92 Original article
Factors associated with the length of stay of patients discharged from emergency department in France Fre´de´ric Capuanoa, Anne-Sophie Lota, Christine Sagnes-Raffyb, Marie Ferruaa, Dominique Brun-Neyc, Henri Leleua, Dominique Paterond, Guillaume Debatyf, Marc Giroudg, Etienne Minviellea and Bruno Rioue Objectives The length of stay in the emergency department (ED) has been proposed as an indicator of performance in many countries. We conducted a survey of length of stay in two large areas in France and tested the hypothesis that patient and ED-related variables may influence it. Patients and methods During 2007, we examined lengths of stay in ambulatory patients, that is, excluding admitted patients. The following variables were considered: (a) at the patient level, age, sex, the day and month of the visit, and the French clinical classification of emergency patients (CCEP) class; (b) at the ED level, annual ED total number of visits, mean age, the proportions of patients less than 15 and more than 75 years, and the proportions of admitted and clinically stable patients with CCEP class 1 and 2. A multilevel hierarchical analysis was carried out. Results We analyzed 988 591 visits in 53 EDs. The ED-specific median length of stay was 98 (IQR: 62–137) min and the ED-specific median proportion of patients with length of stay of more than 4 h was 15 (5–24) %. There was a strong correlation between the ED-specific median length of stay and the ED-specific proportion of patients with a length of stay of more than 4 h (R = 0.96, P < 0.001).
Introduction There is a growing role of the emergency department (ED) in hospital admissions and more globally in the healthcare systems in most countries [1]. Because of overcrowding and the need for a timely response for patients presenting with serious conditions, a triage process has been implemented in the ED and several quality and/or crowding measures have been proposed including waiting times, length of stay, boarding times, the incidence of patients who have left without being seen, and proportion of early revisits [2–4]. The respective influence of these measures on patient outcome may differ markedly. Longer waiting times and longer lengths of stay are associated with an increased risk of short-term death, which is not the case for patients who have left without being seen [4].
Using a multilevel analysis, only three variables were associated significantly with the length of stay: the age and the CCEP class of the patient, and the ED census. Conclusion We observed that the length of stay in the ED needs to be stratified by case mix and the total number of visits of the ED. European Journal of Emergency Medicine c 2015 Wolters Kluwer Health, Inc. All 22:92–98 Copyright rights reserved. European Journal of Emergency Medicine 2015, 22:92–98 Keywords: emergency department management, length of stay, performance, quality a Management in Health Services Research Department, Ecole des Hautes Etudes en Sante´ Publique (EHESP), Institut Gustave Roussy, Villejuif, b Observatoire des Urgences Midi-Pyre´ne´es (ORU-MiP), Toulouse, cMedical Headquarter, Assistance-Publique Hoˆpitaux de Paris (APHP), dEmergency Department, Centre Hospitalo-Universitaire (CHU) Saint Antoine, APHP, Universite´ Pierre et Marie Curie (UPMC), Paris, eEmergency Department, CHU Pitie´-Salpeˆtrie`re, APHP, UPMC, Paris, fPrehospital Emergency Department, SAMU 38, CHU de Grenoble, Grenoble and gPrehospital Emergency Department, SAMU 95, Hoˆpital de Pontoise, Pontoise, France
Correspondence to Bruno Riou, MD, PhD, Service d’Accueil des Urgences, Groupe Hospitalier Pitie´-Salpeˆtrie`re, 47-83 boulevard de l’hoˆpital, 75013 Paris, France Tel: + 33 1 42 17 72 40; fax: + 33 1 42 17 74 12; e-mail: [email protected] Received 12 September 2013 Accepted 9 December 2013
use of advanced imaging techniques [6]. Length of stay has been used as a key target threshold in England where 95% of ED patients must be discharged home or admitted within 4 h [7]. Although some standards have been recommended [7], little information is available on the need for stratification according to some variables such as case mix and/or ED census [8]. We conducted a 1-year survey of length of stay in the ED in two large areas in France and tested the hypothesis that variables at the patient and hospital level may influence it using a multilevel modeling approach [9].
Patients and methods
The length of stay in the ED has been proposed as an indicator of performance or as a measure of crowding in many countries [5]. A trend toward an increase in the median length of stay in the ED has been reported in the USA that is at least partly related to an increase in the
We carried out a retrospective multicenter study of patients who visited the ED in two large administrative areas in France (Ile de France and Midi-Pyre´ne´es) in 2007. These areas were chosen because computerized information on ED visits has been centralized in regional centers (CERVEAU, Paris, and ORU-MiP, Toulouse, respectively). In addition, these two regions are different,
c 2015 Wolters Kluwer Health, Inc. All rights reserved. 0969-9546 Copyright
DOI: 101097/MEJ.0000000000000109
Copyright © 2015 Wolters Kluwer Health, Inc. Unauthorized reproduction of the article is prohibited.
Factors in ED length of stay Capuano et al. 93
representing an essentially urban area in Ile de France and a more rural area in Midi-Pyre´ne´es, with a more important proportion of small ED. In the Ile de France area, the CERVEAU register was connected to 54 (59%) of the 92 ED and in the Midi-Pyre´ne´es area, the ORU-MiP register was connected to all 38 (100%) of the 38 ED. This study was carried out on an anonymous administrative database and was authorized by the Conseil National de l’Informatique et des Liberte´s (CNIL, Paris, France), from whom we received approval not to require patient consent. This study was carried out on the auspices and funding of the French Ministry of Health through the COMPAQH (Coordination pour la Mesure de la Performance et de l’Ame´lioration de la Qualite´ Hospitalie`re) project [10]. The criteria for inclusion in the study were a visit to an ED during 2007 in the two areas of study. Because the databases are anonymous, we could not identify patients with multiple visits, even in different hospitals, and thus in the following, one visit is considered as the observation unit having been made by one patient. Only nonadmitted patients were included because of lack of reliable data on end of ED stay for admitted patients and the marked heterogeneity of admission modalities between hospitals, including the short-duration (< 24 h) hospitalization wards that exist in all ED in France. Moreover, the length of stay of admitted patients was considered to reflect more the efficiency of the hospital itself rather than that of the ED. In practice, length of stay is often separated by admitted and discharged patients [8,11]. Nevertheless, we took into account the proportion of admitted patients in the multilevel model (vide infra). ED length of stay was defined as the interval between time of ED arrival and time of ED departure in ambulatory patients. We also examined the proportion of patients with a length of stay of less than 4 h as this time frame has been already proposed in other countries [7,11]. We analyzed the proportion of patients with a length of stay of more than 4 h according to the hour, weekday, or month of arrival at the ED. Values that differed more than 5% from the proportion in the entire population were considered as outliers. The following variables were considered a priori: (a) at the patient level, age, sex, hour, day and month of the visit, the clinical classification of emergency patients (CCEP class [12]), and the main diagnosis. (b) at the hospital level, total number of visits (visits/year), mean age, proportion of patients less than 15 years (including admitted patients), proportion of patients more than 75 years (including admitted patients), proportion of admitted patients, and proportion of patients with CCEP class 1 and 2 (including admitted patients). CCEP is a French nationwide a posteriori classification of emergency visits according to the final diagnosis and therapeutic resources needed, which has the following five classes: 1: clinically stable patients who did not require any diagnostic
or therapeutic procedure; 2: clinically stable patients who required diagnostic and/or therapeutic procedure; 3: clinical status that might deteriorate in the ED, but without lifethreatening conditions; 4: clinical status that might deteriorate in the ED with possible life-threatening conditions; and 5: life-threatening conditions requiring immediate therapeutic procedures [12]. We excluded ED with only pediatric emergencies and those with more than 20% of missing values regarding length of stay, decision (i.e. admitted vs. nonadmitted), or CCEP class. We also excluded patients with missing values of length of stay or CCEP class, and those with aberrant values (> 24 h) of length of stay. Low annual census was defined as a census lower than the 25 interquartile and high annual census as a census higher than the 75 interquartile. The statistical plan was decided a priori using an expert panel (the author group). Data were entered in a computerized database and inconsistencies between data were systematically checked and solved. Statistical analysis
Data are expressed as mean±SD or median (25–75 interquartile) for non-Gaussian variables. The correlation between two variables was assessed using the Spearman coefficient test. Multilevel modeling is a well-established method for simultaneously study relations among individual-level and group-level variables. Moreover, the hierarchical regression enables us to combine the classical frequentist and Bayesian methods [9]. At the patient level, we assessed the relationship between the individual length of stay and the characteristics of the patients. At the ED level, we assessed the relationship between the median length of stay and the characteristics of the ED. Then, a multilevel multivariate analysis was carried out, first at the patient level then at the ED level. Only variables with a P value of less than 0.20 in the univariate analysis were entered in the model. All statistical tests were two-sided, and a P value of less than 0.05 was required to reject the null hypothesis. Statistical analysis was carried out using SAS 9.1 software (SAS Institute, Cary, North Carolina, USA) and HLM 6.0 (Scientific Software International, Skokie, Illinois, USA).
