.png)
A recent interview with On Amir, a professor at UC San Diego, the chief behavioral science officer for Fiverr, and a Well-Beat Advisory Board member. This conversation explores Well-Beat's solution for the patient adherence issue, the impact of COVID-19 on the healthcare industry, and future directions for Well-Beat and health care.
Professor On Amir’s widespread accomplishments and interdisciplinary academic background have given him a unique perspective on high-tech companies and how they can effectively adapt based on new research in behavioral science and technological advancements. We asked him a few questions to understand what impact he believes Well-Beat can have on the healthcare industry.
Read on to discover how Well-Beat solves the foremost crucial healthcare issue: patient adherence.
It is an honor to have you with us today Professor Amir. To start, can you share a bit about your background?
My career began as an F-16 fighter pilot in the air force. There, I was able to apply my background in computer programming to develop various training programs and systems that helped pilots make improved decisions. To expand my interest in computers, I began studying for a Bachelor's degree in Computer Science while I simultaneously continued to serve as a pilot. Continuing my education, I earned a PhD in Management Science, Marketing, with minors in Psychology, Economics, and Statistics from MIT. Like my work with systems in the air force, my PhD is really about decision making, which falls at the intersection between psychology, economics, and data analysis. I like to call this field behavioral science.
Today, I am the Wolfe Family Presidential Endowed Chair in Life Sciences, Innovation, and Entrepreneurship, and a Professor of Marketing at the UCSD Rady School of Management. As a professor, I have undertaken many research projects relating to consumer behavior, judgement and decision making, and risk and uncertainty. I have specifically studied the high-tech, biotech, defense, and online marketplace industries. Applying this research to practical industry settings, I have worked as the Chief Behavior Scientist for Fiverr Incorporated for the past few years.
Why did you choose to get involved with Well-Beat?
Ravit Ram, the CEO & Co-Founder of Well-Beat, approached me with their idea when she was about to launch the company. She proposed a high-tech solution to solve the patient adherence issue, which often results in increased healthcare costs, medical complications, and even death. She explained to me that she sought to create a more effective patient engagement platform by combining Machine Learning with Behavioral Science.
When asked to help with a new business idea, I evaluate it using three questions:
Does the company bring an important and necessary service to its clients?
Is it a realistically achievable goal?
Do I personally possess the necessary skills needed to add value to their objective?
After evaluating Well-Beat with these criteria, it was obvious that there was potential to create enormous value for the health and wellbeing of patients. Given recent advancements in technology and behavioral science, I believed this idea could be effectively executed. And, being at the intersection of these fields, I knew I could contribute to the company’s success, so we were a good match.
From your work with various health organizations, can you share what challenges they face? How do you see Well-Beat fitting into and changing this industry?
As a professor, I often join and guide students on hands-on projects with companies, and quite a few of these have been in health care. A key commonality to many of these is that they were data-driven. The healthcare world has large amounts of both structured and unstructured data, correlating to a transition from making hand-waving decisions to making data-driven business and medical decisions. However, the problem is that there is way too much data.
In the air, pilots are exposed to a vast amount of information in real time and at a very fast pace. To guide a pilot’s decisions, there needs to be systems that organize, analyze, and present data in a human-centric manner. Ideally, after a decision is made, we would also like to verify that the pilot arrived at the optimal decision, but the quickly changing environment of a fighter pilot leaves no time for verifications. Therefore, we must often rely on heuristics to help make good decisions on the fly.
The healthcare world is much more complex. There is a famous niobate physicist that explained how physics is complicated enough, but at least it is a closed system where all the variables are understood. Now, imagine if instead every electron could think on its own, creating constantly changing variables.
This is akin to the situation that we are presented with when focusing on understanding humans - the challenges are exponentially more complex than any law of physics.
The patient adherence problem is a prime example of the complexity involved in classifying human variables. While effective medical tools based on scientific research often exist, convincing humans to decide to use them is an entirely different story (take for example the challenge of convincing the world to get vaccinated). Businesses have already begun the process of using data to understand their clients and increase sales. Well-Beat replicated the concept to understand patients and enhance health. The information we gather allows us to devise effective methodologies, that will grant us the ability to make calculated medical decisions. We do not yet know if these decisions are “optimal,” but we do know that they create far better results than the status quo. Every step forward into this uncertainty and every piece of new added data helps us update, solidify, and understand a little more about patients and brings us one step closer to reaching an optimal solution for each individual.
