A 34% Override Rate Behind a Wrong-Patient Administration
In a 300-bed hospital, a nurse removed levetiracetam 500 mg from an automated dispensing cabinet for one patient and administered it to the patient in the other bed of the same room. Both patients had similar last names starting with "Mar-." No harm occurred; the dose was within the range both patients could have received. Your review of ADC data shows the override rate for this drug on that unit is 34%, the cabinet holds both 500 mg and 250 mg vials, and the drug was recently added to the formulary in a new concentration. Two weeks earlier the same unit filed a near-miss with the same drug. You are the technician representing pharmacy on the medication safety committee.
1. How do you frame this event for the committee - as an individual performance problem, a system problem, or both? What is the first concrete action, and what is the deepest fix?
Read the worked answer
Frame it as both, in that order: system first, then individual behavior.
Why not blame-first: a 34% override rate is not one nurse's choice; it is the predictable output of a workflow that makes overriding the path of least resistance. If 1 in 3 pulls bypasses pharmacist review, the system has taught staff that the review is an obstacle. Punishing the individual here guarantees under-reporting of the next event, which costs you far more than this one did.
First concrete action (hours):
- Confirm the patients are stable and that clinical monitoring is in place.
- Pull the full ADC transaction log, the eMAR, and the scan history for both patients, to establish the actual sequence (was the drug scanned? was the patient scanned? was the override used because the order wasn't verified yet?).
- Interview the nurse with a just-culture lens - ask what made the shortcut feel reasonable, not who is at fault.
Deepest fix (weeks): the override pathway itself. Three questions:
- Why was this drug on the override list at all? If it is not a true emergency drug (arrest, rapid response, first-dose antibiotic), it does not belong on an override list.
- Why were both strengths in the same drawer? Cabinet configuration is a design decision with error consequences.
- Was barcode scanning actually used, or bypassed? Scanning is the backstop that would have caught this; if it was bypassed, that is the real finding.
The individual conversation still happens - but it is about at-risk behavior (the shortcut), addressed with coaching and clear expectations, not about the outcome.
2. Build the corrective action set with at least four interventions, identify which single intervention is the strongest lever, and state a measurable target.
Read the worked answer
Intervention set (each mapped to a failure point):
| Intervention | Failure point addressed |
|---|---|
| Remove levetiracetam from the override list; require pharmacist verification except for documented rapid-response use | The 34% bypass of review |
| Remove one strength from the unit cabinet; stock only 500 mg with a 250 mg available on request | Look-alike storage of two strengths |
| Tall-man lettering on labels and bin labels; revise the two similar last names with a secondary identifier (MRN or DOB) at the bedside | Similar-name confusion |
| Hard-stop barcode scanning at administration with a documented override reason and pharmacist review of every bypass | The backstop that failed |
| Add a second patient identifier to the ADC pick screen and the eMAR | Wrong-patient selection at the cabinet |
Strongest single lever: removing the drug from the override list and enforcing bedside scanning together. Neither alone closes the loop - a reduced override list without scanning still permits wrong-patient administration, and scanning without reducing overrides still leaves unverified orders in play. If forced to pick one, the override removal is stronger because it eliminates the wrong-patient selection step at the cabinet.
Target: override rate for this drug from 34% to under 5% within 60 days, with zero undocumented scan bypasses; scan compliance ≥ 98% on the unit. Set the target as a rate, not as "zero events" - zero-event targets drive concealment.
3. What data will tell you whether it worked, and what is the trap in relying on override rate alone?
Read the worked answer
Measure a small bundle, not one number:
- Primary process measure: override rate for the drug (target <5%), and scan-bypass count with reasons.
- Secondary process measures: pharmacist verification turnaround time (if you tightened verification, you must confirm you did not create a delay), and number of doses requested from a different location after the strength removal.
- Outcome measure: wrong-patient administration events and near-misses involving this drug, from the reporting system.
- Balancing measures: time-to-first-dose for stat/rapid-response patients, nurse-reported workarounds in a short pulse survey, and the volume of incident reports (a rise in reports after a safety push usually means reporting improved, not that safety worsened).
The trap: override rate is trivially gameable. Staff can lower it by pulling the drug from a different cabinet, by having a colleague with a verified profile pull it, or by delaying the override until the order verifies - none of which means review actually happened. A falling override rate with a flat or rising scan-bypass rate is a red flag, not a success. Pair the process metric with an independent observation (a short audit of 20 administrations with direct observation of scanning) before declaring the intervention effective.