Swedish Medical Center

Swedish Medical Center (http://www.swedish.org/ ) is the largest nonprofit health-care provider in the Greater Seattle area.  In independent research by National Research Corporation, Swedish is consistently named the area’s best hospital, with the best doctors, nurses and care in a variety of specialty areas.

Swedish has three hospital locations in Seattle, a hospital in Edmonds, a freestanding emergency room and specialty center in Issaquah (East King County), and a network of 12 primary care clinics.  It is licensed for 1,245 beds.  Swedish has approximately 8,000 employees and almost 2,300 physicians on staff.  Each of the hospitals has their own emergency department (ED).

The three Seattle hospitals are located at First Hill, Cherry Hill and Ballard.  The First Hill campus specializes in general medicine, orthopedics, women and children’s services, and oncology.  The First Hill ED serves approximately 45,000 patients annually.  The Cherry Hill campus, located less than a mile from the First Hill campus, specializes in cardiac and neurological care.  The Cherry Hill ED serves approximately 21,000 per year, many with high acuity.  The Ballard campus serves as a community hospital to this north end Seattle neighborhood.  The Ballard ED serves approximately 20,000 patients.  In 2005, Swedish opened the state’s first free-standing ED in Issaquah.  It serves approximately 25,000 patients per year.  In 2010, Swedish assumed management of Steven’s Hospital in Edmonds.  The ED there serves 45,000 patients annually.  This section highlights the work at First Hill and Cherry Hill EDs from 2008-2010, even though similar work has been done at all of their facilities.

By the end of 2008 Swedish EDs were in desperate need of improved efficiency.  In 2007 Swedish launched a project to implement electronic medical records (EMR) across their facilities.  This conversion was the figurative straw that broke the camel’s back for the emergency departments.  The EMR added to the existing inefficiencies of the nurses and physicians, and led to increases in length of stay for patients.  The increased length of stay led to patients backing up in the waiting room.  With waiting rooms full, the EDs felt forced to go on ambulance diversion because of concerns for patient safety.  Divert though was not isolated to the Swedish system but was a county wide problem and discussions were starting on a potential county-wide no-divert protocol.  Swedish recognized the need to make improvements before a protocol was imposed on them.

While Swedish was having difficulty getting patients into the system due to ambulance diversion, it was also having difficulties retaining staff.  Both campuses had low levels of staff morale due to constantly feeling overworked and each campus had high turnover of nursing and ancillary staff, relying on agency and traveling staff to fill the voids.  This further added to staff dissatisfaction as there was less staff invested in improving the department.  The staff felt unsupported by hospital administration, who believed the problems experience in the ED were a result of system wide problems and not isolated to the ED.  Additionally, the Cherry Hill physician group experienced a high turnover rate because of changes in patient acuity and limited diversity of the patient population due to strategic changes within the Swedish system.

At the end of 2008 the care of an ED patient involved a series of queues.  An average patient at First Hill went through 11 queues before leaving the ED (see Figure 1) with some queues potentially taking over one hour to move through; Cherry Hill had a similar patient flow.  These queues created average length of stays for all patients (admitted and discharged) of 251 minutes for patients at Cherry Hill and 221 minutes for First Hill patients in 2008.

Figure 1 – First Hill Emergency Department Value Stream Map, 2008

With full waiting rooms, First Hill was on divert for a total of 1,074 hours in 2008 or 12.3% of the time.  Cherry Hill diverted for 936 hours in 2008 or 10.7% of the time.

While diverting patients decreased the number of arrivals, it prevented patients who wanted to come to Swedish from having access.  This not only angered patients but also created ill will between the ED and specialists whose patients were diverted to other hospitals and treated by physicians who were not familiar with their care.  Not only were the patients who were diverted unhappy with their care, but Press Ganey satisfaction scores of patients who were treated in the ED were only average.  Satisfaction raw scores in 2008 for Cherry Hill and First Hill were 81.5 and 82.2, respectively.

In January, 2009, a system-wide education system was implemented beginning with a one-week Advanced ED Operations Course which was attended by key leadership and front line workers from each facility. This was followed by smaller individual education sessions to spread what each team had learned.

Over the course of the next one and a half years, the teams at all sites began to apply lean and flow principles throughout their EDs. The teams at First Hill and Cherry Hill employed Value Stream Mapping to define their current states early on in their journeys. Both locations simplified their processes greatly. For instance, Cherry Hill implemented a low-acuity vertical space, simplified and standardized their suture carts and moved to a geographical management scheme.

They employed 5S and visual management techniques by refining their patient room-level supplies and central supply rooms, implemented procedure carts and implemented an electronic tracking. The y employed rapid changeover techniques in their Code Surge plan in which resources were maximized to turn over beds in the ED.

