Emergency Medicine

The 4-hour target in the emergency department, in-hospital mortality, and length of hospitalization: A single center-retrospective study

a b s t r a c t

Background: The four-hour (4 h’) rule in the emergency department (ED) is a performance-based measure intro- duced with the objective to improve the quality of care. We evaluated the association between time in the ED with in-hospital mortality and hospital length of stay .

Methods: This was a retrospective study performed in one public hospital with over 100,000 ED referrals per year. Hospitalizations from the ED during 2017 were analyzed. We defined time in the ED as either: until a decision was made (DED); or total time in the ED (TED). In-hospital mortality and LOS were evaluated for patients with DED or TED within and beyond 4 h’.

Results: Compared to patients with TED or DED within 4 h’, in-hospital mortality did not increase in patients with TED beyond 4 h’ (2.8% vs. 3.1%, non-significant), or DED beyond 4 h’ (2.1% vs. 3.2%, p < 0.001). LOS did increase in patients with either DED or TED beyond 4 h’ (p < 0.001). In-hospital mortality increased with increasing DED- TED intervals for patients hospitalized in the internal medicine departments: 3.7% (0-1 h’), 5.1% (1-2 h’), 5.7% (2-3 h’), and 7.1% (>3 h’) (p < 0.001).

Conclusions: In-hospital mortality was not associated with time in the ED beyond 4 h’. LOS, however, was in- creased in this group of patients. Decreased LOS observed in patients with time in the ED within 4 h’, does not support patients’ risk as a contributing factor leading to higher trends in mortality observed in this patient group. In-hospital mortality was associated with an increase in DED-TED intervals in patients hospitalized in the internal medicine departments.

(C) 2021 Published by Elsevier Inc.

  1. Introduction

In order to improve the quality of treatment, different performance- based measures have been adopted by various ministries of health and Healthcare organizations worldwide [1-3]. One such measure is the four-hour rule in the emergency department (ED) [4,5]. Within four hours (4 h’), patients attending the ED must be seen, treated, and a de- cision must be reached concerning admission or discharge. Though a popular performance-based measure, whether the four-hour rule in ED is associated with a decrease in in-hospital mortality or a decrease in Hospital length of stay is controversial [4,6,7].

The primary objective of this study was to evaluate the association between time in the ED and in-hospital mortality in one public hospital in Israel. The association of time in the ED and LOS was a secondary ob- jective. Other secondary objectives included the association of time in

* Corresponding author.

E-mail addresses: [email protected], [email protected] (I. Ashkenazi).

1 Both IA and OH previously worked in the Hillel Yaffe Medical Center.

the ED with in-hospital mortality and LOS in subgroups of patients de- fined by specialty, considering that, based on data from previous years, in-hospital mortality is expected to be higher in certain specialties (e.g., internal medicine). Different factors may influence the definition of time in the ED. For the purposes of this study we evaluated two def- initions for time in the ED: one which takes into account the time inter- val between registration in the ED and the decision concerning the patients’ disposition (specialty department); another which takes into account the time interval between registration in the ED and the re- corded time of discharge from the ED in which the patients were trans- ferred to one of the hospital’s wards. While the former definition takes into account procedures dependent on by the ED staff, the latter takes into account the actual time the patients is present in the ED.

  1. Methods
    1. Study setting, design, and selection of participants

The Hillel Yaffe Medical Center (HYMC) is ond of 28 public hospitals in Israel. It is the only hospital in the region, serving a population of over

https://doi.org/10.1016/j.ajem.2021.03.049 0735-6757/(C) 2021 Published by Elsevier Inc.

400,000 residents. Over 100,000 patients are evaluated in the ED yearly. Included in this retrospective study were all the patients hospitalized from HYMC’s ED between January 1st and December 31st, 2017. Pa- tients who died in the ED were excluded as many of these were in extremis when admitted. Patients who were discharged from the ED were excluded since their mortality rate is unknown. This study was au- thorized by the institution’s research ethics committees (HYMC-17- 0123).

Emergency medicine is a relatively new specialty in Israel. Almost all the physicians working in the ED during the study period were residents and specialists who worked in one of the specialties detailed below. Pa- tients admitted to the ED were triaged by a triage nurse according to the CTAS algorithm [8,9]. Depending on the presenting complaint, the triage nurse referred the patients to the physician on-call from the appropriate specialty. Patients were transferred between physicians from different specialties if this was deemed necessary. Up until transfer to the spe- cialty departments, the patients remained under the care of the physi- cians and nurses allocated to the ED. Each specialty was responsible for its patients.

