Artificial intelligence (AI) has become an established presence in the world of healthcare. It’s still a new enough phenomenon, though, that most people might not realize when and how it is being used.
AI is going to have more of an impact as its reliability increases. Here’s a look at how AI is being used in the healthcare field, as well as a couple of studies that show not only the impact AI has on the people using it, but how the technology also relies on those people.
AI in healthcare
Medical professionals are continually discovering ways to use AI to improve care. One of the most common applications of AI is to use it to comb through huge stores of data that would be at least time-consuming and at most overwhelming to people without the tool. This could include everything from basic patient information such as demographics all the way through to certain kinds of test results.
That’s where AI can veer into the life-saving category. AI can examine imaging in a more consistent way than people can. AI supporters often point out that people’s eyes can get tired at the end of a long day of looking at images such as MRIs, but AI removes any variability.
More advanced uses of AI include scanning those same images for rare insights that may not be visible to the naked eye, whether those are predictive factors that could lead to a disease, or to help the AI learn how to identify similar potential warning signs in the future.
How AI should be used effectively
Two recent studies offer some insight into the way AI is used. Researchers from the University of North Carolina, Chapel Hill and the University of Zambia published a study in the Journal of the American Medical Association (JAMA) in August showing how AI could improve care in resource-limited areas. They found that an AI-powered handheld ultrasonography device used by novices could predict gestational age as accurately as experts using ultrasound equipment.
The devices are more affordable than the traditional ultrasound equipment, and require less training to use. The study offers hope that the tool can improve pregnancy care in parts of the world where access to this important diagnostic tool is otherwise not available.
Another study offers insight into just how important that training aspect is. That study had 50 doctors make diagnoses based on case studies, some randomly assigned to use AI assistance, and some using just conventional methods. The doctors using AI scored 76%, the doctors without AI scored 74%, and the AI model working by itself scored 90%.
The study was less instructive in how effective AI could be and more about the way it is used. The doctors using AI assistance were reluctant to allow the technology to change their minds about a diagnosis, even if the AI presented more reasonable evidence for another diagnosis. The study used a small sample size, so further research would be necessary, but there is an important lesson to be learned. Without proper training on how to use the technology, AI can’t be of very much help.
Data is only as good as the people using it
The same is true of data in any form. Data by itself is often meaningless. It takes an analytics solution to turn it into actionable information that can produce results. In the world of healthcare that can mean, among many other things, more efficient processes, better use of hospital space, or the actual medical treatment decisions that directly impact patient care.
Again here, without educating doctors about what they are getting from the analytics, the data will not be as effective as it could be. The data is meant to enhance performance in the same way that AI is meant to supplement the work of medical professionals. Any of its life-saving potential is wasted without proper training or without setting proper expectations.
That needs to happen from the very beginning, whether it’s AI or the data an organization is using to make any decisions. Representatives from all of the stakeholders should be involved in decisions around how to best use data so that it is a collaborative effort. Analytics can be a difference-maker in healthcare and for healthcare organizations, but only if it is used correctly.
- 5 Trends that Will Impact Your Business in 2025 - December 5, 2024
- Using Analytics to Deal with Shifting Alcohol Preferences - December 3, 2024
- AI in Healthcare is Only as Effective as the People Using It - November 25, 2024