Results We included data from 2 254 435 visits in 88 EDs, which represents 14% of all visits in an ED in France during the year studied. We finally retained for analysis 988 591 visits in 53 EDs (Fig. 1). The mean age of the patients was 35±22 years (range 0–110 years) and the proportion of patients older than 75 years was 6%. The main characteristics of the study population are shown in Table 1.
Copyright © 2015 Wolters Kluwer Health, Inc. Unauthorized reproduction of the article is prohibited.
94
European Journal of Emergency Medicine 2015, Vol 22 No 2
Fig. 1
Extracted population 88 EDs 2 254 435 visits Excluded EDs (n = 35) Pediatric EDs (n = 13) >20% missing values CCEP (n = 18) >20% missing values decision (n = 7) >20% missing values LoS (n = 2) Excluded visits Admitted patients (25%) Missing values CCEP (15%) LoS > 24 h (1%)
Studied population 53 EDs 988 591 visits
Flow chart of the study. CCEP, clinical classification of emergency patients; ED, emergency department; LoS, length of stay. The sum of excluded EDs overcomes the total (n = 35) as some EDs had several reasons to be excluded. Decision refers to admission (vs. nonadmission) to the hospital.
Table 1 Univariate analysis of variables associated with the length of stay in the emergency department at the patient level (n = 988 591) Variables Age (years) < 15 16–75 > 75 Missing value Sex Men Women Missing value CCEP class 1 2 3 4 5 Missing value
N (%)
Length of stay (min)
P value
178 090 (18) 750 131 (76) 59 710 (6) 660
80 (45–138) 120 (62–216) 180 (91–302)
< 0.001
538 891 (55) 449 539 (45) 161
108 (56–197) 121 (63–217)
< 0.001
236 544 717 141 34 059 779 68 0
75 124 233 342 167
< 0.001
(24) (73) (3) (0.07) (0.01)
(39–140) (67–219) (135–380) (164–707) (69–352)
Data are N (%) or median (25–75 interquartile). CCEP, clinical classification of emergency patients (see text).
The median total number of visits was 20 640 (10 794–35 652) visits/year. Thus, a low total number of visits was defined as less than 10 794 visits/year and a high total number of visits as more than 35 652 visits/year. The median length of stay in ED with low, intermediate, and high total number of visits was 76 (40–91), 98 (62–105), and 137 (107–159) min, respectively. The median length of stay in the entire cohort was 114 (60–206) min (Fig. 2). The median length of stay of ED-specific medians was 98 (62–137) min. The proportion of patients with length of stay of more than 4 h was 20% in the entire cohort and the median of ED-specific
proportion was 15 (5–24) %. Length of stay varied according to the hour of arrival (Fig. 3a), but not the day (Fig. 3b) or month (Fig. 3c) of arrival. Outliers (difference >5% from the weekday mean) were observed only for some hours of arrival (5:00, 6:00, and 7:00 a.m.), but not for weekday or month of arrival. There was a high and significant correlation between the median length of stay and the proportion of patients with a length of stay of more than 4 h (Fig. 4). In the univariate analysis, three variables were associated significantly with the length of stay at the level of the patient: age, sex, and CCEP class (Table 1). The length of stay was significantly greater in women compared with men (Table 1), but the women studied were also older than men (36±23 vs. 33±20 years, P < 0.001; proportion of patients Z 75 years old 8 vs. 4%, P < 0.001). In the univariate analysis, three variables were associated significantly with the length of stay at the level of the ED: the total number of visits, the proportion of patients below 15 years, and the proportion of patients older than 75 years of age (Table 2). Table 3 shows the multilevel analysis. Only three variables were associated significantly with the length of stay: the age and the CCEP class of the patient, and the total number of visits (Table 3).
Discussion In this observational study, we observed that three variables were associated significantly with the length of stay in an ED, namely, patient age, patient CCEP, and the total number of visits in the ED. Consequently, length of stay was greater in elderly patients and in patients with high CCEP class and in ED with a high total number of visits. Women waited longer than men, but this association observed in the univariate analysis disappeared in the multivariate analysis and is likely to be related to age differences between women and men, and aging has been clearly identified as a patient factor for longer length of stay [13]. We also observed a strong correlation between the median length of stay and the proportion of patients with a length of stay of more than 4 h. In our study, we excluded admitted patients because we considered those patients to be a very different population, whose length of stay in the ED does not reflect the same process [3], although processes are often related and may interfere one with the other. Many previous studies have separated length of stay in admitted and discharged patients: for example a maximum target threshold of 4 h is usually expected in discharged patients, whereas 8 h is expected in admitted patients [11]. Moreover, because a significant proportion of patients are admitted in a dedicated emergency hospital ward within the ED (short-duration hospitalization unit) in France, it was not possible to delineate precisely the length of stay in the emergency room in most admitted patients. This is the main reason why we
Copyright © 2015 Wolters Kluwer Health, Inc. Unauthorized reproduction of the article is prohibited.
