Field Safety Notice on Volumetric Infusion Pumps: Cohort-Specific Withdrawal or Continued Use?

Frame the assessment as severity x probability x detectability, per ISO 14971, not as a single hospital-wide yes/no.

• Nonconformity first. IEC 60601-2-24 requires volumetric accuracy within about +/-5% after stabilisation. A 12% over-delivery is a genuine performance failure, so this is not a labelling issue.
• NICU cohort. At 0.5 mL/h a 12% error is 0.06 mL/h. That sounds trivial, but for vasoactive infusions in a 1.2 kg neonate the therapeutic window is narrow, the error is systematic (always in the same direction), and the nursing workaround of diluting to reach 5 mL/h changes dead volume and start-up accuracy. There is no reliable detection at the bedside at these rates. Recommend immediate withdrawal from NICU and substitution with syringe pumps, with a documented competency check for low-rate syringe delivery.
• ICU cohort. Rates are usually above 5 mL/h, so restriction is achievable, but ICU patients also receive vasopressors. Recommend restricted use with controls: only rates 5 mL/h or above, pharmacy standard concentrations, independent double-check at set-up and at each rate change, and line-patency checks each shift.
• General wards. Mostly maintenance fluids and antibiotics; over-delivery of 12% carries low severity. Continue use with restriction and a 30-day review date.
• Residual risk and sign-off. Record the rationale, the interim controls and the review trigger in the risk management file; obtain clinical co-signature from the ICU and neonatal clinical leads. Never let the engineering team accept residual risk alone.

Run three tracks in parallel: device control, notification, and evidence preservation.

1. Device control. Quarantine the NICU units, tag them 'do not use - FSN pending patch', and update the CMMS status so nobody re-deploys them. Do not return them to the supplier before your own accuracy data are captured.
2. Notification. Report through the hospital's incident system and to the Medical Device Division of the Department of Health under MDACS as a user-facility device problem report, without waiting for the supplier's root cause. Confirm the current reporting form and timeline on the MDD website; a serious injury or death would trigger the shortest window, so escalate now rather than after the investigation. Notify the supplier that your hospital is affected and request written confirmation of the affected serial range and patch date.
3. The nurse reports matter. Two independent observations of the same directional error are a signal, not anecdote. Convert them into a formal complaint: obtain the infusion records, the pump serial numbers and the fluid balance data, and perform gravimetric testing on those two pumps at 1 mL/h and 5 mL/h. If over-delivery is reproduced, the event may meet the threshold for an adverse event report even without patient harm.
4. Communication. Send a one-page clinical alert within the hour: which pumps, which rates, what to do, who to call. Keep it factual and avoid implying the pumps are 'dangerous' in all uses, or staff will bypass the controls.
5. Records. Preserve the FSN, your risk assessment, the testing data and the communications; they will be needed for CAPA verification and any later audit.

A supplier statement that the patch is released is not verification; you must verify at the point of use.

• Independent accuracy testing. After patching, test at least three units at 0.5, 1, 5, 25 and 100 mL/h using a gravimetric method with the correct water density correction, over a sufficient period to reach steady state. Acceptance: within +/-5% across the range, with particular attention to the sub-5 mL/h points that triggered the FSN.
• Configuration control. Confirm the firmware version actually installed by reading it from the device, not from the service report. Record it in the CMMS against each asset number, and check that no unit was missed by the supplier's engineer.
• Documentation of the CAPA. Link the FSN, your risk assessment, the quarantine list, the patch records, the re-test results and the clinical communication in one CAPA file. Include effectiveness checks: no units found at the old firmware at the next safety round, and no recurrence of low-rate discrepancy reports over six months.
• Return to service. Release cohorts in a controlled order (NICU last), with a short post-patch surveillance period and a named clinical contact for reports.
• Learning. Add a purchase-specification clause requiring firmware change notification and validation data before field updates, and add low-rate accuracy to incoming acceptance testing. Update staff training on reporting near-misses, since the nurses' informal reports were the earliest signal.