Where: Beth Israel Deaconess Medical Center (BIDMC), a 500-bed urban university medical center in Boston.
The issue: Responding rapidly when patients become acutely ill to prevent cardiac arrest or death.
“We've all seen cases where you feel like if somebody had noticed something was going off the rails earlier they would have intervened,” said Michael Howell, MD, an intensivist at BIDMC and assistant professor at Harvard Medical School. “I wanted to work in a hospital where we recognized patients were becoming acutely ill, moved them to the intensive care unit [ICU] where they could get optimal care, or had the conversation that they could be DNR [do not resuscitate]. I wanted a hospital where we did very little CPR in uncontrolled settings and we were organized in our response.”
When Dr. Howell and his colleagues first started looking at this issue in 2004, many hospitals were experimenting with rapid response teams (RRTs)—groups of specially trained clinicians who could be summoned to the bedsides of patients who appeared to be declining. They tried something a bit different.
How it works
In 2005, a rapid response system was implemented on all non-ICU wards at BIDMC. There were seven triggers: heart rate under 40 beats/min or over 130 beats/min, systolic blood pressure under 90 mmHg, respiratory rate under eight breaths/min or over 30 breaths/min, oxygen saturation under 90% on oxygen, urine output under 50 cc in four hours, acute change in consciousness or nursing concern.
If any one of these conditions developed in a patient, whoever noticed the change summoned a rapid response team. But unlike the typical rapid response system, this team was composed of the patient's primary nurse and physician, the senior nurse of the floor and, if applicable, the patient's respiratory therapist. When summoned, the team gathered at the bedside and decided what to do for the patient.
The unusual design had one effect from the start: The team was called more often than most intensivist-led rapid response teams are. “Our team activation rate ends up being two or five times higher than has been reported in other studies,” said Dr. Howell. “We think it has to do with the team structure. You're not calling some foreign team, an imposing group coming out of somewhere lese. You're calling people you work with.”
The absence of specialists didn't appear to hurt the team's effectiveness. Results published in Critical Care Medicine in September showed about the same effect on patient outcomes as other rapid response projects. The risk of unexpected mortality (death of a non-ICU, non-DNR patient) dropped significantly: It was 72% lower than before the team was implemented. But the hospital's overall death rate didn't drop significantly.
“It's strange that the unexpected mortality rate went down a lot and that the overall mortality rate either didn't change or only went down a little. What's remarkable is that it's an astoundingly consistent finding across most rapid response team studies,” said Dr. Howell.
There are a couple of potential explanations for the finding. It's possible that the rapid response team activations resulted in patients being transferred to the ICU, where they later died. Or, the rapid response emergencies may have led to patients becoming DNR. “We don't have the data to tell that,” said Dr. Howell. “But my gestalt is that patients decompensate on the floor, people come and take care of them, they stabilize them, and that gives time enough for the patient's son and daughter to come in, or husband/wife, and to have a discussion.”
How patients benefit
Even without a reduction in overall mortality, the drop in unexpected deaths is a good result, according to Dr. Howell. “Everybody who's ever been to a failed code outside of an ICU would agree that it's bad for everybody—it's bad for the patient, it's bad for the family, it's bad for the providers,” he said.
Tracking the rapid response activations also motivated other improvements to patient safety at BIDMC. “We noticed that we were seeing an awful lot of patients have bad things happen because of aspiration, and we looked at that and [asked why].’” The researchers found that many of the aspirating patients had taken advantage of the hospital's order-on-demand food service, and food deliveries weren't being checked with the nurse before patients got them. “Now that mechanism of harm to our patients is eliminated from our hospital,” Dr. Howell said.
Data from the rapid response activations also led clinicians to make use of a non-re-breather mask a trigger for the team.
Dr. Howell and his colleagues are now working on an algorithm to predict at admission a patient's individual likelihood of needing rapid response. Early results are promising, he said.