Prevalence and treatment of pain in EDs in the United States, 2000 to 2010?,??

Hsien-Yen Chang, PhD ^a, Matthew Daubresse, MHS ^b,c, Stefan P. Kruszewski, MD ^c,d,e,

G. Caleb Alexander, MD, MS ^b,c,f,?

^a Department of Health Policy and Management, Johns Hopkins Bloomberg School of Public Health, Baltimore, MD, USA

^b Department of Epidemiology, Johns Hopkins Bloomberg School of Public Health, Baltimore, MD, USA

^c Center for drug safety and Effectiveness, Johns Hopkins University, Baltimore, MD, USA

^d Stefan P. Kruszewski, M.D. & Associates, Harrisburg, PA, USA

^e Department of Mental Health, Johns Hopkins Bloomberg School of Public Health, Baltimore, MD, USA

^f Division of General Internal Medicine, Department of Medicine, Johns Hopkins Medicine, Baltimore, MD, USA

a r t i c l e i n f o

Article history:

Received 11 October 2013

Received in revised form 13 January 2014 Accepted 14 January 2014

a b s t r a c t

Objectives: To describe changes in the prevalence and severity of pain and prescribing of non-opioid analgesics in US emergency departments (EDs) from 2000 to 2010.

Methods: Analysis of serial cross-sectional data regarding ED visits from the National Hospital Ambulatory Medical Care Survey. Visits were limited to patients >=18 years old without malignancy. Outcome measures included annual volume of visits among adults with a primary symptom or diagnosis of pain, annual rates of patient-reported Pain severity, and predictors of non-opioid receipt for non-malignant pain.

Results: Rates of pain remained stable, representing approximately 45% of visits from 2000 through 2010. Patients reported pain as their primary symptom twice as often as providers reported a primary pain diagnosis (40% vs 20%). The percentage of patients reporting severe pain increased from 25% (95% confidence intervals [CI] 22%-27%) in 2003 to 40% (CI 37%-42%) in 2008. From 2000 to 2010, the proportion of pain visits treated with pharmacotherapies increased from 56% (CI 53%-58%) to 71% (CI 69%-72%), although visits treated exclusively with non-opioids decreased 21% from 28% (CI 27%-30%) to 22% (CI 20%-23%). The adjusted odds of non-opioid rather than opioid receipt were greater among visits for patients 18 to 24 years old (odds ratio [OR] 1.35, CI 1.24-1.46), receiving fewer medicines (OR 2.91, CI 2.70-3.15) and those with a diagnosis of Mental illness (OR 2.24, CI 1.99-2.52).

Conclusions: Large increases in Opioid utilization in EDs have coincided with reductions in the use of non- opioid analgesics and an unchanging prevalence of pain among patients.

(C) 2014

1. Introduction

Pain is one of the most common reasons patients visit the emergency department (ED) [1,4], although not all individuals with pain desire pharmacologic treatment [5] pain is nevertheless

? Support. Dr Alexander is supported by the Agency for Healthcare Research and Quality (RO1 HS0189960). These funding sources had no role in the design and conduct of the study, analysis, or interpretation of the data; and preparation or final approval of the manuscript prior to publication.

?? Disclosures. Dr Alexander is an ad hoc member of the FDA’s Drug Safety and Risk

Management Advisory Committee, serves as a paid consultant to IMS Health, and serves on an IMS Health scientific advisory board. This arrangement has been reviewed and approved by the Johns Hopkins University in accordance with its conflict of interest policies. Dr Kruszewski has served as general and case-specific expert for multiple plaintiff litigation involving OxyContin, Neurontin and Zyprexa and has had false claims settled as co-plaintiff with the United States against Southwood Psychiatric Hospital, Pfizer (Geodon) and AstraZeneca (Seroquel).

* Corresponding author. Johns Hopkins Bloomberg School of Public Health Department of Epidemiology, Baltimore, MD 21205, USA. Tel.: +1 410 955 8168;

fax: +1 410 955 0863.

E-mail address: [email protected] (G.C. Alexander).

often undertreated [6,10]. For example, a large prospective multicenter study suggested that after lengthy wait times, only 60% of patients with pain in EDs received analgesics and nearly three-fourths of these individuals were discharged in moderate to severe pain [7].

During the past two decades, these shortcomings have prompted a variety of initiatives by the Joint Commission on Accreditation of Healthcare organizations and professional organizations to improve clinicians’ pain identification and treatment [11,14]. Such undertak- ings have included the “Pain as the Fifth Vital Sign” initiative, sponsored by the American Pain Society [15] the American College of Emergency Physicians’ clinical policy on pain management [16] as well as the novel pain management standards instituted by the Joint Commission on Accreditation of Healthcare Organizations in 2000

[17] and the World Health Organization’s analgesic ladder, designed to educate providers and patients regarding appropriate stepwise pharmacologic treatment of pain [14].

Although laudable, efforts to improve the quality of pain management have coincided with extensive marketing and promo- tion of opioid analgesics [18] sharp increases in opioid prescribing

0735-6757/$ - see front matter (C) 2014 http://dx.doi.org/10.1016/j.ajem.2014.01.015

[1,19] and increases in the morbidity and mortality associated with Prescription opioids [20,21]. Sales of opioids increased 400% from 1999 to 2010 [22]. By 2007, approximately 700 mg of Morphine equivalent doses per person were distributed through the pharma- ceutical supply chain in the United States, enough for every person to receive a routine dose of Vicodin every 4 hours for 3 weeks [23]. In EDs, opioid analgesic use increased more than 60% between 1993 and 2005, from 23% of pain visits in 1993 to 37% of visits in 2005 [11]. Furthermore, national trends suggest a strong positive correlation between prescription opioid sales, admissions for substance abuse treatment and deaths related to opioid use or abuse [20,22].

