Stanford hospital COVID-19 vaccine allocation algorithm triggers backlash

Occurred: December 2020

An algorithmic COVID-19 vaccine distribution system developed and run by Stanford Medicine came under fire for prioritising certain staff members.

The algorithm was designed to consider three categories: employee-based variables (such as age), job-based variables, and guidelines from the California Department of Public Health. However, it resulted in an unequal distribution of vaccines, with only seven out of over 1,300 resident physicians being prioritised for the first 5,000 doses.

Many of these residents worked on the front lines of the COVID-19 pandemic, and they were shocked and angered to see that administrators and doctors working remotely from home were included in the priority list. Only seven of Stanford’s 1,300 patient-facing medical residents made the list. This led to a protest by at least 100 residents during a planned photo op to celebrate the first vaccination.

Stanford Medicine apologised and attributed the errors its 'very complex algorithm'. However, critics argued that the issue was not with the algorithm itself, but with the instructions given to it by humans. They pointed out that the algorithm was not powered by machine learning, but was rule-based, meaning that it simply acted upon a set of instructions provided by humans. 

The incident highlighted the potential pitfalls and complexities involved in using algorithms for critical decision-making processes.

System 🤖

Operator: Stanford Health Care; Stanford School of Medicine
Developer: Stanford Health Care; Stanford School of Medicine
Country: USA
Sector: Health
Purpose: Allocate vaccine beneficiaries
Technology: Rule-based algorithm
Issue: Fairness
Transparency: Governance

Research, advocacy 🧮

News, commentary, analysis 🗞️