Presented at the Society for Academic Emergency Medicine Annual Meeting, New Orleans, LA, May 2009.
The Association between Emergency Department Crowding and Analgesia Administration in Acute Abdominal Pain Patients
Article first published online: 22 JUN 2009
© 2009 by the Society for Academic Emergency Medicine
Academic Emergency Medicine
Volume 16, Issue 7, pages 603–608, July 2009
How to Cite
Mills, A. M., Shofer, F. S., Chen, E. H., Hollander, J. E. and Pines, J. M. (2009), The Association between Emergency Department Crowding and Analgesia Administration in Acute Abdominal Pain Patients. Academic Emergency Medicine, 16: 603–608. doi: 10.1111/j.1553-2712.2009.00441.x
- Issue published online: 1 JUL 2009
- Article first published online: 22 JUN 2009
- Received November 7, 2008; revisions received February 25 and March 13, 2009; accepted March 23, 2009.
- abdominal pain;
- quality of care;
- emergency department
Objectives: The authors assessed the effect of emergency department (ED) crowding on the nontreatment and delay in treatment for analgesia in patients who had acute abdominal pain.
Methods: This was a secondary analysis of prospectively enrolled nonpregnant adult patients presenting to an urban teaching ED with abdominal pain during a 9-month period. Each patient had four validated crowding measures assigned at triage. Main outcomes were the administration of and delays in time to analgesia. A delay was defined as waiting more than 1 hour for analgesia. Relative risk (RR) regression was used to test the effects of crowding on outcomes.
Results: A total of 976 abdominal pain patients (mean [±standard deviation] age = 41 [±16.6] years; 65% female, 62% black) were enrolled, of whom 649 (67%) received any analgesia. Of those treated, 457 (70%) experienced a delay in analgesia from triage, and 320 (49%) experienced a delay in analgesia after room placement. After adjusting for possible confounders of the ED administration of analgesia (age, sex, race, triage class, severe pain, final diagnosis of either abdominal pain not otherwise specified or gastroenteritis), increasing delays in time to analgesia from triage were independently associated with all four crowding measures, comparing the lowest to the highest quartile of crowding (total patient-care hours RR = 1.54, 95% confidence interval [CI] = 1.32 to 1.80; occupancy rate RR = 1.64, 95% CI = 1.42 to 1.91; inpatient number RR = 1.57, 95% CI = 1.36 to 1.81; and waiting room number RR = 1.53, 95% CI = 1.31 to 1.77). Crowding measures were not associated with the failure to treat with analgesia.
Conclusions: Emergency department crowding is associated with delays in analgesic treatment from the time of triage in patients presenting with acute abdominal pain.
Acute abdominal pain is the single most common reason for a visit to the emergency department (ED) and accounted for almost 8 million U.S. visits in 2005.1 While historically the use of analgesia for symptom control in ED patients with abdominal pain had been thought to mask signs of peritonitis and potentially delay care, several studies have shown that this practice does not lead to adverse outcomes.2–9 Therefore, the timely use of analgesia in patients who request symptom control has become a standard part of emergency care. Prolonged time to analgesia in patients with acute painful conditions only produces more suffering in those desiring pain medication.
Oligoanalgesia, or the underuse of pain control, in the ED is common and has been well documented.10–16 Various patient-level factors have been associated with oligoanalgesia including age, sex, ethnicity, triage acuity, and diagnosis.11,17–29 In addition, literature has recently emerged documenting the association between ED crowding, a system-level factor, and the quality of pain control.30–32 Two of these prior studies demonstrated a link between crowding and worse pain control in general ED populations31,32 and another in older adults with hip fracture.30 No studies, to our knowledge, have studied the association between ED crowding and analgesia in patients with abdominal pain.
We sought to evaluate the association between ED crowding and the administration of analgesia in adult ED patients with acute abdominal pain. We hypothesized that at times of higher ED crowding, patients would be less likely to receive analgesia and more likely to experience a delay in time to analgesia.
We performed a secondary analysis of a prospectively collected database of adult patients presenting to the ED from April 5, 2004, to January 4, 2005, with acute abdominal pain of less than 72 hours’ duration.21 The original study was conducted for the detection of new biomarkers for abdominal pain using serum sampling. Institutional review board approval was obtained at the University of Pennsylvania. Written informed consent was obtained from all participants.
Study Setting and Population
The study was conducted at an urban tertiary care ED with an annual census of 55,000 visits. The ED has 25 treatment rooms, 15 hallway treatment spaces, a separate eight-bed fast-track, and an attached three-bed trauma bay. Patients were eligible for inclusion if they presented with acute abdominal pain of less than 72 hours’ duration. Patients who were pregnant or sustained abdominal surgery or abdominal trauma within the prior 7 days were excluded.
Trained research assistants enrolled patients from 8 am to midnight, 7 days a week, and recorded patient demographics, triage pain score (1–10), and triage class (4-point scale from emergent  to nonurgent ), on a standardized collection form. The type and time of analgesic administration were obtained directly from the electronic medical record. The ED uses a computerized charting and order entry system called EMTRAC (University of Pennsylvania).
All clinical decisions, including the administration of analgesics, were determined by the treating physician. A standing order is available to the triage nurses for the administration of acetaminophen without a specific physician order; however, this is typically only performed for patients with fever. A physician order is required for all other analgesics. It is possible a nurse may have requested analgesia and an order placed by a physician prior to his or her examination of the patient. This information was not available in our database. Time to analgesia was defined as time from placement into an ED treatment room or hallway space to time of analgesia administration as documented by electronic time stamp by the nurse administering the medication. Pain medications were defined as any oral (e.g., acetaminophen, ibuprofen, acetaminophen with oxycodone, aluminum and magnesium hydroxide), intramuscular, or intravenous (e.g., ketorolac, morphine, hydromorphone, fentanyl) medication used for analgesia. We chose to include aluminum and magnesium hydroxide as antacids are commonly given for symptom control for acute abdominal pain. Reassessments for pain were not collected.
EMTRAC was used to assign the following overall ED crowding indicators at triage: ED occupancy rate (percentage of overall beds filled including hallway spaces), total patient-care hours (arithmetic sum of the hours of all patients in the ED, excluding trauma and fast-track patients), number of patients in the waiting room, and number of inpatients boarding in the ED (where a bed request had been entered). These crowding variables have been shown to be associated with quality of care in other studies.31,33,34
The main outcomes of this study were whether patients received analgesia during their visit and, if so, whether there were delays in the administration of this medication. In assessing delays to analgesia, we used the outcomes of a delay of greater than 1 hour from triage time to time of analgesia and from room time to time of analgesia. The outcome of a 1-hour delay to analgesia has been used in several prior studies.20,30,31 We also chose the 1-hour cutoff to use binary regression techniques (relative risk [RR] regression), which allows for a more interpretable and comparable risk ratio, compared with using a log transformation of the time.
Descriptive data are presented as means ± standard deviation (±SD) or frequencies with percentages. Times from triage or room placement to analgesia are presented as medians with interquartile ranges (IQR). Each of the four crowding measures was divided into quartiles, and median times to analgesia from triage and room placement were compared among quartiles using an analysis of variance on the ranks to test for either a linear or a quadratic effect. To assess the effect of crowding variables on 1-hour delays in time from triage or room placement to analgesia, while adjusting for potential confounders (age, sex, race, triage class, severe pain [pain score 9–10], final diagnosis of abdominal pain not otherwise specified or gastroenteritis), RR regression using the Gaussian estimating equation was performed.35 The Cochran-Armitage trend test was used to assess trend. Data for these analyses are presented as RR with 95% confidence intervals (CI). All analyses were performed using SAS statistical software (Version 9.1, SAS Institute, Cary, NC). To adjust for multiple statistical tests performed on the same data, the Bonferroni correction was used. A probability of <0.006 was considered statistically significant.
Of 1,000 patients enrolled with acute abdominal pain, 24 patients were excluded for lack of documented pain score leaving 976 patients (98%) for analysis. Patient characteristics are summarized in Table 1. Patients were primarily female (65%), black (62%), and young (mean ±SD age = 41 ± 16.6 years). Median waiting room time was 24 minutes (IQR = 9 to 71 minutes). Analgesia was administered to 649 patients (67%), of whom 436 (45%) received opiates. The median time from triage to analgesia was 107 minutes (IQR = 52 to 190 minutes), and median time from room placement to analgesia was 61 minutes (IQR = 28 to 122 minutes). Of the 649 patients receiving analgesia, 320 (49%) waited longer than 1 hour from room placement until receiving their first dose of analgesia.
|Age (yr), mean ± SD||41 (±16.6)|
|1 (emergent)||46 (5)|
|4 (nonurgent)||7 (1)|
|Pain score (median)||7 (5–9)|
|Received any analgesia||649 (67)|
|First pain medication given|
|Opioid analgesia||436 (45)|
|Aluminum/magnesium hydroxide||47 (5)|
|Other nonopioid analgesia||166 (17)|
|ED crowding measures (median)|
|Total patient-care hours||86 (55–127)|
|Occupancy rate, %||63 (48–75)|
|Number of waiting patients||6 (3–10)|
|Number of inpatients||11 (8–15)|
|Median waiting times (minutes)|
|Triage to room||24 (9–71)|
|Room to analgesia||61 (28–122)|
|Triage to analgesia||107 (52–190)|
ED crowding measures were not associated with the failure to treat with analgesia. Administration of analgesia remained the same regardless of crowding factor quartile (Table 2A). However, with regard to timeliness in receiving analgesia, as crowding increased, time from triage to analgesia increased as well across all crowding measures (test for quadratic fit, p < 0.0001 for all measures; Figure 1). This relationship was not found for time from room placement to analgesia but a threshold effect appeared to be reached by the second quartile of crowding. This relationship persisted when examining 1-hour delays for time from triage to analgesia (Cochran-Armitage trend test p < 0.0001 for all crowding measures; Table 2B), and the same plateau effect was seen for time from room placement to receiving analgesia (Table 2C).
|Patient-hours||Occupancy Rate||Inpatient Number||Waiting Patient Number|
|A. Received Analgesia|
|Q1*||164 (66.9)||172 (64.2)||180 (64.3)||182 (67.4)|
|Q2||160 (64.8)||171 (71.6)||133 (63.3)||180 (64.8)|
|Q3||158 (65.8)||154 (66.1)||189 (73.5)||142 (66.7)|
|Q4||167 (68.4)||152 (64.4)||147 (64.2)||145 (67.4)|
|C-A trend test p value||0.73||0.76||0.42||0.88|
|B. Received Analgesia in <1 Hour From Triage|
|Q1||77 (47.0)||81 (47.1)||82 (45.6)||84 (46.2)|
|Q2||43 (26.9)||49 (28.7)||42 (31.6)||46 (25.6)|
|Q3||37 (23.4)||36 (23.4)||41 (21.7)||30 (21.1)|
|Q4||35 (21.0)||26 (17.1)||27 (18.4)||32 (22.1)|
|C-A trend test p value||<0.0001||<0.0001||<0.0001||<0.0001|
|C. Received Analgesia in <1 Hour From Room Placement|
|Q1||100 (61.0)||100 (58.1)||105 (58.3)||112 (61.5)|
|Q2||70 (43.8)||80 (46.8)||68 (51.1)||77 (42.8)|
|Q3||80 (50.6)||81 (52.6)||85 (45.0)||66 (46.5)|
|Q4||79 (47.3)||68 (44.7)||74 (48.3)||74 (51.0)|
|C-A trend test p value||0.02||0.03||0.02||0.12|
|Waiting patient number||0–3||4–6||7–10||11–43|
After adjusting for possible confounders of the ED administration of analgesia (age, sex, race, triage class, severe pain diagnosis, of abdominal pain not otherwise specified, or gastroenteritis), delays in analgesia from triage were independently associated with all four crowding measures. Comparing the lowest quartile of crowding to the highest quartile, there were increases in delays from triage to analgesia in all crowding measures (Table 3). When examining room placement to analgesia after adjustment for confounders, all four of the crowding measures increased time to analgesia, but these increases reached a threshold by the second quartile (Table 3).
|Patient-hours||Occupancy Rate||Inpatient Number||Waiting Patient Number|
|RR||95% CI||p value||RR||95% CI||p value||RR||95% CI||p value||RR||95% CI||p value|
|>1 Hour From Room to First Pain Medication|
|>1 Hour From Triage to First Pain Medication|
We found that higher crowding levels in the ED were independently associated with delays in analgesia in adults with acute abdominal pain from the time of triage. However, ED crowding did not appear to affect whether they received analgesia, and there were no significant trends in a delay of 1 hour after room placement. These findings are different from another study in the same ED in a broad ED population with severe pain (pain score 9–10) where we found that crowding independently predicted both delays and nontreatment.31 This indicates that the relative resource scarcity created by crowding likely affects some populations more than others when it comes to ED-based symptom control. In abdominal pain, physicians’ decisions to use any analgesia or to delay analgesia after room placement may be less affected by crowding because of concern for more serious causes of pain. In addition, this cohort of patients with abdominal pain was younger, which may have contributed to their ability to advocate for analgesia during times of crowding compared to other populations such as older adults with hip fracture, who may be less demanding despite similar levels of suffering.30
In this study, we found a nonlinear association between crowding and delays in triage to analgesia for several crowding measures (Figure 1). That is, once crowding achieves a certain level (second quartile of patient-hours, third quartile of inpatient number, and second quartile of waiting room number), adults with abdominal pain were more likely to experience a delay once they arrived to a room. This ceiling effect may represent thresholds in the patient-to-nurse or patient-to-resident physician ratio, which increases at low levels of crowding and tends to level off when maximum capacity is reached or, because crowding did not impact the delay once the patients arrived in their rooms, these may reflect time spent in the waiting room. Because of the linear association between room to analgesia in our prior study of a broader population, this may also indicate that abdominal pain patients are prioritized over other patients who are at greater risk for delays during high levels of crowding.31 Occupancy displayed a relatively linear relationship with delays indicating that it may be better able to discriminate quality of emergency care at all levels, providing further evidence that occupancy may be the optimal operational measure of ED crowding.36 By contrast, the effect of patient-hours may saturate at lower levels because it is a measure of overall ED workload—particularly of nursing workload, as nurses care for not only a high number of boarders, but care for them for extended periods of time. Prior studies have reported data with a threshold value of greater than or equal to the 50th percentile; however, these analyses did not investigate whether the association was linear.30,32
This study joins the growing body of literature showing worse quality of care during times of ED crowding. Crowding has also been shown to be associated with delays in antibiotics for pneumonia, higher death rates in a broad cohort of intensive care unit patients, and higher rates of ventilator-associated pneumonia in intubated trauma patients.34,37–39 The main addition to the body of knowledge that this study offers is that crowding has variable effects on different cohorts of patients for the same outcome. This means that patients with particular complaints or disease processes may be protected from the adverse effects of crowding on the quality of pain control, while others may be dramatically affected. Because the Joint Commission has made the treatment of pain a priority, future studies should focus on identifying specific groups at risk for oligoanalgesia during times of crowding to inform targeted interventions aimed at mitigating the impact of crowding on care quality.
Only the initial pain score from triage was documented, and reassessment of pain was not recorded for this study. As pain may vary over time, we did not account for whether this evolution of pain influenced the administration of analgesia or how crowding affected reassessments. We also did not measure patient desire for analgesia. This may account for the relatively low proportion of patients who actually received analgesia in this cohort. In addition, other agents potentially used for analgesia in abdominal pain patients were not assessed (e.g., H2-receptor antagonists, proton pump inhibitors), which may also contribute to the relatively low proportion who received analgesia. Finally, because this study was performed at a single academic institution, our results may not be applicable to other settings.
We found that ED crowding was associated with delays in analgesic treatment from the time of triage in patients presenting with acute abdominal pain in our ED. Efforts to reduce ED crowding may expedite pain management in adults with abdominal pain.
- 1National Hospital Ambulatory Medical Care Survey: 2005 emergency department summary. Adv Data. 2007; (386):1–32., , .
- 22Determinants of variation in analgesic and opioid prescribing practice in an emergency department. J Opioid Manag. 2006; 2:335–40., , , , , .
- 35Relative risk regression in medical research: models, contrasts, estimators, and algorithms. UW Biostatistics Working Paper Series Working Paper 293. July 19 2006., , .