1 | Determine the errors with high risk or high probability that could reach the bedside, using a technique such as failure modes and effects analysis.18 |
2 | Develop specific checklist instructions for each predictable error. Include details of what information to check (eg, dose in mg) and from what sources (physician's order and drug label). Keep the list short27 28 by omitting items with lower risk and lower probability. |
3 | If the possibility of an error is abstract or general (eg, error in physician's dosage choice), but the error itself has a high severity or probability, break the error down into smaller, more specific steps that can be added to the instructions (eg, check dosage on medication order against hospital drug formulary of appropriate adult doses). |
4 | Determine the workflow of the first and second nurses by observing them working in their natural environment using a technique such as contextual enquiry.21 22 |
5 | To encourage efficiency and adoption, assemble the itemised instructions into a checklist that corresponds with their workflow, and use language and terms that match their existing tools such as the infusion pump screen prompts. |
6 | To test and improve the usability of the checklist, recruit a small sample of end users (three to six people) to use the checklist while you observe. If they become confused, use the checklist in a way that is not anticipated, or readily miss errors, refine the design of the form to be more intuitive, and repeat the testing process. |
7 | For each potential error not included on the checklist, develop alternate strategies to prevent it from reaching the bedside. Continue to develop additional strategies for eliminating all possible errors, even those that can be identified with the checklist, since no human checking process is failsafe. |