EDs play an especially important role in efforts to improve the treatment of pain while reducing epidemic rates of prescription drug abuse and addiction. EDs account for 10% of the 117 million visits ambulatory care visits nation-wide [24]. They also deliver care to a large number of patients misusing or abusing pharmaceuticals [25,26]. The number of visits associated with pharmaceutical misuse and abuse increased 115% from 626, 472 visits in 2004 to 1, 345, 645

visits in 2010 [27].

Soaring morbidity and mortality from prescription opioids have prompted recommendations from professional societies in emergen- cy medicine and stricter guidelines for opioid prescribing in EDs. For example, in 2012 the American College of Emergency Physicians released guidance that recommended short-acting opioids for acute Musculoskeletal pain, avoided routine prescriptions of opioids for outpatient management of acute exacerbations of chronic non-cancer pain, reserved opioid analgesics for severe acute low back pain and recommended prescribing only the lowest effective dose in these cases [28]. Additionally, the American Society of Anesthesiologists released updated guidelines recommending multi-modal techniques for pain management, which include the administration of two or more drugs that act by different mechanisms for providing analgesia independent of route of administration [29].

We used nationally representative data from the National Hospital Ambulatory Medical Care Survey to identify and characterize the management of pain among patients seeking care from EDs in the United States between 2000 and 2010. Our focus was to examine changes in the reported prevalence of pain symptoms, diagnoses and severity while characterizing prescribing rates of non-opioid pharmacotherapies and predictors of non-opioid receipt for non-malignant pain.

1. Methods
   1. Study design and setting

We examined data from the 2000 to 2010 NHAMCS ED file, which consists of a nationally Representative sample of visits to hospital EDs in the United States [30]. The NHAMCS, which is designed and conducted by the National Center for Health Statistics (NCHS), provides annual estimates of the number of visits to hospital EDs [31]. Participating EDs systematically select patient visits during a randomly assigned four-week reporting period. For each selected patient visit, data abstractors complete a patient record form that includes patient demographics, up to three patient-reported reasons for visit (ie, symptoms) and physician diagnoses, and up to 8 prescription and over-the-counter medications. In all years, masked sampling design variables were included to allow for adjustment for non-response and non-participation and to allow for statistical projections of national patterns of care. We carefully considered existing guidelines for use of the NHAMCS data in our analyses, by selecting a prevalent outcome, using easily understood variables, examining the frequency of missing responses and addressing coding changes in NHAMCS variables [32].

Selection of participants

We first compiled the survey data for each year from 2000 through 2010. Next, we identified visits with a primary self-reported symptom of pain. To do so, we included select patient-reported symptoms such as those containing the terms “pain”, “ache”, “soreness” or “discomfort.” Next, we used medical coding software

[33] keyword searches, and clinical judgment to identify visits with a primary physician-reported diagnosis related to pain. For pain- related ICD-9 codes, we included terms such as “ache”, “arthralgia”, “headache”, “myalgia”, and “pain”. We combined both types of visits in our final sample, thus we examine visits with a primary patient- reported symptom of pain or a primary physician diagnosis of pain. In order to examine a more homogenous population, we excluded individuals less than 18 years of age (24.1% of all weighted visits) and those with diagnoses of malignancy (0.9% of weighted adult visits) from all analyses. After weighting, an estimated 970 million visits were available for analysis.

Outcomes

Our analyses focused on the use of pharmacologic treatments for pain, such as non-opioid analgesics, opioid analgesics, and adjuvant therapies (eg, gabapentin, pregabalin, triCyclic antidepressants, local injectable anesthetics, Muscle relaxants and topical therapies) (Appendix Table 1). In 2005, the NHAMCS began distinguishing between medications ordered in the hospital and prescribed at discharge. For our trend analyses from 2003 to 2010, we combined these ordered and prescribed medications. However, we also conducted sensitivity analyses from 2005 to 2010 that examined ordered vs prescribed medicines separately. In 2006, the NHAMCS survey adopted the Multum drug ontology [31]. We applied this drug classification system to years prior to 2006 using publicly available SAS code provided by NCHS. We used survey documen- tation from the 2010 NHAMCS to group pharmacological analgesics into mutually exclusive categories, excluding antitussives and expectorants with analgesic properties. As suggested by the NCHS, for trend analyses we included the first six medications in the analyses since in years prior to 2003 only up to six medications were recorded. From 2000 to 2010, the proportion of visits with less than six recorded medications ranged from 93% to 96%. We used a similar strategy to develop categories for products containing non- steroidal inflammatory drugs (NSAIDs), acetaminophen and aspirin. We also examined the use of adjunctive therapies for pain such as anti-convulsants and tricyclic anti-depressants.