The teams also incorporated advanced flow principles into their process redesign. For instance, they used pull systems to redefine the way patients are seen in triage and to improve the process that is used to admit patients to the ICU in a system called the ED-ICU Pull-through Project. They employed load leveling to balance patient flow between campuses and between staff members on each campus, thereby realizing slight, but important increases in productivity that translated into less queuing, reduced length of stays (LOS) and therefore happier patients. They also assumed a mentality that they would seek out and eliminate existing queues where possible. The teams eliminated triage when possible, implementing a direct bedding process and used multi-disciplinary “swarm” teams upon patient arrival to the room.

Cherry Hill ED began by encouraging staff to join shared leadership councils, which drove change from the inside out.  The councils started with trials in the ED which led to implementing strategies such as redefinition of triage, standardization of suture & incision / drainage carts, better alignment of clinician and physician schedules with patient arrival patterns, direct bedding and team-focused care.  By directly bedding incoming patients to available rooms, staff was able to bypass the triage process, leading to shorter wait times to see the physician.  A team-focused approach to patient care meant that a group of nurses, a technician, and a physician were assigned to a specific section of rooms.  This “swarm” approach by the team was encouraged when new, higher acuity patients arrived.

Figure 2 – Cherry Hill Emergency Department Current State VSM, 2010

The councils worked to improve leadership within the department by increasing collaboration with the charge nurses, physicians, and senior leaders.  A key goal was to empower the charge nurses to have more authority within the department and supporting them to manage the flow of patients in order to drive down length of stay, thus avoid divert.  To facilitate these goals, monthly charge nurse meetings were implemented, which included time to meet with the physicians, and enrollment of the charge nurses in a six week leadership class. Senior leaders also supported the departmental changes publicly and in meetings with other departments.

With newfound internal and external support, the councils were able to roll out changes involving different departments.  For example, a strategy called ‘Code Surge’ was implemented for use when the Cherry Hill ED became saturated.  When ‘Code Surge’ was paged out, the nursing supervisor, charge nurses from inpatient units, pharmacy, radiology, housekeeping, and transporters, contacted the ED charge nurse to inquire what they could do to help rapidly move patients through the ED.

First Hill ED began with a multi-disciplinary change team that emerged from the Advanced ED Operations program.  The team started with a comprehensive value-stream mapping effort that created current and future state maps.  From these maps, six initiatives were identified: direct bedding, research into use of scribes for physicians, trial “Care Teams” (1 MD, 2RNs, 1 ED Tech for 6 rooms), trial a Super Fast Track for low acuity patients, analyze all staffing patterns in relation to demand, and develop a ‘Code Surge’ process.  A nine month period of rapid cycle tests produced a consistent drop in patient length of stay and dramatic drop in diversions.  However, staff became disillusioned with Care Teams and the existing change team.

In December of 2009, a new change team was developed consisting of volunteers from the front line staff and MDs.  The new team re-examined the initiatives and elected to focus on: triage re-design, direct bedding, assessment of the MD – Disposition time interval, design of a “flood management” process to deal with surges in patient arrivals, and development of an effective communication process to keep all staff apprised of trials, results and decisions.  The ensuing nine months saw an intensified focus on achieving a strong leadership role for the charge nurse geared towards consistent practice on new initiatives, trialing of a revamped triage process, implementation of a hospital-wide ‘Code Surge’ process, increased collaboration between the ED and radiology to improve turnaround time and improvement of medication handling with pharmacy.  Finally, a 5S process was initiated to improve management of supplies.

Since starting to implement lean processes, the EDs have been recognized both within the Swedish system as well as throughout the county.  The divert problem prior to 2009 has essentially been eliminated.  First Hill has had zero hours of divert since March of 2009 and Cherry Hill has had less than 15 hours of divert during that time period (see graph.)  The monthly reports to Swedish’s board of directors that originally showed ED diverts in a negative light have turned into recognition for the EDs.  Additionally, the King County ED Saturation Project has recognized Swedish as a system to follow.

Figure 3 – Hours on Divert at First Hill and Cherry Hill from January 2008 to August 2010

Through the efforts in the ED and presentations to the hospital, inpatient units and ancillary services have seen the impact they have on the ED and have partnered with the ED to facilitate admissions.  The ICU and ED collaborated to create a septic patient protocol, which has decreased the LOS for an ICU admission from over 2.5 hours to less than 1 hour with a decrease in mortality by over 50%.

First Hill ED has worked with the pharmacy department on appropriate stocking of medications in Pyxis.  After reviewing the medications most commonly used, adjustments were made to stock additional antibiotics within the department.  By stocking one antibiotic in the ED that was used over 70 times in a month, the ED eliminated over 25 hours of waiting time (20 minute average from order placement to medication received).

