Chase Spurlock, PhD, Founder & CEO of Decode Health, on Predictive Data Analytics for COVID-19 & Chronic Disease*

*Including Chronic Inflammatory, Autoimmune & Autoinflammatory Diseases!

A post in our ongoing series of B2B thought leadership interviews with visionary entrepreneurs in digital health. We focus on companies that apply digital tools to improve patient care of chronic immunoinflammatory disorders (CIDs), including autoimmune and autoinflammatory diseases. We were particularly pleased to interview Chase Spurlock, PhD, founder and CEO of Decode Health. The COVID-19 pandemic has given them a novel opportunity to demonstrate the power of Decode Health’s predictive data analytics. 

Chase Spurlock, PhD, founder and CEO of Decode Health
Chase Spurlock, CEO Decode Health

COVID-19 is also an opportunity

The COVID-19 pandemic has been the most disruptive event in healthcare since last century’s HIV/AIDS pandemic. Even more so, since SARS-COV-2 spread so much faster to so many more people. Even so, jet-speed world travel was a big factor in the spread of both diseases. But COVID-19 emerged in a completely different world of digital infrastructure, vast quantities of data and new analytic technology platforms.

In particular, advances in artificial intelligence and machine learning (AI/ML) have enabled a new generation of innovators, including Decode Health. Decode Health’s solution grew from work on diagnostic laboratory testing for autoimmune diseases. Their process used machine learning to match laboratory results to blood patterns of patients known to have target diseases. This enabled potential diagnosis of  those diseases earlier than conventional diagnostic approaches. However, the Decode Health platform has proven just as powerful for detecting populations at high risk of SARS-COV-2 infection and poor outcomes. 

What’s more, apropos of Decode Health’s origins in autoimmune analytics, COVID-19 has dragged immune disorders into the spotlight. Activist COVID “long-haulers” have accomplished in months what took AIDS activists years. Moreover, what the autoimmune community has been struggling to do for decades! That is, to get the medical community, research and clinical, to focus on CIDs, conditions that strongly resemble long-COVID syndrome. Viral infections may also trigger or exacerbate many autoimmune diseases. These include multiple sclerosis (MS), systemic lupus (SLE) and rheumatoid arthritis (RA), to name the most familiar.

Beyond autoimmune to COVID-19 & chronic disease

The COVID-19 pandemic and lockdown responses disrupted healthcare systems worldwide in the biggest way since HIV/AIDS. More importantly, it happened much faster to many more people. This was not just a threat but an opportunity. Chase saw that COVID-19 prediction provided an instantly relevant use case for AI-powered predictive analytics.

COVID-19 & social determinants of health (SDoH)

Early on, stakeholders realized that predictive insights to support mitigation efforts enhanced managing responses to a raging pandemic. It also became apparent early in the pandemic that social determinants of health (SDoH) were useful predictors of risk. This is true both of the risk of SARS-COV-2 infection as well as poor outcomes. SDoH include access to healthcare, education, economic stability, neighborhood and community context, among others. As Chase explained, understanding and managing chronic disease is not just about clinical care: procedures, diagnoses, treatments, medications. Social, physical and environmental factors can trigger disease or influence patient outcomes as much as the clinical journey. 

“What we found to be critically important to understand outcomes, whether in COVID-19 or chronic disease, are patient environments and social determinants of health, including geographic location, access to healthcare, patient resources, demographic trends and ages. All of these factors help to develop risk profiles outside of the clinical journey to predict patient outcomes.”

Chase Spurlock

Through its work with COVID-19 data, the company discovered that the SDoH that most influence poor outcomes change over time, warranting ongoing monitoring of  populations. Additionally, SDoH vary regionally and by population. For example, Decode Health analysed SDoH trends in Utah and Florida populations, searching for multiple sclerosis (MS) adverse events. In Florida they found a strong correlation with poor access to mental health providers. In fact, that appeared in the top ten correlates of bad MS outcomes. Thus, to re-think care strategies in MS, payers should look at anxiety & depression detected by PHQ9 (Patient Health Questionnaire 9–the mental illness screen) assessments. 

The emerging threat of chronic disease 

Chronic diseases (now including post-COVID) loom as the biggest and most costly long-term issue in healthcare. These potentially high-cost conditions lurk in all populations. This in not just a patient care issue, but one influenced yb behavior and SDoH. The big chronic disease categories are cancer, cardiac, pulmonary, metabolic (type 2 diabetes, obesity), then chronic inflammatory diseases (CIDs). Unfortunately, the medical community often overlooks CIDs because they are not in a single category, like cancer. Instead CIDs include some 100 disorders isolated from each other by body part, disease and medical specialty. 

