clinical algorithms
Why it's so hard to get bias out of clinical algorithms
There's a growing movement urging health care providers to stop using race as a factor in clinical tools that predict disease risk or guide treatment plans. But that isn't as simple a solution as it might seem, my colleague Katie Palmer reports.
University of Washington researchers used records from thousands of colorectal cancer patients to test a handful of algorithms predicting the likelihood that cancer might return; the one that included race and ethnicity as a predictive variable performed more equally across patient groups than one that didn't.
"Many groups, including our colleagues in the university, have called for the removal of race in many of the existing clinical algorithms," UW Ph.D. student and lead author Sara Khor said. "I think we need to understand what kind of implications that can have and whether that will actually harm patients of color before just removing all variables of race."
The study, published late last week in JAMA Network Open, doesn't suggest that keeping race as a factor in clinical algorithms actually staves off bias. But the findings indicate that simply removing race doesn't guarantee equitable outcomes.
"Until we know more, there may be circumstances in which including race may be useful," Chyke Doubeni, chief health equity officer at Ohio State University Wexner Medical Center, told Katie. "But we can't do that blindly across the board in all cases." Read the full story here.
Lobbying Datavant wades into lobbying
San Francisco-based Datavant — a health tech company building a data platform designed to expedite clinical research — officially began lobbying the federal government last month. I checked in with lobbying veteran Samantha Segall, who now leads federal affairs for the company. Following a career as a Capitol Hill staffer, Segall tackled policy issues for SAIC/Leidos and later helped start the Washington office for CLEAR, the biometrics company whose kiosks you've likely seen at airports and stadiums.
Datavant hopes to "shape legislation that improves patient outcomes with a privacy first approach," Segall told me, noting that the company felt it had achieved enough scale, and amassed enough expertise, to meaningfully shape issues affecting the health care industry.
She'll focus on health data related issues, including Labor, Health and Human Services, and Education appropriations. Datavant is especially interested in the National Clinical Cohort Collaborative, which is housed within the National Institutes of Health and aims to ensure that researchers can use a common clinical data platform to speed research. "We at Datavant believe this can significantly impact patient care." She's tracking data privacy issues, including the American Data Privacy and Protection Act bill.
Segall said the company is focused on health-related committees, including House Energy and Commerce and Senate HELP and Commerce.
21st century cures act
More on the patient data debate
There's been a years-long debate about whether patients should be able to view their test results before doctors do — an issue that became all the more pressing following the implementation of federal rules requiring health systems to make patient data more easily accessible and shareable. In a First Opinion for STAT, family medicine resident Christopher Medrano calls for more research into how the delivery of lab results actually impacts patients, and warns health systems to be aware of instances in which clinician guidance could be crucial.
It's not a new argument — and it's one some patients advocating for comprehensive control of their own health data have deemed paternalistic. If you have thoughts, or have witnessed these issues firsthand, let us know.
No comments