Concurrent to eliminating diversion, the EDs were able to make substantial improvements in both patient length of stay and “door to doctor” time.  Within Washington state, the average length of stay is over four hours (Press Ganey ED Pulse report).  Cherry Hill went from an average length of stay in 2008 of 251 minutes to 196 in 2009 and 175 in 2010.  First Hill LOS decreased from 221 minutes in 2008 to 188 in 2009 and 175 in 2010. (Figures 4,5). Both teams also saw reductions in their door to doctor times which is the primary correlate to patient satisfaction. Cherry Hill improved its door to doc times from 73 minutes in 2008 to 36 minutes in 2010, and First Hill moved from 80 minutes in 2008 to 27 minutes in 2010 (Figures 6,7).

Figure 4 – Cherry Hill Length of Stay, 2007 to 2010
Figure 5 – First Hill Length of Stay, 2008 to 2010
Figure 6 – Cherry Hill Door to Doc Times, 2007 to 2010
Figure 7 – First Hill Door to Doc Times, 2008 to 2010

In addition to improving the throughput of patients in the ED, patients feel their experience in the EDs has also improved.  Cherry Hill’s Press Ganey patient satisfaction score improved from 81.5 in 2008 to 84.8 in 2009 and 85.6 for 2010 YTD. First Hill’s scores moved from 82.2 in 2008 to 85.0 in 2009 and 85.5 2010 YTD (Figures 8,9). Finally, Cherry Hill ED staff had been surveyed for staff satisfaction by HR in 2009.  A year later, a follow up survey showed significant gains.

Figure 8 – Cherry Hill Press Ganey Patient Satisfaction, 2008-2010
Figure 9 – First Hill Patient Satisfaction, 2008-2010

The change teams faced numerous challenges.  The conversion to an EMR has left physicians spending very large proportions of their time documenting at a computer.  Opportunities for improvement include scribes, better use of templates, and better use of Workstations on Wheels.

At the outset, the existing culture of both EDs was one of maintaining status quo.  The shift to a culture embracing change is not yet complete but great strides have been taken. Both EDs struggled with their learning curve on how to conduct PDSA trials and how to effectively implement changes.  Issues in regard to this included communication, getting sufficient and appropriate staff involved in the change process, tracking changes and achieving consistency and sharing data/outcomes.

Both campuses found their physical geography was not conducive to the type of process changes they wanted to implement.  Lack of funds has prevented anything more than cursory improvements in the physical plant.  This results in work-arounds that are forced to fit into existing space.  There is potential for an ED re-build at Cherry Hill but this is not final at this time.

Cross campus and interdepartmental collaboration was poor at the outset.  Many training efforts, communication initiatives, changes in leadership to people who more strongly supported collaboration (and made it a high priority) have made a major impact on this area.  The current state is one of steadily improving collaboration, which greatly aids patient flow in all directions.

The role of the charge nurse in the EDs had largely deteriorated to one of being a “super float” and the person who assigned lunches.  Charge nurses felt powerless and had no clear idea of their role in a change project and process improvement efforts.  Currently most charge nurses have now had leadership training, clarification of roles, a much greater collaboration with physicians and an elevation to a role of Patient Flow Management.  This role continues to evolve and many charge nurses are actively involved in process improvement committees and teams.

Both campuses have experienced “change fatigue” with staff protesting “enough!”  The shift to a culture of change is arduous but being achieved in increments.

Finally, the challenge of finding or creating effective vehicles of communication has plagued all efforts.  Few staff read email, staff configuration is patterned erratically, and notice boards get little attention as they are jammed full of a myriad of product, education and random information.  At First Hill the process improvement team has formed a “communication tree”.  This tree assigns eight staff to each team member who is then responsible for face to face communication with each of their eight, to share changes, trials and outcomes.  At Cherry Hill, the process improvement team is using daily staff huddles to share information and showing up at quiet times to support on-going discussions, provide opportunities for staff to vent concerns, give feedback and share in process ideation.

In September, 2010 a second session was held and included more members who had become interested in learning about how to apply lean in the ED as well as staff members from the inpatient units and ancillary services as lean had begun to spread throughout the organization. The lean lectures in this second session were actually taught by students from the first session who had applied what they had learned in the true, “See one, do one, teach one” way.

Both Cherry Hill and First Hill EDs are committed to on-going process improvement work as are the other EDs in the Swedish system.  They both have well-established PI teams to drive and monitor the change process and senior leadership is on-board with their support.  Both have also set aggressive goals that will continue to motivate their teams.  They are well positioned to keep their improvement cycles moving forward as evidenced by Figure 10 which represents future projects the teams plan to implement.

Figure 10 – Future Targets for Improvement at Swedish

Dr Brian Livingston, Dr Russell Carlisle, Erica Tuke, BSN, Jeff McAuliffe, MA, Kim Adams, RN