Issue link: https://beckershealthcare.uberflip.com/i/1348345
37 DATA ANALYTICS & INFORMATICS CDC taps Kaiser to lead data-driven COVID-19 vaccine safety surveillance effort By Jackie Drees O akland, Calif.-based Kaiser Permanente's Vaccine Study Center is overseeing a national data effort focused on analyzing the safety of COVID-19 vaccines. The CDC selected Kaiser to lead the Rapid Cycle Analysis through the agency's Vaccine Safety Datalink network, according to a Jan. 6 news release. Organizations that share data through the network include five Kaiser Permanente regions, HealthPart- ners in Minneapolis and the Marshfield (Wis.) Clinic. The surveillance requires searching through a wide collection of electronic patient records to identify potential ad- verse health effects, such as strokes or anaphylaxis, or severe allergic reaction. The data analysts will look for rates of particular outcomes, comparing them with a number that would be typically expected in a given population. "We'll be looking at heart attacks, for example, and it's on the list because no one really knows if there is a risk from the vaccine," said Nicola Klein, MD, PhD, director of Kaiser's vaccine center. "So we'll monitor to see if there are excess cases, and if so, whether an increased risk is associated with a COVID-19 vaccine." Eight healthcare institutions that par- ticipate in the VSD share weekly data reports on the selected diagnoses, and the data analysts search for any possible connection with a COVID-19 vaccination. Dr. Klein said results of the project will be shared within the Vaccine Safety Datalink network. n U of California-Irvine develops model to predict COVID-19 outcomes: 5 details By Laura Dyrda H ealth science researchers at University of California-Irvine developed a machine-learning model to predict outcomes for COVID-19 patients. Five details: 1. The university announced its tool to predict whether COVID-19 patients will need ventilators or ICU has been available online for free since Dec. 17. 2. The model takes a patient's medical history into account when determin- ing whether the patient is at high risk of needing additional care or can be sent home. Based on studies conducted at UCI Health, the tool has a 95 percent accuracy rate. 3. To develop the tool, researchers analyzed data from UCI Health COVID-19 patients dating to January 2020. They finished an initial proto- type of the tool in March 2020 and studied its effectiveness. 4. The machine-learning model used the data to create an algorithm that takes preexisting conditions, test results and demographic data into ac- count to generate a severity score for COVID-19 patients. Researchers also took into account feedback from emergency medicine, hospital medicine, critical care and infectious disease physicians. 5. Researchers plan to expand the tool for other institutions for research and conduct a new study to see which patients will benefit from COVID-19 drug trials. n Epidemiologists launch open- access data platform with 5M+ COVID-19 records By Katie Adams A n international group of epidemiologists on Feb. 24 launched an open-access data platform that houses more than 5 million anonymized COVID-19 records from 160 countries. The website can be reached at global.health. Each record can contain various information about the COVID-19 case, including data on travel history, demo- graphics, testing and outcomes. Google provided $1.25 million in funding for the project. It is led by researchers from Baltimore-based Johns Hopkins Center for Health Security; Boston Chil- dren's Hospital; Boston-based Northeastern University; Cambridge, Mass.-based Harvard University; Washington, D.C.-based Georgetown University; Seat- tle-based University of Washington and University of Oxford. n