The Positives and Negatives of AI in Health Care

AI is spreading all over the place including health care.

The question is that good or bad? 

I want to talk about why AI might be important in health care. The answer may actually be simple,

AI has the ability to prevent misdiagnoses if used properly. Why is that important? It turnsout that misdiagnoses can be deadly. Every year in the US there are around 371,000 deaths and 424,000 cases of permanent disabilities which include brain damage, blindness and loss of limbs due to misdiagnoses.

One recent study had AI come up with the correct diagnosis of 20 different clinical cases. This was compared to medical folks in the study who only had to analyze one case and not 20 different ones.

The AI system and the medical personal were given the triage data, and the patients report of what was bothering them. There was a system review phase, where the AI and the medical folks got additional information from the patient. There was also a physical exam, diagnostic testing and imaging done. 

The results of the diagnosis were that the AI got 10 for 10 correct. The physicians got 9 out of 10 cases correct. The residents got 8 out of the 10 correct. That meant that the residents missed 20% of the cases.

The problem with all of this is no one knows which of the cases were the ones that were misdiagnosed potentially until it’s too late.

That is the positive side of AI.

What about the negative side? Let’s explore it.

One of the most pressing issues with AI in health care is the potential for bias. AI systems learn from vast datasets, and if these datasets reflect existing biases, the AI will inevitably perpetuate them. For instance, if a health care AI is trained predominantly on data from one demographic group, it may not perform accurately for others. This can lead to disparities in diagnosis and treatment, disproportionately affecting minority groups and exacerbating existing health care inequalities.

Health care data is highly sensitive, and the use of AI requires vast amounts of this data to function effectively. There is a risk that patient information could be mishandled or accessed by unauthorized parties. In addition, the risk of breaches remains a significant concern. Just think of all the major data breaches that have happened in the last couple of years.

Health care is not just a science; it is also an art that requires empathy, compassion, and a personal touch. The increasing reliance on AI can lead to a depersonalization of care, where patients might feel like they are interacting with machines rather than humans. 

When I was in the hospital in February, the best part was the care I received from the nursing staff.

While AI can assist health care professionals as I talked about on the positive side, there is a danger of becoming overly reliant on these technologies. This over-reliance can lead to a degradation of human skills and clinical judgment and could result in misdiagnoses or inappropriate treatments, especially if the AI makes an error.

AI in health care also brings about numerous ethical dilemmas. For example, who is accountable if an AI system makes a mistake that harms a patient? Since it is companies that are developing the AI systems the question becomes the accuracy of the data that is being inputted and the accuracy of the algorithms. 

Finally, the cost of Implementing AI in health care is not only technologically challenging but also financially burdensome. This could lead to a gap between well-funded institutions and those struggling to adopt new technologies, potentially creating a two-tiered health care system.

So…it’s probably too soon to tell if AI is going to be a boon or a bane to health care. Stay tuned for future developments.
(Reported WDDTY, June 2024 and by using CHAPGPT)