Currently Viewing:
The American Journal of Accountable Care December 2015
The Need to Level the Playing Field Between Accountable Care Organizations and Medicare Advantage
David Introcaso, PhD
The Debate Over Drug Costs: Instead of "How Much" We Spend, Let's Focus on What We Get in Terms of Health
A. Mark Fendrick, MD Co-Editor-in-Chief, The American Journal of Managed Care Professor of Medicine, School of Medicine Professor of Health Management and Policy, School of Public Health Director, Cen
The Pediatric ACO: A New Frontier in Accountable Care
Anthony D. Slonim, MD, DrPH
Revisiting the Role of Academic Medical Centers in Medicare Shared Savings Program ACOs
Benjamin M. Gerber, JD
The Emerging Business Models and Value Proposition of Mobile Health Clinics
Khin-Kyemon Aung AB; Caterina Hill, MSC, MA; Jennifer Bennet, BS; Zirui Song, MD, PhD; and Nancy E. Oriol, MD
Now That ACOs Are Engaged, What About Those Patients?
Mary K. Caffrey
It Is Time to Ask Patients What Outcomes Are Important to Them
Leif I. Solberg, MD; Stephen E. Asche, MA; John C. Butler, MD; David Carrell, PhD; Christine K. Norton, MA; Jeffrey G. Jarvik, MD, MPH; Rebecca Smith-Bindman, MD; Juliana O. Tillema, MPA; Robin R. Whi
Currently Reading
Continuous Quality Improvement Program, Based on Lean Concepts, Allows Emptying of Emergency Department Corridors
Enrique Casalino, MD, PhD; Christophe Choquet, MD; Mathias Wargon, MD, PhD; Romain Hellmann, MD; Michel Ranaivoson, MD; Luisa Colosi, MD; Gaëlle Juillien, MD; and Julien Bernard, MD
Using Guidelines: When Is It Appropriate?
Maude St-Onge, MD, PHD, FRCPC
The Impact of Direct Oral Anticoagulants in the Managed Care Environment
Charles E. Mahan, PharmD, RPh, PhC; and Robert Lavender, MD, FACP

Continuous Quality Improvement Program, Based on Lean Concepts, Allows Emptying of Emergency Department Corridors

Enrique Casalino, MD, PhD; Christophe Choquet, MD; Mathias Wargon, MD, PhD; Romain Hellmann, MD; Michel Ranaivoson, MD; Luisa Colosi, MD; Gaëlle Juillien, MD; and Julien Bernard, MD
A continuous quality improvement program, based on Lean concepts and including architectural, managerial, and organizational features, allows the emptying of emergency department corridors and the improvement of time interval measurements/quality indicators.

Objectives: To assess the impact on an emergency department (ED) and the sustainability of a multifaceted continuous quality improvement (CQI) program based on Lean concepts.

Study Design: A prospective interventional 6-year study was conducted.

Methods: Interventions included managerial, organizational, and architectural features. Evaluation was conducted using pre-/post intervention ANOVA com­parisons and interrupted time series (ITS) analysis.

Results: We analyzed 413,392 attendances. During the study period, significant (P <.001) trends were found for decreased length of stay (LOS) (–191 minutes; 95% CI, –280 to –101), waiting time before seeing an ED provider (–142 minutes; 95% CI, –225 to –66), and for increased percentage of patients leaving the ED in less than 4 hours (+24%; 95% CI, 12%-38%). ITS analysis demonstrated that the initial quality improvement step and having an ED medical coordinator had a significant impact on LOS, whereas the policy of having 0 patients in ED corri­dors led to significant improvements in LOS in the percentage of patients leaving the ED in less than 4 hours, and in the waiting time before seeing a medical provider. Fast-track pathways were implemented for low- and very low–com­plexity patients.

Conclusions: Our results indicate that a CQI program based on Lean concepts—including successive interventions and a “fast track” for low- and very low–complexity patients—allows for the emptying of ED corridors and the enhancement of time-intervals measurements/quality indicators.
Emergency department (ED) overcrowding has a negative impact on the quality of care and on patient and staff satisfaction.1,2 Operational performance measures have been suggested and various time intervals have been proposed as quality and performance indicators, or crowding markers, in the ED. Longer ED length of stay (LOS) has been associated with a greater risk of death in the short term.3,4 The percentage of patients leaving the ED in less than 4 hours,5 the time spent waiting to see a medical provider,6 and the time spent waiting for a triage nurse7 have all been proposed as time-interval measure­ment/quality indicators (TIMQIs).

