Brief Report

ED crowding is associated with an increased time to pneumonia treatment^?

Rishi Sikka MD ^a,?, Supriya Mehta PhD, MHS ^b, Carleen Kaucky RN^a, Erik B. Kulstad MD, MS ^a

^aDepartment of Emergency Medicine, Advocate Christ Medical Center, Oak Lawn, IL 60453, USA

^bUniversity of Illinois Chicago School of Public Health, Chicago, IL 60612, USA

Received 1 May 2009; revised 9 June 2009; accepted 10 June 2009

Abstract

Objective: We measured the correlation between emergency department (ED) occupancy rate and time to antibiotic administration for patients with pneumonia treated in a community hospital setting.

Methods: We reviewed quality improvement data on patients treated for pneumonia in our ED and admitted over a 5-month period. The outcomes were timeliness of antibiotic therapy (within 4 hours of arrival) and overall time to antibiotic administration. Emergency department crowding was measured as the ED occupancy rate. We calculated (1) the spearman correlation between occupancy rate at time of patient presentation and the time to antibiotic administration, (2) the odds ratio of receiving antibiotics within 4 hours with increasing ED occupancy, and (3) the ability of the occupancy rate to predict failure of achieving the 4-hour goal with the receiver operating characteristic curve.

Results: A total of 334 patients were treated over the study period, of which 262 had complete data available. Occupancy rate ranged from 20% to 245%, and median was 137%. Eighty-one percent received antibiotics within 4 hours; the median time was 150 minutes. time to antibiotics showed a positive correlation with occupancy rate (Spearman ? = 0.17, P = .008). An increasing ED occupancy rate was associated with decreased odds of receiving antibiotics within 4 hours (odds ratio, 0.31; 95% confidence interval, 0.13-0.75). Receiver operating characteristic curve area was 0.62 (95% confidence interval, 0.54-0.70; P = .009).

Conclusion: Emergency department occupancy rate was associated with increased time to antibiotic treatment for patients admitted with pneumonia. Occupancy rate had fair success in predicting failure of treatment within 4 hours.

(C) 2010

Introduction

Background

Presented at the American College of Emergency Physicians Research Forum held in Chicago, IL, October 2008.

^? No outside funding or support was received.

* Corresponding author.

E-mail address: [email protected] (R. Sikka).

The Institute of Medicine has identified a growing crisis in the provision of care within the emergency department (ED). This national problem has its fundamental roots in hospital and ED crowding [1]. Emergency department crowding has been associated with an increased risk of

0735-6757/$ - see front matter (C) 2010 doi:10.1016/j.ajem.2009.06.023

death or disability [2-5], delays in treatment of patients with suspected acute myocardial infarction [6], decreased performance on pneumonia quality-of-care measures [7], transport delays, Ambulance diversion, and patients leaving without being seen by a physician [8]. These quality-of- care and safety issues also directly impact the public perception of care and the decisions of purchasers. publicly reported measures, specifically with respect to treatment of ED patients admitted with pneumonia, have been shown to be affected by ED crowding. For patients with community- acquired pneumonia, a recent analysis performed at a university tertiary care hospital found that increased ED volume was associated with a decreased likelihood of timely antibiotic therapy [9], whereas a study at an Urban academic ED found a substantial impact of ED crowding on time to antibiotics [10]. It is uncertain if this relationship holds in community hospitals, where opera- tional factors may vary and most admitting physicians are in private practice.

The demonstration that ED crowding affects the time- liness of antibiotic treatment for patients with pneumonia in a variety of clinical settings would validate the scope and breadth of the problem of ED crowding. Given the large number of US patients who obtain care at community hospitals, such demonstration in a community hospital ED would provide additional impetus for improvement and change.

Objective

We sought to measure the correlation between the ED occupancy rate and the time to antibiotic administration for patients with pneumonia in a community hospital ED. We hypothesized that increased crowding would show a positive correlation with time to treatment.

Methods

Study design, setting, and selection of participants

We performed a retrospective review of all adult patients older than 18 years treated for pneumonia over a 5-month period from October 1, 2007, through February 29, 2008. The study setting was our tertiary care suburban community hospital that has approximately 85,000 ED visits annually and 700 inpatient beds. Although the hospital has both internal medicine and family practice residency training programs, most patients are admitted to, and cared for by, private practice physicians without resident coverage. All patients with a primary diagnosis of pneumonia as entered in the electronic ED medical record were included. The hospital institutional review board approved the study.

