AI clinical

I didn’t become a physician to do data entry

I opened Ms. Tucker’s chart. There were twenty-one tabs vertically on the left-hand corner of the screen and eighteen tabs horizontally on the top of the screen. I quickly glanced through the cluttered twenty-one vertical tabs; I clicked on the one I am looking for — “transfer medication reconciliation” in the 19th slot. A new grid showing sixteen held orders opened. I selected each of them separately and clicked on “continue.” Select and continue. Sixteen times two: thirty-two clicks.

Source: Mallidi, J. (2018). I didn’t become a physician to do data entry.

Another call for clinicians to be more involved in the design, development, deployment and evaluation of clinician-facing software. There’s evidence that poor software design leads to unreliable data capture, placing patients at risk, as well as being at least partly responsible for physician burnout. Unless you’re a clinician working in the complex (chaotic?) environment of a health system, you’re probably not going to design a user interface that:

  • Is intuitive to use
  • Enables accurate (valid and reliable) data capture
  • Is secure but ALSO user-friendly
  • Degrades gracefully (is tolerant to faults)

See also: