The diagnosis and treatment of rare diseases is often delayed by years, with the average rare disease patient spending nearly seven years visiting physicians before receiving an accurate diagnosis. But a new approach to big data is speeding up time to diagnosis, thereby speeding up time to treatment and saving lives.
The term “rare disease” can be misleading. The Food and Drug Administration (FDA) recognizes more than 7,000 rare diseases which, combined, affect more than 30 million Americans. But these conditions present specific challenges for those dealing with them; medical research is hamstrung by difficulties in recruiting for clinical trials and, even for conditions where treatments exist, physicians may not have the tools to recognize patterns of these diseases which present so infrequently and often have a wide range of symptoms that may not come together into a cohesive diagnosis.
The right data could present a powerful antidote to these challenges, but the traditional players in healthcare data are missing important pieces of the rare disease puzzle.
“The problem with most health data sources today is that they sample only a small slice of the patient journey,” says Aswin Chandrakantan, M.D., chief medical officer at Komodo Health, a healthcare intelligence platform vendor. “They lack a harmonized and complete view of patient journey and, in turn, lack the completeness necessary to address rare disease challenges.”
Mapping the healthcare journey
Healthcare data is heavily siloed. “Outcomes-based research is driven very much by electronic medical records (EMRs) and electronic health records (EHRs),” notes Tabby Khan, M.D., MPH, a senior clinical product specialist at Komodo Health. “Those often only represent interactions from one hospital system. If you’re in New York and you see a physician at NYU and then, two weeks later, you see a physician at Mount Sinai, your EMR/EHR data only shows researchers half of the information.”
Komodo is filling in these gaps by building the first comprehensive Healthcare Map™. With data from a wide variety of sources, they trace healthcare interactions — with primary care physicians, specialists, pharmacies, and other players — of more than 325 million U.S patients. Using machine learning, all of the data for a unique patient is combined and presented under a single patient token, while remaining de-identified.
This linking of patient data exposes previously hidden patterns. “When you don’t have all the pieces, you really can’t solve the puzzle of rare disease — whether at the patient or population level,” notes Chandrakantan. “We see three- to five-times more encounters per patient than any of the legacy data aggregators out there, so we can understand the longitudinal experience of a patient.”
Identifying patterns to improve diagnosis time
This real-world evidence-based approach holds important potential to improve outcomes for rare disease patients.
Real-world evidence data is de-identified information about patients that can be aggregated to draw conclusions about a broader group. It can include medical claims, disease state symptoms, patient history captured in ICD-9 and ICD-10 codes, procedural history captured in current procedural terminology (CPT) codes, prescriptions, hospitalizations, and granular information about the number of dosages and refills.
That data can be analyzed for patterns, like prescriptions or reported symptoms that often precede the diagnosis of a rare disease. Recognizing these patterns can speed up the time to diagnosis for a person living with a rare disease.
“Let’s look at this through the lens of an example, like transthyretin-mediated amyloidosis,” Khan continues. “This rare genetic disease impacts around 50,000 people worldwide and takes an average of 10-15 years to diagnose. By understanding what happens during those 10-15 years — the specialists a patient is visiting, the symptoms they are experiencing, the false diagnoses they may receive, the patterns in their interactions with the healthcare system — we can help the healthcare system drastically improve their ability to spot this rare disease and treat it accordingly, reducing morbidity and mortality.”
Streamlining clinical trials
Komodo has built a patient-alerting product called Pulse to leverage their data for physicians. “Pulse is the ‛livestream’ of the real-world evidence market,” says Chandrakantan. “It’s constantly updating to show you precisely where patients meeting a very specific set up criteria are, the providers they are seeing, and the healthcare systems they are interacting with.”
Not only can this make a tremendous impact on improving physician education — to make sure the right doctors are on the lookout for patterns that could indicate rare diseases — but it also has the potential to radically speed up and iron out the design and recruitment of clinical trials. Recruitment for trials for rare diseases has long presented challenges, as identifying potential participants from these small disease pools of patients can be time-consuming, expensive, and inefficient.
“About half of all of clinical trials fail because of recruitment challenges,” explains Chandrakantan. “With the Healthcare Map, life science companies can pinpoint the right patient at the right time to offer trial participation and get those patients to the potentially life-saving therapies they need. Ultimately, these technologies reduce disease burden, accelerate recruitment, and increases the quality and speed of trials.”
And Chandrakantan notes that the benefits go far beyond individual clinical trials. “Looking at it from an efficiency perspective, if real-world data sources, free from the traditional siloes of healthcare data, can bring down your failure rates by half, a biotech or pharmaceutical company could be running more trials for the same amount of money and the same amount of time — or less. That means more successful therapies making it to patients in less time and, eventually, for a lower cost.”
Chandrakantan says part of Komodo’s role is educating the healthcare market. “Many people in this market are still operating under the traditional models but we know there is so much potential to be realized by applying technology and data effectively to the healthcare industry,” he says. “If you want to get life-saving therapies into the hands of patients, this is the way that you need to operate. This is the new world.”
To learn more about healthcare mapping and rare disease diagnosis, visit https://www.komodohealth.com/.
