Simulating the people aspect of care with AI technology

Every medical practitioner encounters a variety of patient “personas.” Each requires a unique approach to discussing and providing care that encourages the best possible outcome. One patient may be inclined to hear their doctor out and proceed with the recommended course of treatment as outlined. Another patient may be guarded, hesitant, skeptical, or even combative. This particular “people aspect” of practicing medicine can be unpredictable or challenging to navigate, especially for those who are early in their career.

Yan Chow, MD: Image courtesy of Automation Everywhere - ©Automation Everywhere

However, emerging technologies have the potential to give medical professionals more experience with different patient personas—before they even interact with a patient seeking diagnosis and treatment. The introduction of chatbots powered by artificial intelligence (AI), including ChatGPT, Bing, and Bard, inspired me to consider the creative ways that medical professionals might use these technologies to support their practice.

Key to determining applications of AI chatbots in medicine is understanding how this technology works. At a very high level, AI chatbots are trained on an incredible amount of data to help them determine the most likely response to a given question or prompt. This training, and their capability to infer, gives them the ability to interact with the technology user in a more natural, conversational manner.

As you may already be imagining, compelling use cases of AI chatbots include offering insights, such as social determinants of health (SDOH) and other demographics, that can inform faster diagnosis and the best course of treatment; simplifying data entry and processing while reducing errors; and supporting the revenue cycle. Of note, AI chatbots are only half the solution, as intelligent automation is needed to actually remove the associated tasks from staff’s to-do lists.

These technologies can also give medical professionals much-needed experience with a range of different patient personas through simulations. Here's what this could look like.

Consider a medical resident who has had limited patient interactions to date. In preparation, the resident may prompt an AI chatbot to play the role of various personas: a parent or guardian who is overseeing the care of a minor, a patient who has a history of not following through on prescribed treatments, a patient who tends to be dismissive of diagnoses altogether, and so on. Working with this chatbot can provide the resident with the foundation needed to be comfortable in speaking with and getting through to a variety of patients.

In another example, consider a doctor who has begun dedicating their time and services to a global nonprofit organization that provides care in regions with limited medical resources. Although they have the experience of communicating with a variety of patient personas, they may lack the cultural familiarity to discuss diagnosis and treatment with patients in a particular region. Practicing with a chatbot can help the doctor get up to speed with local best practices and sensitivities that enable the optimal outcomes for patients.

Reflecting on my own time as a new medical professional, I can see immense value in using these technologies to gain the experience required to be confident and successful during patient interactions, among other use cases. Simulations could help simplify the complexities of the people aspect of practicing medicine. They could help us avoid those awkward, early-career fumbles that, while valuable learning experiences for us, can sometimes make it more challenging for patients to trust their provider, understand and comply with the prescribed treatment, and obtain the compassionate, efficient, and effective care they seek.

Yan Chow, MD, is a Global Healthcare Industry Leader and Strategist at Automation Anywhere.