Analysis

We used descriptive statistics to compare patient, provider and practice characteristics of patients visiting the ED with and without a pain-related primary symptom or diagnosis. We then examined the annual prevalence and severity of reported pain. Next, we explored the proportion of patients who received different pharmacotherapies. To do so, we first stratified subjects into three groups: (1) those receiving at least one opioid analgesic; (2) those receiving only non-opioid analgesics; (3) those not receiving any analgesic pharmacotherapy. We then further characterized the treatments received by these individuals. For example, among visits where an opioid was used (“opioid visits”), we examined the proportion in which a variety of different treatment combinations occurred such as the use of fixed- dose combination therapies, NSAIDs, acetaminophen or adjunctive treatments such as anti-convulsants and anti-depressants. We analyzed trends for all opioid visits, opioid only visits, opioid and NSAID visits, and non-opioid only visits using the Cochran-Armitage trend test.

In our regression models, we defined our primary outcome as the receipt of a non-opioid rather than opioid analgesic for pain and restricted to visits between 2003 and 2010 with: (1) no cancer, (2)

Table 1

Patient and physician characteristics stratified by presence of non-malignant pain, 2000-2010

18 + years old, (3) having pain as the primary diagnosis or symptom, and 4) receiving at least one analgesic medication. The 2000 to 2002 NHAMCS surveys did not include pain severity, thus we limited our primary analysis to years 2003 to 2010. Between 2003 and 2008, the NHAMCS included an ordinal measure of pain severity with four categories ranging from “none” to “severe” and an

No pain symptoms or diagnosis

(N = 530 Million)

Total, % Age, %	54.6	45.4	100
18-24 years	15.9	16.9	16.3
25-44 years	35.1	42.1	38.3

Pain symptoms or diagnosis

(N = 440 Million)

Total

(N = 970 Million)

ten-point numerical pain scale into four ordinal categories ranging

Female sex, % 53.8 59.7 43.5

Race/ethnicity, %

additional item for “unknown” severity. In 2009 and 2010, an	45-64 years	25.5	26.7	26.0
alternative measure was used that asked patients to rate their pain	65-74 years	9.0	6.5	7.8
on a level from one to ten. In our main models, we converted the	N 75 years	14.6	7.8	11.5

from “none” to “severe” based on a standard conversion of the ten- point Visual analogue scale [34]. We conducted bivariate analyses to examine the association between patient (pain severity, age, sex, race, ethnicity, morbidity, number of medications), provider (type, shift) and visit (payment source, MSA, hospital ownership, region, triage level, admission status, visit length, day of the week, year)

White

(Non-Hispanic)

Black

(Non-Hispanic) Other race/ ethnicity Ethnicity, %

75.6 74.4 75.0

21.2 22.5 21.8

3.3 3.1 3.2

characteristics of interest and this outcome. Next, we used multivariate logistic regression to obtain the predicted probability of non-opioid receipt for the years 2003, 2007, and 2010. In the first round of model building we included Patient sex as well as other variables used in previous studies [1,7,9,10]. Next, we used backward selection to remove variables that became statistically insignificant (P N .05) to construct the final model. Predicted probabilities were calculated at the mean level of these variables.

Hispanic 10.4 11.2 10.8

Non-Hispanic 89.6 88.8 89.2

Pain Severity, %

None 22.7 3.7 14.1

Mild 10.6 9.8 10.2

Moderate 13.7 23.3 18.1

Severe 11.8 30.6 20.4

Unknown 41.1 32.7 37.3

Mental illness, %

No 88.0 96.0 91.6

We conducted all analyses using SAS version 9.2 with survey	Yes	12.0	4.0	8.4
weights and masked sampling design variables to account for the Number of medications, %
complex survey design of the NHAMCS.	Zero	26.6	15.7	21.6
	1	25.6	23.7	24.7

We conducted sensitivity analyses to confirm the robustness of

2.5. Sensitivity analyses	2 3	20.3 12.5	25.0 16.0	22.4 14.1
	4-6	15.0	19.7	17.1

our results and examine how they varied under different assump- tions. First, we repeated our trend analyses excluding over-the-

Payment source, % Blank/Other/ Unknown

12.6 11.4 12.1

counter medications due to potential underestimation of these drugs in the NHAMCS survey as well as the possibility of important shifts in their use of the decade examined. Second, we examined medications ordered and prescribed separately from 2005 to 2010. Third, we repeated our multivariate logistic regression using the entire period of observation, 2000 to 2010, by excluding pain severity from the model, since we only had data on pain severity from 2003 to 2010. Fourth, we repeated our multivariate logistic regression using the period when the pain severity was available and consistently measured, 2003 to 2008.

1. Results
   1. Patient, provider, and hospital characteristics associated with pain visits

From 2000 to 2010, approximately 45.4% of ED visits (440 of 970 million visits) were associated with a primary symptom or diagnosis of pain (Table 1). Although pain visits had similar characteristics as their non-pain counterparts overall, pain-visits were more likely than non-pain visits (42% vs 35%) to involve patients between the ages of 25 and 44.

Trends in prevalence and severity of pain

During the time period examined, the number of pain visits increased from 35 million visits in 2000 to 48.2 million visits in 2010. The proportion of pain visits as a patient-reported symptom or physician diagnosis remained stable and consistently represented approximately 45% of ED visits (Table 2). Overall, patients reported

Private 32.0 36.2 33.9

Medicare 24.7 17.2 21.3

Medicaid 14.3 17.0 15.5

Self-Pay 16.5 18.2 17.3

Type of provider(s) seen, %

Staff physician	86.7	87.9	87.2
Resident/intern	9.4	8.8	9.1
Other physician	8.0	7.0	7.5
Physician assistant	8.3	8.3	8.3
Nurse	90.6	91.6	91.1
Other	17.8	18.2	18.0

Source: National Hospital Ambulatory Medical Care Survey, 2000-2010.

pain as their primary reason for visit twice as often as providers reported a primary diagnosis of pain (~ 40% vs ~ 20%).