US healthcare systems have trouble managing chronic disease patients because they do not respond well to the sick-care, fee-for-service (FFS) model. One reason is many chronic diseases are not diagnosed until patients are obviously sick and much more expensive to treat. Type 2 diabetes is the classic example; many people do not even realize they are ill until advanced symptoms emerge. This is one of the drivers towards value-based care (VBC). In this model, payments by patient (capitation) or shared-risk case management try to better align incentives of payers and providers. A parallel goal is improving care outcomes for patients.

Chronic inflammatory and autoimmune diseases

This is especially true of CID, autoimmune and autoinflammatory patients. Initial symptoms are often vague and easily confused with other diseases. Care practices may dismiss patients’ complaints or refer patients out to mental health treatment. That can lead to delayed diagnoses (sometimes by years!) while disease progresses. The same issues have emerged with post-COVID syndromes. This shares similarities with many earlier post-viral syndromes. These are conditions that have been woefully under-researched and generally dismissed by clinicians for decades. 

Decode Health’s predecessor IQuity focused on autoimmune diseases and CIDs. Such diseases in total are at least as costly as oncology, but are much less visible to payers. Once payers start using population AI analytics for targets they are already aware of: COVID, COPD or diabetes 2, the use-case for autoimmune/CID becomes much more obvious.

Predictive analytics using many types of data

So, how to find these chronic disease patients earlier? Predictive analytics approaches that leverage artificial intelligence/machine learning methods are no longer just promising, but proven, to enhance detection of chronic disease patients earlier than previously possible. Chase told us Decode Health uses a variety of data to feed their predictive framework, depending on which data their customers actually have. 

“Healthcare data comes in many forms. As an analytics provider we need to meet our customers where they are from a data perspective and be flexible in the data requirements necessary to produce predictive insights.” 

So, Decode Health utilizes a wide variety of data:

  • Claims (payments),
  • PHR (patient health records),
  • ICDs (International Classification of Diseases–digital codes for diagnoses),
  • CPT (Current Procedural Terminology),
  • HCPCS (Healthcare Common Procedure Coding System–”hick-picks”),
  • NDC (National Drug Codes) and more. 

COVID-19-specific data sets

For COVID-19, Decode Health uses data from point-of-care questionnaires. For instance, from SARS-COV-2 testing stations that collect data at the point of test. Decode Health pairs that with SDoH data at state, county, and community levels.

Detecting SDoH and long-term patterns

Chase told us he thinks there’s also an opportunity to define personas of patients who are more or less likely to suffer post-COVID or other post-viral syndromes, and track the emergence or exacerbation of chronic disease. Obviously, researchers are only starting to have access to a longitudinal record in the context of COVID-19. However Decode Health and other analytics organizations can also utilize historical data sets from other outbreaks, such as seasonal influenza, to uncover SDoH and clinical patterns leading to poor outcomes in viral respiratory illness outbreaks.

Anonymized public datasets, especially those with a longitudinal view, help Decode Health identify disease trends alongside the SDoH. Such SDoH trends change over time. In the COVID-19 world, the rapid movement of people from one area to another could dramatically change the composition of regional SDOH features making it important to monitor for those changes over time.

Cost control through chronic disease detection and management

Speciality pharmaceuticals (e.g., biologicals), including infusion-delivered medications, are major cost drivers in payer populations. Health plans and self-insured employers are looking for strategies to optimize these treatments due to their high cost. MS patients, like oncology patients, are among the most expensive to treat. Use of Decode Health analytics can help identify patients who may be misdiagnosed and potentially on wrong, very expensive, treatments. 

Decode Health can also provide directional, insights-based, risk trend analyses. Most payers must deal with the fact that they have thousands of patients, but only resources to care-manage a few hundred. Often, the only data they have is how much they spent on patients. Decode Health found that trailing average costs aren’t good predictors of a looming bad event, so they seek out additional sources of data, as with COVID.

After population analysis, identifying and stratifying patients guides payers toward better population management. Looking at the history and projecting forward, Chase told us Decode Health has seen 40-60% cost savings possible by engaging patients sooner. For example, Medicaid aims to squeeze 4-5% out of the system every year. Analytics can help them understand where to add services to get that savings.

“Analytics companies need to work alongside care management providers with demonstrated utility. Predictive analytics delivered alongside providers who can pull these insights through to impact patient care will achieve the best outcomes.”

Chase Spurlock

Realizing savings through case management

More importantly, to achieve savings, payers must pursue more case management. For example, Chase brought up Mymee’s efforts to better manage, even reduce, use of biological therapies through tailored, individualized trigger identification and avoidance coaching programs for autoimmune patients. Management software platforms like SonarMD in IBD can get patients back to specialists in time to reduce hospitalizations and length of stay, both big cost drivers. SonarMD has done their own analytics, but Decode Health can find high-risk-of-hospitalization patients in advance to work hand-in-hand with care management software companies who can help keep them out of emergency care.

“Autoimmune populations are low prevalence, high cost diseases that may not always be top of mind. Patients are often faced with a diagnostic odyssey that lasts for years before reaching a diagnosis. In payer populations right now are patients with undiagnosed chronic illness including autoimmune disease.”