Factors in ED length of stay Capuano et al. 95
Fig. 2
(b)
20
40
30 Percent of ED
Percent of patients
(a)
10
20
10
0
0
1 2 3 4 5 6 7 8 9 10 Length of stay (h)
1
2 3 4 Length of stay (h)
5
Distribution of length of stay in patients (a, n = 988 591) and in emergency department (ED) (b, median, n = 53). Each bar represents an interval of 30 min (i.e. 0–30 min).
10
10
0
0
0
M
Tu
o
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23
0 Hour of admission
Day of admission
30
10
20 5 10
ec D
ov N
p
g
ct O
Se
Au
l
n
Ju
r
ay
Ju
M
ar
Ap
M
Fe
Ja
b
0 n
0
Length of stay > 4 h (%)
Proportion of patients (%)
(c)
Su
20
Sa
2
20
Fr
10
30
Th
4
Proportion of patients (%)
20 6
Length of stay > 4 h (%)
Proportion of patients (%)
8
30
e
(b) 30
10
W
(a)
Month of admission
Variation of length of stay according to the hour (a), day (b), or month (c) of arrival at the emergency department in the entire population (n = 988 591). Scale bars refer to the proportion of visits and dot to the proportion of visits with a length of stay of more than 4 h.
Copyright © 2015 Wolters Kluwer Health, Inc. Unauthorized reproduction of the article is prohibited.
Length of stay > 4 h (%)
Fig. 3
96
European Journal of Emergency Medicine 2015, Vol 22 No 2
length of stay in contrast to high-complexity cases (i.e. admitted patients). In our study, the median length of stay was 114 min, higher than that observed in England (99 min) [15], where a threshold target has been implemented and lower than that observed in the USA (138 min) [16]. The median French ED discharged 85% of its patients within 4 h, a result very close to the median US ED (87%) [16], but lower than that recorded in England (98%) [15].
Fig. 4
Median length of stay (min)
200
150
100 R = 0.96 P < 0.001
50
0
0
10 20 30 40 Length of stay > 4 h (%)
50
Correlation between the median length of stay and the proportion of visits with a length of stay of more than 4 h in the different hospitals (n = 53).
Table 2 Univariate analysis of variables associated with the length of stay in the emergency department at the hospital level (n = 53) Variables
Spearman’s correlation coefficient
P value
0.54 – 0.39 – 0.35 0.017 0.048 – 0.101
< 0.001 0.004 0.011 0.90 0.73 0.47
Annual census of the ED Proportion of patients < 15 years Proportion of patients > 75 years Mean age Proportion of patients admitted Proportion of patients CCEP 1 and 2 Proportion of patients CCEP 2
0.161
0.25
CCEP, clinical classification of emergency patients (see text); ED, emergency department.
Table 3 Multilevel analysis of factors associated with the length of stay in the emergency department Variables Patient level Intercept Age < 15 years (vs. 16–75 years) Age > 75 years (vs. 16–75 years) CCEP class 2 (vs. class 1) CCEP class 3, 4, 5 (vs. class 1) Hospital level Low annual census of the ED (vs. high) Moderate annual census of the ED (vs. high)
Coefficient 132 – 36 61 51 193 – 90 – 68
SE
P value
147 9 6 5 5 33 43 14 12
< 0.001 < 0.001 < 0.001 < 0.001 < 0.001 < 0.001 < 0.001 < 0.001
Low annual census was defined as a census lower than the 25 interquartile (i.e. 35 652 visits/year). CCEP, clinical classification of emergency patients (see text).
only considered ambulatory patients in our study. Nevertheless, when considering the ED level, we took into account the proportion of admitted patients as this proportion was considered to reflect a large burden of clinical work carried out in the emergency room. In fact, Schull et al. [14] observed that low-complexity cases (i.e. nonadmitted patients) had a relatively low impact on the
The length of stay in the ED has been used recently as a key target threshold of performance in several countries [5,17]. Increased length of stay has been shown to be associated with delays in analgesia [18], worse adherence to guidelines [19], and an increased risk of short-term mortality [3]. However, the length of stay in the ED is only one of the available indicators and cannot reflect completely the performance of an ED. Other indicators may be more crucial, particularly in critically ill patients such as time to triage or the waiting time to see a physician [20,21]. However, this indicator may be interesting, provided that exogenous variables are taken into account. We believe that our study provides useful information about the variables associated with this indicator, and suggest that its value should be interpreted with other important variables at both the patient and the ED levels (Table 3). Performance measures, particularly those linked to payment, may have deleterious consequences. Belorgey [22] has suggested a link between the reduction in the length of stay and an increase in the proportion of unscheduled return visit to the ED. Mason et al. [15] observed that the introduction of the 4-h target led to an expected decrease in the proportion of prolonged stays, but with an increasing proportion of patients leaving the ED in the last minutes before threshold, particularly in elderly patients. This suggests that ED may be achieving performance but without improving care, which is a known unintended consequence of targeted approach [23]. We observed that the median length of stay and the proportion of patients with a length of stay of more than 4 h were significantly and strongly correlated (Fig. 4). Of course, some degree of correlation between these two variables was expected, but its strength was remarkable. As the first one is less susceptible to manipulation, we recommend the use of the median length of stay as a quality measure. At the patient level, we did not observe significant changes in the median length of stay between days of the week (Fig. 3b) or months (Fig. 3c). In contrast, there was a marked difference between hours of arrival (Fig. 3a). It has been shown previously that length of stay measured at a daily level may mask much of the variation that occurs within a 24-h period [24]. It is of interest to note that the more important proportions of prolonged stay were observed at the end of the night (from 4:00 to
Copyright © 2015 Wolters Kluwer Health, Inc. Unauthorized reproduction of the article is prohibited.
Factors in ED length of stay Capuano et al. 97
7:00 a.m.) when the proportion of new patients is the lowest (Fig. 3a). Downing et al. [25] also observed that arrival at night was associated with a higher risk of prolonged stay in the ED. Several hypotheses may explain this difference: first, there may be a higher cumulative proportion of patients with an admission decision but without possibility to discharge them from the ED, leading to an increased burden of work for the emergency team; second, there may be a reduction in the number of physicians available to take care of the new patients during this period; and finally, there may be a decreased efficiency of the emergency team during these late hours either because of fatigue or decrease in the flow, or closure of electronic files a long time after patients had actually left the ED, or decrease in overall hospital efficiency such as access to radiology and laboratory tests. Our study did not enable us to verify these hypotheses. These differences at a daily level may not be important when assessing the performance of an ED, but may be of paramount importance when assessing the impact on individual outcome [24]. However, improving the length of stay during this period may be a useful objective, although it is relevant for only a relatively small proportion of patients.
Haute Autorite´ de Sante´ (HAS) through the National COMPAQH program (Coordination pour la mesure de la Performance et l’Ame´lioration de la Qualite´ Hospitalie`re). The authors thank Dr David J. Baker, DM, FRCA (Dept. of Anesthesiology, CHU Necker-Enfants Malades, Paris, France) for reviewing the manuscript, and Marc Levaillant (Engineer, CERMES3, UMR CNRS8211INSERM 988, Villejuif, France) for methodological advice. They also thank Perrine Rame´-Matthieu and Vincent Beaugrand (DGOS, Paris France), Vale´rie Salomon (Ministry of Health, Paris, France), and Catherine Grenier (Haute Autorite´ de la Sante´, Paris, France) for continuing support. Conflicts of interest
There are no conflicts of interest.
