The Hype Around Artificial Intelligence
Artificial intelligence has a bright future in healthcare.
Many of us have some understanding of artificial intelligence (AI) and its increasing prevalence in modern business and society. Whether it’s your navigation app, personalized shopping, fraud detection calls from your bank, or an assistant like Siri or Alexa – AI is already an integral part of our daily lives. AI is also being applied to healthcare with the potential to transform many aspects of patient care and administrative processes.
We spoke with Richard Greenhill, DHA, FISQua, FACHE about the hype and reality surrounding AI in healthcare. Greenhill is an Assistant Professor and Assistant Program Director in Healthcare Management & Leadership at the TTUHSC School of Health Professions.
He is the creator and host of the Improve Healthcare Podcast, with listeners in 73 countries. He and his guests share national and international perspectives on a range of topics that impact healthcare quality and the care delivery continuum.
Applications of AI in Healthcare
Greenhill mentioned one of many interesting ways that AI is currently being used in healthcare delivery today: image analysis in radiology.
“Radiologists use AI to compare tumors and look at masses on x-rays,” he says. Studies have shown that “AI can perform as well as or better than humans at key healthcare tasks, such as diagnosing disease.”
Greenhill notes that this is just one example, and there are applications ranging from administrative processes, analyzing clinician notes and care coordination, to decision-support. These are only a handful within direct care delivery. The applications change when we step outwards into population health, bio-surveillance and clinical research. It is key to remember that each of these examples can bring its own challenges, tools and techniques, and achievement of technical success at one may not always translate to skill in another.
Is the AI Hype Really Worth Hyping?
According to Greenhill, the hype is that AI methods aren’t new. They have been around since the 50s.
“It’s not new – but our ability to use it is growing and changing,” he says.
In healthcare specifically, Greenhill sees a house not in order when it comes to data.
“Many people don’t understand that you must have the right data,” he explains. “The data we have in healthcare is vast and not usually in an easily manipulatable format. That’s mostly because the variances in how we collect, store, and use it.”
Across industries like banking, data is collected in similar formats; making the use of AI easier. In healthcare, we have several different electronic healthcare record (EHR) systems, and many organizations often use them slightly differently even in the same facility. So, comparability and integrity across healthcare systems is challenging. This should be a cautionary note when using AI features in EHRs to diagnose and treat patients. Data in AI algorithms must be trained or shown an example before it can be used in decision making, and this training is normally an ongoing process.
“You usually can’t use patient population data to train AI algorithms from Northern California and apply it directly to treatment pathways in Texas,” Greenhill says. “That doesn’t mean data from another area isn’t valuable for some purpose, but we need to step back and make sure we have the right data in terms of the structure, quality-level and applicable population. Otherwise, we run the risk of perpetuating, or even worsening, disparities in care delivery and outcomes that we see today.”
The Future of AI in Healthcare
Greenhill points to the metaverse [a virtual reality network of sorts] that has been described by Mark Zuckerberg as the future of data collection and sharing.
Currently in healthcare, data is spread out in different systems and not always in usable formats. Greenhill says a health metaverse, [a virtual means for care delivery, e.g., virtual hospitals etc.] may have utility in helping healthcare leaders and providers paint a more complete picture of care delivery, interoperability and thus predictions using AI tools.
While AI and the concept of an AI-enabled metaverse are exciting, we must ensure that the data that represents us accurately represents all of us. AI will mirror us, and it is imperative to accurately reflect the diverse society that exists in order to become the society we want to be.
The transition into fall can lead to a form of depression known as seasonal affective disorder, or SAD, a low level of depression that's common in the northern hemisphere during this time of year.
Ninh (Irene) La-Beck, Pharm.D., with the TTUHSC Jerry H. Hodge School of Pharmacy, received a five-year, $2.49 million grant to investigate how nanoparticles interact with the immune system and cancer.
To help investigate the influence basal sex hormone alterations may have on chronic post-op pain, the NIH recently awarded a grant to Jenny Wilkerson, Ph.D., from the Jerry H. Hodge School of Pharmacy.