    1. Measurements

Data collected included age, gender, time in the ED, specialty depart- ment, the decision to admit or discharge, LOS, and in-hospital mortality. Time until the decision in the ED (DED) for each patient was defined as the interval between the electronically-recorded admission time to the ED (the exact time when the patient was registered) and the electronically-recorded time of the decision made by the clinician in the ED of the patient’s disposition concerning target hospitalization (specialty department). Actual time spent in the ED (TED) for each pa- tient was defined as the interval between the electronically-recorded admission time to the ED and electronically-recorded time of discharge from the ED to one of the hospital’s wards. Thus, TED includes the time defined by DED (registration to decision) and the interval between the decision and the patient’s actual transfer from the ED to the specialty department (DED-TED interval). The DED-TED intervals represent Boarding times. Specialty department was defined by the specialty ad- mitting the patient from the ED, whether internal medicine (internal medicine, cardiology, and neurology), pediatrics, general surgery, gyne- cology, orthopedic surgery or other (ENT, psychiatry, ophthalmology, and urology). Specialty department for patients transferred from the ED to the intensive care unit was determined by the specialty of the physician who signed the decision to hospitalize.

    1. Outcomes

In-hospital mortality and LOS were the primary and secondary end- points measured. Mortality did not include patients who might have died immediately after their discharge, whether at home or at another hospital. LOS of those hospitalized did not include their time in the ED.

    1. Sample size calculation

In this study, all the patients who were registered in the ED during 2017 were included and all the patients who ended up hospitalized were analyzed. Overall in-hospital mortality was the primary outcome assessed. Prior data indicated that in-hospital mortality rates among pa- tients hospitalized from the ED approximate 3%. We assumed that in order to identify a >= 0.25 difference in in-hospital mortality in patients with either DED or TED beyond 4 h’ with an 80% power, using Chi- square test, assuming a two-sided ? of 0.05, with approximately two controls per case, a total sample size of 20,775 patients would be needed.

Assessment of in-hospital mortality in different speciality depart- ments was a secondary objective. In-hospital mortality in the internal medicine departments was assumed to be the highest within all

departments. Prior data indicated that in-hospital mortality in patients hospitalized from the ED to the internal medicine departments is in the order of 5%. We assumed that in order to identify a >=0.25 difference in in-hospital mortality in patients with either DED or TED beyond 4 h’ with an 80% power, using Chi-square test, assuming a two-sided ? of 0.05, with approximately two controls per case, a total sample size of 11,830 patients would be needed.

Assessment of LOS was another secondary objective. Prior data indi- cated that median LOS for patients hospitalized from the ED was ap- proximate 2 days. We assumed that in order to identify a 0.25 day difference in LOS with either DED or TED beyond 4 h’ with an 80% power, assuming a two-sided ? of 0.05, with approximately two con- trols per case, a total sample size of 2650 patients would be needed.

    1. Data analysis

Two models were constructed. These differed whether DED or TED was chosen as the representative time for the 4-h rule. DED was ana- lyzed since it includes a time period that is mostly dependent on the ED staff and the ED procedures. TED, on the other hand, represents the actual time the patients spent in the ED. TED is also the method used by other studies that evaluated the 4-h rule [6,10].

Differences in the proportion of sex distribution and in-hospital mor- tality were assessed with Chi-square test. Comparison between patients’ age, DED, TED, the time interval between DED and TED, and LOS were analyzed with the Mann-Whitney test. Association between age and LOS was investigated with the Spearman nonparametric correlation test. A post-hoc analysis of in-hospital mortality rates in patients with DED-TED intervals of up to 1 h, 1-2 h, 2-3 h, and more than 3 h were compared with Chi-square for trend. The combined effect of variables that were significantly related to in-hospital mortality was investigated twice using logistic regression, once for DED, and once for TED. Possible interactions between the variables were incorporated into the model. Results are presented as odds ratios and 95% confidence intervals. The combined effect of variables that were significantly related to LOS was investigated using negative binomial regression with the log link func- tion and an estimation of the dispersion parameter for hospitalized pa- tients. Results are presented as Incidence rate ratio and 95% confidence intervals. Data were analyzed using dedicated statistical software pro- grams (GraphPad Instat 3.06 and GraphPad Prism 6.00 versions for Win- dows, GraphPad Software Inc., San Diego, CA; SPSS Statistics for Windows, version 25.0, IBM Corp. Released 2017. Armonk, NY: IBM Corp.; R Core Team (2018). R: A language and environment for statistical computing. R Foundation for Statistical Computing, Vienna, Austria, URL http://www.R-project.org/). p-Values less than 0.05 were considered significant. Numbers, percentages, and interquartile ranges (IQR) were approximated to the nearest tenth and significant p-Values, odds ratio and 95% confidence intervals (95%CI) values to the nearest thousandth.