References 1 2
3
This study has several limitations. First, because of a high proportion of missing values, we could not include the characteristics of the final diagnosis of the patients who visited the ED. Second, our results may not apply to specialized pediatric EDs as we excluded them. Third, we did not take into account the number of staff, both medical and nonmedical, available in these EDs to take care of the patients. These missing variables may play an important role [26] that should be assessed in future studies. Conversely, we did not take into account the architectural characteristics of the ED and the bed capacity of the short-duration hospitalization unit in the ED or of the hospital. Finally, we did not take into account socioeconomic factors that are known to interfere with ED use [27]. All these limitations should be considered as having implications for future research and potential changes in clinical practice.
10
Conclusion
11
In this observational study carried out in two large areas in France, we observed that the length of stay in the ED needs to be stratified by case mix (age, CCEP class) and total number of visits of to the ED. We observed a strong correlation between the median length of stay and the proportion of patients with a length of stay of more than 4 h and thus we recommend the use of the median length of stay as a quality measure as it is less susceptible to manipulation.
Acknowledgements The study was sponsored by Direction Ge´ne´rale de l’Offre de Soins (DGOS), French Ministry of Health, and
4
5
6
7
8
9
12
13
14 15
16
Schuur JD, Venkatesh AK. The growing role of emergency departments in hospital admissions. N Engl J Med 2012; 367:391–393. National Health Quality Forum. National Quality Forum endorses measures to address care coordination and efficiency in hospital emergency departments. Available at: http://www.qualityforum.org/News_And_ Ressources/Press_Release/2008/National_Quality_Forum8Endorses_ Measures_to_Adress_Care_Coordination_And_Efficiency_in_Hospitals_ Emergency_Department.aspx. [Accessed 1 January 2013]. Guttmann A, Schull MJ, Vermeulen MJ, Stukel TA. Association between waiting times and short term mortality and hospital admission after departure from emergency department: population based cohort study from Ontario, Canada. BMJ 2011; 342:d2983. Sauvin G, Freund Y, Saı¨di K, Riou B, Hausfater P. Unscheduled return visits to emergency department: consequences for triage. Acad Emerg Med 2013; 20:33–39. Pines JM, Hilton JA, Weber EJ, Alkemade AJ, Al Shabanah H, Anderson PD, et al. International perspectives on emergency crowding. Acad Emerg Med 2011; 18:1358–1370. Herring A, Wilper A, Himmelstein DU, Woodhandler S, Espinola JA, Brown DFM, et al. Increasing length of stay among adults visits to U.S. emergency departments, 2001–2005. Acad Emerg Med 2009; 16: 609–616. Weber EJ, Mason S, Carter A, Hew R. Emptying the corridors of shame: organized lessons from England’s 4-hour emergency throughtput target. Ann Emerg Med 2011; 57:79–88. Pines JM, Decker SL, Hu T. Exogenous predictors of national performance measures for emergency department crowding. Ann Emerg Med 2012; 60:293–298. Greenland S. Principles of multilevel modeling. Int J Epidemiol 2000; 29:158–167. Corriol C, Grenier C, Coudert C, Daucourt V, Minvielle E. The COMPAQH: research on quality indicators in hospitals. Rev Epidemiol Sante Pub 2008; 56 (Suppl):S52–S60. Fee C, Burstin H, Maselli JH, Hsia R. Association of emergency department length of stay with safety-net status. JAMA 2012; 307:476–482. Fourestie´ V, Roussignol E, Elkharrat D, Rauss A, Simon N. Clinical classification of emergency patients. Definition and reproducibility [in French]. Re´an Urg 1994; 3:573–578. Freund Y, Vincent-Cassy C, Bloom B, Riou B, Ray P. the APHP Emergency Database Study Group. Association between age older than 75 years and exceeded target waiting times in the emergency department: a multicenter cross-sectional survey in the Paris metropolitan area, France. Ann Emerg Med 2013; 62:449–456. Schull MJ, Kiss A, Szalai JP. The effect of low-complexity patients on emergency department waiting times. Ann Emerg Med 2007; 49:257–264. Mason S, Weber EJ, Coster J, Freeman J, Locker T. Time patients spend in the emergency department: England’s 4-hour rule. A case of hitting the target but missing the point? Ann Emerg Med 2012; 59:341–349. Horwitz LI, Green J, Bradley EH. United States emergency department performance on wait time and length of visit. Ann Emerg Med 2010; 55:133–141.
Copyright © 2015 Wolters Kluwer Health, Inc. Unauthorized reproduction of the article is prohibited.
98
17
18
19
20
21
European Journal of Emergency Medicine 2015, Vol 22 No 2
Jones P, Schimanski K. The four hour target to reduce Emergency Department ‘waiting time’: a systematic review of clinical outcomes. Emerg Med Australas 2010; 22:391–398. Mills AM, Shofer FS, Chen ES, Hollander JE, Pines JM. The association between emergency department crowding and analgesia administration in acute abdominal pain patients. Acad Emerg Med 2009; 16:603–608. Diercks DB, Roe MT, Chen AY, Peacock WF, Kirk JD, Pollack CV, et al. Prolonged emergency department stays of non-ST-segment elevation myocardial infarction patients are associated with worse adherence to the American College of Cardiology/American Heart Association guidelines for management and increased adverse events. Ann Emerg Med 2007; 50:489–496. Horwitz LI, Bradley EH. Percentage of US emergency department patients seen within the recommended triage time. Arch Intern Med 2009; 169:1857–1865. Ng D, Thomas S, Schmidt N. Applying the Lean, principles of the Toyota Production System to reduce wait times in the emergency department. CJEM 2010; 12:50–57.
22
23 24
25
26
27
Belorgey N. Reduction of waiting time and length of stay in emergency department. A reform process in a public service and its social effects [in French]. Actes Recherches Sci Soc 2011; 189:16–33. Mason S, Nicholi J, Locker T. Four hour emergency target. Targets still lead care in emergency departments. BMJ 2010; 341:c3579. McCarthy ML, Ding R, Pines JM, Zeger SL. Comparison of methods for measuring crowding and its effects on length of stay in the emergency department. Acad Emerg Med 2011; 18:1269–1277. Downing A, Wilson RC, Cooke MW. Which patients spend more than 4 hours in the Accident and Emergency Department. J Public Heath 2004; 26:172–176. Ekelund U, Kurland L, Eklund F, Torkki P, Letterstal A, Lindmarker P, et al. Patient throughput times and inflow patterns in Swedish emergency departments. A basis for ANSWER, A National Swedish Emergency Registry. Scan J Trauma Resusc Emerg Med 2011; 19:37–46. Moore G, Gerdtz M, Manias E, Hepworth G, Dent A. Socio-demographic and clinical characteristics of re-presentation to an Australian inner-city department: implications for service delivery. BMC Public Health 2007; 7:320.
Copyright © 2015 Wolters Kluwer Health, Inc. Unauthorized reproduction of the article is prohibited.
