Artificial intelligence-drive platforms and chatbots are already helping hospitals across the U.S. connect with patients and optimize operational efficiency. But they can do more with the large clinical datasets to make clinicians' lives even easier.
Jordan Dale, MD, chief medical information officer and hospitalist at Houston Methodist, shared his perspective on how generative AI can support clinical workflows for significant improvement in patient care and experience.
Note: This response was lightly edited for clarity and length.
Q; When you think about the intersection of clinical care and digital technologies, where are the big areas for growth and development over the next few years?
Dr. Jordan Dale: Generative AI has taken leaps and bounds over the last six months. A lot of people focus on how it can do content creation because that's how most of the tools have been set up as a chat interaction to create content. I have a little bit of a different angle. The struggle of a clinical informatics or health IT team has always been the amount of data that we create as clinicians isn't normalized or set up appropriately so we can actually unlock the potential for future uses.
That's one of the main ways we can really unlock the data, by applying generative AI that can drift across data requirements to understand things in a much more seamless or more human way than how we've been able to apply things like natural language processing in the past. One example, there's a way you can get to know your patients as you're taking over their care in the hospital a little bit more quickly as a physician or nurse. I'm still a practicing hospitalist, so I do inpatient care. It's tedious to review a lot of notes and find the critical values and if I'm missing something, it could impact how I deliver care even if I was trying to do my best.
We've actually applied some generative AI technology that will help generate a patient summary for me to tee-up the key past medical history, current problems, medications they're receiving and what consultants are following. Before, some of those data elements were really difficult to find because they were not discreet and this technology allows us to leverage that.
Another benefit of that is we can use the information summarized in charts and encounters, and we can actually use large language models and translate that to different levels of comprehension across the healthcare experience. You could translate that to a different clinical role, for example, that is more pertinent to a physical therapist. You could also translate that report to a patient or their family members based on their reading level or preferred language very quickly.
As a clinician, everyone wants to provide a great experience for patients but we're not able to do that on the fly as well as some of these technologies could do for our patients. It's a great way that we'll be able to better serve our patients with technology and unlock some of the massive amounts of data in areas that have been locked away in our system to really explain and get to know our patients and take care of them better.