  1. Results
    1. Study population

During 2017, 106,842 patients were admitted to HYMC’s ED. Seventy-six patients who died during their ED stay and 78,658 patients who were discharged were excluded. The 76 excluded patients who died in the ED represent 8.4% (76/901) of the total deaths observed. Their median stay in the ED was 1.2 h’ (IQR 0.7, 1.8).

A total of 28,108 patients who were hospitalized were included in the analysis (Table 1). By four hours, DED was accomplished in 76.4% of the patients, while TED was accomplished in only 46.7%. This is inher- ent to the definition of TED which is longer than DED since it also in- cludes the time between the decision and the actual transfer of a patient from the ED to the specialty department. The rates of patients with DED and TED at four hours for the different specialties are pre- sented in Table 2.

Table 1

Demographic data, time until a decision was made in the ED (DED), and total time in the ED (TED), of 28,108 hospitalized patients included in this study.

Variable Hospitalized (n = 28,108)

Male% 53.1

Median Age (IQR) 54 (24, 73)

All (825/28,108)

3.2% (685/21,488)

2.1% (140/6620)???

Gynecology (6/1265)

0.4% (4/958)

0.7% (2/307)ns

Table 3

Mortality, both absolute and relative, according to specialty, as a function of time until a decision was made in the ED (DED).

Specialty (mortality/all) % mortality of hospitalized patients in each

subcategory (subgroup mortality/subgroup total)a

Median DED (IQR) 2.6 (1.6, 3.9)

Median TED (IQR) 4.2 (3.0, 5.7)

Age and sex attributes of patients with DED or TED within and beyond 4 h’

In patients with DED within 4 h’

In patients with DED beyond 4 h’

DED

TED

Orthopedic surgery (20/2438)

0.5% (8/1673)

1.6% (12/765)?

Within 4 h’

Beyond 4 h’

Within 4 h’

Beyond 4 h’

Pediatrics (3/4708)

Internal Medicine (720/14,412)

0.1% (3/4051)

5.5% (622/11,364)

0% (0/657)ns

3.2% (98/3048)???

Median age (IQR)

54 (22, 73)

55 (30, 74)

51 (19, 71)

57 (28, 75)

General Surgery (73/3835)

1.9% (45/2319)

1.8% (28/1516)ns

Male%

78.2

21.8

48.5

51.5

Others (3/1450)

0.3% (3/1123)

0% (0/327)ns

Female% 74.4 25.6 44.8 55.2

IQR – interquartile range.

Table 2

Cumulative percent DED and TED within 4 h’ by specialty.

DED within 4 h’ TED within 4 h’

p-Values for differences in mortality rates between patients with DED beyond 4 h’ and pa- tients with DED within 4 h’: ns nonsignificant; ?p < 0.05; ??p < 0.01; ???p < 0.001.

a % mortality is calculated by dividing the number of patients who died within each of the DED subgroups and the total number of patients within the same subgroup (numera- tor/denominator).

Table 4

Mortality, both absolute and relative, according to specialty, as a function of total time in the ED (TED).

Specialty (mortality/all) % mortality of hospitalized patients in each

subcategory (subgroup mortality/subgroup total)a

In patients with TED within 4 h’

Gynecology (1265)

75.7

53.8

Orthopedic Surgery (2438)

68.6

45.0

Pediatrics (4708)

86.0

54.8

Internal Medicine (14,412)

78.9

44.8

General Surgery (3835)

60.5

38.1

Others (1450)

77.4

59.3

All (28,108)

76.4

46.7

In patients with TED beyond 4 h’

All (825/28,108)

3.1% (411/13,132)

2.8% (414/14,976)ns

Gynecology (6/1265)

0.4% (3/681)

0.5% (3/584)ns

3.2

. Association of prolonged DED or TED with in-hospital mortality

Orthopedic surgery (20/2438)

0.5% (5/1097)

1.1% (15/1341)ns

Pediatrics (3/4708)

0.1% (2/2578)

0.0% (1/2130)ns

Eight hundred twenty-five patients died (Tables 3 and 4). Of these,

Internal Medicine (720/14,412)

5.7% (371/6453)

4.4% (349/7959)???

