Inquiries for journalistic consideration when discussing the interface of health care and AI technology
Artificial Intelligence (AI) holds the potential to revolutionise modern medicine, but it also presents several challenges that need to be addressed.
One of the key concerns is the risk of racial or gender disparities when algorithms are trained using limited data sets. This issue was highlighted in a 2015 study published in the peer-reviewed medical journal "Blood" by the International Warfarin Pharmacogenetics Consortium, which showed that the influence of known genetic variants on warfarin dose differs by race.
Another challenge is the issue of overfitting, where AI algorithms learn too well on one dataset and perform poorly on new data. This can lead to misleading information, as seen in an example where an algorithm incorrectly concluded that people with asthma who get pneumonia are less likely to die, when in reality, people with asthma and lung problems have a higher risk of dying.
Providers have expressed concerns about the usability of AI systems, with many reporting difficulties and glitches. However, they are generally open to discussing these issues and working towards improving the systems.
The value of new AI technology should be evaluated by comparing it to the old system, ensuring that any benefits outweigh the potential risks.
In the realm of business journalism, it's worth investigating how much hospitals are paying for artificial intelligence. The costs associated with implementing and maintaining AI systems can have a significant impact on healthcare budgets.
The Food and Drug Administration (FDA) has recently approved the first AI software that can interpret images and determine whether a patient has diabetic retinopathy without the assistance of a physician. However, the FDA's recent guidance on AI has failed to answer big questions about the safety and effectiveness of these systems.
The FDA needs to be aware of how algorithms are trained and whether it has the staff or capacity to understand them. This is crucial in ensuring that AI systems are safe and effective before they are implemented in healthcare settings.
At the Association of Health Care Journalist's 25th conference in Phoenix, a panel discussed key takeaways for journalists, moderated by Stat managing editor Gideon Gil. The panel highlighted the importance of being open to the benefits of AI while also being vigilant about its potential risks and disparities.
An example of the potential benefits of AI can be seen in its ability to find operational efficiencies in healthcare systems. In 2017, Watson for Oncology, IBM's AI system for cancer treatment, fell short of the expectations set for it. However, with continued development and refinement, AI has the potential to significantly improve the efficiency and effectiveness of healthcare systems.
In conclusion, while AI holds great promise for modern medicine, it is crucial that we address the challenges and potential risks associated with its use. By being vigilant and proactive, we can ensure that AI is implemented in a way that benefits all patients, regardless of their race or gender.