Factors associated with the length of stay of patients discharged from emergency department in France Fre´de´ric Capuanoa, Anne-Sophie Lota, Christine Sagnes-Raffyb, Marie Ferruaa, Dominique Brun-Neyc, Henri Leleua, Dominique Paterond, Guillaume Debatyf, Marc Giroudg, Etienne Minviellea and Bruno Rioue Objectives The length of stay in the emergency department (ED) has been proposed as an indicator of performance in many countries. We conducted a survey of length of stay in two large areas in France and tested the hypothesis that patient and ED-related variables may influence it. Patients and methods During 2007, we examined lengths of stay in ambulatory patients, that is, excluding admitted patients. The following variables were considered: (a) at the patient level, age, sex, the day and month of the visit, and the French clinical classification of emergency patients (CCEP) class; (b) at the ED level, annual ED total number of visits, mean age, the proportions of patients less than 15 and more than 75 years, and the proportions of admitted and clinically stable patients with CCEP class 1 and 2. A multilevel hierarchical analysis was carried out. Results We analyzed 988 591 visits in 53 EDs. The ED-specific median length of stay was 98 (IQR: 62–137) min and the ED-specific median proportion of patients with length of stay of more than 4 h was 15 (5–24) %. There was a strong correlation between the ED-specific median length of stay and the ED-specific proportion of patients with a length of stay of more than 4 h (R = 0.96, P < 0.001).
Introduction There is a growing role of the emergency department (ED) in hospital admissions and more globally in the healthcare systems in most countries [1]. Because of overcrowding and the need for a timely response for patients presenting with serious conditions, a triage process has been implemented in the ED and several quality and/or crowding measures have been proposed including waiting times, length of stay, boarding times, the incidence of patients who have left without being seen, and proportion of early revisits [2–4]. The respective influence of these measures on patient outcome may differ markedly. Longer waiting times and longer lengths of stay are associated with an increased risk of short-term death, which is not the case for patients who have left without being seen [4].
Using a multilevel analysis, only three variables were associated significantly with the length of stay: the age and the CCEP class of the patient, and the ED census. Conclusion We observed that the length of stay in the ED needs to be stratified by case mix and the total number of visits of the ED. European Journal of Emergency Medicine c 2015 Wolters Kluwer Health, Inc. All 22:92–98 Copyright rights reserved. European Journal of Emergency Medicine 2015, 22:92–98 Keywords: emergency department management, length of stay, performance, quality a Management in Health Services Research Department, Ecole des Hautes Etudes en Sante´ Publique (EHESP), Institut Gustave Roussy, Villejuif, b Observatoire des Urgences Midi-Pyre´ne´es (ORU-MiP), Toulouse, cMedical Headquarter, Assistance-Publique Hoˆpitaux de Paris (APHP), dEmergency Department, Centre Hospitalo-Universitaire (CHU) Saint Antoine, APHP, Universite´ Pierre et Marie Curie (UPMC), Paris, eEmergency Department, CHU Pitie´-Salpeˆtrie`re, APHP, UPMC, Paris, fPrehospital Emergency Department, SAMU 38, CHU de Grenoble, Grenoble and gPrehospital Emergency Department, SAMU 95, Hoˆpital de Pontoise, Pontoise, France
Correspondence to Bruno Riou, MD, PhD, Service d’Accueil des Urgences, Groupe Hospitalier Pitie´-Salpeˆtrie`re, 47-83 boulevard de l’hoˆpital, 75013 Paris, France Tel: + 33 1 42 17 72 40; fax: + 33 1 42 17 74 12; e-mail: [email protected] Received 12 September 2013 Accepted 9 December 2013
use of advanced imaging techniques [6]. Length of stay has been used as a key target threshold in England where 95% of ED patients must be discharged home or admitted within 4 h [7]. Although some standards have been recommended [7], little information is available on the need for stratification according to some variables such as case mix and/or ED census [8]. We conducted a 1-year survey of length of stay in the ED in two large areas in France and tested the hypothesis that variables at the patient and hospital level may influence it using a multilevel modeling approach [9].
Patients and methods
The length of stay in the ED has been proposed as an indicator of performance or as a measure of crowding in many countries [5]. A trend toward an increase in the median length of stay in the ED has been reported in the USA that is at least partly related to an increase in the
We carried out a retrospective multicenter study of patients who visited the ED in two large administrative areas in France (Ile de France and Midi-Pyre´ne´es) in 2007. These areas were chosen because computerized information on ED visits has been centralized in regional centers (CERVEAU, Paris, and ORU-MiP, Toulouse, respectively). In addition, these two regions are different,
c 2015 Wolters Kluwer Health, Inc. All rights reserved. 0969-9546 Copyright
DOI: 101097/MEJ.0000000000000109
Copyright © 2015 Wolters Kluwer Health, Inc. Unauthorized reproduction of the article is prohibited.
Factors in ED length of stay Capuano et al. 93
representing an essentially urban area in Ile de France and a more rural area in Midi-Pyre´ne´es, with a more important proportion of small ED. In the Ile de France area, the CERVEAU register was connected to 54 (59%) of the 92 ED and in the Midi-Pyre´ne´es area, the ORU-MiP register was connected to all 38 (100%) of the 38 ED. This study was carried out on an anonymous administrative database and was authorized by the Conseil National de l’Informatique et des Liberte´s (CNIL, Paris, France), from whom we received approval not to require patient consent. This study was carried out on the auspices and funding of the French Ministry of Health through the COMPAQH (Coordination pour la Mesure de la Performance et de l’Ame´lioration de la Qualite´ Hospitalie`re) project [10]. The criteria for inclusion in the study were a visit to an ED during 2007 in the two areas of study. Because the databases are anonymous, we could not identify patients with multiple visits, even in different hospitals, and thus in the following, one visit is considered as the observation unit having been made by one patient. Only nonadmitted patients were included because of lack of reliable data on end of ED stay for admitted patients and the marked heterogeneity of admission modalities between hospitals, including the short-duration (< 24 h) hospitalization wards that exist in all ED in France. Moreover, the length of stay of admitted patients was considered to reflect more the efficiency of the hospital itself rather than that of the ED. In practice, length of stay is often separated by admitted and discharged patients [8,11]. Nevertheless, we took into account the proportion of admitted patients in the multilevel model (vide infra). ED length of stay was defined as the interval between time of ED arrival and time of ED departure in ambulatory patients. We also examined the proportion of patients with a length of stay of less than 4 h as this time frame has been already proposed in other countries [7,11]. We analyzed the proportion of patients with a length of stay of more than 4 h according to the hour, weekday, or month of arrival at the ED. Values that differed more than 5% from the proportion in the entire population were considered as outliers. The following variables were considered a priori: (a) at the patient level, age, sex, hour, day and month of the visit, the clinical classification of emergency patients (CCEP class [12]), and the main diagnosis. (b) at the hospital level, total number of visits (visits/year), mean age, proportion of patients less than 15 years (including admitted patients), proportion of patients more than 75 years (including admitted patients), proportion of admitted patients, and proportion of patients with CCEP class 1 and 2 (including admitted patients). CCEP is a French nationwide a posteriori classification of emergency visits according to the final diagnosis and therapeutic resources needed, which has the following five classes: 1: clinically stable patients who did not require any diagnostic
or therapeutic procedure; 2: clinically stable patients who required diagnostic and/or therapeutic procedure; 3: clinical status that might deteriorate in the ED, but without lifethreatening conditions; 4: clinical status that might deteriorate in the ED with possible life-threatening conditions; and 5: life-threatening conditions requiring immediate therapeutic procedures [12]. We excluded ED with only pediatric emergencies and those with more than 20% of missing values regarding length of stay, decision (i.e. admitted vs. nonadmitted), or CCEP class. We also excluded patients with missing values of length of stay or CCEP class, and those with aberrant values (> 24 h) of length of stay. Low annual census was defined as a census lower than the 25 interquartile and high annual census as a census higher than the 75 interquartile. The statistical plan was decided a priori using an expert panel (the author group). Data were entered in a computerized database and inconsistencies between data were systematically checked and solved. Statistical analysis
Data are expressed as mean±SD or median (25–75 interquartile) for non-Gaussian variables. The correlation between two variables was assessed using the Spearman coefficient test. Multilevel modeling is a well-established method for simultaneously study relations among individual-level and group-level variables. Moreover, the hierarchical regression enables us to combine the classical frequentist and Bayesian methods [9]. At the patient level, we assessed the relationship between the individual length of stay and the characteristics of the patients. At the ED level, we assessed the relationship between the median length of stay and the characteristics of the ED. Then, a multilevel multivariate analysis was carried out, first at the patient level then at the ED level. Only variables with a P value of less than 0.20 in the univariate analysis were entered in the model. All statistical tests were two-sided, and a P value of less than 0.05 was required to reject the null hypothesis. Statistical analysis was carried out using SAS 9.1 software (SAS Institute, Cary, North Carolina, USA) and HLM 6.0 (Scientific Software International, Skokie, Illinois, USA).
