This is for Every Kidney Discarded Due to Unseen Bias
The Hidden Dangers of Implicit Bias in Clinical Algorithms
I’ll admit that not much has genuinely “shocked” me lately. However, a recent reel by medical student Joel Bervell, aka The Medical Mythbuster, had me double-take to ensure my eyes were not deceiving me. I wasn’t the only one shocked that up until the end of June 2024, due to a clinical algorithm called the Kidney Donor Profile Index (KDPI), kidneys donated from Black Americans were more likely to be discarded than kidneys donated from donors of any other race.
If you’re like, you’re shocked, even if you’re not that surprised that this disturbing practice was due to a false assumption that all Black people’s kidneys were likely to function for a shorter amount of time than those of other races. Why it’s taken until last month for this false assumption to be addressed is anyone’s guess. However, it’s another painful example of the profound impact of implicit bias on healthcare outcomes. It underscores the urgent need for healthcare organizations to recognize and address implicit biases embedded within their systems.
The Hidden Dangers of Implicit Bias in Clinical Algorithms
Clinical algorithms like the KDPI are designed to improve decision-making in healthcare. However, when implicit biases influence these algorithms, they can perpetuate inequalities and harm marginalized communities. Here are some key points to consider:
1. Understanding Implicit Bias in Healthcare
Implicit bias is unconscious attitudes or stereotypes that affect our understanding, actions, and decisions. These biases can influence diagnosis, treatment plans, and patient outcomes in healthcare. The KDPI algorithm's assumption about Black patients' kidney function is an example of how implicit bias can be built into clinical tools, leading to discriminatory practices.
2. The KDPI Algorithm's Flawed Assumptions
The KDPI was intended to predict the longevity of donated kidneys. However, its criteria included race-based factors, assuming that kidneys from Black donors were inherently less viable. This assumption was not based on biological differences but on flawed interpretations of data, reflecting broader societal biases. As a result, potentially life-saving kidneys from Black donors were discarded at higher rates, exacerbating health disparities.
3. Impact on Black Patients
The consequences of the KDPI algorithm's bias are severe. Black patients, already facing higher rates of kidney disease and longer wait times for transplants were further disadvantaged by the unnecessary discarding of viable organs. This reflects a broader pattern of healthcare inequities where implicit bias directly impacts patient care and outcomes.
3 Practical Steps for Healthcare Organizations to Address Implicit Bias
Step 1: Training and Education
Implementing comprehensive training programs is an effective strategy for addressing implicit bias. Through training and education, healthcare organizations can focus on recognizing and addressing implicit bias in their daily operations. By this, I mean explicitly considering the inherent biases that may be embedded in the policies and practices of our healthcare systems. Unless we’re open to recognizing that, we will only continue perpetuating the harm we are fighting to address.
Step 2: Algorithm Audits
Regularly reviewing and updating clinical algorithms is essential to ensure they are free from biased assumptions. According to Friis and Riley (2023), Algorithmic bias occurs when certain algorithmic decisions systematically disadvantage certain groups, and it is common among Black populations. Therefore, regularly reviewing clinical algorithms ensures we can make necessary adjustments and take steps to improve health outcomes.
Step 3: Data Transparency
Advocating for transparency in the data and criteria used in clinical decision-making tools is non-negotiable. A lack of transparency in our healthcare system, specifically when it comes to Black populations, leads to a lack of trust between Black patients and providers. Therefore, transparency in how data is being collected and used is essential. In addition, utilizing unbiased is critical to ensure we’re not drawing the wrong conclusions.
The KDPI algorithm's biased assumptions are a stark reminder of the pervasive impact of implicit bias in healthcare. By understanding these biases and actively working to mitigate them, healthcare organizations can ensure fairer, more equitable care for all patients.
Book me to speak at your organization if you’re ready to foster an inclusive and equitable healthcare environment. Together, we can identify and address implicit biases, create effective training programs, and build a healthcare system that truly serves everyone.
Featured Product
The Dose Podcast by Joel Bervell
If you want to hear more from Joel Bervell, I highly recommend listening to his podcast, The Dose. The podcast features emerging and established voices in health equity who are changing how health professionals and patients interact with the healthcare system.
Listen on Apple Podcasts or wherever you listen to podcasts.
(Note: While I earn a commission if you purchase using my link, I only recommend products that I use, love, and trust are a good option for you.)
Best regards,
Tomesha
Founder and Chief Education Officer
Enhance Black Women’s Health
Reference
Friis, S., & Riley, J. (2023, September 29). Eliminating algorithmic bias is just the beginning of equitable AI. Harvard Business Review. https://hbr.org/2023/09/eliminating-algorithmic-bias-is-just-the-beginning-of-equitable-ai