Various strategies to improve ED patient-flow management, performance, and quality of care have been proposed.8,9 Continu­ous quality improvement (CQI) is probably the most established approach,10 with “Lean thinking”— first implemented in man­ufacturing, later extended to healthcare organizations including EDs11—as one approach to CQI. Various other tools have been proposed to improve the quality of the EDs,12 including fast-track13 triage area organization,14 reorganization of patient flow, and clinical information system and process redesign. However, evaluations of their success were frequently carried out by using short-term pre-/post intervention measurements.15 The long-term impact of interventions is difficult to evaluate in quality im­provement evaluations. However, a central idea in improvement is to make changes incrementally, by learning from experience.16 We have found very few studies that have reported a continuous evaluation over time.17

The aim of this study was to measure the impact of the re­organization of an ED based on CQI and Lean thinking on TIMQIs. Thus, we used pre- and post intervention comparisons for short-term evaluation. To evaluate the long-term impact of successive interventions, we used interrupted time series (ITS) analysis, assessment technique, developed to analyze work flows and processes within an organization (reengineering).18


Patients and Data Collection

This anonymous data set did not contain any identifiable person­al health information, and it is currently used as an ED quality and performance measurement as part of an ongoing emergen­cy activity and performance evaluation. The study was approved by the emergency ethical committee of the Assistance Pub­lique-Hôpitaux de Paris.

Study Design

We report a CQI project conducted over a 6-year period. Specific interventions were introduced in a sequential manner during the study period. To evaluate the impact of each specific interven­tion, we conducted a prospective pre-/post intervention 1-way analysis of variance (ANOVA),19 comparing the means of several quality and performance indicators from the 28 days before to 42 days after each intervention.

An ITS analysis was also conducted, which included monthly average values for the performance indicators as a way to eval­uate each specific intervention in the overall CQI project. The time series included 72 measured points from January 2006 to December 2011, which spanned a period prior to CQI imple­mentation to 24 months after the last intervention.20


The study was carried out in an academic hospital (Bichat-Claude Bernard [BCB]) located in the Paris metropolitan area. Over the course of the study, there were no significant changes in the number or composition of the medical staff (trainees, residents, or staff emergency physicians), in the paramedical staff, in the hospital policies on admissions, or in access to radiology or lab­oratories. ED renovation had already been planned, including a temporary move to a provisional structure, and finally, into our permanent structure.

Interventions Description

Each of the interventions included managerial, organizational, and architectural modifications. Briefly, they included: a) initial quality improvement step: the creation of a CQI team project that includes a change of culture promoting performance and quality; b) implementation of an ED medical coordinator re­sponsible for the good functioning of the ward through lead­ership and formal authority; c) a triage nurse who streams the patients into 2 different sectors according to their acuity level; d) the move to a provisional ED structure (ED area reduced by 20%) with a new policy of 0 patients in the ED corridors; e) the move to our permanent ED structure (40% larger than original ED area) with all sectors on 1 unique floor, meaning personnel can be deployed according to patient flow in real time; and f) a dedicated physician for fast-tracking low-complexity patients.

Details of each intervention are shown in Table 1. We did not evaluate intervention “a”—the initial quality improvement step as a pre-/post intervention)—but it was integrated into the ITS model because it represents the first modification to the culture of the service and accompanied the first organizational modifi­cations to our ED.


Daily and monthly values for time-interval measurements were extracted automatically in real time from the electronic med­ical and hospital database records. All visits to the BCB’s ED during the study period were included. The following variables were studied: number of daily and monthly visits; acuity level; patient’s final outcome (ie, not admitted, admitted in our obser­vation ward, admitted to another ward, or transferred to another hospital). Acuity was measured according to the Canadian Triage scale from 1 to 5 (1 = high acuity or complexity, 5 = low acuity or complexity).