Beyond drug costs are other less visible factors. For instance, difficult and delayed diagnoses, especially in MS, lupus, or RA where expensive diagnostic and monitoring procedures are now standard of care. Furthermore, the longer the patient is untreated, the more the disease progresses. 

Chronic disease detection and stratification

Chase told us Decode Health’s analytics can tune models to sub-segments of populations. Pick any garden variety chronic disease, like type 2 diabetes (T2D) or COPD. Decode Health has developed unique ways to map clusters of patient journeys. For example, sub-populations with T2D plus asthma show particularly bad outcomes, so that’s a target group for the payer to focus resources on to improve outcomes while realizing cost savings. 

A geographic example is to compare populations in Tennessee vs Vermont. What are the top five cost factors for each of those groups? It might be one set of five in Tennessee and a different set in Vermont insofar as there are SDoH and demographic differences. Decode Health analytics looks at those breakdowns and can predict risk tuned specifically to that given payer’s population.

“Decode’s analytics can predict outcomes in common chronic illnesses including heart disease, diabetes and COPD. Decode’s approach can spotlight risk payers may not be aware of and is extensible to a variety of disease indications.”

Chronic disease management through analytics

Chase described Decode Health’s ideal position as being a keystone embedded within ongoing population management, connecting analytics to specific patient personas. By building personas, essentially stand-in patients with characteristics typical of those with a condition, Decode Health can provide insights into potential poor outcome risk for such patients. The next step will be linking patients to social services or other elements in a partner’s network to address SDoH deficiencies. 

There is still work needed on understanding the common threads that lead to bad outcomes so as to tackle communities’ responses. For instance, Unite Us is developing menus of SDoH services community by community. If a healthcare system can determine which of these work in particular communities they can do a better job of controlling everyday, non-COVID respiratory infections as well as COVID-19. Whether infectious or chronic disease, if payers and providers can identify and stratify, they can begin to apply case management more effectively. 

SonarMD’s focus on IBD means they can take analytics and apply them to their own case management programs. Mymee targets autoimmune, a broader focus over multiple diseases with immunological dysfunction in common. They are at the top of the list as best-in-class partners for autoimmune disease analytics because they know what they are doing and know how to touch these patients, that is they already have good ideas on what to do with the information provided by data analysis. 

Three things to know about Decode Health in Chase’s own words

“Our chronic care management platform finds patients trending toward a diagnosis, those with uncontrolled disease, who are likely to experience an adverse clinical event, and those who are potentially misdiagnosed and may be receiving expensive and incorrect care.”

“At the core we want to help healthcare become more proactive, anticipate outcomes and use our framework to help understand which patients will have good or bad outcomes to better manage care. We want to embed our solution as a keystone in a partner’s network to help them orchestrate the tools at their disposal that impact care.”

“Decode Health’s analytics not only works for payers that have access to specialized care management programs like Mymee or SonarMD, but can also help others know which programs to include as part of their plan design. It helps them benchmark their population and find opportunities to improve patient outcomes and generate cost savings opportunities.” 

Contact Decode Health

Decode Health is a Nashville-based AI solutions provider. Using multiple data sources including public health information, claims databases and proprietary social determinants of health data, Decode Health deploys unique data modeling techniques and machine learning to impact care on two critical healthcare fronts: COVID-19 mitigation and chronic disease management. Decode Health’s technology provides actionable insights early, enabling proactive care, better outcomes and significant cost savings. For more information, visit https://www.decodehealth.ai.

Contact DrBonnie360

We approach these interviews from our two different multi-lens perspectives

  • DrBonnie360: clinical dentist, Wall Street analyst, patient advocate, and digital health consultant. 
  • Ellen M Martin: evolutionary life science, finance & investor relations, marketing, communications and writing/editing.

DrBonnie360
Strategic Consulting & Professional Services 

We provide professional consulting and services to companies working to bring the best of digital, conventional and functional medicine to patients with chronic inflammatory diseases. 

  • We are thought leaders helping our clients to apply digital health innovations to chronic inflammatory, autoinflammatory & autoimmune disorders. 
  • Our subject matter expertise includes oral health and microbiome, autoimmune patient advocacy, digital health, self-hacking and more.
  • We have decades of experience in finance, marketing, and communications for healthcare and life sciences.
  • Our backgrounds include clinical dentistry, osteology, biotech investor and public relations, marketing communications, content creation, strategic consulting, autoimmune advocacy and much more.

Contact us for help defining and articulating your marketing position and strategy, including conducting virtual facilitated brainstorming and planning sessions. We excel at creating content, including articles, blog posts, collateral materials, web site copy and white papers. Our Your Autoimmunity Connection website showcases our own content.

Disclosure–we have done paid consulting work for IQuity and Mymee.

Leave a Reply

This site uses Akismet to reduce spam. Learn how your comment data is processed.

0
%d bloggers like this: