Dr. Chatbot Will See You Now

While it is an interesting exercise to imagine the role chatbots can have in subbing in as your treating physician, the team here at Baxter believes there are far more powerful – and perhaps realistic – opportunities for AI to make healthcare more efficient and effective. Spoiler alert, none of those involve replacing the teams of nurses, physicians and techs who are at the center of every patient interaction. Here are the best opportunities for AI in healthcare:

1. Do More, With Less: "For me, the ability to do more with less is a key driver for technological advancements. Post pandemic, we know that staff shortages have led to challenging nurse to patient ratios that can compromise patient care and lead to high attrition rates among nurses. AI can assist healthcare professionals in making more accurate and timely decisions by providing real-time data analysis, flagging potential risks or issues, and suggesting treatment options based on patient data. An example could be an electronic health record (EHR) algorithm that identifies patients at risk for falls or pressure injuries. With this information, the care team can then initiate a prevention protocol," said Traci Reed, Baxter's director of data science and AI. 
2. Reducing Administrative Burden: "Fifteen to twenty-five percent of healthcare delivery costs is administrative. AI can be deployed to drastically streamline the administrative processes that sit behind our healthcare system. For payors, using AI in the pre-authorization process can reduce the manual human processing that happens today. For hospitals, AI can help optimize patient scheduling, clinician scheduling and equipment management to help ensure a hospital is running as efficiently as possible," said Chad Malone, M.D., Baxter's vice president of enterprise connectivity. 
3. AI as Liaison: "We know how impactful AI can be with large volumes of information and the ability to understand, interpret and act upon the information quickly. Generative pre-training transformers (GPTs) are designed to communicate with the user, opening new pathways for question and answers not only for clinicians, but for patients and their families as well. AI will have the ability to act as a liaison between patient care, clinical process, and information dissemination. I think the 'liaison' aspect of a chatbot, coupled with text to speech capabilities, has real potential for care communication and clinical decision support tools," said Patrick Harrison, Baxter's director of data science.
4. Aiding Early Intervention: "Anywhere we can automate information collection and provide robust insights into a patient's condition, we can remove manual steps in the workflow and make clinical intervention timelier. Take for example, patient vital signs. Every hospitalized patient has vital signs collected, either continuously or at set intervals. You can combine AI with large data sets to build a robust model to predict patient deterioration. If a hospitalized patient is under continuous monitoring, care teams can use the model and real-time data to identify patients who may be heading toward serious complications," said Meredith Hannah, Baxter's director of digital partnership and innovation.
5. More Effective Screening & Referrals: "The opportunity to have more effective referrals and better use of specialists can improve our healthcare system by reducing waste. AI combined with medical devices enable clinicians to screen patients more effectively in lower cost settings, ensuring those with the most critical needs are referred on to specialists sooner. It also increases the speed by which you can assess and provide feedback to the patient," said Kristin Russell, Baxter's vice president of enterprise connectivity.
6. Predicting Chronic Diseases: "AI can be used to identify patients that are developing or at a risk of developing chronic diseases that currently overburden our healthcare system. This early prediction would allow an opportunity for preventative treatment and lifestyle change, which can stave off complications and high cost of adverse event treatments down the road. An example that I have personally seen is using voice characteristics as a method of determining disease risk. With a simple phone call, a payor or provider could use AI to predict depression, anxiety, Alzheimer's or other diseases in real time, allowing for interventions to happen right then and there," said Dr. Malone.