Methods of measurement

Occupancy rates, defined as the number of patients in the ED divided by number of licensed ED beds, were obtained 24 hours a day using a VBScript (Microsoft) running on a remote server that is programmed to sample and archive our electronic patient tracking board (Picis ED PulseCheck, Wakefield, MA) at 20-minute intervals. Although our ED is licensed for 50 beds, because of increasED patient volume and a lack of inpatient beds to place admitted patients, we frequently use hallway space for treatment of patients. Consequently, our occupancy rate is often above 100%.

The major outcome measure was timely antibiotic therapy defined as initial antibiotic administration within 4 hours of ED arrival [11]. A secondary, continuous outcome measure was the time to initial antibiotic administration. Charts were identified for inclusion by search of the diagnosis assigned and entered into the electronic medical record by the treating ED physician. Charts were reviewed by a trained abstractor as part of our ongoing quality improvement initiatives. The abstractor performs these duties without knowledge of any further study hypotheses and was blinded to the current study hypothesis. Because this is an ongoing collection of data for quality improvement initiatives, no formal regular checks are performed on abstracted data.

Primary data analysis

We assessed the relation between occupancy rate at the time closest to (within 20 minutes of) patient presentation and the time to antibiotic administration by calculating the Spearman rank correlation coefficient. We performed logistic regression analysis to calculate the odds ratio (OR) of receiving timely antibiotic therapy using ED occupancy rate as the explanatory variable and performed a Linear regression analysis on the time to antibiotics as a continuous outcome variable (log-transformed to correct for non-normality) with ED occupancy rate as the independent variable. We additionally included the particular shift in which the patient presented as a covariate (7 AM-3 PM, 3-11 PM, or 11 PM-7 AM) to account for the possible effects of presentation time. We assessed the model goodness of fit with the Hosmer and Lemeshow test. We determined the ability of occupancy rate to predict failure to administer timely antibiotic therapy using the area under the receiver operating characteristic curve. Data were entered into an Excel spreadsheet (Microsoft), and analysis was performed with SPSS version 15.0 (SPSS Inc, Chicago, IL).

Results

A total of 334 patients were treated for pneumonia over the study period and were therefore eligible for this analysis.

Out of these 334 patients, 262 had complete data available. The most commonly missing data element was the occupancy rate at time of admission, which was missing for 20 days in October and 20 days in November because of a server malfunction.

Out of 262 patients, 214 (81.7%; 95% confidence interval [CI], 76.5%-86.2%) received antibiotics within 4 hours, whereas 48 patients did not. This percentage did not differ from that for the patients for whom occupancy data were missing. The median time to antibiotic administration was 150 minutes (95% CI, 133-168 minutes). Occupancy rate ranged from 20% to 245% throughout the study period, with a mean and median of 136% and 137%, respectively.

The time to antibiotics showed a positive correlation with occupancy rate (Spearman ? = 0.17, P = .008). Fig. 1 shows a scatterplot of the time to antibiotics for each patient as a function of the occupancy rate at the time of arrival. An ED occupancy rate greater than the median was associated with a decreased OR of receiving antibiotics within 4 hours (OR, 0.31; 95% CI, 0.13-0.75), with the Hosmer and Lemeshow significance level of 0.37 showing no significant difference between observed and predicted values in the model. Inclusion of presentation time, expressed as either first, second, or third shift, in the logistic regression model showed a nonsignificant change (OR, 0.996; 95% CI, 0.63-1.6), which did not affect the influence of occupancy rate. Linear regression, using log- transformed time to antibiotics as a continuous dependent variable, showed occupancy rate to have a significant effect (standardized ? = 0.17; 95% CI, 0.09-0.50), which was likewise not affected by inclusion of presentation time (standardized ? = -0.05; 95% CI, -0.15 to 0.06). The area under the receiver operating characteristic curve (Fig. 2), constructed using occupancy rate as the test variable and administration of antibiotics in under 4 hours as the state variable, was 0.62 (95% CI, 0.54-0.70; P = .009).

Fig. 1 Scatterplot of the time to antibiotic administration for each patient as a function of the occupancy rate at the time of arrival.

Fig. 2 Receiver operating characteristic curve, constructed using occupancy rate as the test variable, and administration of antibiotics in less than 4 hours as the state variable.

To evaluate potential effects from missing data, we performed an analysis using imputation of missing occu- pancy rate data. Using the missing value analysis and multiple imputation function of SPSS version 17, we imputed missing values of occupancy rate using a monotone imputation method and repeated our logistic regression analysis. The results yielded a similar decreased OR of receiving antibiotics within 4 hours (OR, 0.38; P = .03), with no significant effect seen from the shift of patient presenta- tion (OR, 1.05; P = .82).