Changes in patients’ presenting level of pain over time varied based on severity. From 2003 to 2008, the percentage of patients reporting severe pain increased from 25% (95% confidence intervals [CI] 22%-27%) to 40% (CI 37%-42%) (Table 2).

Pharmacological treatment of pain in the ED

Table 3 depicts trends in the use of pharmacological analgesics in the ED for non-malignant pain from 2000 to 2010. During this time period, the proportion of pain visits treated with pharmaco- therapies increased from 56% (CI 53%-58%) to 71% (CI 69%-72%), whereas the proportion of visits with no pharmacological analge- sics decreased from 44% (42%-47%) to 30% (28%-31%). In 2000, there were an estimated 35 million ED visits for non-malignant pain and nearly an equal proportion of these visits were treated

Table 2

National trends in provider and patient reported non-malignant pain in EDs, 2000-2010?

2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010

Presence of pain

Pain reported from patient or provider

As primary code	44.0	44.3	43.5	43.9	45.0	45.0	45.3	46.6	45.2	46.7	48.6
As any code Provider reported pain	52.4	52.9	52.2	52.8	53.8	53.9	54.3	54.9	54.2	55.4	57.3
As primary diagnosis	16.7	17.2	17.8	18.5	18.7	18.9	19.2	19.7	19.8	19.6	21.0
As any diagnosis	9.6	20.5	21.6	22.1	22.6	22.4	23.4	23.7	24.1	23.9	25.7
Patient reported pain As primary symptom	41.4	41.1	40.3	40.6	41.6	41.5	41.6	43.3	41.4	43.1	45.3
As any symptom	49.8	50.1	49.1	49.9	50.7	51.0	51.1	51.9	51.0	52.1	54.2
Level of pain (2003-2008) None	—	—	—	3.5	3.9	4.2	4.4	5.1	4.5	—	—
Mild	—	—	—	15.9	14.7	12.5	12.0	10.8	10.9	—	—
Moderate	—	—	—	28.6	34.4	32.3	33.6	30.5	29.2	—	—
Severe	—	—	—	24.5	26.8	34.7	35.3	39.7	39.5	—	—
Missing or Unknown	—	—	—	27.5	20.2	16.4	14.8	13.9	15.9	—	—
Level of pain (2009-2010) 0	—	—	—	—	—	—	—	—	—	5.3	4.8
1-3	—	—	—	—	—	—	—	—	—	6.1	6.1
4-6	—	—	—	—	—	—	—	—	—	20.1	19.5
7-10	—	—	—	—	—	—	—	—	—	51.2	55.2
Missing or unknown	—	—	—	—	—	—	—	—	—	17.3	14.4

Source: National Hospital Ambulatory Medical Care Survey, 2000 to 2010.

* Values represent column percents; pain assessed from 2003 to 2008 using the Numeric rating scale (4 points; 0-3) and 2009 to 2010 using the Numeric Rating Scale (11 points; 0-10).

exclusively with non-opioid (28%; 27%-30%) and opioid (27%; 25%- 29%) pharmacotherapies, however by 2010, the proportion of ED visits treated exclusively with non-opioids decreased to 22% (20%- 23%) of visits.

Although the exclusive use of non-opioids for non-malignant pain decreased during the study period, the concomitant use of opioid analgesics with NSAIDs or acetaminophen nearly doubled. From 2000 to 2010, the proportion of visits treated with an opioid-containing analgesic (fixed-dose combo or opioid-only) and a separate NSAID medication increased from 9% to 17%. Visits treated with an opioid- containing medication and acetaminophen separately increased from 1% to 2%. During this time period, the use of fixed dose combinations also increased from 22% to 31%. Trend tests for all opioid visits, opioid only visits, opioid and NSAID visits, and non-opioid only visits

achieved statistical significance (P b .0001). In analyses stratified by severity, the proportion of visits treated with opioids increased, whereas the proportion of visits treated with non-opioids decreased (Appendix Table 2).

Patient, provider and practice characteristics associated with non-opioid use

Overall, many patient, physician, and visit characteristics were associated with the likelihood of receiving a non-opioid rather than an opioid medication (Table 4). For example, bivariate analyses showed rates of non-opioid receipt were higher among patients with diagnoses of mental illness (52% vs 35%), African Americans (43% vs 34% among whites), no primary symptom or diagnosis of pain (65% vs

Table 3

Trends in treatment of non-malignant pain among ED visits in the United States, 2000 to 2010?

