Network Solutions for Medical Bias
Link to Article: https://penntoday.upenn.edu/news/Penn-study-clinician-peer-networks-remove-race-gender-bias
Throughout my time at Cornell, I have been in several classes that have discussed bias within the medical field. We examined patterns of bias and the long-term effects they have on both individual and community health. Almost everyone can recognize that bias in the medical is a huge issue and that it has been plaguing communities for decades. Despite our ability to converse about medical bias and point out endless examples– we could not come up with a great solution other than to just try to continue to educate those in the medical field on implicit bias. So, when I found this article, I was struck by how closely their proposed solution mirrors the material we have covered in class.
The article titled, “Clinician peer networks remove race and gender bias” covers the research of Damon Centola of the Annenberg School for communication. It discusses the “striking evidence that network science can be used to remove race and gender bias in clinical settings”. The research specifically provides a better way to provide women and minorities with equitable health care through managing clinician peer networks. The researchers found that by changing the structure of information networks, they could change doctors’ perceptions of their client’s information. Essentially the idea is that doctors think differently in groups than they do alone.
I found the experiment to be interesting in itself. After establishing that doctors do have bias through one experiment, they moved forward and divided clinicians into two different groups. The control group watched a clinical history video on their own and then solely made a decision. This resulted in there being a large amount of disparity between the medical advice given to black women and that given to white men. In the experimental group, the clinicians were connected to 40 other clinicians and could see the evaluations made by their peers after watching the video. The clinicians in the experimental group were given the opportunity to change their recommendations depending after learning of their peer’s decisions. They found that the network led to improved clinical accuracy and eliminated bias– white male patients and black female patients were receiving the recommended care at the same rate.
This study ties in extremely close to the network effects we have been studying in class. Specifically, in Chapter 16, we discussed how following a crowd can influence people’s decisions and behavior. In cascades, individuals are not actually imitating the behavior of their peers but rather creating rational decisions from limited information. The study that they conducted here mirrors that of the people deciding whether or not to pick majority red or majority blue when picking marbles out of the bag. I found this study to be striking because we expect doctors to be extremely rational and not biased when it comes to providing patients care one on one. What we find instead, from this study, is that doctors are able to make far better and less biased decisions when they are drawing conclusions from other doctors’ decisions. I originally thought that doctors working together would perpetuate bias but it is clear that it is quite the opposite. In general, I thought this article and study was incredibly fascinating as it applied concepts I would typically attribute to computers or games and applied it to the medical field to help solve an issue that has been negatively impacting communities for decades.
12/2/2021 edit to include another interesting article:
After I posted this blog, I found another article that discussed how the University of Maryland discarded a race-based algorithm that discriminated against Black Patients who needed kidney transplants. This article shows how algorithms can have a negative impact on people receiving medical care. Through reviewing networks and algorithms, we can create better medical equality.
link: https://wtop.com/health-fitness/2021/11/u-md-medical-system-discards-race-based-algorithm-that-discriminates-against-black-kidney-patients/