Alarm reductions don’t improve ICU response times
TORONTO – It will take more than a reduction in alarms to address the issue of alarm fatigue in the ICU; a change in the ICU staff culture is needed, suggests new research.
“It may take years to recondition clinicians [to realize] that alarms are actionable and must get a response,” Afua Kunadu, MD, said during her presentation on the study at the CHEST annual meeting. Results from prior studies had suggested that as many as 99% of clinical alarms do not result in clinical intervention, noted Dr. Kunadu, an internal medicine physician at Harlem Hospital Center in New York.
Mitchel L. Zoler/Frontline Medical News
Dr. Afua Kunadu
A program run at Dr. Kunadu’s hospital showed that cutting back in alarm number alone did not lead to better response times to alarms. Counterintuitively, response times worsened as the total number of alarms fell. “This was a big surprise,” Dr. Kunadu said. Dealing with this issue will “require a shift of focus from alarm fatigue to response time. Even though we made the alarms more actionable the conditioning remained” that most alarms are not actionable.
She described the program, which started in the 20-bed adult ICU of Harlem Hospital Center, following a 2014 National Patient Safety Goal issued by The Joint Commission to improve the safety of clinical alarm systems by reducing unneeded alarms and alarm fatigue. The Harlem Hospital task force that ran the program began with an audit of alarms that went off in the ICU and used the results to identify the three most common alarms: bedside cardiac monitors, infusion pumps, and mechanical ventilators. The task force arranged to reset the default settings on these devices to decrease alarm frequency and boost the clinical importance of each alarm that still sounded. Concurrently, they ran educational sessions about the new alarm thresholds, the anticipated drop in alarm number, and the increased urgency to respond to the remaining alarms very quickly for the ICU staff.
The raised thresholds effectively cut the number of alarms. The average number of alarms per patient per hour fell from 4.5 at baseline during September 2016 to about 2 after 1 month, during December 2016. Then the rate further declined to reach a steady nadir that stayed at about 1.3 alarms per patient per hour 4 months into the program.
But timely responses, measured as the percentage of alarm responses occurring within 60 seconds after the alarm went off, fell from 60% at 1 month into the program down to 12% after 4 months, Dr. Kunadu reported.
She had no disclosures.
On Twitter @mitchelzoler
