Is 2025 the Year AI Becomes a Point-of-Care Tool in Your Facility?

As a solo practitioner in a SNF for 40+ years, I have developed a set of decision support tools accessible on my smartphone that address many of the questions that arise while caring for increasingly complex post-acute rehabilitation patients. I have been following the evolving literature on adaptation of AI (Artificial Intelligence) to clinical medicine in medical journals and podcasts, but have not seen practical uses for clinical problem-solving in the SNF setting.

In our setting, licensed nurses still seem to be practicing using memory-based tools from in-services, internet teaching modules, and clinical algorithms. I have seldom seen a licensed nurse consult their smartphone or computer for real-time information pertinent to a clinical problem of a patient under their care. I have previously written an article for the WAVE on the development of a digital library on their facility computers like what is available in acute hospitals but customized to the SNF setting, with files for accessing VIS (Vaccine Information Statements) from ACIP for informed refusals, and Patient/Family Education handouts for common clinical problems in SNFs. This library could support critical thinking and consistent “person- centered” clinical care.

A colleague at my Continuing Care Retirement Community (CCRC) has recently begun using the AI app, “Perplexity,” to seek quick answers to specific clinical questions on his smartphone. With that nudge, I have begun to use that app as well as Google’s new AI app, “Gemini”, Gemini is a click away on google chrome when you click on the app grid. Both are available with free versions for Android and Apple phones and tablets, and both have voice-driven search engines with quick answers to specific questions. For instance, I recently had a patient with onset of cholestatic jaundice without an apparent cause. I had recently prescribed methimazole for a new diagnosis of hyperthyroidism (multinodular goiter) and used the Epocrates app to find that this medicine is associated with hepatotoxicity. After some time in the UpToDate App, I was able to find information on this subject helpful to the management of this problem. Out of curiosity, I asked both AI apps for information on hepatotoxicity from this medicine and immediately had a detailed answer consistent with what I had learned via my search on UpToDate.

In another situation, a nurse called regarding a critically low calcium level, which I have not seen in years. I asked “Gemini” for the common causes and clinical manifestations of hypocalcemia and then for basic laboratory workup and therapeutic options. Again, I found quick information consistent with what I had previously read that helped me to develop an appropriate care plan at that moment.

Another website with AI geared toward literature review answers to physician queries is Open Evidence (https://wwwl.openevidence.com/). This is available as an app, “Open Evidence” and quickly gave me a response to my question about managing hyperthyroidism in a patient allergic to Methimazole along with links to pertinent guidelines and articles.

I am impressed with the quality of real-time clinical information available with AI tools on computers, smartphones, and tablets. I believe our licensed nurses can similarly be nudged to explore clinical uses of these apps as they see and address clinical problems in the patients under their care.

Perplexity and Gemini apps end each search with a disclaimer that some of the content could be unreliable. This acknowledges that the database for AI may have implicit biases and may come to conclusions that are based on imaginary literature, or so-called “hallucinations.”
That said, my limited experience has been very favorable and warrants further use for future clinical problem solving.

Are you ready to jump into the water?

Share this post:

Comments on "Is 2025 the Year AI Becomes a Point-of-Care Tool in Your Facility?"

Comments 0-5 of 1

- Monday, February 03, 2025
2009462652

I admire your openness to using new tools and appreciate the clinical caution in endorsing wholesale adoption. Reading your piece, I remembered when hand held calculators first came into fashion. Coming from the era of memorizing times tables, being expected to "make change" in my head, and showing off math tricks at parties, I at first found myself double checking the calculators answers by hand. This gave over to just trusting that it could and would do the tasks asked of it reliably and without error (operator error excepted). I find myself doing similar things with AI. I need to learn how to craft my queries to obtain the answers that seem trustworthy. I am also aware that some of this may be generational. Younger professionals have grown up with an implicit trust of tech. This is good and bad, especially if there are gaps in experience and/or knowledge. As you suggest, all this warrants further use . . .

Please login to comment