Truveta aims to use massive data repository to aid research
The new company is aggregating de-identified data from 20 member healthcare organizations with the goal of using it to support evidence-based care.
Twenty healthcare organizations are making a big bet on the power of big data. They’ve invested millions in Truveta, a new company that’s building a massive database of de-identified patient data to support research that uses artificial intelligence.
Seattle-based Truveta hopes to use that data aggregated from electronic health records of patients across the U.S. to answer a variety of questions, such as which treatments are most effective for rare diseases, how therapies can be tailored based on a patient’s genomic, clinical and social profile, or which patients might benefit most from a new drug.
Unlike many other data aggregation research initiatives that involve one large provider organization, Truveta incorporates input from a broad array of major providers across the nation that each have an ownership stake. That wide swath of data offers promise for future evidence-based research augmented by computer intelligence capabilities.
“If Truveta realizes its potential, it will allow us to do a better job of identifying patients for clinical trials, help with bedside clinical decision-making and maybe even help us decide where to locate a clinic,” says Ari Robicsek, MD, chief medical analytics officer at Providence, a 52-hospital system that was one of Truveta’s first investors. “There is huge potential to improve the care we provide to communities by using data intelligently.”
Laura Wilt, CIO at Ochsner Health System, one of Truveta’s newest members, says Truveta offers “an opportunity to collaborate and apply collective de-identified data to improve patient care, advance health equity and expedite understanding of rare diseases.”
Ochsner, which operates medical centers and clinics throughout Louisiana, expects research through Truveta could help its “Healthy State” initiative, designed to improve healthcare access and quality in the state.
Early efforts
Truveta, launched late last year, is still in the preliminary stages of aggregating data, refining its platform and identifying research projects, says Providence’s Robicsek, who serves on the company’s clinical and scientific advisory committee.
As of early December, de-identified data from 35 million patients’ records had been aggregated in the “Truveta Platform” database, and more records are being added daily, a Truveta spokesperson says. The database is designed to support a wide variety of research at provider organizations as well as life sciences firms, including pharmaceutical companies.
Based on the 20 healthcare organizations that have invested in Truveta, the company contends on its website that its platform will “represent more than 16 percent of U.S. patient care from tens of thousands of clinical care sites in 42 states.”
Truveta reports that it’s raised almost $200 million in investments from its 20 “member” healthcare provider organizations and other investors. “Truveta continues to be governed and majority-owned by our health system members,” a company spokesperson says.
Microsoft is making a “strategic investment” in Truveta, and Truveta will become a Microsoft Cloud for Healthcare partner, the company reported in a recent press release. Terry Myerson, Truveta’s CEO, formerly served as executive vice president at Microsoft.
Truveta is using Microsoft Azure as its cloud platform. “Truveta will run its platform on Microsoft Azure and combine its own AI services with Microsoft AI and ML [machine learning] to scale and accelerate the pace of healthcare discovery,” its spokesperson says.
Microsoft has “no rights to the Truveta data,” according to an earlier announcement from the start-up firm.
Truveta facilitates research by transforming data for faster insights using AI and machine learning, structuring clinical data into a data model that employs the Fast Healthcare Interoperability Resources, or FHIR, standard, and leveraging SNOMED CT, RxNorm and LOINC ontologies.
Data supports COVID-19 research
In its first effort to demonstrate how researchers can use its platform, Truveta last month released the results of a study designed to gain some insights on COVID-19; the research was based on aggregated patient records in its database.
Results highlighted the ability of the aggregated data to quickly provide insights from a large population, the company says. “In just under two weeks, our team was able to ask and answer important medical questions using one of the largest comprehensive real-time datasets of fully vaccinated Americans,” says Nick Stucky, MD, director of clinical research at Truveta and a practicing infectious diseases physician at Providence Portland Medical Center.
The study determined that 9 percent of those who have received the COVID vaccine manufactured by Pfizer-BioNTech and had a “breakthrough” infection were hospitalized, while the hospitalization rate for patients with breakthrough infections was 15 percent for those who received the one-dose vaccine developed by Johnson & Johnson.
Among patients with high-risk conditions, those with chronic kidney disease required hospitalization most often when they had a breakthrough COVID-19 infection, the study determined.
The study also showed that fewer than 1% of those vaccinated have experienced adverse events, with those getting the Moderna vaccine reporting the highest rate of such events.
Future research
With the COVID-19 research serving as an appetizer, member organizations are hopeful that Truveta can help open the door to a wide range of big data research by providers, such as determining which patients would most benefit from telehealth, says Providence’s Robicsek.
In addition, Truveta expects that pharmaceutical companies will use its platform for several purposes, including monitoring the safety and effectiveness of drugs and identifying patients who could participate in clinical trials.
“Health systems will receive compensation when their de-identified data is used in research” by such outside entities, the Truveta spokesperson says.
Truveta’s platform eventually will contain a wide variety of de-identified EHR data, including lab results; vitals; diagnosis and procedure codes; physicians’ notes; pathology reports; diagnostic images and genomics, the company says. The data is de-identify using ML technology to help ensure compliance with HIPAA, the company says.
But the prospect that Truveta may partner with other data processing companies raises concerns about how it will ensure that its de-identified patient data can’t be matched up with identifiable data, potentially violating privacy.
The Truveta spokesperson offers an indirect answer to this concern: “We are open to partnership opportunities … [but that] comes with an enormous responsibility to protect the health data we manage and maintain the highest standard of patient privacy. We have developed some of the most advanced data security and patient privacy systems and processes in the industry, including multiple de-identification steps, to ensure we are taking advanced measures to protect patient data.”
Wilt, the CIO at Ochsner, is hopeful Truveta can achieve its goals of enabling researchers to build on each other’s work.
“The sky’s the limit on what diseases, drugs or devices could be studied and researched with new levels of depth and accuracy because representative data updated daily is available to us,” she says.
Howard Anderson, contributing editor, was the founding editor of Health Data Management.