	2000	2001	2002	2003	2004	2005	2006	2007	2008	2009	2010
All pain visits?? Treated with pharmacotherapies???	55.7	58.7	62.2	63.1	65.9	65.6	66.0	67.2	68.9	68.5	70.5
No pharmacotherapies	44.3	41.3	37.9	36.9	34.1	34.4	34.0	32.8	31.1	31.5	29.5
Visits treated with opioids+	27.4	29.9	34.3	36.4	38.1	40.6	42.2	43.9	45.9	46.3	48.9
Fixed-dose combination visits	22.1	24.3	26.9	26.9	27.5	29.1	28.9	29.0	29.7	30.4	30.5
Opioid only visits+	6.9	7.0	10.2	12.8	15.0	16.7	19.0	21.7	23.2	23.7	26.8
NSAIDs visits+	9.1	10.2	12.1	12.1	12.4	13.9	14.0	13.9	15.1	14.9	16.8
Acetaminophen visits	0.9	1.1	1.2	1.3	1.8	1.7	1.7	1.9	1.9	1.9	1.8
Adjunctive treatment visits	3.6	4.4	4.9	5.3	6.0	5.5	6.5	6.8	7.2	7.4	8.2
Visits treated with non-opioids only+	28.2	28.8	27.8	26.7	27.8	25.1	23.8	23.3	23.0	22.3	21.6
NSAIDs visits	20.9	21.9	21.9	20.7	21.6	19.6	18.5	18.3	18.1	17.5	16.8
Acetaminophen visits	6.5	6.0	5.4	5.4	6.2	5.0	5.0	4.9	5.0	4.6	4.4
Adjunctive treatment visits	5.5	5.8	5.8	5.2	5.5	5.5	4.3	4.8	4.5	4.5	5.0
Pain visits by therapy
Opioid (millions), N	9.6	10.7	12.2	13.4	14.1	15.6	17.6	18.2	19.7	22.0	23.6
Non-opioids only (millions), N	9.9	10.3	9.8	9.8	10.3	9.7	9.9	9.7	9.9	10.6	10.4
No pharmacotherapies (millions), N	15.5	14.7	13.4	13.6	12.6	13.2	14.2	13.6	13.4	15.0	14.2
All pain visits (millions), N??	35.0	35.7	35.4	36.9	37.0	38.5	41.7	41.5	42.9	47.6	48.2

Source: National Hospital Ambulatory Medical Care Survey, 2000 to 2010.

* Except where otherwise indicated values represent column percentages.

?? Pain visits include visits with a primary patient-reported symptom of pain or a primary physician diagnosis of pain.

??? Pharmacotherapies include opioids, non-steroidal Anti-inflammatory agents (NSAIDs), acetaminophen, and adjunctive treatments (gabapentin, pregabalin, tricyclic antidepressants, local injectable anesthetics, muscle relaxants and topical therapies).

⁺ Cochran-Armitage trend test (P b .0001).

Table 4

Multivariate association between patient, physician and hospital characteristics and non-opioid use for pain patients visiting the ED

% Receiving Predicted probability of non-opioid rather than opioid receipt (95% CI)

	Non-opioids	Odds ratio	95% CI	2003	2007	2010
Average visit Patient characteristics Age, years 18-24	40.7	1.35	1.24-1.46	.61 (.56-.66) .43 (.41-.46)	.52 (.49-.54) .39 (.37-.41)	.46 (.43-.49) .36 (.34-.38)
25-44	34.5	1.06	.99-1.13	.37 (.35-.39)	.34 (.32-.35)	.31 (.29-.32)
45-64	34.0	REF	REF	.36 (.34-.38)	.32 (.31-.34)	.30 (.28-.31)
65-74	37.8	1.08	.96-1.21	.38 (.35-.41)	.34 (.32-.36)	.31 (.29-.34)
75 or more Sex Male	37.9 35.7	1.03 REF	.93-1.15 REF	.37 (.34-.40) .37 (.35-.39)	.33 (.31-.35) .33 (.32-.34)	.30 (.28-.33) .31 (.29-.32)
Female	35.8	1.06	1.01-1.12	.38 (.36-.41)	.35 (.33-.36)	.32 (.30-.33)
Race White	33.5	REF	REF	.35 (.33-.37)	.31 (.30-.33)	.29 (.28-.30)
Black	42.6	1.61	1.48-1.75	.47 (.44-.49)	.42 (.40-.45)	.40 (.37-.42)
Other Ethnicity Hispanic	39.7 40.8	1.45 1.43	1.22-1.73 1.31-1.55	.44 (.39-.49) .45 (.43-.48)	.40 (.36-.44) .41 (.39-.44)	.37 (.33-.41) .38 (.36-.41)
Non-Hispanic Pain Severity	35.1	REF	REF	.37(.35-.39)	.33 (.32-.34)	.31 (.29-.32)
None	65.1	4.78	4.09-5.57	.67( .63-.71)	.63 (.60-.67)	.61 (.57-.64)
Mild	50.1	2.38	2.17-2.61	.51 (.48-.53)	.46 (.44-.49)	.43 (.41-.46)
Moderate	40.5	1.71	1.61-1.80	.42 (.40-.44)	.38 (.37-.40)	.35 (.34-.37)
Severe	26.6	REF	REF	.30 (.28-.32)	.27 (.25-.28)	.24 (.23-.26)
Unknown Morbidity (# of CCS) None	40.6 43.0	1.73 —	1.60-1.87 —	.43 (.40-.45) —	.39 (.37-.41) —	.36 (.34-.38) —
1	36.1	—	—	—	—	—
2	34.9	—	—	—	—	—
3 Mental Illness No	35.4 35.1	— REF	— REF	— .37 (.35-.39)	— .33 (.32-.35)	— .31 (.29-.32)
Yes Number of medications 1	52.4 51.8	2.24 2.91	1.99-2.52 2.70-3.15	.57 (.54-.60) .53 (.51- .56)	.53 (.50-.56) .49 (.48-.51)	.50 (.47-.53) .46 (.44-.48)
2	37.5	1.67	1.56-1.79	.40 (.38-.42)	.36 (.34-.37)	.33 (.31-.35)
3	31.1	1.30	1.21-1.39	.34 (.32-.36)	.30 (.29-.32)	.28 (.26-.29)
4-6 Provider reported pain No	24.9 35.1	REF REF	REF REF	.28 (.27-.30) .36 (.34-.39)	.25 (.24-.26) .33 (.31-.34)	.23 (.21-.24) .30 (.29-.32)
Yes Patient report pain No	36.5 34.2	1.15 —	1.09-1.21 —	.40 (.38-.42) —	.36 (.35-.37) —	.33 (.32-.34) —
Yes Provider characteristics Type of provider(s) seen Staff physician	35.6 35.4	— —	— —	— —	— —	— —
Resident/intern	36.4	—	—	—	—	—
Other physician	36.0	—	—	—	—	—
Physician assistant	35.9	—	—	—	—	—
Nurse	35.6	—	—	—	—	—
Other Shift Day	33.9 36.6	— —	— —	— —	— —	— —
Evening	35.2	—	—	—	—	—
Night	37.3	—	—	—	—	—
Missing Visit characteristics	34.8	—	—	—	—	—
Primary payment source Private	37.6	—	—	—	—	—
Medicare	35.7	—	—	—	—	—
Medicaid	34.7	—	—	—	—	—
Self-Pay	36.6	—	—	—	—	—
Other Metropolitan area No	34.9 40.3	— 1.32	— 1.11-1.56	— .43 (.39-.48)	— .40 (.36-.43)	— .37 (.33-.40)
Yes Hospital ownership Voluntary	34.9 35.4	REF REF	REF REF	.37 (.35-.39) .37 (.35-.39)	.33 (.32-.34) .33 (.32-.34)	.30 (.29-.32) .31 (.29-.32)
Government	40.8	1.35	1.14-1.58	.44 (.40-.49)	.40 (.36-.44)	.37 (.33-.41)
Proprietary geographic location Northeast	31.6 46.0	.99 1.61	.86-1.13 1.45-1.79	.37 (.33-.40) .48 (.46-.50)	.33 (.30-.36) .49 (.42-.45)	.30 (.27-.33) .41 (.39-.43)
Midwest	35.5	1.08	.94-1.25	.38 (.35-.41)	.35 (.32-.37)	.32 (.29-.34)