Results We included data from 2 254 435 visits in 88 EDs, which represents 14% of all visits in an ED in France during the year studied. We finally retained for analysis 988 591 visits in 53 EDs (Fig. 1). The mean age of the patients was 35±22 years (range 0–110 years) and the proportion of patients older than 75 years was 6%. The main characteristics of the study population are shown in Table 1.
Copyright © 2015 Wolters Kluwer Health, Inc. Unauthorized reproduction of the article is prohibited.
94
European Journal of Emergency Medicine 2015, Vol 22 No 2
Fig. 1
Extracted population 88 EDs 2 254 435 visits Excluded EDs (n = 35) Pediatric EDs (n = 13) >20% missing values CCEP (n = 18) >20% missing values decision (n = 7) >20% missing values LoS (n = 2) Excluded visits Admitted patients (25%) Missing values CCEP (15%) LoS > 24 h (1%)
Studied population 53 EDs 988 591 visits
Flow chart of the study. CCEP, clinical classification of emergency patients; ED, emergency department; LoS, length of stay. The sum of excluded EDs overcomes the total (n = 35) as some EDs had several reasons to be excluded. Decision refers to admission (vs. nonadmission) to the hospital.
Table 1 Univariate analysis of variables associated with the length of stay in the emergency department at the patient level (n = 988 591) Variables Age (years) < 15 16–75 > 75 Missing value Sex Men Women Missing value CCEP class 1 2 3 4 5 Missing value
N (%)
Length of stay (min)
P value
178 090 (18) 750 131 (76) 59 710 (6) 660
80 (45–138) 120 (62–216) 180 (91–302)
< 0.001
538 891 (55) 449 539 (45) 161
108 (56–197) 121 (63–217)
< 0.001
236 544 717 141 34 059 779 68 0
75 124 233 342 167
< 0.001
(24) (73) (3) (0.07) (0.01)
(39–140) (67–219) (135–380) (164–707) (69–352)
Data are N (%) or median (25–75 interquartile). CCEP, clinical classification of emergency patients (see text).
The median total number of visits was 20 640 (10 794–35 652) visits/year. Thus, a low total number of visits was defined as less than 10 794 visits/year and a high total number of visits as more than 35 652 visits/year. The median length of stay in ED with low, intermediate, and high total number of visits was 76 (40–91), 98 (62–105), and 137 (107–159) min, respectively. The median length of stay in the entire cohort was 114 (60–206) min (Fig. 2). The median length of stay of ED-specific medians was 98 (62–137) min. The proportion of patients with length of stay of more than 4 h was 20% in the entire cohort and the median of ED-specific
proportion was 15 (5–24) %. Length of stay varied according to the hour of arrival (Fig. 3a), but not the day (Fig. 3b) or month (Fig. 3c) of arrival. Outliers (difference >5% from the weekday mean) were observed only for some hours of arrival (5:00, 6:00, and 7:00 a.m.), but not for weekday or month of arrival. There was a high and significant correlation between the median length of stay and the proportion of patients with a length of stay of more than 4 h (Fig. 4). In the univariate analysis, three variables were associated significantly with the length of stay at the level of the patient: age, sex, and CCEP class (Table 1). The length of stay was significantly greater in women compared with men (Table 1), but the women studied were also older than men (36±23 vs. 33±20 years, P < 0.001; proportion of patients Z 75 years old 8 vs. 4%, P < 0.001). In the univariate analysis, three variables were associated significantly with the length of stay at the level of the ED: the total number of visits, the proportion of patients below 15 years, and the proportion of patients older than 75 years of age (Table 2). Table 3 shows the multilevel analysis. Only three variables were associated significantly with the length of stay: the age and the CCEP class of the patient, and the total number of visits (Table 3).
Discussion In this observational study, we observed that three variables were associated significantly with the length of stay in an ED, namely, patient age, patient CCEP, and the total number of visits in the ED. Consequently, length of stay was greater in elderly patients and in patients with high CCEP class and in ED with a high total number of visits. Women waited longer than men, but this association observed in the univariate analysis disappeared in the multivariate analysis and is likely to be related to age differences between women and men, and aging has been clearly identified as a patient factor for longer length of stay [13]. We also observed a strong correlation between the median length of stay and the proportion of patients with a length of stay of more than 4 h. In our study, we excluded admitted patients because we considered those patients to be a very different population, whose length of stay in the ED does not reflect the same process [3], although processes are often related and may interfere one with the other. Many previous studies have separated length of stay in admitted and discharged patients: for example a maximum target threshold of 4 h is usually expected in discharged patients, whereas 8 h is expected in admitted patients [11]. Moreover, because a significant proportion of patients are admitted in a dedicated emergency hospital ward within the ED (short-duration hospitalization unit) in France, it was not possible to delineate precisely the length of stay in the emergency room in most admitted patients. This is the main reason why we
Copyright © 2015 Wolters Kluwer Health, Inc. Unauthorized reproduction of the article is prohibited.