Discussion

This analysis identified a positive association between time to antibiotic treatment for patients admitted with pneumonia and ED crowding, as measured by ED occupancy rate. An ED occupancy rate greater than the median was associated with an increased likelihood of failure to administer antibiotics within 4 hours of arrival.

The results of this study, within a community hospital, confirm findings from urban, academic medical centers. In one study of patients admitted with community-acquired pneumonia at a single, urban academic ED, the crowding measures of waiting room number and recent ED length of stay for admitted patients were associated with failure to receive antibiotics within 4 hours [10]. In a study within a university tertiary care hospital of patients also admitted with community-acquired pneumonia, total ED volume and total number of admitted patients were associated with failure to receive antibiotics within a timely fashion [9].

Our analysis builds on these previous studies by demonstrating the effect of ED crowding within the

community hospital setting and with patients admitted with all etiologies of pneumonia, not just community- acquired disease. Because we found that crowding correlates with failure to receive antibiotics within a timely manner, our results also help to establish the content validity of ED occupancy rate as a measure of ED crowding. By also examining antibiotic timing data as a continuous variable, our results overcome the limitation of dichotomized times to antibiotics cited by Fee et al [9] and highlight the association between crowding and treatment delays.

Limitations

This is a retrospective review of patients with a primary admitting diagnosis of pneumonia. Individuals with a secondary admitting diagnosis of pneumonia could not be captured in this analysis and are not included because secondary diagnoses are not entered into the electronic registry. The time to treatment for patients with secondary admitting diagnoses of pneumonia is unknown.

We did not attempt to control for patient characteristics in our analysis. This analysis did not examine the impact of timeliness of antibiotic therapy by day of week or time of year. Such variables, in addition to ED occupancy rate, may have a correlation or interaction effect with the time to initial antibiotic treatment. We examined delayed antibiotic treatment in association with ED occupancy greater than the median. Although the median occupancy in our institution may differ from that of other institutions, the measure has content validity in that it represents the usual occupancy status in our ED: more than half the time, we have more patients than we technically have beds for.

These results represent a single institution and may therefore not generalize to other settings. Although abstraction was not performed in a blinded fashion, both time of antibiotic administration and occupancy are objectively documented in our electronic charting system, eliminating potential bias in the assessment of these values.

Conclusions

In summary, our analysis identified a significant correla- tion between ED occupancy rate and failure to receive antibiotics within 4 hours of patient arrival. These results add to the growing evidence of the negative impact of ED crowding on quality of care. Future research should examine specific process steps that are affected by increased ED occupancy so barriers may be anticipated and removed.

References

1. Committee on the Future of Emergency Care in the United States Health System. Hospital-based emergency care: at the breaking point. Washington, DC: Institute of Medicine (U.S.) Committee on the Future of Emergency Care in the United States Health System; 2006.
2. Sprivulis PC, Da Silva JA, Jacobs IG, et al. The association between Hospital overcrowding and mortality among patients admitted via Western Australian emergency departments. Med J Aust 2006;184(5):208-12.
3. Richardson DB. Increase in patient mortality at 10 days associated with emergency department overcrowding. Med J Aust 2006;184(5):213-6.
4. Derlet RW, Richards JR. Emergency department overcrowding in Florida, New York, and Texas. South Med J 2002;95(8):846-9.
5. Miro O, Antonio MT, Jimenez S, et al. Decreased health care quality associated with emergency department overcrowding. Eur J Emerg Med 1999;6(2):105-7.
6. Schull MJ, Vermeulen M, Slaughter G, et al. Emergency department crowding and thrombolysis delays in acute myocardial infarction. Ann Emerg Med 2004;44(6):577-85.
7. Pines JM, Hollander JE, Localio AR, et al. The association between emergency department crowding and hospital performance on antibiotic timing for pneumonia and percutaneous intervention for myocardial infarction. Acad Emerg Med 2006;13(8):873-8.
8. Hoot NR, Aronsky D. Systematic review of emergency department crowding: causes, effects, and solutions. Ann Emerg Med 2008;52(2): 126-36.
9. Fee C, Weber EJ, Maak CA, et al. Effect of emergency department crowding on time to antibiotics in patients admitted with community- acquired pneumonia. Ann Emerg Med 2007;50(5):501-509, 509 e501.
10. Pines JM, Localio AR, Hollander JE, et al. The impact of emergency department crowding measures on time to antibiotics for patients with community-acquired pneumonia. Ann Emerg Med 2007;50(5):510-6.
11. The Joint Commission. Current specification manual for national Hospital quality measures. http://www.jointcommission.org/Performance Measurement/PerformanceMeasurement/Current+NHQM+Manual.htm [Accessed: March 11, 2009].