According to a letter sent from Stanford Medicine’s chief resident council on December 17, an algorithm specifying that Stanford Medicine workers would receive their 5,000 initial doses of the COVID-19 vaccine included only seven medical residents/fellows on the list. The leadership of Stanford Medicine has since apologized and vowed to re-evaluate the proposal. Stanford’s representatives clarified that an algorithm was used to distribute their first allotment of the vaccine, according to an email sent by a chief resident to other residents. Along with factors such as age and the place or unit where they work in the hospital, the algorithm was said to have prioritized those health care staff at the highest risk for COVID infections. Residents obviously did not have an assigned venue, and they were dropped low on the priority list along with their usually young age.

A rules-based formula designed to assess the order in which thousands of medical staff should be vaccinated at Stanford was at the core of the fiasco. According to MIT Technology Review, the instrument took into account employee-based variables such as age, job-based variables, and advice on public health. But flaws in that estimate meant that hospital managers and other personnel working from home were at the front of the line, although the list was made by just seven of the 1,300 Stanford medical residents.

STAT was advised by experts that this was a human issue from start to finish. Critically, the Stanford algorithm was not motivated by machine learning, in which, without explicit human programming, the computer learns from the data. Instead, as explained by the MIT Technology Review, it was rule-based, which implies that humans wrote out a series of instructions that the instrument actually operated upon. Representatives of the hospital did not address questions as to who would be involved in the current planning phases or whether the algorithm would continue to be used. An internal email summarising the medical school’s response, shared with MIT Technology Review, states that neither program heads, department chairs, attending physicians, nor nursing staff was involved in the original algorithm design. Currently, however, some faculty are pushing to have a greater role, totally removing the effects of the algorithms and instead of giving the authority to make decisions on their own teams to division chiefs and chairs.