(continued on next page)

Table 4 (continued)

% Receiving

Predicted probability of non-opioid rather than opioid receipt (95% CI)

	Non-opioids	Odds ratio	95% CI	2003	2007	2010
South	34.8	REF	REF	.36 (.34-.39)	.33 (.31-.35)	.30 (.28-.32)
West Immediacy/triage level No triage/unknown	28.1 34.5	.79 .71	.67-.94 .64-.79	.31 (.28-.35) .37 (.35-.40)	.28 (.25-.31) .34 (.32-.36)	.25 (.23-.28) .31 (.29-.33)
Less than 15 min	41.6	REF	REF	.46 (.43-.48)	.42 (.40-.43)	.39 (.36-.41)
15-60 min	33.9	.70	.64-.76	.37 (.35-.39)	.33 (.32-.35)	.31 (.29-.32)
1-2 h	34.9	.70	.63-.78	.37 (.35-.39)	.33 (.31-.35)	.31 (.29-.32)
2-24 h Admission to hospital No	35.5 35.4	.62 REF	.55-.70 REF	.34 (.32-.37) .37 (.35-.39)	.31 (.29-.33) .33 (.32-.34)	.28 (.26-.30) .34 (.29-.32)
Yes Length of visit Short	38.0 38.8	1.41 REF	1.31-1.53 REF	.45 (.43-.48) .39 (.37-.41)	.41 (.39-.43) .35 (.34-.37)	.38 (.36-.40) .33 (.31-.34)
Medium	36.7	1.01	.94-1.07	.39 (.37-.41)	.35 (.34-.37)	.33 (.31-.34)
Long	32.0	.86	.79-.93	.36 (.34-.38)	.32 (.31-.33)	.29 (.28-.31)
Missing Weekend visit No	41.3 36.1	.96 —	.84-1.09 —	.38 (.35-.41) —	.34 (.31-.37) —	.32 (.29-.35) —
Yes Procedures provided No	34.9 38.8	— REF	— REF	— .39 (.37-.42)	— .36 (.34-.37)	— .33 (.31-.34)
Yes Year 2003	32.4 42.3	.87 REF	.82-.93 REF	.36 (.34-.38) —	.33 (.31-.34) —	.30 (.28-.31) —
2004	42.2	.96	.95-.97	—	—	—
2005	38.2			—	—	—
2006	36.0			—	—	—
2007	34.7			—	—	—
2008	33.3			—	—	—
2009	32.5			—	—	—
2010	30.6			—	—	—

— Indicates variables excluded from regression model.

27% among those with severe pain) and those receiving one medication (52% vs 25% among those receiving four or more medicines). The proportion of patients receiving non-opioids did not differ significantly based on provider type, shift, or source of payment. Results from multivariate analysis mirrored those of the bivariate analysis. For example, the odds of non-opioid receipt rather than opioid receipt were greater among visits for patients 18 to 24 years old (odds ratio [OR] 1.35, CI 1.24-1.46), visits without a primary symptom or diagnosis of pain (OR 4.78, CI 4.09-5.57), those where fewer medicines were used (OR 2.91, CI 2.70-3.15), and those associated with a diagnosis of mental illness (OR 2.24, CI 1.99-2.52) (Table 4). The probability of an average pain visit receiving non-opioid medication decreased from 61% in 2003 to

46% in 2010.

