Computer In Heels

Added on Aug 21, 2015

Katherine Johnson is known for her contribution to America’s aeronautics and space advances. She is an American physicist, space scientist, mathematician, and a human computer who influenced America’s aeronautics and space programs.

Hospital Safety Procedures

Katherine was a NASA pioneer performing mathematical computations for trajectories of satellites and manned space flights. Before the electronic digital computer was popular – people performing mathematical computation were called computers. And Katherine Johnson was a very good human computer.

Katherine was already in high school at age 10. She graduated from what is now West Virginia State University with degrees in French and mathematics. After college she taught elementary school and high school in West Virginia and Virginia. But when a relative shared with her that NASA was looking for people who excelled in math, she resigned her teaching job to work for NASA in the Guidance and Navigation Department at the Langley Research Center.

Katherine calculated the trajectory for Alan Shepard’s first mission into space in 1959. And she calculated John Glenn’s first orbit. Apollo 11 flight to the moon? She did that. Apollo 13 with the equipment failures causing the mission abort and rescue? She did that, too.

She was so good with the numbers that when NASA switched to electronic digital computers for mathematical computation - she was regularly asked to verify the computers’ output. Her ability and reliability helped to establish confidence in the new computer technology, an excellent example of resilience engineering through checks using an independent method.

A second computer using the same method could check the first computer. That is called redundant. A person using manual methods to check the computer is better because this check is both redundant and diverse.

Diversity is a resilience engineering tactic. Two of the same type could experience common mode failure (make the same mistake for the same reason). Two of different types are less likely to experience a common mode failure.

Checks are not nearly as reliable as we think, mostly because we ask for a check without specifying the right check for the job. Checks can be independent in person, time, thought, and method. (See inset left.) Katherine Johnson checking the computer’s math is a good example of a check that is independent in person, time, thought, and method – very reliable.

Hospital Safety

Concurrent checks are not as independent and therefore not as reliable as other checks. Concurrent checks are only used when the check is required before the doer can accomplish their task. Imagine a process where the surgeon makes an incision and the scrub nurse later decides if the surgical site was correct! And even later the scrub tech decides! That check – the time-out before the procedure - is better as a concurrent check before the incision, even though concurrent checks are more susceptible to doer/checker co-dependency.

Self-checks, where the doer is also the checker, are the least effective. The checker is not independent in thought or method or time. Self-checks typically happen immediately after the action. (Think of self-checking with STAR technique (stop, think, act, and review) - the review typically follows the act by one second.)

Make your checks Katherine-Johnson-reliable by choosing the right kind of check and by being specific in describing the check in the protocol. The four elements of a reliable check spell the word SPAM. Be specific in these four elements when writing the check into the protocol:

  • Scope
    What is to be included in the check?
  • Purpose
    Why is the check important – what is the potential for risk and the consequence of failure?
  • Accept/Reject Criteria
    What are the pass/fail criteria? What actions are required if the task fails the check?
  • Method
    How is the check to be performed, who is to perform the check, and when is the check to be performed?


Questions to consider:

  • 1. Do we have the right checks? Do we sometimes ask our people for concurrent checks – which are less reliable – when we should be asking for independent checks?

  • 2. Do we teach the SPAM of the check to the doers and the checkers? Most healthcare protocols specify a check without any detail on the scope, purpose, acceptance criteria, and method. The protocol just says – for example – “the pharmacist should always check the…”

  • 3. Do we have consistent use of the words shall, should, and may? The aforementioned “should always” is a poor practice. Say shall if the action is always required.


Lesson: Go beyond the task at hand; ask questions; be inquisitive. Let yourself be heard. ~Katherine Johnson

Katherine is now retired and living in Hampton, Virginia. She will celebrate her 97th birthday on August 26, 2015.