Log in

Countless advancements in health care data collection and the use of artificial intelligence to analyze that data are opening up virtually limitless possibilities for mining patient information for new clinical insights. But the process remains very difficult.

A data aggregation and analytics company affiliated with multiple Catholic health systems has taken on this challenge and is helping to produce vital research on health outcomes. In the four years since its founding, the for-profit Truveta has amassed a dataset of more than 100 million patient records. The company has been using quickly evolving artificial intelligence tools to process that data and make it usable for researchers.

Dr. Michael Simonov, Truveta vice president of product, says, "It used to take about two decades for research to impact patient care, and that was deplorable. We are shrinking down that timeframe, and we also are expanding the type of data used."

"We are changing the way researchers study and analyze data," he says.

Born at Providence
Truveta was the brainchild of leaders at Providence St. Joseph Health, who started exploring around 2018 how best to use the power of big data to answer pressing questions about health care delivery and outcomes. Providence, Trinity Health, Advocate Health and Tenet Health together launched Truveta as an independent company in September 2020.

Truveta's collective membership has grown to 30 health systems that provide 18% of the daily clinical care across all 50 states from 800 hospitals and 20,000 clinics. Catholic systems that are members of Truveta are Bon Secours Mercy Health, CommonSpirit Health, Providence and Trinity Health.

Truveta member systems participate at various levels, with some serving on the Truveta board, some contributing as investors in Truveta, and all contributing data to Truveta's dataset and getting access to the data for use in research. They get access to that data as part of their paid membership.

Daily updates
The member health systems provide data from patient medical records to populate Truveta's dataset. According to information from Truveta, this "daily updated data delivers the most complete, timely, and clean data on U.S. health."

Simonov says that Truveta has been evolving its processes of using artificial intelligence tools and other technologies to de-identify the patient data, clean it and organize it for optimal use. Simonov says the methods are so advanced that researchers can use the deidentified data to conduct longitudinal studies, correlate maternal health to fetal health, track trends in public health, and conduct other types of analysis.

He says Truveta also has greatly expanded the types of medical record data that it can pull and use in its dataset. This now includes genomic information, clinician notes on patients, and data from electrocardiograms, CT scans, PET scans and other imaging. It also can include information relating to the social determinants of health, which can allow researchers to study the connection between health disparities and social determinants.

Researchers at life science and pharmaceutical companies, government agencies and academic institutions can buy access to Truveta's dataset.

'Juicy' findings
Simonov, who was a medical researcher in academia before he joined Truveta, says the richness of Truveta's data and the great capabilities of the processing tools have allowed for some "juicy" findings from the medical records.

"You pick a disease, and we can study it," he says.

A recent area of focus has been maternal and fetal health. Truveta announced in the spring that it has developed the largest and most complete mother-child electronic health record dataset available. According to information from Truveta, "with more than 1 million mother-child pairs, these data will enable researchers to discover insights into the continuum of care from pre-pregnancy through childbirth and beyond, while upholding the highest standards of privacy protection and regulatory compliance."

Truveta expects researchers to use this specialized dataset to better understand the connection between maternal health and neonatal outcomes, monitor the safety of medications and vaccines used by pregnant moms, and explore the correlation of various childhood conditions with mothers' demographic characteristics and other factors.

Already, analysis conducted by Truveta's internal research team has found, through the study of more than 500,000 cases in the Truveta dataset, a higher risk of heart failure in the 90 days following delivery for women who experience preeclampsia. The researchers found that Black women with preeclampsia are twice as likely to have heart failure as white women with the condition. The study sheds light on maternal health outcomes in the U.S., and how race may factor into related disparities.

Simonov says it is because of the smart use of large complete datasets and data-processing tools that this important information can be gleaned.

He says that such information can help clinicians and patients. He notes that without early intervention, disparities can worsen and poor health outcomes can result.

He says, "We're doing all this to advance our mission of saving lives with data. It's impactful work — it's reaching patients."

Related Articles:

Truveta says its platform could speed medical research, inform treatment decisions

Systems say records platforms have ethical, security safeguards

De-identified medical records data is big asset for health systems