Race, Medicine and AI
Healthcare professionals still argue the impact of race in determining medical diagnosis.   One camp suggest that racial and ethnic categories are important components in determining clinical usefulness – while the other camp believes that race itself couldn’t possibly be a determinant factor.  For instance,  Do Black people contract diseases like Diabetes more readily than other races because of some genetic trait that exist within their DNA or is it because history has proven than our eating habits by far is a more contributing factor than any inherent factor?  But even those who consider race and ethnicity as a valuable component,  many of those would agree that those small benefits ARE overshadowed by the potential harms that has arisen from the tainted examples of racism and the abhorrent history of race-abuse in medicine. google TUSKEGEE EXPERIMENT.  But even with the systematic abuses that is prevalent in healthcare, race-based medicine is still a predominant practicing technique by the industry.

What does AI have to Do with Race?

Despite mounting evidence and research studies that proves that race is not a reliable proxy for genetic difference, this data continues to find itself into the training data of many healthcare algorithms.  Those technologist amongst us who have worked with healthcare systems, know how race subtly injected into medicine feeds diagnostic algorithms that adjust or “correct” their outputs on the basis of a patient’s race or ethnicity.  Healthcare professionals use these algorithms to individualize risk assessment and guide clinical decisions. By embedding race into the basic data and decisions of health care, these algorithms propagate race-based medicine.

I provide several examples of this in my book “Hidden In White Sight”, where I dedicate a chapter to this.

I recently read that the American Heart Association (AHA) published Risk Score and accompanying algorithms predict the risk of death in patients admitted to the hospital.   It assigns three additional points to any patient identified as “nonblack,” thereby categorizing all black patients as being at lower risk.  The AHA provides little rationale for this adjustment.   This is race-based medicine at it’s best.   Why does this matter?  A recent study showed that Black and Latinx patients who show up at emergency rooms with heart failure are less likely than white patients to be admitted to the cardiology service.
As a Black technologist – I refuse to accept race-base medicine as an option for my health. Treat my symptoms not my race. By the way, isn’t “race” a protected class – which means that most AI ethically-infused scholars would argue that using a protected attribute like race, gender, religion etc. etc. is a no go and non-starter.  If so, why do we stand by idly and allow institutions to do this. hashtag