720 (87.3%) died in the internal medicine departments. Patients hospi- talized in the internal medicine departments, general surgery, and or- thopedic surgery accounted for 98.5% (813/825) of the total mortality. Patients hospitalized in these three specialty departments accounted for 98.5% (675/685) of the mortalities observed in patients with DED within 4 h’ and 98.6% (138/140) of the mortalities observed in patients with DED beyond 4 h’. They accounted for 98.5% (405/411) of the mor- talities observed in patients with TED within 4 h’ and 98.6% (406/408) of the mortalities observed in patients with TED beyond 4 h’.

In absolute terms, most of the in-hospital mortality occurred in pa- tients with DED within 4 h’. When in-hospital mortality was examined in relation to the number of patients in each subgroup, the in-hospital

General Surgery (73/3835) 2.0% (29/1463) 1.9% (44/2372)ns Others (3/1450) 0.1% (1/860) 0.3% (2/590)ns

p-Values for differences in mortality rates between patients with TED beyond 4 h’ and pa- tients with TED within 4 h’: ns nonsignificant; ?p < 0.05; ??p < 0.01; ???p < 0.001.

a % mortality is calculated by dividing the number of patients who died within each of the TED subgroups and the total number of patients within the same subgroup (numera- tor/denominator).

Table 5

Univariate analyses for possible risk factors for mortality. Univariate analysis

mortality rate did not increase in those with DED made beyond 4 h’. Subgroup analysis by specialty reveals some in-hospital mortality rates did increase in patients with decisions made beyond 4 h’. The rel- ative weights of these increases are offset by a decrease in in-hospital mortality observed in internal medicine.

TED produced different results compared to DED. In absolute terms, the number of patients who died in each time period was almost equal. In both the orthopedic and surgery departments, an increase in the ab- solute number of deaths was noticeable. When translated into percent-

Variable Died

(n = 825)

Median age in years (IQR) Sex

81 (70, 88)

53 (23, 72)

<0.001

Male (%)

431 (52.2)

14,486 (53.1)

0.65

Median DED (IQR)

2.2 (1.4, 3.3)

2.6 (1.6, 3.9)

<0.001

Median TED (IQR)

4.0 (2.8, 5.5)

4.2 (3.0, 5.7)

0.04

IQR – interquartile range.

Survived hospitalized

(n = 27,283)

p-Value

ages, in-hospital mortality rates in both time periods were relatively similar. This is true in most subspecialties, except in internal medicine, where in-hospital mortality decreased. Thus, no increased in-hospital mortality was observed in patients with TED beyond 4 h’.

Table 5 analyzes the association of DED, TED, and other possible co- variates with in-hospital mortality. Both DED and TED were shorter in patients who died. Of age and sex, only older age was associated with in-hospital mortality. When analyzed together with age, both DED and TED became non-significant while age remained significant (Table 6).

While longer DED was strongly associated with decreased

in-hospital mortality, longer TED was not. This led us to perform a post-hoc analysis of the association of in-hospital mortality with the

time interval between DED and TED (DED-TED). DED-TED interval was associated with an increase in mortality (<0.001) (Table 7). When the association between DED-TED intervals was examined for each of the three subspecialties that were the major contributors to in-hospital mortality, a real increase in in-hospital mortality was no- ticed only in internal medicine (<0.001).

3.3. Association of prolonged DED or TED with LOS

The total number of days of hospitalization was 111,762. These were distributed between the different specialties in the following way:

Table 6

Logistic regression for mortality evaluating age and time in the ED.

Age

1.071

1.062, 1.081

<0.001

DED

0.989

0.794, 1.232

0.92

2nd Model: Time defined according to TEDa

Age

1.076

1.064, 1.088

<0.001

TED

1.082

0.909, 1.288

0.37

Variable Odds ratio 95% CI p-Value 1st Model: Time defined according to DEDa

CI – confidence interval.

a Logistic regression was performed twice, depending if DED or TED were used to define time in the ED.

Table 7

Mortality according to DED-TED interval.

