Ryan Kelly (CTO, Capital Rx) was skiing with his daughter when she came down with a cold. One urgent care visit and two lost days at the mountain later, his daughter’s cold had turned to strep. She was prescribed Augmentin, which she is allergic to, but somehow the doctors had overlooked the note on her chart. Standing in line at the pharmacy counter waiting for the medical office to answer the phone, Ryan knew there was a better way to get the prescription right the first time. Now his daughter would have to suffer another day without the correct medication. And then it hit him: We need a faster, smarter way to clinically assess his daughter’s personal medical information to resolve these issues in real-time – or better yet, ensure they don’t happen to begin with.
Healthcare information is being collected every time someone visits the doctor, has an operation or fills a prescription. However, most companies today haven’t taken on the opportunity to integrate patient data and analyze it in order to improve patient care. Given the increasing use of artificial intelligence and machine learning in healthcare, there’s a considerable opportunity to review the data and provide insights on the overall health of the population, with the potential of leading to fewer procedures and fewer hospital visits just by providing better prescription drugs.
Looking Beyond Prescription Drugs
Capital Rx is a next generation Pharmacy Benefit Manager (PBM) with a purpose-driven mission to change the way pharmacy benefits are procured, administrated, and processed for employer groups. In 2017, the three founders of Capital Rx, AJ Loiacono (CEO), Joseph Alexander (COO), and Ryan Kelly (CTO), recognized the existing PBM industry is shadowed by complex and opaque pricing and contracting terms, but see that the future holds potential for PBMs to deliver untapped value that improves patient outcomes.
Ryan Kelly is unique in the PBM industry, possessing expertise in modern software architecture, claim adjudication, and prescription benefit services. With over a decade of experience in software development, Ryan successfully designed and implemented systems that have processed and stored tens of millions of claim transactions.
Capital Rx was built with the desire and intention to deliver value in the form of high-touch clinical services such as real-time prior authorization, data-driven programs analyzing how genes affect person’s response to drugs, and deeper insights on the health of employee populations. This approach to proactive management is far ahead of competitive PBMs today, yet Ryan and his team knew at the heart of this revolution was completely rebuilding the claim processing infrastructure.
“We know that the current systems used by PBMs are sunsetting within the next few years,” says Kelly, “But it’s not just technology updates to adjudication and reporting that’s going to change the industry – it’s the integration of AI.”
Reduction of Human Error
Humans are fallible and are therefore prone to mistakes and oversights. It is impossible to reduce all aspects of unwanted variability in human behavior, so errors like what Ryan and his daughter experienced are to be expected. But at what point does the need for proactive clinical assessment surpass the acceptance of human error? Technology gives us the ability to engage in proactive clinical assessment and to resolve bottlenecks in real-time without manual intervention. By reducing the fatigue that results in human error, patients will see an improvement in problem resolution and overall service.
Artificial intelligence and machine learning algorithms have the capacity to assess a problem and streamline the workflows in real-time. Resolving problems quickly and effortlessly ultimately lead to an improved patient experience. Humans are too busy performing their tasks in the “proven” way to have the time necessary to analyze data the way AI does. Part of what makes AI so efficient is the pattern recognition and learning behavior that the human brain does not have the capacity to replicate. Over time, these AI algorithms will understand the protocols in proactive prescribing, fulfillment and follow-up protocols to minimize issues in the patient journey.
“We recognize the significance of AI in healthcare and are investing significant R&D into these algorithms,” says Kelly, “We want to utilize AI to improve workflows and build better plan designs in a faster and safer way than humans can.”
Investing in Healthcare is Investing in Humans
Healthcare produces incredible amounts of data. Dozens of supply chain participants produce petabytes of information (basic research, clinical trials, lab work, medical diagnoses, claims utilization, financial reimbursements, etc. ) that require an unimaginable amount of resources to analyze with today’s standard practices.
Partly due to the sheer amount of data, most healthcare companies have failed to look beyond their operational silos and combine their information with additional data sources. Available and shareable data is one of the most important aspects of integrating AI and machine learning into measuring the health of a population. Learning algorithms can become more precise and accurate as they interact with different kinds of data, uncovering unprecedented insights into plan design and optimization.
Healthcare is also a highly personal and individualized. Currently, it would be impossible to assess and treat every human on the planet with a personalized approach, however AI is getting us closer to that possibility. The problem is that healthcare companies and plan sponsors today are run by humans, who don’t have the time to ask a simple question: what is the plan trying to solve for?
Capital Rx looks to improve plan designs by targeting the right mix of cost sharing and drug adherence, which should ultimately lead to improved medical spend. For example, better drug adherence can leads to fewer emergency room visits, drives down medical costs and direct people-related costs like payroll due to sick days. The program that Capital Rx is building aims to make program recommendations based up utilization data and underwriting, with the goal of a fully predictive, automated recommendation process.
“AI is a key component of our technology roadmap, promising incomparable efficiency in advancing patient outcomes,” says Kelly, “From forward-looking analytics that deliver actionable insights to utilizing pharmacogenomics to create custom formularies, we are looking to AI as the lever needed to move our industry from a focus on cost to, instead, one of value.”
Data is the Language of AI
With a large data set and a complex plan, it becomes extremely difficult to scale workflows and determine the best outcomes. Healthcare is personal, and without an answer to the question of what the sponsor is attempting to solve for, designing a plan becomes a bit like throwing darts while blindfolded.
Too many plan sponsors stay focused on “lower costs”, but the number of additional variables to consider is extensive and perhaps even more significant. Given the plan size, demographics, utilizing population, copays and coinsurance and trying to find the right combination of patient access, customer service, formulary design, patient outcomes, this ever-growing list of variables can produce a factorial nightmare. Depending on the weight of each variable, which is different for each plan sponsor, a manual analysis of an employer’s health benefit options, is more of an artform, compared to a precise science.
Health care companies must confront an uncomfortable truth, that working with human administrators will never scale with vast amounts of available data. This isn’t to say that humans cannot perform the necessary calculations and produce reasonable results, but to deliver a precise recommendation for an entire company’s employee population is like using a toothpick to paint the exterior of the Empire State Building. Technology, and particularly AI, is allowing us to scale and take on more complex factors which are necessary to model different approaches to improve healthcare on every level.
Programmatic vs. Consultative Approach
It’s no secret that the use and investment in artificial intelligence has been spreading like wildfire throughout the healthcare industry. From a patient’s diagnosis to medical imaging, AI offers a revolutionary way to reduce errors, enable earlier detection of disease, and improve overall patient care. At both the patient level and the plan level, AI is offering opportunities for faster and smarter analyses, streamlined workflows, and the ability to exponentially increase scale.
Next winter when Ryan Kelly is back in mountains skiing, he’ll still be concerned if he or his daughter experiences another medical emergency. However, he can’t help thinking about the promise of the future and the quality of care improvements that will be afforded through AI.