Emergency Department Crowding Is Associated With Poor Care for Patients With Severe Pain

Introduction 

Acute pain is one of the most common reasons for seeking emergency care. According to the 2004 National Hospital Ambulatory Medical Care Survey data, approximately 32% of emergency department (ED) patients presented with either moderate or severe pain.1 However, oligoanalgesia, which is defined as the undertreatment of pain, is common in EDs.2, 3, 4, 5, 6 Factors associated with oligoanalgesia include older age, ethnicity, provider perceptions of a patient’s pain, concern about drug-seeking behavior, and reluctance to mask symptoms to ensure an accurate diagnosis.7, 8, 9, 10, 11, 12, 13, 14, 15, 16 In the era of ED crowding,17, 18, 19 another potential reason for oligoanalgesia may be that providers are too busy to appropriately assess and treat ED patients with painful conditions. A recent study demonstrated that older adults with hip fracture are less likely to receive analgesia and more likely to experience delays in analgesia at high levels of ED census.20 However, to our knowledge, there have been no studies reporting the impact of ED crowding on a general sample of ED patients with severe pain.

We sought to study the impact of ED crowding on both treatment and delays in treatment in a broad cohort of ED patients who presented at triage with a complaint of severe pain. We hypothesized that when the ED is crowded, patients will experience both failure to receive treatment and delays in treatment for pain.

Results 

Throughout the study period, 13,758 patients reported severe pain at ED triage. The median waiting room time was 28 minutes (IQR 8 to 72 minutes). A total of 6,746 (49%) patients received pain medication while in the ED. Of those, 4,052 (60%) were treated with narcotic analgesia. Of the patients who were treated, the median time from triage to administration of first pain medication was 74 minutes (IQR 34 to 142 minutes) and median time from room placement until first pain medication was 23 minutes (IRQ 8 to 49 minutes). A total of 3,965 (59%) treated patients had to wait for at least an hour from triage for medications, and 1,319 (20%) patients had to wait for more than 1 hour from room placement until administration of their first dose of pain medication. Table 1 provides demographic characteristics, ED crowding levels, and characteristics of pain medication.

Table 1. Patient demographics and ED crowding measures (n=13,758).

	Patient demographics	Crowding measures
	Age, y, mean (SD)	39(16)
	Female sex	64%
	Race, %
	Black	73
	White	18
	Asian	1
	Other	8
	Triage level, %
	1 (Most severe)	4
	2	40
	3	38
	4	17
	ED crowding measures
	Median total patient-care hours	120 (IQR 82–164)
	Median occupancy rate, %	65 (IQR 50–78)
	Waiting room patients, No.	8 (IQR 4–13)

In multivariable analysis, increasing levels of ED crowding measures were significantly associated with failure to treat or delayed treatment, after adjustment for patient-level confounding. Table 2 reports the results of multivariable analysis by the 3 primary outcomes.

Table 2. Adjusted models using ED crowding measures, adjusting for patient-level factors to predict primary outcomes (n=13,758).

	Adjusted models	Adjusted OR (95% CI)
	No analgesia in the ED
	Waiting room number⁎	1.03(1.02–1.03)
	Occupancy rate†	1.01(0.99–1.04)
	Age ≥65 y	1.61(1.40–1.86)
	Female sex	1.30(1.20–1.41)
	White race	0.89(0.80–0.98)
	More severe triage level(1 or 2)	1.10(1.02–1.18)
	Delay >1 h in analgesia from triage
	Waiting room number⁎	1.05(1.04–1.06)
	Occupancy rate†	1.18(1.15–1.21)
	Age ≥65 y	1.47(1.22–1.77)
	Female sex	1.37(1.25–1.50)
	White race	0.84(0.75–0.94)
	More severe triage level(1 or 2)	0.57(0.52–0.62)
	Delay >1 h in analgesia placement in room
	Waiting room number⁎	1.02(1.01–1.03)
	Occupancy rate†	1.06(1.04–1.08)
	Age ≥65 y	1.61(1.39–1.87)
	Female sex	1.26(1.16–1.36)
	White race	0.90(0.81–0.99)
	More severe triage level(1 or 2)	0.89(0.83–0.95)

⁎ For an increase in waiting room number of 1 patient. † For an increase in occupancy rate of 10%.

