Patient flow scorecards show complexity
Approach relies on multidisciplinary team
A common refrain of ED administrators is that when it comes to patient flow, there is only so much they can do to eliminate wait times when the upper floors cannot quickly accommodate admissions from the ED. Further, numerous studies have shown that capacity problems of this nature can impact care quality and patient satisfaction scores.
Recognizing that such problems are difficult to resolve without fully appreciating the inter-dependence of multiple departments, and the many different factors that ultimately impact the flow of patients through the hospital, investigators at Children’s Hospital of Philadelphia (CHOP) developed a five-domain patient flow scorecard, designed to capture the complexity that is inherent in the patient flow process and to highlight specific areas where ripe opportunities for improvement exist.
While the approach is being continually refined, leaders of the effort note that it has helped them home in on the specific reasons for hold-ups so resources can be focused in the most effective way while also giving hospital administrators and staff a larger, holistic view of the patient flow process.
Establish a multidisciplinary team
The new approach was initiated in the summer of 2009 after hospital administrators identified patient flow improvement as a priority, explains Evan Fieldston, MD, MBA, MSHP, medical director of care model innovation and an attending physician in the Division of General Pediatrics at CHOP, and the lead author of a research paper on the project.1
"I have been involved for many years in our patient flow quality improvement activities and thought that we needed better systems of measurement," says Fieldston. "While the ED in many places is a large microsystem, it is also in a concentrated place with concentrated staff. For our hospital, we have 500 beds [with 9,000] people distributed in over 21 physical locations, each with their own dynamics and operational flows."
Among the problems administrators encountered when measuring performance was that individual measures, such as the time between arrival in the ED and admission to an inpatient floor, did not adequately reflect the complexity of the patient flow process. Further, administrators were concerned that trying to improve one isolated measure could adversely impact another, simply moving the problem rather than making the overall process better.
Consequently, researchers sat down with a multidisciplinary team to develop a patient flow scorecard that captured data from five domains, including:
• ED and ED-to-inpatient transition;
• bed management;
• discharge process;
• room turnover and environmental services department activities; and
• scheduling and utilization.
Within these domains are several more component measures that are assigned one to four points. For instance, there are eight individual component measures within the "ED and ED-to-inpatient transition" domain. These include various time intervals, such as "arrival to physician evaluation," with the goal of 80% of patients seeing a physician within 60 minutes; "ED physician evaluation to decision to admit," with the goal of 80% occurring within 240 minutes; and "decision to admit to MD report complete," with the goal of 80% occurring in 120 minutes. Each of these components receives four points on the patient flow scorecard, with five other component measures in the ED domain receiving fewer points.
For example, three points are aligned with "RN report to patient floor," with the goal of 80% occurring within 60 minutes; two points are aligned with "ED length-of-stay (LOS) for non-admitted patients," with the goal of 80% of patients having an LOS less than 300 minutes; and one point is aligned with a "leave without being seen (LWBS) rate," with the goal of less than 3%.
Also included in the ED domain are "ED admission rate" and "ED volume" adjusting measures, each of which is aligned with one point. These enable administrators to adjust the scoring based on the severity of patients visiting the ED and high volumes.
While the number and type of components differ within each domain, the total number of points associated with each domain is the same at 20 points, adding up to a maximum patient flow composite score of 100 points.
Tie entry points to workflows
When selecting metrics to be included on the patient flow card, developers looked not just for overall relevance to patient flow, but also for items that were automated, explains Fieldston. "We wanted something that would be relatively easy to use, so [all the measures] come from any number of the variety of electronic systems in the hospital," he says.
While there was not much disagreement about which time intervals were important in the emergency setting, coming up with the best metrics to use was still challenging, according to Nicholas Tsarouhas, MD, medical director of the emergency transport team and associate medical director of the ED at CHOP, and a co-author of the research. Tsarouhas notes that the ED at CHOP sees nearly 90,000 patients a year, so it is a very busy department.
"You need a starting point and an ending point, and then you need the users to be compulsive about entering those time points," he says. "So the challenge [revolved around] potentially disrupting someone’s workflow and making them go to a computer to hit a button."