Time intervals (minutes) were calculated automatically on a daily and monthly basis and were defined as follows: a) LOS— minutes between the time the patient was identified in the ED and the time the patient left the ED; b) percentage of patients leaving in less than 4 hours—calculated from the LOS for each visit; c) waiting time to triage nurse—minutes between the time the patient was identified in the ED and the time the patient was seen by the triage nurse; d) waiting time to ED provider—min­utes between the time the patient was identified in the ED and the time the patient was first treated by the ED provider. The final decision of the ED provider is registered by the ED com­puter program.

Statistical Analysis

The pre-/post intervention analysis compares pre- and post in­tervention means of daily time-interval measurements (28 days be­fore and 42 days after each intervention) by using 1-way ANOVA.19

ITS analysis at multiple time points before and after interven­tions were conducted in order to detect whether or not the dif­ferent interventions had a significantly greater effect than any underlying trend.20-22 Autoregressive integrated moving average (ARIMA) models, using Box-Jenkins methodology,20 were used to evaluate whether relationships existed between time intervals (LOS, percentage leaving ED in under 4 hours, wait time to see provider) and the different interventions introduced. Time be­tween interventions were longer than 3 months (with 3 measures points) to guarantee a stable mean, and moving averages were calculated if necessary to correct trend.20-22

The model was identified by determining the ARIMA model orders (p, d, q) using auto-correlation and partial auto-correla­tion. Finally, the adequacy of the model was checked and statisti­cal significance of the parameters was determined. The best type of impact was evaluated as previously described.20,21 The gradual permanent impact was used; gradual permanent pattern includes 2 parameters estimated from the ITS analysis, which allows us to evaluate the level of the disturbance event and its duration. Omega estimates the difference between the data series before and after the intervention; delta indicates the duration of any change and the rate of recovery if present. Then, omega (ω) indicates the overall shift where the sign represents the direction of the change, and delta (δ) represents the speed at which a grad­ual increase or decay of these initial changes occurred over time. Both parameters should be statistically significant so as not end up with paradoxical conclusions.21 Omega and delta asymptotic parameters ± SD values are presented as well as their 95% CIs.

Statistica software (Statsoft, Tulsa, Oklahoma) was used for analysis. We deemed statistical differences to be significant for P <.05.


Characteristics of the Study Subjects

During the 6-year study period, 413,392 attendances were reg­istered in the ED. Monthly ED attendances increased over the study period from 4540 ± 214 in 2005 to 5928 ± 299 in 2011 (+31%; 95% CI, 20%-42%; P <.001) (Figure). No significant change in admission rate (20.3%), transfer to intensive care (3.2% of the entire population), or patients’ typology (medicine [60%], surgery [36%], psychiatric [4%]; patients ≥75 years [10.5%]) was found during the study period.

Main Results

The trends for main TIMQIs and monthly number of visits are presented. The chronology of the different interventions studied is indicated as well.

During the 6-year study period, findings included decreases in trends for LOS (–191 minutes; 95% CI, –280 to –101; P <.001) and waiting time to provider (–142 minutes; 95% CI, –225 to –66; P <.001), and a significant increase in the trend for the per­centage of patients leaving the ED in less than 4 hours (+24%; 95% CI, 12%-38%; P <.001).

Pre- and Post Intervention Analysis

Table 2 shows the results of ANOVA comparisons for the mea­sured time intervals in the pre-/post intervention analysis. Hav­ing an ED medical coordinator significantly reduced ED LOS, notably for high-acuity patients The interventions of having a dedicated physician for fast-tracking what is considered to be low-complexity also significantly improved the percentage of pa­tients leaving the ED in less than 4 hours.

Interrupted Time Series Analysis

If we look at the progress of the measurements of the different time intervals in the Figure and Table 2, we can see improvements between post intervention measures and the following pre-inter­vention, ie continuous improvement unrelated to interventions. Thus, another evaluation method was necessary to evaluate the impact of interventions over the study period. Table 3 shows the results of ITS models for ED TIMQIs.

We found that the interventions of an initial quality improve­ment step and having an ED medical coordinator had a signifi­cant impact on ED LOS, and that the interventions of moving to ED provisional structure and having 0 patients in the ED corri­dors had a significant impact on the percentage of patients leav­ing ED in less than 4 hours and on wait time to see the provider.


Copyright AJMC 2006-2019 Clinical Care Targeted Communications Group, LLC. All Rights Reserved.
Welcome the the new and improved, the premier managed market network. Tell us about yourself so that we can serve you better.
Sign Up