Factors in ED length of stay Capuano et al. 95
Fig. 2
(b)
20
40
30 Percent of ED
Percent of patients
(a)
10
20
10
0
0
1 2 3 4 5 6 7 8 9 10 Length of stay (h)
1
2 3 4 Length of stay (h)
5
Distribution of length of stay in patients (a, n = 988 591) and in emergency department (ED) (b, median, n = 53). Each bar represents an interval of 30 min (i.e. 0–30 min).
10
10
0
0
0
M
Tu
o
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23
0 Hour of admission
Day of admission
30
10
20 5 10
ec D
ov N
p
g
ct O
Se
Au
l
n
Ju
r
ay
Ju
M
ar
Ap
M
Fe
Ja
b
0 n
0
Length of stay > 4 h (%)
Proportion of patients (%)
(c)
Su
20
Sa
2
20
Fr
10
30
Th
4
Proportion of patients (%)
20 6
Length of stay > 4 h (%)
Proportion of patients (%)
8
30
e
(b) 30
10
W
(a)
Month of admission
Variation of length of stay according to the hour (a), day (b), or month (c) of arrival at the emergency department in the entire population (n = 988 591). Scale bars refer to the proportion of visits and dot to the proportion of visits with a length of stay of more than 4 h.
Copyright © 2015 Wolters Kluwer Health, Inc. Unauthorized reproduction of the article is prohibited.
Length of stay > 4 h (%)
Fig. 3
96
European Journal of Emergency Medicine 2015, Vol 22 No 2
length of stay in contrast to high-complexity cases (i.e. admitted patients). In our study, the median length of stay was 114 min, higher than that observed in England (99 min) [15], where a threshold target has been implemented and lower than that observed in the USA (138 min) [16]. The median French ED discharged 85% of its patients within 4 h, a result very close to the median US ED (87%) [16], but lower than that recorded in England (98%) [15].
Fig. 4
Median length of stay (min)
200
150
100 R = 0.96 P < 0.001
50
0
0
10 20 30 40 Length of stay > 4 h (%)
50
Correlation between the median length of stay and the proportion of visits with a length of stay of more than 4 h in the different hospitals (n = 53).
Table 2 Univariate analysis of variables associated with the length of stay in the emergency department at the hospital level (n = 53) Variables
Spearman’s correlation coefficient
P value
0.54 – 0.39 – 0.35 0.017 0.048 – 0.101
< 0.001 0.004 0.011 0.90 0.73 0.47
Annual census of the ED Proportion of patients < 15 years Proportion of patients > 75 years Mean age Proportion of patients admitted Proportion of patients CCEP 1 and 2 Proportion of patients CCEP 2
0.161
0.25
CCEP, clinical classification of emergency patients (see text); ED, emergency department.
Table 3 Multilevel analysis of factors associated with the length of stay in the emergency department Variables Patient level Intercept Age < 15 years (vs. 16–75 years) Age > 75 years (vs. 16–75 years) CCEP class 2 (vs. class 1) CCEP class 3, 4, 5 (vs. class 1) Hospital level Low annual census of the ED (vs. high) Moderate annual census of the ED (vs. high)
Coefficient 132 – 36 61 51 193 – 90 – 68
SE
P value
147 9 6 5 5 33 43 14 12
< 0.001 < 0.001 < 0.001 < 0.001 < 0.001 < 0.001 < 0.001 < 0.001
Low annual census was defined as a census lower than the 25 interquartile (i.e. 35 652 visits/year). CCEP, clinical classification of emergency patients (see text).
only considered ambulatory patients in our study. Nevertheless, when considering the ED level, we took into account the proportion of admitted patients as this proportion was considered to reflect a large burden of clinical work carried out in the emergency room. In fact, Schull et al. [14] observed that low-complexity cases (i.e. nonadmitted patients) had a relatively low impact on the
The length of stay in the ED has been used recently as a key target threshold of performance in several countries [5,17]. Increased length of stay has been shown to be associated with delays in analgesia [18], worse adherence to guidelines [19], and an increased risk of short-term mortality [3]. However, the length of stay in the ED is only one of the available indicators and cannot reflect completely the performance of an ED. Other indicators may be more crucial, particularly in critically ill patients such as time to triage or the waiting time to see a physician [20,21]. However, this indicator may be interesting, provided that exogenous variables are taken into account. We believe that our study provides useful information about the variables associated with this indicator, and suggest that its value should be interpreted with other important variables at both the patient and the ED levels (Table 3). Performance measures, particularly those linked to payment, may have deleterious consequences. Belorgey [22] has suggested a link between the reduction in the length of stay and an increase in the proportion of unscheduled return visit to the ED. Mason et al. [15] observed that the introduction of the 4-h target led to an expected decrease in the proportion of prolonged stays, but with an increasing proportion of patients leaving the ED in the last minutes before threshold, particularly in elderly patients. This suggests that ED may be achieving performance but without improving care, which is a known unintended consequence of targeted approach [23]. We observed that the median length of stay and the proportion of patients with a length of stay of more than 4 h were significantly and strongly correlated (Fig. 4). Of course, some degree of correlation between these two variables was expected, but its strength was remarkable. As the first one is less susceptible to manipulation, we recommend the use of the median length of stay as a quality measure. At the patient level, we did not observe significant changes in the median length of stay between days of the week (Fig. 3b) or months (Fig. 3c). In contrast, there was a marked difference between hours of arrival (Fig. 3a). It has been shown previously that length of stay measured at a daily level may mask much of the variation that occurs within a 24-h period [24]. It is of interest to note that the more important proportions of prolonged stay were observed at the end of the night (from 4:00 to
Copyright © 2015 Wolters Kluwer Health, Inc. Unauthorized reproduction of the article is prohibited.
Factors in ED length of stay Capuano et al. 97
7:00 a.m.) when the proportion of new patients is the lowest (Fig. 3a). Downing et al. [25] also observed that arrival at night was associated with a higher risk of prolonged stay in the ED. Several hypotheses may explain this difference: first, there may be a higher cumulative proportion of patients with an admission decision but without possibility to discharge them from the ED, leading to an increased burden of work for the emergency team; second, there may be a reduction in the number of physicians available to take care of the new patients during this period; and finally, there may be a decreased efficiency of the emergency team during these late hours either because of fatigue or decrease in the flow, or closure of electronic files a long time after patients had actually left the ED, or decrease in overall hospital efficiency such as access to radiology and laboratory tests. Our study did not enable us to verify these hypotheses. These differences at a daily level may not be important when assessing the performance of an ED, but may be of paramount importance when assessing the impact on individual outcome [24]. However, improving the length of stay during this period may be a useful objective, although it is relevant for only a relatively small proportion of patients.
Haute Autorite´ de Sante´ (HAS) through the National COMPAQH program (Coordination pour la mesure de la Performance et l’Ame´lioration de la Qualite´ Hospitalie`re). The authors thank Dr David J. Baker, DM, FRCA (Dept. of Anesthesiology, CHU Necker-Enfants Malades, Paris, France) for reviewing the manuscript, and Marc Levaillant (Engineer, CERMES3, UMR CNRS8211INSERM 988, Villejuif, France) for methodological advice. They also thank Perrine Rame´-Matthieu and Vincent Beaugrand (DGOS, Paris France), Vale´rie Salomon (Ministry of Health, Paris, France), and Catherine Grenier (Haute Autorite´ de la Sante´, Paris, France) for continuing support. Conflicts of interest
There are no conflicts of interest.
