Q&A with Dr. Nehama Lewis
A whole lot of AI power, mathematical brilliance, behavioral science theory, and research, and emotional intelligence goes into designing the tools and recommendation models for our SaaS platform. We e-met with Dr. Nehama Lewis, Medorion’s Chief Behavioral Officer and expert in behavioral science, who addresses some of your most frequently asked questions, such as how our SaaS platform combines members’ clinical data and behavioral science to communicate with members to persuade them at scale, and of course - why it works.
1. How is Medorion’s technology better at predicting and influencing decision-making than what data scientists are already doing?
The typical approach to predicting decision-making or behavior in data science is to identify the data elements that are associated with the outcome you are trying to predict. For example, if I was to try to predict who will be vaccinated for COVID-19, I might use a database of individuals and see what variables are significantly associated with the likelihood of being vaccinated. The problem with this approach is that it’s not scalable, and produces insights that can be invalid or unhelpful (for example, that people who own pets are more likely to be vaccinated). This approach doesn’t identify the human (cognitive and emotional) factors that influence people’s decisions, which you can apply to other contexts or use in a persuasive message. In contrast, Medorion uses models and constructs from behavioral science theory that have been empirically tested and found to predict human behavior across a range of contexts. Medorion’s technology combines this behavioral layer with engagement data to make a smarter prediction model that targets real obstacles to behavior.
2. How do you combine behavioral science and machine learning?
Until recently, the field of behavioral science was largely of interest to researchers, who applied theories and concepts to the design and evaluation of large-scale interventions. Machine learning technology has developed rapidly, becoming increasingly more sophisticated. It has opened up new opportunities for reaching large populations at scale and tracking engagement. As machine learning technology and computational data analytics developed, it became possible to combine the two. We can now collect and analyze enormous amounts of data and apply concepts from behavioral science to identify the factors that influence decision making that can be applied to messages tailored to the individual.
3. How do you appeal to people’s emotions to influence decision-making?
People make decisions about health (and other contexts) based on a range of factors. These factors can be rational or may be based on emotion, or a mix of both. In order to understand and influence the way in which people make decisions that can enhance their health, it is important to consider all of the elements and obstacles that might come into play. For example, the behavioral layer in our technology can incorporate message design elements that are based on theories of fear appeals to design messages that are likely to be more persuasive. As fear can be a relevant factor that drives many behaviors, this approach can be applied across contexts and populations.
4. How accurate is Medorion’s Behavioral AI tech compared to our traditional patient profiling?
Traditional patient profiling doesn’t distinguish between significant data predictors and significant predictors that are also meaningful elements that may influence behavior. When your analysis is data-driven, the end result may seem to be useful insights but may not be useful, and might even be spurious and misleading. The risk is that this approach can skew focus away from what is really important for decision making toward data elements that seem to be important but are simply the result of ‘noise’’ or proxies for unmeasured meaningful predictors. Carl Bergstrom and Jevin West do a great job of covering this in their book ‘Calling Bullshit’.
5. How much data do you need to get the job done?
As much as I can get! Data is an essential element in the process. When we have data that are relevant to the context we are working in, it increases the probability that the technology will identify the meaningful factors that drive decision making. Data is also an important element in ensuring that we target different people with messages that emphasize different decision-making factors, based on data elements that relate to that factor.
The most advanced health communication campaigns made simple, scalable, and clickable.
Learn about Medorion’s technology.
Contact us for more information about how our Behavioral AI SaaS can empower your team.
Why members’ most recent experience determines their attitude towards your health plan and what you can do about it.
Medorion’s Persuasion AI shows how to predict tolerance levels in older adults and what you can do about it.
Why health plan managers need to understand the barriers behind medication adherence before they can effectively address them.