The Questions to Ask When Making It Real: Generative AI and Adherence, Part II
Three Essential GenAI Qs
GenAI's ability to generate user-friendly content from learned patterns and existing data has the potential to educate and motivate patients. Even though substantial fine-tuning and legal clearance must take place, such a capability would empower patients to take an active role in their care journey.
In this post, the second installment of our series on generative AI (GenAI) and patient adherence, we suggest concrete ways of creating a human-centered patient experience using GenAI. Read on and you’ll find the necessary questions you’ll need to ask as you proceed.
Building the Right Team. Asking the Right Questions. Taking Responsibility.
If GenAI is to tag along on the patient adherence journey, it must be accompanied by a comprehensive understanding of human emotional and functional needs. It takes a broad spectrum of perspectives and specialties to build solutions that effectively assist patients.
Capturing implicit user needs is crucial. Jonathan Rioux, Director of Data Analytics Consulting, believes that responsible innovation comes from “identifying opportunities for improving your product” by collecting user feedback for iterative design.
Achieving this demands a collaborative, multi-disciplinary effort, including mutual education across the team to produce effective solutions. Rioux says that data scientists need to be “more sensitive to the reality of how their models are going to be applied,” while designers and developers also “need to be more familiar with technology.” A team assembled to develop solutions that can enhance patient adherence and outcomes should ask the following questions throughout the process:
- How can we create a credible and relatable GenAI solution that motivates patients?
Jonathon Swersey, Senior Director, Innovation Consulting, says: “What GenAI means to patient adherence will depend on how people perceive the GenAI and on their motivation.” Perceptions are often linked to our sense of selves — so having a GenAI that “sounds like me” may help create a positive perception. For many patients, the primary motivation is to get better for loved ones, so we can use AI to create an agent that is responsive to an individual patient’s lifestyle and preferences. Regarding both perception and motivation, having access to the necessary data will be essential. GenAI could train minds to appreciate simple pleasures and help them work towards health goals with family and support networks. - How can we responsibly scale up GenAI tools?
According to Naomi Korn Gold, Senior Director of Innovation Consulting, deploying GenAI tools ethically involves considering “access, cost and presumption of fast reliable internet access.” Without this lens, GenAI solutions may risk further obscuring social determinants of health. For instance, while some have multiple devices, such as Apple Watch or Alexa, that would enable interactions with GenAI healthcare solutions, 23% of US households do not have internet access. Failing to intentionally include these groups may exacerbate the social challenge of healthcare access and healthcare inequity. - How can we develop a technology that naturally steers patients to consistently adhere to the right health practices and decisions?
If GenAI is to continuously assist patients with adherence, we must develop an in-depth understanding of “how the users are going to be interacting with the system,” Rioux says, emphasizing that designing LLMs to “degrade gracefully” is a critical challenge for the inevitable moments when the system doesn’t have the correct answers or is simply dysfunctional. Patient interactions should be designed so that doing the right thing is “the easiest path to follow.” Achieving such precision in design takes a multidisciplinary effort.
Second Thoughts, Second Opinions
Of course, there is much more to say about GenAI and patient adherence. It’s obvious that we must adhere to a focus on patient needs and ethical diligence. It’s our collective responsibility to identify and manage the inherent risks of this delicate, nascent field. Let’s take it on!
But the truth is, this work will require more conversations. Informed ones. Our hope is that these blog posts have suggested some questions to you and your institution. If you’d like to speak more about it, meet me on LinkedIn.
