1. Door to Doc (D2D) Reduces ED Patient “Walkout” Rate
Twila Burdick
Banner Health

2. Objectives Recognize ED patient flow as a patient safety problem
Describe improved patient flow using ED Door to Doc (D2D) Split Patient Flow model
Discuss approach for implementing D2D in diverse ED settings
Show impact of D2D on patient safety

3. ED Patient Safety Challenge
Capacity constrained EDs
Increased visits
Holding inpatients
Long waits for patients arriving
Patients deteriorating in the waiting room
Patient complaints, patient dissatisfaction
Patients leaving without treatment
self-diagnosis is not safe!
Long waits to see an ED physician related to LWOTs
Issues with ED crowding are pretty well known (increasing visits, not enough ip beds, pts getting sicker as they wait, being unhappy) and patients leaving without being seen
Banner began to understand it not just as an “inconvenience” but as a ‘patient safety’ issue—the literature confirmed our conclusion.
“A study of the consequences of leaving the emergency department prior to a medical evaluation at one public hospital found that 46 percent of those who left were judged to need immediate medical attention, and 11 percent who left were hospitalized within the next week.”[1]
“At follow-up, patients who left without being seen were twice as likely as those who were seen to report that their pain or the seriousness of their problem was worse.”[2]
“Of the children who left without being seen, 24 (15%) were triaged as "urgent," and none had a CTAS score of less than 3…Our finding that 15% of patients who left without being seen had been triaged as "urgent" is of concern.”[3]
“Forty-six percent of those who left were judged to need immediate medical attention, and 29% needed care within 24 to 48 hours.”[4]
“Overall, 60% of LWBS (Left Without Being Seen) cases sought medical attention within one week; 14 patients were hospitalized, and one required urgent surgery.”[5]

4. Improving ED Safety Challenge: Idea:
How to get patients seen by ED physicians sooner in overcrowded, busy EDs across Banner Health
Idea:
Apply “science of throughput” to change patient flow and reduce delays
Start with D2D “straw model”
Implement in diverse EDs
Once we recognized the impact on patient safety, our challenge was clear—get patients to docs sooner
Our IDEA was to use “science” of throughput (based on what we were learning from ASU), start with a model developed at one of our facilities, and see if we could implement this in diverse ED settings.

5. Designing D2D ED Improvement/Design Oversight Team
Care
Transformation
Clinical
Risk
Intervention
Throughput
AHRQ Grant
Productivity
Common Launch:
Learning Session
Clarification of Multiple Outcomes
Oversight Process Defined
Phasing Described
Work Design
Work Team with Representation from all Stakeholders
Dedicated, nearly full time effort for up to 4 weeks
Start with “straw model” based on best practices
Physician Work Group
Consistent
application
of automation
Discharge
Process to reduce
risk and returns
Process and
Measures
Process
Description
for Layout
ED Improvement/Design Support Team
Stakeholder
Review
Strategic Svs
Leadership
Conf
ED Call
Coverage
Facility
Issues,
Designs
Behavioral
Health
Patients
Once we recognized our need to address this, the question was how—
As we began our work, we began to find that there was a great deal of organizational interest—in lots of different silos!
Our first step was to bring the silos together—
The Design Process was collaborative with appropriate support
As we began our door to doc project, a team of physicians and nurses from all Banner facilities began to look at our current process and discovered we were able to fill a wall with all of the steps needed to bring a patient into our system. By eliminating all of the redundancies in the system we found we could get the patients to the doctor within – minutes……but we didn’t want to move the problem from the front of the ED to the back.

6. Door to Doc Care Process
Quick Registration
Quick Look (ESI)
Split Patient Flow
Intake area for Less acute patients
Joint medical screening
Patient moves to testing and treatment
Informed discharge
IP care for admitted patients
Along the way, We visited several other facilities and learned about a design that split the patients into “sick” and “not so sick” groups.
With our less sick patients, we decided to keep them moving through the system to make room for more patients…this is our “bed ahead” concept. In our sicker area, we designated space for patients who were waiting for testing or needed longer treatments. By this continuous movement of patients we were able to continue seeing patients regardless of our volume…or so we thought. When admitted patients were not moved to the inpatient areas, we felt the pain of being unable to bring more patients in so we designed space for this group of patients and decided best practice would dictate nurses who are best trained to care for them would be assigned to the inpatients.

