Based in Seattle, WA, Swedish Medical Center is the largest private non‐proﬁt medical center in the six‐state Northwest region of the United States. The facility cares for 43,000 inpatients annually, and counts 8,500 employ‐ees, 3,000 physicians, and 1,200 volunteers working at four hospital campus‐es. The Emergency Department includes six EDs with two major urban ERs and four freestanding suburban EDs serving over 107,000 emergency roomvisits each year.
Dr. Brian Livingston, Director of the Swedish Emergency Department, and Dr. Bob Seale, a former ED physician turned medical informatics specialist, had long been convinced that the throughput of the medical center’s emergency department needed to improve. Like many administrators, they initially ap‐proached the problem in logistical terms—if they just had one more exam room and a few more staﬀ members, they would reach greater eﬃciency.
In an eﬀort to explore solutions, they hired a ﬁrm to conduct a lean tech workshop with 25 to 30 members of the Emergency Department. The con‐sultants had the staﬀ set up a mock ED, and then ran a series of simulations hoping to identify bottlenecks in procedures. In the ﬁrst simulation, they ran the ED using their current procedures and were able to complete about 60% of the cases in the allotted time. In the second simulation, they were told to operate the same way, but without talking. They found that they completed 80% of the cases in the same amount of TIme. In the ﬁnal simulation, they were to operate with no talking, one less nurse, and one less room. Again, they were amazed at how many more patients they cared for in the department during that time.
The exercise was an eye‐opener. Dr. Livingston and Dr. Seale now knew it was not a logistics issue. However, they still didn’t know where to begin making changes to increase ED throughput, and still needed data to identify botllenecks. Then they could use the training exercises to make changes and measure the eﬀects.
Swedish Medical Center had successfully implemented an EPIC electronic health records system in the emergency department the previous year. They asked their IT department for a report showing their length‐of‐stay times. After several weeks of communication back and forth with IT, they received a Clarity report on length of stay. But by then, they decided they also needed to look at door‐to‐doc times. Several more weeks later, they had a Clarity report showing door‐to‐doc times. But by the time they received that report, the data was no longer current. Getting the information they needed out of their EHR system was just too cumbersome, too generic, and too limited for them to make timely changes, rapidly evaluate them, and make more modiﬁcations.
Dr. Seale heard about Deep Domain from colleagues within the Swedish system who were using the healthcare IT company’s technology to extract data from their EPIC database for diabetes management. Deep Domain’s solution allowed end‐users to deﬁne their own reports and generate multiple variations in just minutes, enabling clinical managers to look at a patient’s critical data, including LDLs, blood glucose, and a dozen other parameters. It also allowed them to measure and monitor physician performance and patient compliance. The reports could be created separately, or stacked and deﬁned in a dozen diﬀerent ways.
Dr. Seale and Dr. Livingston asked Deep Domain to apply its technology to their emergency department. But instead of patient data, they wanted to track a variety of time measures, including door‐to‐room, door‐to‐doc, doc‐to‐discharge, and length‐of‐stay. They also wanted to monitor performance against those measures in as close to real time as possible. With Deep Domain’s ED Analyzer, physician administrators were able to discover hidden bottlenecks at diﬀerent times of the day and at diﬀerent stages of care. They produced in‐depth reports on a variety of parameters, including eﬃciency by individual physician, which they shared with their doctors. Being able to see the data was a powerful moti‐ vator. Combining the Deep Domain data with their lean tech experiments, they were also able to systematically try diﬀerent procedural changes, and then use Deep Domain’s ED Analyzer software to quickly evaluate whether the modiﬁcation they made had an eﬀect on performance. The result was a dramatic improvement in ED throughput.
“With the ability to look at several one‐month graphs in a row – clicking June, July, and August – it almost looked animated,” explains Dr. Seale. “What jumped out was a trend toward longer stays beginning about 9:00 am that tended to improve around 6:00 to 9:00 pm, then lengthened again around midnight. This is why we nickname this the Discovery Tool because now we can ask: What are the department issues causing this?”
The Business Results
In the ﬁrst year of using Deep Domain’s ED analyzer tools, Swedish Medical Center’s Emergency Department has been able to reap the following beneﬁts:
“In 2008, we had 1,086 hours of divert. Since March 26, 2009, we have not had a single hour of divert time,” concludes Dr. Seale. “One of the obstacles of using the data contained in our EPIC EHR is the language barrier between clinical people and IT people. Not knowing what to ask is a huge hurdle. With Deep Do‐main’s ED Analyzer, our end‐users can ask the questions themselves and instantly deﬁne and generate reports to answer their own questions. We are able to identify problems that we didn’t know we had. The results are clear: Patient satisfaction is dramatically higher, and costs are signiﬁcantly lower.