Specialty Mortality for different DED-TED intervals (%)

0-1 hour

1-2 hours

2-3 hours

3 or more hours

P value

All

212 (2.0)

319 (3.0)

179 (3.9)

115 (4.4)

<0.001

Internal Medicine

167 (3.7)

285 (5.1)

161 (5.7)

107 (7.1)

<0.001

General Surgery

32 (1.9)

22 (1.5)

15 (3.3)

4 (1.7)

0.56

Orthopedic Surgery

7 (0.7)

8 (0.8)

3 (1.1)

2 (1.8)

0.24

internal medicine 56.6%; general surgery 14.7%; orthopedic surgery 11.6%; pediatrics 9.1%; gynecology 3.1% and others 4.8%. Supplemental Table S1 presents the median LOS for the different specialties as a func- tion of DED or TED within 4 h’ or beyond 4 h’. Overall, LOS increased in patients with either DED or TED beyond 4 h’. Nevertheless, this was not true in all specialties. LOS decreased in internal medicine patients with DED or TED beyond 4 h’.

Beyond time until the decision, two other possible predictors were evaluated for their association with LOS (Supplemental Table S2). Fe- males had higher LOS compared to males. Correlation between age and LOS was positive and significant. When the three predictor vari- ables were modelled together, only age and ED time, whether DED or TED, predicted LOS (Supplemental Table S3).

  1. Discussion
    1. Principal findings

In this study, complying with the four-h-rule was not associated with lower in-hospital mortality. Furthermore, shorter stay in the ED (within 4 h’) was associated with increased in-hospital mortality in pa- tients hospitalized in internal medicine in both absolute and relative terms. Increasing time interval between DED and TED was also associ- ated with increased in-hospital mortality in these patients. LOS, how- ever, did increase in all patients hospitalized beyond 4 h’ in the ED.

    1. Interpretation within the context of the wider literature

In 2014, 2.9 million ED admissions were recorded in 28 public hospi- tals in Israel, commonly leading to occupancy exceeding 100% [11]. ED patient stay was on the rise, exceeding 5 h’ in 20%. Time spend in the ED became a national performance measure. However, in this study, no increased in-hospital mortality was noted in patients with either DED or TED beyond 4 h’.

The implementation of the 4-h-rule was achieved in 76.4% or 46.7% of the patients depending on if DED or TED were used to define time stay in the ED. This is inferior to the 85% target defined by the NEAT pro- ject in Australia [4]. In Australia, diminishing length of stay in the ED was associated with improved 30-day mortality in Western Australia while in other regions no improvement in mortality rates was noted

[10]. Of note is the observation that reports linking in-hospital mortality and ED crowding were reported by authors working in Western Australia before the implementation of the NEAT project [7].

Overcrowding has been shown to have a negative impact on differ- ent patient-related treatment factors other than in-hospital mortality [6,12]. Pines and Hollander demonstrated that emergency room crowding was associated with poor quality of care in patients with se- vere pain, whether this was manifested by a lack of treatment or delay until treatment [13]. It has also been associated with adverse cardiovas- cular outcomes other than mortality [14]. These studies suggest that overcrowding is related to problems in providing Optimal treatment to patients, and we cannot assume different performance in non-life- threatening problems and life-threatening ones. Alternate endpoints to evaluate possible beneficial effects of the 4-h rule should be set that are not necessarily linked to survival. These could include timing of an- tibiotic treatment in patients with infection, timing of surgery in pa- tients in need of acute care surgery, among others.

Mortality was associated with increased DED-TED intervals in pa- tients hospitalized in internal medicine. Several other studies using dif- ferent methodologies evaluated this interval. Singer et al. found that increasED boarding time was associated with both increased mortality and increased LOS [15]. Still, a significant increase in mortality rate (p < 0.05) only manifested itself when patients hospitalized beyond 12 h were compared to those hospitalized within 2 h. The number of pa- tients hospitalized beyond 12 h accounted for only 6% of their cohort. Reznek et al. analyzed mortality in patients hospitalized in two hospitals with similar overall bED capacity and annual admissions to the medical center evaluated in our study [16]. They stratified patients according to if these were admitted to the intensive care unit (ICU) or not. Longer boarding times were 1.2-fold higher in non-ICU patients who died in-hospital compared to survivors. A significant proportion of their pa- tients were hospitalized in ICU beds (21.3%). Assuming lower-risk pa- tients were admitted to non-ICU beds further strengthens the association between prolonged boarding time and mortality.

The finding that LOS was mildly increased in patients with an ex-

tended length of ED stay has been reported by others [4,17]. Though LOS is associated with the patients’ individual medical condition, Liu et al. reported factors other than ED length of stay that may be associ- ated with increased LOS such as age, payment classification, source of referral, specialty, and ethnic group [18]. In this study, only age and time in the ED, whether DED or TED, consistently remained significantly associated with LOS in a model analyzing their combined effect.

