Defibrillator 'Failures' and a Proposal to Double the PM Interval
A health system owns 84 monitor/defibrillators across six sites: three models, with the oldest units more than 12 years old. The PM interval is 6 months. During the last quarter, 5 units failed the delivered-energy test: at a 200 J setting, measured delivered energy was 172–184 J. Three of the five were more than 12 years old and two were 4 years old. No adverse events were reported. The manufacturer's service manual specifies annual PM and does not support extending the interval. Finance has asked clinical engineering to cut PM labor by 20%, and the director proposes moving all 84 units to a 12-month interval, citing 'AEM literature.' You have three years of PM records and a test database with delivered energy for every unit. Two of the 84 units have no recorded test in 14 months because they live in satellite clinics.
1. Before accepting that 5 units 'failed,' what would you verify - and does the data support extending the interval to 12 months?
Read the worked answer
The word 'failed' is doing a lot of work in this scenario and must be defined before any decision.
First verify the acceptance criterion. Defibrillator delivered-energy tolerance is typically stated as a percentage of the set energy into a specified test load (commonly ±15% at 50 Ω). At a 200 J setting, 172 J is −14% and 184 J is −8%. If the manufacturer's tolerance is ±15%, none of these five devices is out of tolerance, and the real findings are a too-tight internal threshold, an uncalibrated analyzer, or a wrong test load. If the facility's criterion is ±10%, then 172–180 J units are genuinely out of tolerance. Check the analyzer calibration certificate and the load resistor before concluding anything.
Second, disaggregate. The director's proposal treats 84 devices as one population. The two 4-year-old units failing argues against a pure age explanation and toward a model-specific or component issue; the three old units argue for retirement, not for interval juggling.
Decision: do not extend all 84 to 12 months now. An AEM interval is defensible only with documented evidence that the failure rate and severity do not increase. One quarter containing failures, on an unverified measurement basis, is not that evidence, and the manufacturer explicitly does not support extension. A defensible path is: correct the test basis, retest, then consider 12 months only for the model with a clean multi-year record, keeping 6 months for the aging units and any unit used on a code cart. Offer finance the real alternative - retire the oldest units and reduce fleet size - rather than trading test frequency for a number.
2. What immediate actions would you take on the fleet and on the two units missing tests?
Read the worked answer
Containment first, documentation second.
- Take the five flagged units out of service. Do not simply retest and return them: retest with a calibrated analyzer and a verified load, repair or replace the energy-delivery components as indicated, then retest and document.
- Test the two units with no record within 24 hours, and audit the satellite clinics for units that bypassed the schedule. Any unit without a current documented test does not belong on a code cart.
- Test the remaining fleet on a compressed schedule rather than waiting for the next PM cycle, prioritizing units in high-acuity areas.
- Check what the energy test does not check: battery capacity and runtime (test on battery, not just AC), charged-and-ready indicators, pad and cable condition and expiry, synchronizer function, and the self-test log.
- Look for a common cause: plot delivered energy by serial number and by model over three years. A step change in one model points to a component or a service-parts change; scattered outliers point to age or test technique.
Then close the loop with clinical users: tell code-committee and nursing education which units were affected and when, and confirm no unit was in use during a resuscitation while out of tolerance. If one was, that changes the reporting analysis in the next step.
3. What reporting and documentation obligations apply, and what must the AEM file contain if the interval is ever changed?
Read the worked answer
Reporting: with no death or serious injury, the mandatory FDA user-facility reporting duty is not triggered here. The right actions are still to file a complaint with the manufacturer (which feeds its trending), keep an internal safety report, and monitor for any later injury that would change the analysis. If a unit malfunctioned during a resuscitation and a patient was harmed, the reporting duty and timeline change, and risk management must be involved immediately rather than after the fact.
Documentation of the interval decision must show reasoning, not just a conclusion:
- The acceptance criterion and its source, the analyzer calibration records, and the corrected test results.
- The failure data by model, age, and use location, and the rationale for treating models differently.
- The manufacturer's recommendation and why the organization is or is not departing from it.
- The approved policy, the interval per model, and a defined review date with the data that would trigger reverting to 6 months.
- Records retention consistent with the facility's CMS and accreditation requirements.
The two units with no test in 14 months are a separate finding and should be written up as a scheduling failure with a corrective action, not folded into the AEM discussion. If the AEM decision is later reviewed by a surveyor, that gap is what will be examined first.