The Apervita Approach to Health Analytics
By Sarianne Gruber
“Tossed away is the idea of a data model and relational databases” avowed Dr. Brad Ryan, as he began comparing Apervita’s new technology platform to the traditional health data solutions. Dr. Brad Ryan, Chief Commerce Officer of Apervita, describes its core functionality as an “enabler”, but to the user, it is market place exchange on a cloud-based platform. Apervita has conceived a common way for sharing information, and coined it “a democratization of health knowledge”. They have unlocked analytics as a self-service and removed the need to hard code algorithms to a given data model. A health knowledge base can function with anything that can be made computable, you name it: algorithms, analytics, pathways, protocols or measures. Prominent health institutions have joined Apervita’s rapidly growing analytics and data community. Current members include the Cleveland Clinic and the Mayo Clinic, with the shared prospect that portable, computable knowledge will improve patient outcomes. Having recently finished an $18 million Series A funding round, Apervita continues to reach medical institutions, hospitals, payers and pharmaceuticals with the benefits of health knowledge via a marketplace venue as an exchange.
I had the opportunity meet with Apervita’s Chief Commerce Officer, Dr. Brad Ryan while attending the September MedCity Converge Conference in Philadelphia. A market exchange for healthcare knowledge intrigued me. If I was researcher or modeler, I may want to test the performance of my algorithm using another institution’s de-identified patient sample. Participating institutions on the “community site” can publish their datasets with payment terms. Dr. Ryan clarified that publishing a dataset does not mean the data gets transmitted between parties. The data is private and secure, and reporting might not be on a patient or record level. He also mentioned they have a lot of enterprises wanting to publish and monetize their data. This scenario could also work the other way around. I would publish my analytic, in this case a predictive model algorithm. A “community site” institution could then request my analytic to be applied on their data. None of the code, data coefficients, or other intellectual property has to be published to the recipient, just the predicted scores. Dr. Ryan explained that in order for the two institutions to link together on the platform, the “authoring side” sets up concepts and values that are packaged with the analytics, and as long as the institution from “the user side” has that same set of concepts, value sets that can “extend or add to, or map” to the user’s own health information. Hence, the self-service facility is established by linking the ontology of all the shared knowledge elements maintained and exchanged on the platform.
Perhaps my next question may seem pretty intuitive. Would one then consider this platform a work around for interoperability? Dr. Ryan replied yes and no. He feels that a part of him doesn’t want to be associated with interoperability because of its legacy, which has all been about central planning, based on the definition and imposition of those standards. At Apervita, they are trying to solve those problems, but are doing so in an open market. “A common way of sharing can evolve and build upon itself “, explained Dr. Ryan, “our platform is a two sided market place and we want an open market business”. It is his conjecture that Apervita’s platform may be a “side step” to interoperability by letting demand drive the work that goes into mapping and building bridges between ontologies and value sets at a community level.
Getting back to the data; I asked Dr. Ryan how the platform computes health knowledge. I learned the environment is NoSQL, and a traditional ETL process is not used. The data tends to be structured and does not have a required format. All data is abstracted into facts and metadata about those facts. A fact becomes its own piece of information, and it can be used in a variety of different purposes. Dr. Ryan explained that “data is taken from different systems and a variety of formats, forms and fidelities”. It is with the implied ontology, which is the way data is represented or the relationship between pieces of data, a value-set of codes can be created and linked to another dataset’s ontology. If I wanted to publish my algorithm on the Apervita platform, I would implement it using Python code.
On a final note, Dr. Ryan shared his thoughts on the value of the Apervita platform on a consumer level. Based on his own experience working at IMS Health, he witnessed many requests from researchers for licensed data sets. However, there were times when the customer’s questions didn’t necessitate the licensing of the whole dataset. It would have been ideal to provide a custom report or just pay for an answer to a question. This is an attempt to solve that problem. Apervita has a radically different approach. They have set out to create scalable solutions. And they are succeeding.
For more information on Apervita and demo is available on the website.
Brad is an industry thought leader in digital health and intelligence solutions, focusing on innovation at the intersection of clinical medicine, data, and technology. At Apervita, Brad leads the commercial, product, and market development efforts to democratize health analytics & data through Apervita’s unique distribution platform and marketplace model. Prior to Apervita, Brad was General Manager of Payer and Provider Solutions at IMS Health, leading development and commercialization of data and analytics-based technology solutions in areas such as value-based care, pharmacy cost and utilization management, transparency, and referral management. Previously, he was a leader in the Healthcare Practice at McKinsey & Company. Brad received his M.D. from the Johns Hopkins School of Medicine and his B.S. in Chemical Engineering from the University of Alabama.
MedCity Converge Speaker Bio September 1, 2015