads-up than something is happening with their patients;
- Research on brain tumors and other cancers is progressing thanks to big data approaches;
- Sepsis detection rates have improved dramatically, leading to fewer deaths;
- Sophisticated propensity models can tell healthcare providers what patients are likely to be re-admitted.
These are just some examples of how the AI and data science revolution is impacting healthcare. One could extrapolate from this and conclude that as the data sets grow bigger and the algorithms get smarter and the computers get faster, that AI eventually exceeds the ability of humans in the medical field. We’re at the early stages of precision medicine, but if things work out, humans could soon live to be 150 year’s old on a regular basis, Cloudera co-founder Mike Olson surmised a couple of years back.
But there’s another result of the AI revolution that could be just as impactful to the healthcare industry, if not more so: just getting basic care to those who need it.
While people in developed nations already have a decent baseline of medical options, those in poorer developing countries often lack access to any professional care. That’s the outlook that one provider of big data and AI technology to the healthcare field, John Snow Labs, recently provided.
“If you look long-term the biggest benefit is about access to healthcare,” said a technical advisor to John Snow Labs who asked not to be quoted. “Because what you can do right now, in some cases you can do better than humans. But even if you can do as well as humans or close to it, now you can actually get to every person.”
There are many people around the world who fall through the cracks in the healthcare field. They either lack the resources to see a doctor, or there simply is no doctor for them to see. This is a common situation in developing areas, such as rural Africa, where medical facilities aren’t computerized yet and just tracking the names of people is a major undertaking. But this phenomena is also impacting the United States, where a recent study found more than 40% people postposed or avoided medical care because of the high cost.
Now folks in the healthcare industry are excited about the prospect of leveraging AI to dramatically expand the number of people who can get access to basic care. It’s all about taking the advances in healthcare that AI can bring and applying them broadly to bring a large swath of the population up to a minimum standard of care, as opposed to boosting those who already receive decent care into even higher realms.
“If you think about most people in the world, even in the US but definitely outside the US, there’s no access to psychiatrist or oncologists or ophthalmologists,” the John Snow advisor says. “Those people are just not there. I think the promise here is you can have technology that can actually give everyone the best care that resources can offer, and do it more cheaply and at scale.”
John Snow Labs provides an array of big data and AI-based software, services, and data to healthcare companies. The Delaware company developed a data science platform that utilizes technology like Spark, Tensorflow, and Elasticsearch to helps clients like Kaiser Permanente develop machine learning models for healthcare use case and deploy them into production. It also has a team of specialists who help maintain high quality for about 800 datasets across 18 areas of healthcare-related fields.
In terms of AI, healthcare is still playing catch up to less regulated industries. It would be hard to confuse the sort of data analytics work that marketers are doing with Facebook data with the data science work that’s going on in hospitals, where HIPAA and other regulations require strict handling of data and keep close tabs on privacy controls.
While healthcare has a ways to go in terms of data science sophistication, it’s also poised to capitalize on AI advances at a potentially faster clip. We can see how neural networks, such as the one Stanford AI researcher Andrew Ng helped devise to automatically detect pneumonia in chest X-Rays, could help to bend the cost curve down and thereby make quality healthcare accessible to a wider number of people.
There’s a lot of work to do yet before a generalized approach to AI-powered healthcare can be hammered out – both on the technology side as well as on the business side of healthcare. But if the experts are right, AI’s biggest contribution to the healthcare industry could be a dramatic democratization of a basic level of care for everybody.