Limitations 

There are a number of limitations in this study. The largest limitation is that it is a retrospective study, and like other studies on the treatment of pain in the ED, although pain score was measured, patient desire for pain medication and reasons for nontreatment or delays in treatment were not directly measured. We did not include any reassessment of pain and measured patients’ subjective pain at only 1 point (triage). Because pain can change over time without any intervention, this may have biased results by excluding patients who reported a low pain score at triage but may have had severe pain at room arrival. In addition, this may have biased results for patients whose pain improved spontaneously. Because nurses manually enter time of administration after administering medications, there may have been discrepancies between the recorded time and the actual time of medication administration. This study was performed at only 1 hospital, which limits the generalizability of these results to other settings. It has been shown in previous studies that a significant predictor of oligoanalgesia is the effect of specific providers’ practice patterns.8 We were unable to cluster on the provider and thus could not control for provider effects.

Discussion 

We detected a significant association between measures of ED crowding and oligoanalgesia as defined by delays in treatment and nontreatment with pain medications in a general population of ED patients reporting severe pain at triage. These findings were robust in multivariable analysis after adjusting for other known risk factors for oligoanalgesia. The assessment and treatment of pain is an important outcome in emergency care and a significant quality issue. Many studies have reported that pain is often undertreated in the ED.2, 3, 4, 5, 6 We found similar relationships between age, race, and sex that have been reported in the literature on ED oligoanalgesia.9, 10, 11, 12, 13, 14, 16 In addition, we report that severity is associated with a higher odds of nonreceipt of any pain medication, possibly because providers focus more on diagnosis than symptom control. However, when patients with more severe conditions were actually treated, they were less likely to experience delays in treatment.

The Joint Commission for Accreditation of Healthcare Organizations has made assessment and management of pain a quality standard on which hospitals are judged.23 Recent guidelines by the American Pain Society recommend that efforts be made to improve the quality and specifically the timeliness of pain management.24 Previous studies have reported associations between ED crowding and poorer quality of patient care.25, 26 The associations reported in this study provide more evidence that when the ED is crowded, patients do not receive optimal care. The relationship between ED crowding and oligoanalgesia is not unexpected. When the ED gets busier, staff may be less responsive to the needs of individual patients, and as a result, patients have a higher likelihood of nontreatment and also delays in treatment.

These findings indicate that ED crowding is not only associated with oligoanalgesia in older adults with hip fracture19 but also with oligoanalgesia in the broad ED patient population. Although Hwang et al20 reported a specific threshold effect (>120% occupancy) for ED census on oligoanalgesia, we did not find a specific threshold effect of crowding measures and oligoanalgesia in the unadjusted and adjusted models. A potential reason for this is that care of patients with hip fracture, because they have a definite injury, may not be affected only at relatively high levels of crowding. In a broad ED patient population with less clear-cut diagnoses, care may be affected at lower levels of ED crowding. These data challenge the concept of a specific threshold level to define a crowded ED, in which high-quality care is delivered up to a specific crowding threshold, above which care quality declines.

In the multivariable models, we found independent effects of waiting room number and ED occupancy delays in treatment. When the variable total patient-care hours was added to the regression model that already contained occupancy and waiting room number, it did not add predictive value to the model and was not an independent predictor of delays and nontreatment, because of the collinearity with occupancy and waiting room number. That is, the more parsimonious model (fewer crowding variables) was actually better, which indicates that because of collinearity between the myriad ways to measure crowding, using only a few of the simplest measures (waiting room number and occupancy) is highly predictive of poorer quality care. In addition to waiting room number, occupancy rate was predictive of delays in administration but was not a significant predictor of nontreatment. A potential reason for this is that when there are many undifferentiated patients waiting, ED staff may feel rushed to move existing patients and may be less responsive to the needs of individual patients such as pain control. Waiting room number seemed to be highly predictive of failure to treat, as well as treatment delays, and independent of occupancy, indicating that waiting room number may serve as one reasonable criterion standard for ED crowding. It is easily measured, has good face validity, and is generalizable. Walking in the front door of any ED and counting waiting room patients may represent a good barometer for how crowded it is. Future studies should consider using waiting room number and other simple measures of crowding as components in an index for ED crowding. As certification and quality organizations propose and implement process measures for emergency care and pay-for-performance, they may consider using ED crowding measures themselves as quality measures and defining crowding by an individual patient’s exposure to elements such as waiting room number when the patient is triaged. Because ED crowding has now been shown to be associated with poorer quality of care for pneumonia, acute myocardial infarction, and now pain, it may make more sense to consider broad fixes to health care quality than to only focus on specific conditions.

We demonstrate a significant association between ED crowding and oligoanalgesia (both delays and nontreatment) in a large academic ED. These results were independent after adjusting for other known risk factors for oligoanalgesia. Efforts to improve health care system functioning and reduce ED crowding may improve the quality of ED pain management.