Data Sharing Impacts Human Health
Like all other industries, the medical field is continuously challenged—and supported—by emerging technologies. With the use of ethical and responsible data strategies, health care professionals can use modern innovations to take the best care of U.S. citizens.
For Belinda Seto, who is a trained biochemist and serves as the deputy director at the Office of Data Science Strategy at the National Institutes of Health (NIH), the focus of her role is the interoperability of data.
Seto shared her thoughts with an audience of health care professionals at the annual AFCEA Bethesda Chapter Health IT Summit 2024 in Bethesda, Maryland.
At NIH, the requirement to share meaningful data in a standardized format helps collect foundational, discovery, clinical and other types of research. Working with data from across the United States and international locations, the NIH must comply with data sharing policy and management, Seto explained. “If taxpayers fund your research, you have the obligation to share your data.”
“Part of my responsibility is to oversee how we exchange patient clinical data,” she went on. The COVID-19 pandemic played a large role in ramping up data exchange tactics, while keeping privacy safeguards in mind.
An ongoing challenge within the field is the use of different systems among health care providers. “How do we make these systems communicate and exchange data?” Seto continued. At NIH, the launch and practice of the Fast Healthcare Interoperability Resources (FHIR) standard helped drive data in a meaningful way to best understand underlying pathophysiology of COVID infections, she shared.
Seto additionally shared an upcoming request for information to be launched by NIH. “How would you like NIH to think about minimum core data elements?” she asked. The question will be similar to the collection of demographic data, seeking healthcare professionals’ opinions on whether patients should be asked for their age, sex assigned at birth, preferred gender, race and ethnicity.
Upon review of all responses, the NIH will decide whether there will be a set of minimum core data that all NIH-supported clinical trials will have to collect. With that type of collected data, electronic health records will make it easier to search databases hosted by NIH to build cohorts. “That's where we’re going in terms of standardizing common data elements,” Seto stated.
Regarding artificial intelligence (AI) and machine learning (ML), Seto explained that the technologies have been widely used in the health care industry—especially in radiology—for the last 10 years. "When I go get my annual mammogram, no matter which manufacturer for the machine is embedded in it, it’s a computer assisted diagnosis, so that the radiologist is not reading every single scan.” AI has allowed health care professionals to examine and analyze patient case history to plan the most appropriate treatments and solutions, Seto told the audience.
“When we talk about artificial intelligence and models ... we’re thinking about the whole life cycle from the very beginning in design what the algorithm and the model will be like.”
Seto additionally noted the importance of algorithmic models with diverse communities in mind. “My access to care is different than those who don’t have insurance.”
There are much deeper design questions when it comes to AI/ML model algorithm developments, she concluded.
