The Road to AI in Healthcare Is Still Long.


Artificial intelligence (AI) technology has long held out the promise of significantly enhancing the healthcare sector. Healthcare commentators have enthusiastically anticipated the widespread application of AI, whether it be due to the promise of improving access to and understanding of data, offering methods to better navigate patient care, or better comprehending new research and development activities. Several businesses have committed billions of dollars in the goal of enhancing the effectiveness and feasibility of AI in their respective fields. And with good reason—these efforts have undoubtedly produced a lot of beneficial outcomes, a lot of which have served as the cornerstone for ongoing construction and innovation in this field. However, there is still a long way to go with the technology.

Finding quality data sets to utilise as teaching models has been one of the main hurdles in the development of AI technologies for the healthcare industry. Generally speaking, “AI” technology leverages enormous data sets to identify trends and provide recommendations accordingly. However, the quality of these suggestions and pattern recognition results depends on the data sets used, which can be problematic in a variety of situations, but is particularly true when working with patient care data.

Key leaders have talked extensively about this potential introduction of bias in AI-based care. According to Dr. Paul Conway, the American Association of Kidney Patients’ Chair of Policy and Global Affairs, “devices leveraging AI and ML technology will alter healthcare delivery by enhancing efficiency in crucial patient treatment processes.” Pat Baird, Regulatory Head of Global Software Standards at Philips, stated that in order to better support our patients, “we need to become more familiar with them, their medical conditions, their environment, and their needs and wants in order to be able to better understand the potentially confounding factors that drive some of the trends in the collected data.” The latter points to a very particular issue that numerous AI enthusiasts frequently run into: bias due to very small, very segmented, or very inaccurate data-sets.

For instance, it would probably not make sense for an AI system designed to prescribe pain-relieving drugs to apply to the general population if it was based on a data collection that only included cases of cancer patients. After all, the recommendations would be severely skewed because the pain drugs required for cancer patients are significantly different and probably more potent than those needed by the general population. This bias is only one kind; extending the same potential bias and inaccuracy across ethnicities, races, socioeconomic position, and other characteristics can lead to clinical judgments that are dangerously wrong.

Why is this crucial? Because AI has the potential to be a tremendous force in the healthcare setting if used properly. I’ve previously discussed how AI can be a useful tool in a range of industries, from radiology to cancer treatment. There may be a place for AI modalities as a tool to supplement clinical work processes, even though it may not have the strength to completely replace the complexities, expertise, and wisdom of physician-led patient care.

However, in order for this technology to truly provide value, systems must generate high-fidelity suggestions, making sure that they account for accurate and representative data. Only then will doctors be able to really benefit from this technology and influence healthcare delivery without bias. With this technology, entrepreneurs, healthcare executives, and care providers will have a challenging challenge ahead of them in the years to come.

Next Post

2023 Symposium Update in Allergy and Immunology

This activity is designed to meet the educational needs of Internal Medicine physicians, Pediatric physicians, Family Medicine physicians, Emergency Medicine physicians, and Allergy/Immunology physician specialists. 20-21 January 2023 | 7:30 AM- 2 PM | Tampa, Florida Read More