    1. Implications for policy and practice

Improved performance should be expected whenever a specific ac- tivity is being sequentially evaluated. However, the practice of targeting particular performance-based health measures has been criticized for distorting clinical priorities since the excess emphasis on one target means that other important aspects of care, especially those not easily measured, may be ignored [19-21]. Implementation of the four-h-rule in EDs should be associated with observations concerning important outcomes, such as decreased in-hospital mortality, decrease in LOS, re- duction in ED readmissions, increased patient satisfaction, among others. Whether these outcomes are improved remains controversial. Increased DED-TED association with mortality in the subgroup of pa- tients hospitalized in internal medicine demands attention. Different hospitals should explore which subgroups require shorter boarding times and other procedures to decrease mortality.

    1. Strengths and limitations

The results of this study rely on a large database that included 106,842 patients registered over 1 year in the ED of a public hospital open for all in the public. Since this is the only hospital in the region

providing emergency services, almost all urgent cases in this region are referred to this hospital. This setting limits selection bias. Of those regis- tered, 28,108 were hospitalized and included in the analysis. The study provides information concerning the association between different ED times (DED, TED, and DED-TED interval) and either mortality or LOS, further stratified by specialty.

The study was based on a large database that includes information

on patients registered in the ED. The database includes information concerning identifiers, demographic data, times, and disposition. Thus, our evaluation of in-hospital mortality and disposition were limited to these variables.

The study was performed in one hospital. Time in the ED may be de- fined differently in other hospitals with different treatment pathways [16]. Thus, over 4 h’ stay in one hospital in which sophisticated diagnos- tic exams are allowed in the ED will have a different impact on in- hospital mortality if these exams are not permitted in an other hospital. Though this may limit the possibility to compare one hospital with an- other, it should not restrict comparisons made between two patient groups in the same hospital defined by time spent in the ED.

We lacked information concerning the patients’ triage category, lim-

iting our possibility of adjusting for patients’ risk. One could claim that patients with a higher risk for death were attended first, and their work up in the ED was terminated early (e.g. cerebrovascular disease and acute coronary syndromes). However, such a claim undermines the justification of the 4-h rule. If higher-risk patients are indeed treated first, what is the rationale behind limiting the time a patient needs to stay in the ED? Shifting low-risk patients from one group (ED stay be- yond 4 h’) to the other group (ED stay within 4 h’) by expediting their workup will not decrease the overall mortality. Furthermore, if such a bias indeed exists, it should have affected in-hospital mortality and LOS similarly. This study shows that in-hospital mortality showed a trend to decrease with time in the ED, while LOS increased.

We also lacked information on whether a change in the specialty during the ED’s workup affected the DED, TED, and mortality. Though no association was observed between prolonged DED or TED and mor- tality, this does not rule out a possible association between changes in specialty and adverse outcomes. This issue should be further explored. Organizational differences between the different specialties might have impacted DED and/or TED dissimilarly. Differences include the number of patients attended in the ED daily by each specialty. These in- clude the number of physicians allocated to the ED from each specialty and their experience. Different specialties treat different pathologies. Thus, patients need to undergo different workup in the ED depending on their underlying pathology. Differences in capacity to hospitalize pa- tients from the ED between the specialties will impact TED. These and other organizational differences will manifest as differences in DED and TED between specialties. However, these should not impact the as- sociation of either mortality or LOS with prolonged DED and TED of pa- tients treated by one specialty, since all these patients are treated under

similar constraints.

We stratified patients according to specialty regardless if these pa- tients were eventually treated in the ICU or not. The number of ICU beds in HYMC is limited. At the time of the study, there were only 13 beds (5 general ICU, eight cardiac ICU) out of 495 hospitalization beds in the hospital. Many of the patients in need of intensive care were ei- ther admitted to the postoperative unit or to the specialty departments until an ICU bed became available. This limits our possibility to compare our study results to others. Similar to other constraints discussed above, limited ICU availability impacted equally patients with DED or TED within 4 h’ and beyond 4 h’.

In-hospital mortality rates are commonly low and very large databases are needed to evaluate whether differences in mortality rates between two patient groups are indeed significant. One could argue that our sample was underpowered to evaluate whether differ- ences observed in in-hospital mortality were significant or not.

