Text-message-driven intervention, real-time feedback slashes time-to-treatment for stroke patients

Build quality-improvement interventions around hospital-specific challenges

Sometimes just making people aware of their performance is all that is necessary to significantly improve care. Investigators at the University of California at San Francisco (UCSF) found this to be precisely the case when they attempted to use this approach to improve door-to-needle times for stroke patients who presented to the ED for care at UCSF Medical Center.

As a teaching hospital, it can be challenging to reinforce care guidelines with residents who continually rotate through the stroke care team, explains Molly Burnett, MD, who is a resident in the Department of Neurology at UCSF. “We rotate so frequently that a lot of times we are not [with the stroke team] long enough to get feedback about how long we are taking, or even to realize that the goal for [time-to-treatment] is 60 minutes,” she says.

Administrators observed that when stroke patients presented to the ED, it was taking the hospital longer than some of the other hospitals in the area to deliver brain-saving tissue plasminogen activator (tPA). “Our average door-to-needle times were over 80 minutes, and definitely not in compliance with the guidelines,” explains Burnett.

To address the issue, Burnett and colleagues decided to test whether immediate notification about the door-to-needle time for each patient, delivered to all members of the stroke team in real time, via text message, would be enough to boost awareness of the guidelines, and expedite care to patients.

Monitor performance

The intervention was designed to follow the traditional “code stroke” page that is always triggered by an ED provider whenever a patient presents with symptoms of stroke. That page goes to the entire treatment team, including representatives from radiology, bed control, the lab, and several others, explains Burnett. Investigators added a “reply” page to this sequence that would go to the same group of people as the initial “code stroke” page. It would indicate via text message whether or not tPA was given and, if so, what the door-to-needle time was.

Initial results from the test of this intervention, which was completed in 2011, were clear-cut. The average door-to-needle time for 95 patients treated before the intervention was implemented was 82 minutes. The 45 patients who received tPA after the intervention was employed were treated within 61 minutes. Further, investigators report that a significantly higher percentage of intervention-group patients were treated within the recommended 60-minute time frame (50%) than was the case in the pre-intervention group (16%). Burnett notes that since 2011, results have continued to improve at UCSF Medical Center, as the text-messaging intervention is now a standard of care at the facility. “On average, our door-to-needle times are in the 50- to 60-minute range now,” she says.

Burnett emphasizes that one key to the intervention’s success is the constant monitoring of the text messages by Andy Kim, MD, MAS, an assistant professor in neurology and medical director of the Stroke Center at UCSF. “If a door-to-needle time is more than 60 minutes, he will contact the treating team and ask what the impediment was to making the 60-minute time frame,” says Burnett. “He also sends out stroke performance results to the entire multidisciplinary team about every two weeks. These data show us how well we are doing.”

Deliver real-time feedback

While it would seem that the Stroke Center director serves as the task master in this intervention, he explains that the approach was actually very easy to implement. “I spent the first month or two sending reminder pages at all hours of the day and night, but quickly it became self-sustaining even without the additional reminder pages,” explains Kim.

Kim also emphasizes that ED staff are crucial to making the intervention work. “Providing objective feedback to the entire team, including the ED, in real-time rather than a week, a month, or a quarter later, only serves to enhance our existing relationship because it allows us to celebrate our successes together rather than just interacting when things do not go as smoothly,” he says.

Other hospitals interested in experimenting with a similar strategy need to design their approach around the specific challenges that they face, advises Kim. “Here, we knew that our existing system was capable of delivering tPA quickly because in individual cases we were able to achieve our goal,” he says. “We focused on the problems of short institutional memory due to rotating residents in our training programs and the diffusion of responsibility, given the large team of people involved.”

Having the technology already in place to implement the intervention was certainly helpful, says Kim. But he suggests that success of the approach had more to do with what the technology enabled the UCSF team to accomplish. “It is the real-time nature of the feedback, clear and immediate accountability, and frequent and sustained effort to improve care that are the key components of any successful quality improvement initiative,” he says.


• Molly Burnett, MD, Resident in the Department of Neurology, University of California at San Francisco. E-mail: molly.burnett@ucsf.edu.

• Anthony Kim, MD, MAS, Assistant Professor of Neurology, University of California, San Francisco, and Medical Director, UCSF Stroke Center. E-mail: anthony.kim@ucsf.edu.