Best Practices: Sustainable Quality Improvement at Thomas Jefferson University Hospital

September 1, 2013

All at once? Or one change at a time? This was the key choice we faced two years ago at Thomas Jefferson University Hospital in Philadelphia. We were charting a course to implement many of the changes that other academic EDs had employed successfully to improve ED patient flow. These were changes we’d read about in journals and heard about at conferences, and we were ready to go. But did we want to bolus these changes through, or run a slow IV infusion?

It was a long list of substantial changes in how we would do our work: immediate bedding until full; bedside registration; bedside triage; expanding fast track to take higher acuity “mid track” patients; conversion of rooms from containing a single stretcher to having multiple chairs for increased “vertical” space; and physician-in-triage. At conferences we’d see the results from other centers, with walkouts down, diversion down, reduced time-to-provider and length-of-stay.

We knew we wanted to make these changes. What we didn’t know was whether to bunch them into one big change, or pace ourselves. And we had seen our share of stalled and stymied changes. We had tried to create vertical space the previous year. Mysteriously, each night when nursing and physician administrators went home, the chairs would disappear from the vertical rooms, and the stretchers would find their way back. The same thing happened on weekends.

There were respectable arguments for both approaches to change. In favor of a piecemeal, one-change-at-a-time approach, was the idea that we would not ask for too much wrenching change from our faculty, housestaff, nurses and other staff at once. People could retain familiar processes and valued skill sets. When people lose an old way of doing things, there can be a sense of mourning for what is left behind, for the knowledge and confidence in what to do and how to do it, and the investment of time and effort made to gain that knowledge. Did we want to plunge everyone into deep mourning?

In favor of the all-at-once approach was the thought that people don’t say goodbye to the old ways until they have completed the stages of mourning. If we did things piecemeal there would be no clean break, no finality to the change. The period of disruption would be longer, and the time when people are truly engaged with new ways of doing things pushed further into the future.

Which path did we choose? Actually, we chose both. We had a firm date for a complete, all-at-once implementation: July 5th, 2011. But in the time leading up to that date, we piloted most of the component parts of the change. As mentioned, we piloted vertical space. We also piloted having immediate bedding during “low utilization” states (times at which we had empty beds, roughly from midnight to noon). We piloted bedside registration, abbreviated triage out front, full triage at the bedside, putting a nurse practitioner out front, and expanding the acuity of our fast track. The immediate results were mixed. There would be reversion to familiar processes at off hours or at times of stress, and inconsistent implementation based on who was working. Each component of change had different bases of support and opposition, and there was no unifying end-point in sight to let people know they had arrived and could engage with the new way. There were cumulative stresses, without any singular big wins.

However, these pilots paved the way. They proved in concept each of the component changes, and showed they would not be disastrous. They also got our staff used to change itself. When it came to the final months preparing for wholesale change, the base of support was larger and more confluent than the pockets of resistance.

In those final months, we employed several strategies which proved useful. We had a core “flow” group dedicated to developing and implementing changes. It was an open meeting where staff could present any ideas or problems that came along. We invited people from other departments. We used multiple lines of communication: meetings at different times of the day to cover different shifts, emails, presentations outside the department, and huddles during shifts. We always emphasized the “why” behind the changes. We “sold” the problem of long waits and walkouts in terms of safety, risk, patient satisfaction and staff morale. We showed quotes from our satisfaction surveys, quantified the relationship between waits and legal risk, and shared the financial details of what is lost from walkouts and ambulance diversion.

When the time came to go all in, the changes stuck. We have never gone back. (We also made sure that the vertical equipment was too heavy to be moved during night shift.) Our walkouts halved, ambulance diversion is down to less than 10% of what it was, time to provider has been cut in half, and other metrics such as length of stay have also improved. Did we choose the all-at-once approach? Maybe we should call our approach hop, skip and jump, with an emphasis on the final jump.

Paris LovettParis Lovett, M.D. M.B.A, Medical Director, Department of Emergency Medicine, Medical Director, Patient Flow Management Center, Thomas Jefferson University Hospital