The Journal: Gill, can you describe the role that artificial intelligence can play within physician practices?
Gill Eapen: Yes, but first, let me take a minute to discuss what artificial intelligence is. Artificial Intelligence is the idea of taking predictive analytics and applying it to economics to make outcome-based decisions that improve a company’s value. As a financial advisory firm with roots in valuation, Stout helps CEOs and CFOs make better economic decisions by analyzing raw data within a company and identifying trends that can influence key financial levers and the economics of the company. Unlike organizations that provide technology and analytics around data, we start by identifying the business problem and then using analytics to solve for it. We’ve coined the name, artificial intelligence, because this is a new capability not offered in the marketplace. In the case of physician practices, we help answer two fundamental questions:
We start the process by meeting with CEOs and CFOs to identify areas of highest priority. Then, with assistance from the organizations’ information technology and data specialists, we identify the most important data held in electronic medical records (EMR) systems that could help us create models to improve decision-making in specific areas. Our clients then continually use our models to identify optimal interventions and timing, triage patients to the best provider, and reduce overall costs of provision while simultaneously improving outcomes for their patients.
The Journal: Discuss some of the specific problems you’ve helped physician practices overcome.
Eapen: A few specific examples come to mind:
We are practicing personalized medicine in every sense of the word, with better outcomes for patients, efficient usage of physician time, and higher economics for the practice.
The Journal: Why is this segment of the healthcare industry ripe for this type of innovation?
Eapen: Generally, industries with dynamic data and complex decisions will benefit the most from artificial intelligence. I have worked with many companies in the oil and gas, aerospace, and pharmaceutical industries – which all face similar challenges as healthcare – and they have been implementing predictive analytics and economic modeling together for decades.
Previously, I also led research and development at a major pharmaceutical company so I bring a unique perspective and understanding to how patients respond to treatments and interact with clinicians. From my experience, I see a huge opportunity to improve treatment and economic outcomes at physician practices.
The Journal: What’s the process of working with physician practices?
Eapen: Every project is different, but we always start each one with a conversation with the CEO and CFO on the issues and challenges the practice faces. Depending on the problem, we also meet with experts both on the clinical side and business side to fully understand the status quo and any constraints that may exist.
We then come up with a data plan in coordination with the information technology professionals. Our systems are capable of using very large amounts of data, but in physician practices, the data sets, particularly from a decision making perspective are much smaller. Our processes can be used effectively by a practice of any size, from hundreds of thousands of patients to a single physician.
We use a proprietary software, Decision Options®, that has been developed over the last two decades, solving challenging artificial intelligence problems in many industries. This highly automated technology has a library of over 100 well known machine-learning techniques and applies sophisticated methods to organize and analyze data to reach decisions. Because of this, we can go from data to robust models in weeks.
Decision makers can use these models in a number of ways through their existing systems, including EMR systems. Information can also be pushed to the decision makers by email or text. Our goal is to ensure that the users can take advantage of the information with no additional work or data and no new systems implemented. The models will provide information to the decision makers at the right time and place, seamlessly.
The Journal: What is your vision for artificial intelligence over the next five years?
Eapen: I have been in this area for nearly 30 years, starting with my graduate thesis in 1986. Over the years, there have been many ups and downs in this area, which is now being called Artificial Intelligence (AI). Since we have been embedded here for a long time, we have a strong understanding of and experience in artificial intelligence and AI. Now that computing power has increased tremendously, we can solve many problems we simply could not even approach 10 years ago. So, the next five years will be very exciting in this field and Stout is at the forefront of it all.
In fact, in healthcare, I believe physician practices not incorporating artificial intelligence will have a hard time staying in business. Because we can solve these problems quickly and efficiently, we have had a lot of interest from smaller practices. We can typically engage with practices without any upfront investment from them and once we build usable models that produce actionable information, we can enter into a subscription based arrangement that scales with the practice as it grows organically or through acquisitions.
Contact Gill Eapen at +1.860.536.9990 or firstname.lastname@example.org.