ARTIFICIAL INTELLIGENCE, OR AI, has been garnering a lot of press on how it is being applied in healthcare. This is not without its challenges.
AI is a technology that uses computers to learn and make decisions. The engine behind AI is algorithms. These algorithms depend on data — plenty of data, and I should add, quality data — that is shoveled into computers. AI gives insight into vast amounts of data.
Artificial Intelligence Challenges
In healthcare the challenge is that data is largely unstructured, based on doctors’ notes. This is what’s termed language-based information. A March 25 article in The Wall Street Journal reported on how the Mayo Clinic is testing Google Healthcare Natural Language API to turn this unstructured data into discrete fields that become structured data. This would improve the results of data searches.
But we are still in the learning stage. Evidence of this is the disappointing results of IBM’s Watson, which was heralded as helping doctors diagnose, treat, and potentially cure cancer. IBM decided to bow out of this project. It was proving to be a bigger undertaking than anticipated, and the results did not measure up.
AI can help in diagnosing sepsis before it is too late. It can be used to predict who will be readmitted to the hospital within a specific number of days since discharge. And using access to social media, AI can gain better insights into adverse drug events. What we have here is a picture of how technology is being deployed to assist in diagnosing, treating, and getting better patient outcomes.
What does all this mean to pharmacy? Pharmacy can be a data source on allergic reactions and adverse drug events. This will depend, however, on pharmacists documenting this information in their pharmacy management systems. As with electronic medical records, what is now captured is unstructured data. But as pointed out earlier, this limitation is being addressed.
There are a lot of technology companies investing in AI. These companies see a bright financial future in this technology. As a result, we are certain to see progress in machine learning as we move forward that is going play out well in reducing hospital admissions, getting better outcomes from drug therapy, and, in the end, increasing longevity. The cover story in this issue takes a look at what is taking place with the deployment of AI in pharmacy. CT