Patient engagement programs should continually evolve in order to meet patient needs and behaviors, explained Linda Schultz, PharmD, vice president, clinical and customer success, AllazoHealth, during a session at the Pharmacy Quality Alliance 2021 Online Annual Meeting. Without this element of continual evolution, these engagement programs may fail to achieve their desired outcomes.
In the past, engagement programs were adapted to meet patients’ changing engagement behaviors by first testing intervention and then assessing whether the intervention should be added to the program or should replace an existing one.
Today, Schultz explained that such efforts are ineffective and outdated due to the ever-changing advancements occurring in digital platforms and technology. In light of this, there are new artificial intelligence (AI) engines that can analyze data—including social vulnerability indices—which can make it possible for program owners to analyze more complex vectors in order to not require binary decisions around advancing patient engagement behavior.
Through the use of AI, interventions can target patients as individuals to assess the greatest impact method for each patient. Additionally, such a personalized approach allows for considerations such as the impact of social vulnerability on intervention response to be accounted for.
Without AI, technology has to follow preset rules or a sequence of steps in a linear process. This process is not necessarily outcomes-based and is not patient-specific, Schultz explained.
“Whereas with AI, we can focus on desired outcomes, and through varied and large datasets, technology can make decisions on how to best achieve that desired outcome. It can focus on what is going to resonate with the targeted individual, be it a patient or prescriber, when it comes to the health care arena. We’ve seen that in social media galore,” Schultz said during the session.
By using AI, the engine can be told where to go instead of what to do. Then, on its own, the AI engine can figure out the best route to get there. In order to do this, the engine can test hypotheses for future actions and learn from the results gleaned from large and varied datasets. This method of testing can make the process more efficient and effective when learning the appropriate approach to gathering information.
“There are large and varied datasets that AI engines leverage and learn from across the health care spectrum. These are taken in from various places, from program historical data, including pharmacy and medical claims data, as well as patient demographic data. It also is taken in from various resources that track things like consumer behavior and social determinants of health, for instance,” Schultz said.
From these data, the AI engine can then learn and predict who and what is best for each individual patient in order to effectively influence that person. Furthermore, the coordination of internal and external programs is possible at the same time and does not end up overlapping each other.
This coordination can also help to eliminate the need for predefined business rules, which are normally used to deliver the same message to all patients at the same time. Through the use of AI, messages are specific to patients to promote engagement.
“It can even get down to the detailed level of the subject lines of an email or email content blocks, as well as other personalization methods that would help to influence a particular person’s engagement. All of this supports the ability to update existing interventions or even provide insights to new intervention ideas that would be worth adding to a patient engagement program,” Schultz said.