Verato Offers Patient Matching Challenge to Vynca
MCLEAN, VA–(Marketwired – Nov 28, 2017) – Verato, Inc., a leading provider of SaaS-based patient matching solutions including the revolutionary Verato Universal™ MPI, today announced that it is, indeed, ready to rumble — and has challenged Vynca, the winner of the Patient Matching Algorithm Challenge sponsored by the Office of the National Coordinator for Health Information Technology (ONC), to a head-to-head competition to see which company’s matching approach is more accurate. Should Vynca take up the gauntlet, the ensuing competition will forever be remembered in the history books as the Very Exciting Real-world Algorithm Test Of 2017 (or “VERATO 2017″).
Verato currently weighs in at a whopping one billion identity resolution queries answered. Even more impressive is that more than 15% of the US population is currently being matched by Verato’s next-generation SaaS-based master patient index (MPI) solution, which leverages a “Referential Matching” architecture to make matches conventional algorithms can never make.
The ONC’s Patient Matching Algorithm Challenge was open to anyone to participate. All participants were given the same set of one million synthetically-generated patient identities (with synthetic names, addresses, birthdates, etc., for each identity) to run through their matching algorithms. The synthetically-generated data was based on real-world data in an MPI, and correct matches within the dataset were designed to mimic real-world scenarios. Matching accuracy was calculated based on how many correct matches an algorithm found without falsely discovering incorrect matches. Participants got to retest their algorithms up to 100 times and would see their accuracy scores after each try.
“First and foremost, we’d like to congratulate Vynca on winning the Patient Matching Algorithm Challenge and commend the ONC for running it,” said Mark LaRow, CEO of Verato and well-known mixed martial artist whose street-fighting name is The Macho Matcher. “The healthcare industry needs greater clarity on the capabilities and limits of patient matching algorithms and this was a good first step. However, all conventional algorithms have a mathematical limit to how accurate they can be, which is why today’s MPIs have anywhere from an 8% to 20% duplicate rate. We think our innovative Referential Matching approach can win outright against any conventional algorithm. We’d like to put this hypothesis to the test using a dataset of real patient identities with real-life errors and anomalies. Only when put to the test against real data in the real world can a patient matching approach truly be measured.”
Verato did not enter the Patient Matching Algorithm Challenge because the challenge used a synthetically-generated dataset of patients to test participants’ matching approaches. Basically, the Patient Matching Algorithm Challenge was akin to a boxing match between a boxer and a punching bag, because any conventional algorithm can match synthetically-generated data that has been doctored with predictable patterns of errors, typos, and inconsistencies — especially when savvy participants can adjust their algorithms 100 times. The Very Exciting Real-world Algorithm Test Of 2017, on the other hand, would require matching real-world data. And Verato’s Referential Matching approach is extremely accurate precisely because it matches real patient data.
Conventional algorithmic approaches directly compare two patient identities to see if their demographic data is the same or very similar. Verato’s Referential Matching approach, on the other hand, compares each patient identity to Verato’s comprehensive reference database of highly-curated demographic data spanning the entire US population. This reference database includes not just current and correct data for each person, but also out-of-date and frequently-occurring incorrect data — like previous addresses, maiden names, and common typing errors. In essence, this reference database is a pre-built “answer key” for demographic data for every patient in the US. By referencing this answer key during the matching process, Verato can make matches that conventional algorithms can never make.
“Ultimately, conventional algorithms can only ever be as accurate as the underlying patient data they are matching,” continued LaRow, whose signature mixed martial arts move is called The Dirty Data Destroyer. “But 30% or more of patient demographic data is out-of-date, incorrect, errored, or incomplete — and this number is rising each year. Referential Matching is not limited by the quality of the underlying data it is matching, and in fact it thrives in low-quality data environments. We look forward to putting our Referential Matching approach to the test against using real-world data in the Very Exciting Real-world Algorithm Test Of 2017.”
Verato offers a cloud-based matching platform that links and matches identities across disparate databases or organizations with the highest accuracy rates in the industry. Verato leverages an extensive self-learning database of U.S. identities as a reference, or universal “answer key.” And because it is cloud-based, the Verato platform is less expensive, faster to implement, and more scalable than traditional matching technology. Verato is based in McLean, VA. For more information, visit www.verato.com.