To make such entries or time stamping as easy and efficient as possible, developers endeavored to tie the key time entries to parts of the ED workflow that made it practical for someone to be at the computer when the entries needed to be made. For example, the nurses were instructed to always note in the computer whenever a patient is leaving the ED to go to an upper floor, but the challenge was getting the staff to do this consistently. "Sometimes [these types of entries] are made before the patient leaves; sometimes the nurses forget and they wait until they take the patient upstairs and they come back down and do the time stamp when they return," says Tsarouhas. "So even though the job is actually done, when you time stamp it is very important."
Another challenge with the patient flow cards was trying to make sure that hospital staff fully appreciated the results. "As much as there is a problem with single measures not giving you a holistic picture of what is happening, a multi-component patient flow scorecard also has its complexities in the ability of people to understand it," says Fieldston. "Getting used to it, and getting people to understand the overall flow as well as how the various pieces connect together, is a communications and a culture issue."
To help with understanding, developers used color descriptors. For example, when staff met or exceeded a particular component goal, the results were in the "green" category. Results that were 10 percentage points below the goal fell into the "yellow" category, and performances below this level were in the "red" category. Domain and overall composite scores were likewise color-coded, with performances of 16 or above on the 20-point domain scale or 80 and above on the overall composite score presented in green, scores between 70% and 79% in yellow, and scores below that level in red.
Fieldston notes that the patient flow team also communicated directly with the teams and sub-teams for the particular domains, sending out scores or sub-scores with explanations about what was happening with respect to the different parts of the patient flow process.
Engage frontline staff on metrics
Even with the added level of complexity, developers say the approach has helped the hospital better home in on the specific areas that need improvement. "All of our metrics are broken down into sub-metrics, and the sub-metrics let us look at the areas that need focus," says Tsarouhas.
For example, from the scorecard results, administrators were able to discern that the bed-cleaning process was not commencing quickly after patients left beds, delaying the availability of the beds to new patients. "Examining our metrics enabled us to see that there was a lot of down time when the rooms were empty and nobody knew that someone should start cleaning," says Tsarouhas. "That was an example of where we tried to close that gap."
In response to those metrics, the improvement team decided to link the process of removing a patient from the ED tracking board with a notification to environmental services that the bed needed to be cleaned. "What we realized is if you tie workflows together, that makes the data better because when you tie one operation to the next, you get more efficiency," says Tsarouhas.
Also in response to these data, environmental services reorganized their workforce, and the changes produced improvement. "In that way, rather than just looking at one global number [showing] how long patients spend in the ED or what time they get discharged, which doesn’t really cause other members of the team to connect to the process and the improvement work, we were able to provide all of the key stakeholders in the patient flow progression with information that they could respond to," says Fieldston.
The metrics utilized in the patient flow cards are continually refined, and developers are always looking for new and better ways to present the information in the most concise and meaningful way, explains Fieldston. His advice to others interested in employing a similar approach is to engage frontline staff on what is important to patient flow and what metrics to follow.
"We have identified some important time stamps in patient progression, but by no means are these the only steps in the process, and our values may be different than what other places have," says Fieldston. "I also think that pairing the [patient flow card] approach with things like process mapping and value stream mapping, so that organizations can understand the key steps in their patient flow processes and identify where there are opportunities for improvement, are crucial steps in patient flow improvement."
Tsarouhas observes that the overall focus on patient flow was well received in the ED. "We always feel that it is so important to our optimal functioning that we move patients out efficiently so that we can bring other patients in," he says. "The scorecard itself is just an objective way for us to measure some of our processes to hopefully provide objective data to help drive improvement work."
However, Tsarouhas adds that a multidisciplinary group needs to drive the improvement process. "It doesn’t work if there are just doctors or nurses or administrators," he says. "The success of our programs has really been predicated on every one of our meetings including doctors, nurses, nurse practitioners, people from environmental services, people from bed management, and people from administration who are all in the same room, and all committed to the work."
Further, for any improvement to work, there needs to be high-level executive support, says Tsarouhas. To win this support, you have to present good data, he says. "If you can give data to the executives and they can see where the areas for improvement are, that makes it easier for them to support the work."
- Fieldston E, Zaoutis L, Agosto P, Guo A, et al. Measuring patient flow in a children’s hospital using a scorecard with composite measurement. Journal of Hospital Medicine, 2014 April 18. [Epub ahead of print]