Komodo Health Hacks COVID-19
It’s not every day that you are given an opportunity to work on meaningful solutions for—quite literally—the biggest problem in the world.
But when COVID-19 began wreaking havoc across all industries, we knew health data would play a key role in bringing the pandemic under control. At Komodo Health, patient-level health data is our expertise. We believe better decisions begin with data, and the insights therein to help us move toward the right solution.
We recently had the opportunity to deliver critical insights at the Pandemic Response Hackathon, co-sponsored by Komodo Health, and convened by Datavant. Twenty Komodo Health dragons, ranging from data scientists to clinical experts and engineers, formed three teams to collaborate with esteemed health policymakers, public health experts, and technologists from around the globe to create tech-enabled COVID-19 solutions. The five of us made up one of those teams.
Amid the adjustments to quarantine, our team collectively logged more than 40 hours during the last weekend in March to “hack” COVID-19. We debated priorities, defined cohorts, and crunched numbers, culminating in recommendations to get ahead of COVID-19-indicated drug supply shortages by projecting the usage of these therapies over time across geographic regions with various levels of COVID-19 disease burden.
Our project, COVID-19 Drug Demand and Supply Curves, defines the COVID-19 patient population based on an overlap of relevant medical and pharmacy claims in order to identify spikes in therapies prescribed to this population and underscore potential supply shortages. These types of projections will be critical as strained healthcare resources are directed to the locations where patients need them most.
How We Did It
We looked at a wide range of both medical and pharmacy claims to narrow down the population of COVID-19 patients in the U.S. As a result, we were able to predict regional drug demand and supply outages impacting our population of interest. By leveraging our patient definition, physicians and corporations are able to get ahead of the curve to proactively reduce the disease burden in the U.S. healthcare system.
Building out our project was not without its challenges. With so much attention on COVID-19, the pandemic has generated an abundance of information. While incredibly valuable, it was difficult to filter through the noise and find the key information we needed as data inputs for the project. The inconsistent diagnosis coding for COVID-19 made this project more challenging still. Even though the CDC introduced a new COVID-19-specific diagnosis code in April, many cases were being coded as bronchitis, pneumonia, or other respiratory conditions that were not necessarily specific to COVID-19. We had to be creative and came up with a proxy to find what we believed to be the relevant patients and created a definition given a lack of a diagnosis or procedure code.
Out of 230 teams, our project was selected as one of five “Spotlight Projects.” We were blown away! Spotlight Projects were selected based on their potential to be highly impactful in mitigating the effects of COVID-19, as well as feasibility of implementing the project in the real world. Project ideas were evaluated by a panel of healthcare luminaries, including:
- David Shulkin, Former VA Secretary
- Tom Daschle, Former Senate Majority Leader
- Don Berwick, Former CMS Administrator
- Andrew von Eschenbach, Former FDA Commissioner
- Ed Simcox, Former CTO of HHS
- Stephen Konya, Senior Innovation Strategist, ONC
- Georges Benjamin, Executive Director of American Public Health Association (APHA)
- Priyanka Surio, Director of Data & Informatics, Association of State and Territorial Health Officials (ASTHO)
- Tom Leppert, Former Mayor of Dallas
- Peter Neupert, Healthcare Technology Leader
In a closing ceremony, healthcare luminaries commented on our project:
“Using this type of tool that was built for predictive planning, I think, is absolutely incredible, and is going to help not only for this pandemic, but for future pandemics.”
– David Shulkin, Former VA Secretary
“This gives us an opportunity to really track [therapies] in the course of the pandemic… we can’t lose sight of the ancillary drugs [involved].”
– Andrew von Eschenbach, Former FDA Commissioner
We want to expand our universe definition of COVID-19 to include new diagnoses and potential therapies as they are rolled out to the market. Our analysis focused solely on specific regions, but it could be applied to other geographies, depending on how the virus spreads. We plan to build a more robust model that will help us identify unusual drug demand activity in order to quickly alert appropriate health industry entities and suppliers of any spikes in utilization. In addition to COVID-19, we also believe we can expand the scope of this tool to other therapeutic areas in order to more proactively anticipate drug demand spikes and supply shortages.
Priyanka Surio, director of data and informatics at ASTHO, provided our team with guidance. She advised us to think in terms of how we can collaborate with existing public health information systems to share these types of alerts and build our tool to scale. Needless to say, we are excited to explore this.
COVID-19 has weighed heavily on all our minds, from a population health perspective as well as personal concerns for loved ones. Thanks to the commitment of so many leaders nationwide, we are grateful to have been given the opportunity to channel our energy and concern toward building solutions.