Artificial intelligence could alleviate access and resource challenges in small and rural hospitals by streamlining administrative tasks, identifying high-risk patients, improving disease management, facilitating telehealth, and generally allowing these institutions to do more with less, said presenters at the ASHP Midyear 2024 Clinical Meeting and Exhibition, in New Orleans.

Gretchen Brummel, PharmD, BCPS, the director of the Professional Experience Program and a clinical associate professor in the Department of Pharmacy Practice and Science at the University of Iowa College of Pharmacy, in Iowa City, described several AI applications validated in clinical trials that could be used in a small or rural hospital.

Insulin Titration

In the ADVICE4U trial, published in 2020, 108 patients with type 1 diabetes who used insulin pumps were randomized to receive standard care management, with physicians from specialized diabetes centers titrating their insulin, or an AI algorithm that performed the insulin titration (Nat Med 2020;26[9]:1380-1384). Based on the primary outcome (the percentage of time spent within the target glucose range), AI titration was noninferior to that done by physicians (P<0.0001). Additionally, three severe adverse events related to diabetes were reported in the physician arm and none in the AI arm.

“This [strategy] has some great potential in the rural setting because we know that primary care physicians as well as specialists are in shortage,” Dr. Brummel said. “If we are able to leverage [AI] to help us care for this patient population, that would be very beneficial.”

Asthma Management

Another trial reviewed the implementation of an AI “Asthma-Guidance and Prediction System (A-GPS)” for pediatric primary care patients, who were randomized to either A-GPS or usual care. During the study, the A-GPS system generated four reports containing a patient information summary covering 32 variables, a machine learning algorithm predicting adverse event (AE) risk, and an asthma management plan (PLOS One 2021;16[8]:e0255261).

“The intervention group had no difference in AEs, cost or time to patient follow-up,” Dr. Brummel said. “But the intervention group did see a higher enrollment in asthma management plans by providers in the intervention group, and a significant decrease in charting time, going from around 11 minutes to around three minutes. Increased disease state awareness and significantly reduced charting time could have significant positive implications for rural pharmacy practice.”

Blood Pressure Control

Top-line findings from a study presented at the American Heart Association Scientific Sessions, in September 2024, found that a Bluetooth-enabled remote monitoring system combined with pharmacist interactions allowed up to 74% of participants with resistant or difficult-to-control hypertension to significantly improve their blood pressure control. “They leveraged AI with predictive analytics to manage these very challenging patients via telehealth,” Dr. Brummel said. “They demonstrated that 67% of the patients achieved [blood pressure] control within six months, and 74% within the first year, while they also saw fewer hospitalizations within the study period compared with the previous 12 months.”

Medication Review

Many small and rural hospitals do not have 24-hour pharmacy coverage, Dr. Brummel noted. She pointed to a recent French study that found that an AI tool could prioritize medication review by categorizing medication orders as low- or high-risk (Int J Clin Pharm 2022;44[2]:459-465).

Prescriptions identified by the decision support tool as low-risk ultimately had a 2.8% incidence of severe drug-related problems (DRPs), while those identified as high-risk had a 15.3% incidence of severe DRPs. “They also compared traditional clinical decision support [CDS] to a hybrid model that partnered CDS with an expert pharmacist, and that hybrid system had much higher capture of severe DRPs [94% vs. 20%; P<0.001] compared with CDS alone,” Dr. Brummel said. “In a resource-limited setting, application of a tool like this could help you decide which orders are going to be higher priority for your pharmacists to pay closer attention to.”

Create a Road Map

Small and rural hospitals often lack the financial resources to invest in costly AI tools, noted Kyle Johnicker, PharmD, the pharmacy clinical coordinator at Northwestern Medicine Kishwaukee Hospital, in DeKalb, Ill. “These systems can run the gamut from $20,000 for a plug-and-play system with no extra support, up to more than a million dollars if you’re going to work side by side with the programmers and develop something from scratch,” he said. “At most of the facilities I’ve worked in, even the $20,000 conversation with the C-suite is going to be difficult.”

Nonetheless, he continued, “there are opportunities that you can bring to that conversation. For example, if you save time utilizing an AI tool to identify drug diversion patterns, because you’re no longer looking through all those reports, can that pharmacist then provide some additional patient care that might decrease readmissions?”

Create a road map of what you want to accomplish with AI, Dr. Johnicker advised, whether that is identifying high-risk patients, predicting disease progression or analyzing genetic data. “You need to know where you want to go, so you can be prepared when having these conversations, to find the solution that is appropriate for your facility.”

https://www.pharmacypracticenews.com/Pharmacy-Technology-Report/Article/06-25/AI-Benefits-Rural-Hospitals/77057