Nevertheless, it should be taken into account that the primary assump- tion behind the 4-h rule is that prolonged stay is associated with in- creased in-hospital mortality. Even if differences in this study did not always reach significance, in-hospital mortality observed in this study in patients TED extended beyond 4 h’ was actually lower compared to patients with TED stay within 4 h’. Increasing the sample size would not have reversed this trend. The observation that in-hospital mortality was significantly lower in patients with DED beyond 4 h’, reinforces our conclusion that ED stays beyond 4 h’ does not increase in-hospital mortality.

The last limitation we wish to discuss is the exclusion of patients who were discharged from ED and were not hospitalized. One could rightfully propose that these patients add to the workload and should have been included in the analysis, rather than excluded. However, our data reveals that the proportion of patients discharged was highest in those with TED within 4 h’, followed by those with DED within 4 h’ (see Supplemental Digital Content). Since the mortality for discharged patients is unknown, including them in the denominator would have falsely elevated the survival in patients with time in the ED of 4 h’ or less. Including only patients who were hospitalized in this analysis, bet- ter served to define the association between time in the ED and mortal- ity. Thus, caution should be exercised when evaluating the association between mortality and time in the ED, whether TED or DED, when all patients are included in the denominator.

  1. Conclusions

The four-h-rule in the ED was introduced as a performance-based measure in various health systems worldwide. In this study, in- hospital mortality was not associated with time in the ED beyond 4 h’. LOS, however, was increased in this group of patients. Decreased LOS observed in patients with time in the ED within 4 h’, does not support patients’ risk as a contributing factor leading to higher trends in mortal- ity observed in this patient group. Increased in-hospital mortality asso- ciated with increasing DED-TED intervals in internal medicine patients should be further investigated and appropriate procedures placed to avoid excess mortality.

Ethics approval

This study was approved by the Hillel Yaffe Medical Center’s re- search ethics committee (REC) (reference number 0123-17-HYMC), and the study was performed in accordance with the ethical standards as laid down in the 1964 Declaration of Helsinki and its later amend- ments. Statement concerning approval by the institutional REC and pro- tocol number are included in the Methods section. This was a retrospective study with personal identifiers removed. The need for consent from individuals was waived by the institutional REC.

Conflicts of interest

The authors declare that they have no conflict of interest to declare.

Data sharing statement

The datasets generated during and/or analyzed during the current study are not publicly available as these are deposited within the hospital’s electronic file with identifiers. A copy of the datasets without identifiers will be made available from the corresponding author on rea- sonable request.

Previous presentations

Data on surgical patients was presented in the Israeli Association of General Surgery annual meeting, Kfar Blum, May 2019.

Funding

This research did not receive any specific grant from funding agen- cies in the public, commercial, or not-for-profit sectors.

Authors’ contributions

Itamar Ashkenazi and Ohad Hochman conceived the study and de- signed the trial. Itamar Ashkenazi and Lital Gefen managed the acquisi- tion of the data. Itamar Ashkenazi, Lital Gefen, and Elias Tannous analyzed the data. Itamar Ashkenazi drafted the manuscript, and all au- thors contributed substantially to its revision. All authors authorized the final version being submitted for possible publication. Itamar Ashkenazi takes responsibility for the paper as a whole. Patients, or the public, were not involved in the design, or conduct, or reporting, or dissemina- tion plans of this research.

Appendix A. Supplementary data

Supplementary data to this article can be found online at https://doi. org/10.1016/j.ajem.2021.03.049.

References

  1. Schneider EC, Riehl V, Courte-Wienecke S, Eddy DM, Sennet C. Enhancing perfor- mance measurement: NCQA’s road map for a health information framework. Na- tional committee for quality assurance. JAMA. 1999;282(12):1184-90. https://doi. org/10.1001/jama.282.12.1184.
  2. Bryndova L, Bar M, Herzig R, Mikulik R, Neumann J, Sanak D, et al. Concentrating stoke care provision in the Czech Republic: the establishment of stoke centres in 2011 has led to improved outcomes. Health Policy. 2021. S0168-8510(21)00028-2. Epub 2021 Jan 30. https://doi.org/10.1016/j.healthpol.2021.01.011.
  3. Peled R, Porath A, Wilf-Miron R. Improving the quality of primary care by allocating performance-based targets, in a diverse insured population. BMC Health Serv Res. 2016;16(1):668. https://doi.org/10.1186/s12913-016-1920-6.
  4. Staib A, Sullivan C, Griffin B, Bell A, Scott I. Report on the 4-h rule and the National Emergency Access Target (NEAT) in Australia: time to review. Aust Health Rev. 2016;40(3):319-23. https://doi.org/10.1071/AH15071.
  5. Weber EJ, Mason S, Freeman JV, Coster J. Implications of England’s four-hour target for quality of care and resource use in the emergency department. Ann Emerg Med. 2012;60(6):699-706. https://doi.org/10.1016/j.annemergmed.2012.08.009.
  6. Plunkett PK, Byrne DG, Breslin T, Bennet K, Silke B. Increasing wait time predict in- creasing mortality for emergency medical admissions. Eur J Emerg Med. 2011;18

(4):192-6. https://doi.org/10.1097/MEJ.0b013e328344917e.

