BERG, AdventHealth Use AI to Scan Patient Records, Improve Outcomes
By Bio-IT World Staff
November 24, 2020 | BERG and AdventHealth have launched a collaboration to build a patient registry biobank intended to enhance COVID-19 patient care by rapidly identifying the right path to improve patient outcomes.
AdventHealth is a non-profit healthcare system with locations across the US. Berg and AdventHealth initially formed a partnership in the fall of 2019 to improve outcomes and precision medicine for those with Non-Alcoholic Fatty Liver Disease (NAFLD) and Sarcopenia but shifted the focus to COVID-19 in the spring and launched a pilot in Florida.
AdventHealth will release health data to Berg in two phases. First, the healthcare system will release data on patient demographics, COVID-19 clinical information and personal medical history from Florida. BERG will leverage its AI-enabled Interrogative Biology Platform to review AdventHealth’s comprehensive electronic medical records for all patients tested for COVID-19 to identify the best treatments for each patient based on their biology. Among critical data that will be considered include hospital or ICU length of stay, all administered medications, personal medical history, and patient outcomes.
In the second phase of the project, Berg will work with enterprise-wide data from across AdventHealth’s nationwide locations, conducting an exploratory analysis of chronically administered meds that could be linked with a better outcome or lower probability of SARS-CoV-2 infection.
Currently, “BERG is collaborating on the types of data going into BERG’s AI system and organizing the initial algorithms where they will be applied to the type of molecular and laboratory data, in addition to clinical data that's analyzed by the AI system,” explained Niven R. Narain, Co-Founder, President & CEO, BERG, in an email to Bio-IT World. “The initial insights will be shared with Advent leadership and scientific and clinical staff as a calling out point as to the areas we want to focus on.”
Steven Smith, Senior Vice President and Chief Scientific Officer, AdventHealth, explains the advantages to AdventHealth of this return of data. “We have the opportunity to go back and do additional analysis with samples that we have stored in our bio repository from hospitalized patients with COVID-19 to drill down into specific hypotheses that may come out of that,” he said, also by email.
After Berg’s Interrogative Biology Platform identifies the hypotheses that are likely to have the most impact on patient care, the second phase of the collaboration, “will focus on validating some of the data, either the EMR and the anonymous or aggregate of clinical features, and an outcome of data medications,” Narain continued. The focus is on “iterative learning,” Smith added.
A pilot is already underway in Florida and will serve as a model for the rest of the USA. Narain said that Berg is processing data in real time and expects to share the first dataset with in the next 30 - 45 days. “Once the first model is built and early validation is complete, the second phase of project,” will launch, he said.
Both Narain and Smith see great potential for benefits in the program—both for patients and the healthcare system. They expect a more efficient and predictive triage process, potential refinement of the evidence-based treatment algorithm, and better patient outcomes.
Berg and AdventHealth do plan to circle back to NAFLD and Sarcopenia. They expect the framework they are now building for COVID-19 to be applied to other diseases.