Better mental health treatments with synthetic patient data
GettyImages/Yuichiro Chino
Synthetic data will allow the Virginia Department of Behavioral Health and Developmental Services to research mental health treatments without exposing patients' identity.
The Virginia Department of Behavioral Health and Developmental Services (DBHDS) will be creating an anonymized digital twin of patient data to more securely explore artificial intelligence applications to advance patient care.
Working with AI firm Diveplane and Iron Bow Healthcare Solutions, DBHDS will first deploy Diveplane's GEMINAI synthetic data engine, which creates a duplicate dataset of patient information. That digital twin will have the same statistical properties, nuances and characteristics of a population of interest, but it will contain no personal information associated with patients that might reveal their identity.
With GEMINAI, DBHDS can generate “synthetic patients” with specific medical conditions that fit certain demographic profiles, all without the personal health information of the original dataset, and with no one-to-one relationship back to the production data or any way to reverse-engineer the data to tie it back to a real person.
According to the Department of Health and Human Services' HealthIT.gov website, synthetic health data can help researchers, health IT developers and informaticians test theories, data models, algorithms or prototype innovations.
In a Dec. 21 solicitation, DBHDS said it wanted Diveplane's AI software so it could "study the feasibility of developing a secure, deidentified, renewable, and relational database of criminal justice, behavioral health, and other human services records to facilitate development of more effective interventions."
The department said the data it had been using in its test and development environment did not meet security baselines for protection patient data. For those less secure applications, DBHDS needed synthetic, or properly de-identified and HIPAA compliant data.
In addition to providing the synthetic data, the department said it wanted capabilities for machine learning prediction, data characterization, decision reasoning, transparency and auditability.
With data gathered and synthesized from 12 facilities, 40 community service boards and 52 treatment centers, the AI software will help DBHDS share data with partners develop new mental health treatments before patients turn to illegal activity, drug use or self-harm, the solicitation said. Using synthetic medical records would also reduce data quality errors caused by manual processes and thereby improve research outcomes.
The next phase of the project will explore the broader use of Diveplane's REACTOR Understandable AI platform, which will help DBHDS explore how it can influence care strategies to support enhanced patient outcomes. The technology is built on top of a human-understandable machine learning platform and enables supervised, unsupervised and reinforcement learning techniques, the companies said in their announcement.
"Data privacy and security is at the heart of everything we do, as we are entrusted by the public to keep personal information private," said DBHDS Chief Information Security Officer Glenn Schmitz, "GEMINAI will enable us to securely analyze and share our data, both internally and with other government agencies."
"We are on an amazing journey with the Virginia DBHDS, and we are excited about the possibilities this collaboration can deliver," Diveplane CEO Mike Capps said. "Their ambition to use AI that is designed to be natively explainable at every step is a testament to the culture and values of the organization that put the patient first."