  1. Geelhoed GC, de Klerk NH. Emergency department overcrowding, mortality and the 4-hour rule in Western Australia. Med J Aust. 2012;196:122-6. https://doi.org/10. 5694/mja11.11159.
  2. Grafstein E, Bullard MJ, Waren D, Unger B, CTAS National Working Group. Revision of the Canadian Emergency Department Information System (CEDIS) presenting complaint list version 1.1. CJEM. 2008;10(2):151-73. https://doi.org/10.1017/ s1481803500009878.
  3. Bullard MJ, Chan T, Brayman C, Warren D, Musgrave E, Unger B, et al. Revisions to the Canadian Emergency Department Triage and Acuity Scale (CTAS) guidelines. CJEM. 2014;16(6):485-9.
  4. Forero R, Man N, Ngo H, Mountain D, Mohsin M, Fatovich D, et al. Impact of the four- hour National Emergency Access Target on 30 day mortality, access block and chronic emergency department overcrowding in Australian emergency depart- ments. Emerg Med Australas. 2019;31(1):58-66. https://doi.org/10.1111/1742- 6723.13151.
  5. Haklai Z, Meron J, Appelbaum Y, Aburbeh M, Shlichkov G. Emergency room visits: Summary 2014. State of Israel, Jerusalem: Health Information Division, Ministry of Health; 2014.
  6. McKenna P, Heslin SM, Viccellio P, Mallon WK, Hernandez C, Morley EJ. Emergency department and Hospital crowding: causes, consequences, and cures. Clin Exp Emerg Med. 2019;6(3):189-95. https://doi.org/10.15441/ceem.18.022.
  7. Pines JM, Hollander JE. Emergency department crowding is associated with poor care for patients with severe pain. Annu Emerg Med. 2008;51(1):1-5. https://doi. org/10.1016/j.annemergmed.2007.07.008.
  8. Pines JM, Pollack Jr CV, Diercks DB, Chang AM, Shofer FS, Hollander JE. The associa- tion between emergency department crowding and adverse cardiovascular out- comes in patients with chest pain. Acad Emerg Med. 2009;16(7):617-25. https:// doi.org/10.1111/j.1553-2712.2009.00456.x.
  9. Singer AJ, Thode Jr HC, Viccellio P, Pines JM. The association between length of emer- gency department boarding and mortality. Acad Emerg Med. 2011;18(12):1324-9. https://doi.org/10.1111/j.1553-2712.2011.01236.x.
  10. Reznek MA, Upatising B, Kennedy SJ, Durham NT, Forster RM, Michael SS. Mortality associated with emergency department boarding exposure. Are there differences between patients admitted to ICU or non-ICU settings? Med Care. 2018;56(5): 436-40. https://doi.org/10.1097/MLR.0000000000000902.
  11. Derose SF, Gabayan GZ, Chiu VY, Yiu SC, Sun BC. Emergency department crowding predicts admission length-of-stay but not mortality in a large health system. Med Care. 2014;52(7):602-11. https://doi.org/10.1097/MLR.0000000000000141.
  12. Liu X, Phillips M, Codde J. Factors influencing patients’ length of stay. Aust Health Rev. 2001;24(2):63-70. https://doi.org/10.1071/ah010063.
  13. Nolte E, McKee M, Wait S. Describing and evaluating health systems. In: Bowling A, Ebrahim S, editors. Handbook of health research methods. Investigation, measure- ment and analysis. McGraw-Hil, Glasgow, UK: Open University Press; 2005.

p. 12-43.

  1. Orr J. The good, the bad, and the four hour target. BMJ. 2008;337:a195. https://doi. org/10.1136/bmj.a195.
  2. Hughes G. The four hour target; problems ahead. Emerg Med J. 2006;23:2. https:// doi.org/10.1136/emj.2005.031948.

Leave a Reply

Your email address will not be published. Required fields are marked *