Time's up for respiratory rate mistakes

A team effort increased the accuracy of respiratory rate measurements and saved time.


Patients' respiratory rates are often inaccurately recorded in hospitals. A study published in the October 2017 BMJ Quality & Safety found that far more patients had documented rates of 18 to 20 breaths/min than would make sense with normal distribution. “One of the dirty secrets in hospital medicine and inpatient internal medicine is that respiratory rates, when you see them in the computer, you don't really know how accurate they are,” said ACP Member Neil Keshvani, MD. “It's definitely not just a local hospital problem, it's a national problem.” To improve the accuracy of one of the fastest vital signs to change during illness, his team went straight to the source.

How it works

As a first step, the team worked with patient-care assistants (PCAs), who perform the majority of vital signs measurements in the hospital, on a single 28-bed medical inpatient unit. “Once I started following them around, I realized first and foremost that there was no [analog] clock in the room, and not everyone wore a watch,” said Dr. Keshvani, a chief medical resident at UT Southwestern Medical Center in Dallas. To address this issue, the team bought three stopwatches and attached them to the vital signs carts. “You can't expect someone to do something if you don't give them the proper tools,” he said.

In addition, the PCAs didn't realize that respiratory rate was important and that the default respiratory rate of about 20 was not helpful, Dr. Keshvani said. So the team retrained them to use the stopwatches and count chest rises for 30 seconds, multiply by 2, and log the respiratory rate. “We retrained people pretty quickly. It was within 15 minutes during their shift handoff because it's not that hard of an intervention,” he said. To avoid interruptions in workflow, the PCAs did their respiratory rate measurements during the 30 or so seconds when patients' blood pressures were being measured automatically.


Overall, the project improved the absolute respiratory rate accuracy by 22% while saving PCAs 41 seconds on average per vital sign measurement, according to results published online in June by the Journal of Hospital Medicine. Prior to the intervention, the median PCA respiratory rate was 18, compared to 12 for gold-standard measurements by trained clinicians. Only 36% of the PCAs' measurements were considered accurate. After the intervention, the median PCA and gold-standard respiratory rates matched up at 14, with a PCA respiratory rate accuracy of 58%. “Basically, for every single PCA, there was a lot more variability in the recorded respiratory rates” after the intervention, Dr. Keshvani said.

The study also assessed changes in the number of patients meeting criteria for systemic inflammatory response syndrome (SIRS) based on elevated respiratory rate. After the intervention, there was a 7.8% decrease in the proportion of hospitalized patients with tachypnea-specific SIRS, showing that the intervention may have improved the accuracy of electronic alerts for SIRS, Dr. Keshvani said. “You need to have good data to have good early warning systems,” he said.


The biggest challenge was the labor involved in figuring out why respiratory measurements were inaccurate, Dr. Keshvani said. “I actually had to do a lot of shadowing and watching of PCAs while also realizing that whenever you're watching, people tend to act a little bit differently,” he said. “I will say, though, that our PCAs were amazing collaborators throughout this whole process.”

Words of wisdom

“I think this could get done at any other hospital that would invest the time to improve their respiratory rate accuracy,” Dr. Keshvani said. “And if a certain hospital has the same issues we do, then it's not even going to take that much time because we already isolated some of the problems: a lack of timekeeping capabilities, inefficient use of time at the bedside, and a lack of understanding of what should be a priority.”

Next steps

The goal is to do a full-hospital rollout of the project, which was a team effort among residents, faculty, nursing, and administration, Dr. Keshvani said. “We've passed this along to nursing with an effort of trying to eventually hit the whole hospital,” he said.