7. Common ED Characteristics
ED arrival volume patterns are predictable by hour of the day.
9 am-9 pm peak (30% higher)
Relative Length of Use by acuity level has a similar pattern across facilities
While we were designing the D2D SPF process, we started learning about some characteristics that EDs seem to have in common that are important in thinking about taking a single design to different settings.
--ED arrival patterns are predictable by time of date
--Relative length of use (or time in the ED) by acuity has a similar pattern

8. Systems View: Queuing Model
Inpatient Transitional Care
Hospital Exit
rID
rIH = 1-rID
rOD = 1-fRE = 80%
= fA /(rRO*fRE+rRI)
rOW = (f3+f4) / (f3+f4+f5)
IPED
fRE = 20%
Intake/ Discharge
Results Waiting
rWO = 100%
From a systems engineering view, here’s how the D2D SPF decomposes into a queuing model
rRI = 1-rRO
rRO
Quick Look
LWOT
Ambulance Diversion 0%

9. Choosing ‘Patient Safe’ Capacities
Using info specific to each facility along with the Queuing model, the SPF D2D model can be sized appropriately.
Recommendations for discussion are in bold

10. Improvement in Patient Safety through Reduced LWOTs
As a result of implementing D2D, Banner facilities have seen improvement in LWOT rates of 35 to 65%--regardless of whether the facility had high LWOT rates or already reasonably low rates.
42% Improvement
63% Improvement
35% Improvement

11. Here are line graphs that show the changes over time
Emergency Department Patients That Leave Without Treatment
Sample Size = 24 months
Banner Health: Banner Baywood Medical Center
Mesa, Arizona, United States of America
21%
20%
19%
Door to Doc
18%
Implementation
17%
16%
15%
14%
13%
12%
11%
% of Patients that Left Without Treatment
10% 9% 8% 7%
Emergency Department Patients That Leave Without Treatment
Sample Size = 24 months
Banner Health: Banner Desert Medical Center
Mesa, Arizona, United States of America 0% 1% 2% 3% 4% 5% 6% 7% 8% 9%
10%
11%
12%
13%
14%
15%
16%
17%
18%
19%
20%
21%
Jul-05
Aug-05
Sep-05
Oct-05
Nov-05
Dec-05
Jan-06
Feb-06
Mar-06
Apr-06
May-06
Jun-06
Jul-06
Aug-06
Sep-06
Oct-06
Nov-06
Dec-06
Jan-07
Feb-07
Mar-07
Apr-07
May-07
Jun-07
Month-Year (month)
% of Patients that Left Without Treatment
Line = 30% Improvement
Door to Doc
Implementation 6%
Line = 30% improvement 5% 4% 3% 2% 1% 0%
Jul-05
Aug-05
Sep-05
Oct-05
Nov-05
Dec-05
Jan-06
Feb-06
Mar-06
Apr-06
May-06
Jun-06
Jul-06
Aug-06
Sep-06
Oct-06
Nov-06
Dec-06
Jan-07
Feb-07
Mar-07
Apr-07
May-07
Jun-07
Month-Year (month)
Here are line graphs that show the changes over time

12. Emergency Department Patients That Leave Without Treatment
Sample Size = 24 months
Banner Health: Banner Mesa Medical Center
Mesa, Arizona, United States of America 0% 1% 2% 3% 4% 5% 6% 7% 8% 9%
10%
11%
12%
13%
14%
15%
16%
17%
18%
19%
20%
21%
Jul-05
Aug-05
Sep-05
Oct-05
Nov-05
Dec-05
Jan-06
Feb-06
Mar-06
Apr-06
May-06
Jun-06
Jul-06
Aug-06
Sep-06
Oct-06
Nov-06
Dec-06
Jan-07
Feb-07
Mar-07
Apr-07
May-07
Jun-07
Month-Year (month)
% of Patients that Left Without Treatment
Line = 30% Improvement
Door to Doc
Implementation
Emergency Department Patients That Leave Without Treatment
Sample Size = 24 months
Banner Health: North Colorado Medical Center
Greeley, Colorado, United States of America 0% 1% 2% 3% 4% 5% 6% 7% 8% 9%
10%
11%
12%
13%
14%
15%
16%
17%
18%
19%
20%
21%
Jul-05
Aug-05
Sep-05
Oct-05
Nov-05
Dec-05
Jan-06
Feb-06
Mar-06
Apr-06
May-06
Jun-06
Jul-06
Aug-06
Sep-06
Oct-06
Nov-06
Dec-06
Jan-07
Feb-07
Mar-07
Apr-07
May-07
Jun-07
Month-Year (month)
% of Patients that Left Without Treatment
Line = 30% Improvement
Door to Doc
Implementation

13. LWOTs and D2D Time
There is a strong relationship between LWOT% and D2D time.
A linear model (shown) explains 93% of the LWOT% /D2D data variation (correlation coefficient = 0.96).

14. Lessons Learned ED Patient Safety can be improved Collaboration is key
The D2D SPF design reduces LWOTS 
Collaboration is key
ASU Industrial Engineering
Physicians, Staff
Implementation is the hardest part
Changing minds along with process
Leadership matters
High level organizational commitment
Keep measuring and monitoring
Ongoing improvement should occur

15. The SPF D2D model can apply in many EDs
A toolkit with interactive modeling tools is available at Innovations.org

16. Acknowledgements
This project was funded under grant U18 HS from the Agency for Healthcare Research and Quality, U.S. Department of Health and Human Services.
Special thanks to the many people who have made this improvement possible! Especially Professor Jeff Cochran!