Sensitivity analyses

Sensitivity analyses supported our analytic approach and the robustness of our results. Despite a reduction in magnitude, the decreasing trend in non-opioid prescribing persisted in analyses excluding over-the-counter medications. From 2005 to 2010, changes in the proportion of visits treated with opioids and non- opioids at discharge were consistent with changes in overall analgesic use (Appendix Table 3). Results from our multivariate logistic regression remained stable and substantively unchanged using data from 2000 to 2010 that excluded pain severity and from 2003 to 2008 that included measures of pain severity (Appendix Table 4).

1. Discussion

In this serial cross-sectional study of nationally representative data from EDs in the United States, we found that the prevalence

of patient-reported pain has remained stable, the proportion of visits where patients reported severe pain increased and the use of analgesics increased substantially during the past decade. During this period, the proportion of visits treated with non-opioid medications decreased considerably. Although these data do not provide sufficient clinical information to judge the clinical appropriateness of analgesic prescriptions, these results are important since efforts to curb the undertreatment of pain and improve the identification and treatment of pain in EDs and other clinical settings have coincided with larger increases in opioid use and ED visits associated with pharmaceutical misuse and abuse [27].

Although several prior studies have described increasing trends in opioid use in EDs [1,9], fewer have examined how these changes in opioid utilization have corresponded with the prevalence and severity of pain or the use of multi-modal therapies and non-opioid pharmacotherapies such as NSAIDs, acetaminophen and other analgesic agents. In contrast to one study [9], our results indicate substantial declines in non-opioid use; these differences may be due to differences in the study sample, design and analytic periods examined. In addition, these results suggest an over-reliance on opioid monotherapy among ED physicians. Prior studies have also tended to focus on specific subpopulations of patients, such as those with orthopedic fractures, or have focused exclusively on the aggregate use of prescription opioids. In contrast to these studies, we evaluated specific patterns of analgesic treatment such as the use of opioid monotherapy, fixed-dose combination products and the combination of prescription opioids with other analgesics such as NSAIDs. The proportion of pain visits receiving any analgesia and specifically Opioid analgesia in our analyses are consistent with another NHAMCS study; however, our results suggest patients ages 18 to 24 were more likely to receive Non-opioid analgesia than any other age group [35].

Our results suggest a notable increase in pain severity occurred among ED visits from 2003 to 2008. The contributors to this increase remain unclear and warrant further research. As described in the 2011 Institute of Medicine report, there are multiple complex societal- and individual-level determinants of pain [11]. Although pain severity increased during this five-year period, results from our regression analyses suggest these changes are insufficient to explain shifts in opioid and non-opioid prescribing that we describe.

Our analyses were not designed to evaluate the comparative safety or effectiveness of opioid vs non-opioid therapies. There are numerous pharmacologic and non-pharmacologic therapies available to patients in pain and despite their risks, both opioids and non- opioids play important roles in pain treatment. Despite this, professional societies and other stakeholders are increasingly recog- nizing the limitations of opioid therapy (e.g. respiratory depression, delirium, sedation, opioid induced hyperalgesia, analgesic ceilings) and reevaluating the appropriate role of opioids given the morbidity and mortality associated with an epidemic of their overuse [28]. For example, in January 2013, EDs in New York City’s eleven public hospitals adopted voluntary guidelines limiting prescriptions for short-acting opioid analgesics to a maximum three-day supply and halting dispensation of long-acting opioids such as oxycontin and fentanyl [36]. EDs in Michigan and Utah have established similar guidelines, which completely prohibit and delay opioid prescribing for patients with chronic conditions [37]. These Policy changes are especially important since there is little evidence to suggest that opioids are safer or more effective than many other alternative analgesics, especially in terms of functional outcomes and long-term use [38,39], and there is some evidence that opioid utilization is associated with higher Healthcare costs than alternative analgesic agents [40,41]. Although we excluded children from our analyses, opioid use among children is an important area of future research.

A number of factors limit our study. First, the NHAMCS is cross-

sectional, and thus does not capture longitudinal outcomes for individual patients. Second, these data may underestimate mentions of over-the-counter therapies prescribed at discharge, such as acetaminophen and naproxen. This under capture would pose a threat to validity if it were systematically impacted by secular

changes such as the market withdrawal of rofecoxib in 2004. However, sensitivity analyses that excluded over-the-counter med- ications yielded results consistent with our original findings. Third, prior to 2005 the NHAMCS survey did not distinguish between medications ordered in the hospital and prescribed at discharge. Our sensitivity analyses examining trends in analgesic use in the hospital and at discharge yielded results consistent with our original analysis. Finally, providers prescribe additional therapies off-label to treat pain, such as the prophylactic use of ?-blockers, amitriptyline, or topiramate for recurrent headaches. We conservatively excluded these therapies from our analyses given their many indications unrelated to pain.