References 1 2
3
This study has several limitations. First, because of a high proportion of missing values, we could not include the characteristics of the final diagnosis of the patients who visited the ED. Second, our results may not apply to specialized pediatric EDs as we excluded them. Third, we did not take into account the number of staff, both medical and nonmedical, available in these EDs to take care of the patients. These missing variables may play an important role [26] that should be assessed in future studies. Conversely, we did not take into account the architectural characteristics of the ED and the bed capacity of the short-duration hospitalization unit in the ED or of the hospital. Finally, we did not take into account socioeconomic factors that are known to interfere with ED use [27]. All these limitations should be considered as having implications for future research and potential changes in clinical practice.
10
Conclusion
11
In this observational study carried out in two large areas in France, we observed that the length of stay in the ED needs to be stratified by case mix (age, CCEP class) and total number of visits of to the ED. We observed a strong correlation between the median length of stay and the proportion of patients with a length of stay of more than 4 h and thus we recommend the use of the median length of stay as a quality measure as it is less susceptible to manipulation.
Acknowledgements The study was sponsored by Direction Ge´ne´rale de l’Offre de Soins (DGOS), French Ministry of Health, and
4
5
6
7
8
9
12
13
14 15
16
Schuur JD, Venkatesh AK. The growing role of emergency departments in hospital admissions. N Engl J Med 2012; 367:391–393. National Health Quality Forum. National Quality Forum endorses measures to address care coordination and efficiency in hospital emergency departments. Available at: http://www.qualityforum.org/News_And_ Ressources/Press_Release/2008/National_Quality_Forum8Endorses_ Measures_to_Adress_Care_Coordination_And_Efficiency_in_Hospitals_ Emergency_Department.aspx. [Accessed 1 January 2013]. Guttmann A, Schull MJ, Vermeulen MJ, Stukel TA. Association between waiting times and short term mortality and hospital admission after departure from emergency department: population based cohort study from Ontario, Canada. BMJ 2011; 342:d2983. Sauvin G, Freund Y, Saı¨di K, Riou B, Hausfater P. Unscheduled return visits to emergency department: consequences for triage. Acad Emerg Med 2013; 20:33–39. Pines JM, Hilton JA, Weber EJ, Alkemade AJ, Al Shabanah H, Anderson PD, et al. International perspectives on emergency crowding. Acad Emerg Med 2011; 18:1358–1370. Herring A, Wilper A, Himmelstein DU, Woodhandler S, Espinola JA, Brown DFM, et al. Increasing length of stay among adults visits to U.S. emergency departments, 2001–2005. Acad Emerg Med 2009; 16: 609–616. Weber EJ, Mason S, Carter A, Hew R. Emptying the corridors of shame: organized lessons from England’s 4-hour emergency throughtput target. Ann Emerg Med 2011; 57:79–88. Pines JM, Decker SL, Hu T. Exogenous predictors of national performance measures for emergency department crowding. Ann Emerg Med 2012; 60:293–298. Greenland S. Principles of multilevel modeling. Int J Epidemiol 2000; 29:158–167. Corriol C, Grenier C, Coudert C, Daucourt V, Minvielle E. The COMPAQH: research on quality indicators in hospitals. Rev Epidemiol Sante Pub 2008; 56 (Suppl):S52–S60. Fee C, Burstin H, Maselli JH, Hsia R. Association of emergency department length of stay with safety-net status. JAMA 2012; 307:476–482. Fourestie´ V, Roussignol E, Elkharrat D, Rauss A, Simon N. Clinical classification of emergency patients. Definition and reproducibility [in French]. Re´an Urg 1994; 3:573–578. Freund Y, Vincent-Cassy C, Bloom B, Riou B, Ray P. the APHP Emergency Database Study Group. Association between age older than 75 years and exceeded target waiting times in the emergency department: a multicenter cross-sectional survey in the Paris metropolitan area, France. Ann Emerg Med 2013; 62:449–456. Schull MJ, Kiss A, Szalai JP. The effect of low-complexity patients on emergency department waiting times. Ann Emerg Med 2007; 49:257–264. Mason S, Weber EJ, Coster J, Freeman J, Locker T. Time patients spend in the emergency department: England’s 4-hour rule. A case of hitting the target but missing the point? Ann Emerg Med 2012; 59:341–349. Horwitz LI, Green J, Bradley EH. United States emergency department performance on wait time and length of visit. Ann Emerg Med 2010; 55:133–141.
Copyright © 2015 Wolters Kluwer Health, Inc. Unauthorized reproduction of the article is prohibited.
98
17
18
19
20
21
European Journal of Emergency Medicine 2015, Vol 22 No 2
Jones P, Schimanski K. The four hour target to reduce Emergency Department ‘waiting time’: a systematic review of clinical outcomes. Emerg Med Australas 2010; 22:391–398. Mills AM, Shofer FS, Chen ES, Hollander JE, Pines JM. The association between emergency department crowding and analgesia administration in acute abdominal pain patients. Acad Emerg Med 2009; 16:603–608. Diercks DB, Roe MT, Chen AY, Peacock WF, Kirk JD, Pollack CV, et al. Prolonged emergency department stays of non-ST-segment elevation myocardial infarction patients are associated with worse adherence to the American College of Cardiology/American Heart Association guidelines for management and increased adverse events. Ann Emerg Med 2007; 50:489–496. Horwitz LI, Bradley EH. Percentage of US emergency department patients seen within the recommended triage time. Arch Intern Med 2009; 169:1857–1865. Ng D, Thomas S, Schmidt N. Applying the Lean, principles of the Toyota Production System to reduce wait times in the emergency department. CJEM 2010; 12:50–57.
22
23 24
25
26
27
Belorgey N. Reduction of waiting time and length of stay in emergency department. A reform process in a public service and its social effects [in French]. Actes Recherches Sci Soc 2011; 189:16–33. Mason S, Nicholi J, Locker T. Four hour emergency target. Targets still lead care in emergency departments. BMJ 2010; 341:c3579. McCarthy ML, Ding R, Pines JM, Zeger SL. Comparison of methods for measuring crowding and its effects on length of stay in the emergency department. Acad Emerg Med 2011; 18:1269–1277. Downing A, Wilson RC, Cooke MW. Which patients spend more than 4 hours in the Accident and Emergency Department. J Public Heath 2004; 26:172–176. Ekelund U, Kurland L, Eklund F, Torkki P, Letterstal A, Lindmarker P, et al. Patient throughput times and inflow patterns in Swedish emergency departments. A basis for ANSWER, A National Swedish Emergency Registry. Scan J Trauma Resusc Emerg Med 2011; 19:37–46. Moore G, Gerdtz M, Manias E, Hepworth G, Dent A. Socio-demographic and clinical characteristics of re-presentation to an Australian inner-city department: implications for service delivery. BMC Public Health 2007; 7:320.
Copyright © 2015 Wolters Kluwer Health, Inc. Unauthorized reproduction of the article is prohibited.