After learning about these challenges within the healthcare industry and the Well-Beat solution, I was confident that we can easily create value by developing this platform and help every organization. Today’s healthcare companies are virtually swimming blind, but we can give them some light. Even now, before we have collected a vast amount of data, we provide immediate results because of the automated personalized recommendation for each patient interaction. As we continue to learn and improve the system, we will come even closer to a general solution.
In one of our pilot runs at Sheba Medical Center, our results showed a 300% improvement to patient adherence with Well-Beat’s system, just by providing basic tools. I call them basic, because our system was only at its beginning, rudimentary stages. We can and will do much better over time because our Machine Learning system is, by definition, designed to learn. And, it can learn on a much larger scale than any human or existing system. A very skilled nurse can likely learn their patients’ behavior and decide how to work with them on an individual level. However, the nurse cannot teach others how to work with any patient in the world, cannot generalize their solutions so they can be applied elsewhere, and they may get it wrong sometimes.
If you want a system that can teach, generalize, and get it right at a significantly higher rate, you need something like the Well-Beat solution.
Each professional learns how to solve a specific problem or identify a specific aspect of their patients, but they may miss other points. On the other hand, Well-Beat can aggregate and learn from existing information in order to provide ‘decision aid tools’ and automated systems that suggest customized treatments and personalized interventions for patients which will only get more accurate in the future.
How does Well-Beat’s AI for patients differ from that of AI for business customers?
The healthcare industry faces an AI dilemma. Usual business methods involve targeting customers’ desires and the factors that initiate a purchase. Solving the patient adherence issue requires a much more sophisticated approach. Each person has their own specific dynamic of states and emotions with respect to their condition, and these are much more relevant in medical treatment than in consuming everyday products because of heightened sensitivity. This emotional aspect demands that the caretaker use specific tones, instructions, and reminders to ensure patient compliance.
Well-Beat is tasked with understanding the deep psychology behind patient communication. This task proves challenging because, more often than not, the patient themselves are not even cognizant of which tactics work best for themselves. However, this task has important implications because using the wrong approach could lead patients to reject the treatments necessary for their own health. This emphasis on individual psychology is part of the reason why Well-Beat’s algorithms are so innovative. They successfully provide tools and even the vocabulary that healthcare providers should use in order to understand and assist each patient according to their specific needs. It is important to note that the solution is not independent of everything professionals are doing, but rather integrates with existing methodologies to work with whatever data they currently have and enhance caregiver’s outcomes.
What long-term impact do you think COVID-19 will have on the healthcare world, and how will it specifically affect Well-Beat?
When healthcare organizations integrate a new solution into their existing methods they need all sides to “buy-in.” So, if the industry is unwilling to change, the platform simply will not work. Even the best possible solution will not work without the chance to prove its efficacy. The pandemic has changed the willingness of the healthcare world to try new tools and embrace innovative ideas.
In addition, the mandatory quarantine and social distancing guidelines that were implemented made it crucial for providers to offer virtual care, encouraging organizations to try out the available digital tools. Many healthcare professionals previously believed that, without face-to-face meetings, it was impossible to provide effective patient support. Now there is a willingness to change, brought on in part by necessity. In the post-pandemic era, there is a huge opportunity for digital tools to prove their efficacy.
What value does Well-Beat provide for resource management?
The healthcare system has found itself in dire economic strain, creating a need to quickly become more efficient and effective. Well-Beat offers much-needed optimization. It eliminates overspending, improves existing tools’ efficacy, enhances job satisfaction, and reduces unnecessary costs caused by lower adherence.
What future directions do you see for Well-Beat, especially in regards to integration with futuristic technologies such as facial recognition and virtual reality?
The Well-Beat solution was planned from the beginning to be able to take all input types into consideration. It is source-agnostic which means that, no matter where information comes from, we can integrate it within our system. The fact that it is an open solution makes it flexible to any new updates and technologies without requiring excessive efforts.
Do you think the market is currently ripe enough for this solution?
Health care is a very broad market with many needs, and each company has their own requirements for success. Well-Beat was built to be dynamic and able to provide solutions for all forms of organizational needs. The demand for Well-Beat is increasing dramatically.
I have no doubt that, within ten years, every respectable healthcare organization will be using a Well-Beat styled solution.
Comments