The NHAMCS does not provide detailed clinical information regarding patient diagnoses, or preferences for treatment or medication dosing; thus, we were unable to determine the appropriateness of analgesic use. As we expected, at an aggregate level, as pain severity increased, the proportion of pain visits without analgesics or with non-opioid analgesics decreased; however, the proportion of changes in pain severity may reflect changes in nurse charting. Despite this, approximately one-third of visits with moderate pain and one-fifth of visits with severe pain remained untreated. This is consistent with findings from a recent Institute of Medicine report describing inadequate treatment of pain in the United States, although our estimates of untreated visits likely include a combination of patients whose preference was not to receive treatment [5] as well as those whose non-treatment reflects a breach in the quality of care provided [6,7].

While the role of non-opioid analgesics in EDs has waned during the decade examined, the overall prevalence of pain visits has remained stable, the prevalence of visits with severe pain increased, and providers increasingly prescribed opioid analgesics despite substantial increases in morbidity and mortality associated with these drugs. Guidelines and recommendations encouraging multi- modal therapies and more conservative use of opioid analgesics, as well as proactive interventions by regulators, payers and other healthcare delivery stakeholders, may assist in reestablishing a safer and more effective balance between a variety of analgesics available for the treatment of pain.

Appendix

Appendix Table 1

Drug classification

Opioid-only Hydrocodone Dihydrocodeine Buprenorphine-naloxone

Opioid Fixed-dose Combinations	Butalbital-codeine Codeine Codeine-papaverine Aspirin-codeine preparations Aspirin-propoxyphene	Morphine Propoxyphene Hydromorphone oxycodone Fentanyl Acetaminophen-dextropropoxyphene Acetaminophen-codeine preparations	Buprenorphine Methadone Oxymorphone Tramadol Acetaminophen-oxycodone Acetaminophen-propoxyphene
Acetaminophen	Aspirin-hydrocodone Asa-dihydrocodeine Aspirin-oxycodone Acetaminophen preparations	Acetaminophen-hydrocodone Acetaminophen-dihydrocodeine Acetaminophen-chlorzoxazone	Acetaminophen-tramadol Hydrocodone-ibuprofen Ibuprofen-oxycodone Tramadol-aspirin Acetaminophen-pamabron
	Acetaminophen-aspirin preparations Acetaminophen-butalbital preparations	Acetaminophen-diphenhydramine Acetaminophen-isometheptenemucate	Acetaminophen-phentolamine Acetaminophen-phenylpropanolamine
NSAIDs	Acetaminophen-caffeine Nonsteroidal anti-inflammatory agents COX-2 inhibitors Celecoxib Rofecoxib Valdecoxib Ibuprofen Naproxen	Acetaminophen-miscellaneous analgesics Aspirin-carisoprodol Aspirin-meprobamate Aspirin-methocarbamol Ketoprofen Diclofenac Diclofenac-misoprostol Indomethacin	Acetaminophen-phenyltoloxamine Meclofenamate Mefenamic acid Nabumetone Oxaprozin Piroxicam Sulindac Tolmetin
Muscle relaxants	Naproxen-preparations Aspirin Aspirin preparations Aspirin-butalbital Baclofen Carisoprodol Chlormezanone	Meloxicam Diflunisal Etodolac Fenoprofen Flurbiprofen Lorazepam Mephenesin Meprobamate	Ketorolac Sodium salicylate Choline salicylate Asa/caffeine/orphenadrine Asa/caffeine/salicylamide Nitrazepam Orphenadrine Quinine
Anti-convulsants	Chlorphenesin Chlorzoxazone Cyclobenzaprine Diazepam Donepezil Pregabalin	Meprobamate-pentaerythritol Meprobamate-tridihexethyl chloride Metaxalone Methocarbamol Gabapentin	Quinine-urea Tizanidine Trazodone Tybamate
TCAs	Amitriptyline	Clomipramine	Maprotiline
Topical preparations Injectables	Amitriptyline-chlordiazepoxide Amitriptyline-perphenazine Amoxapine Benzocaine topicals Camphor topicals Capsaicin topicals Diclofenac topical Local injectable anesthetics	Desipramine Doxepin Imipramine Lidocaine topicals Menthol topicals Methyl salicylate topicals	Nortriptyline Protriptyline Trimipramine Pramoxine topicals Phenylephrine topicals Trolamine salicylate topicals

Appendix Table 2

Trends in treatment of non-malignant pain by self-reported pain severity?

2003 2004 2005 2006 2007 2008 2009 2010

All Pain Visits

Appendix Table 3

Trends in analgesic use in-hospital and at discharge, 2005-2010?

2005 2006 2007 2008 2009 2010 Percent

Change

Visits treated with opioids

Visits treated with

36.4 38.1 40.6 42.2 43.9 45.9 46.3 48.9

26.7 27.8 25.1 23.8 23.3 23.0 22.3 21.6

All pain visits?? Visit with any pain medication???

65.6 66.0 67.2 68.9 68.5 70.5 7.4

non-opioids

Visits untreated 36.9 34.1 34.4 34.0 32.8 31.1 31.5 29.5

Missing or Unknown Pain

At discharge 39.6 41.0 40.8 41.9 42.9 42.9 8.5

In hospital 44.5 45.3 47.0 49.8 49.0 52.7 18.6

Visits treated with opioids

Visits treated with

non-opioids

31.6 30.2 30.9 34.2 33.9 36.2 35.7 39.7

25.0 29.9 24.9 20.5 21.3 20.3 20.6 21.9

Opioid Visits

Visits treated with opioids

At discharge 25.6 26.8 27.4 28.7 30.0 30.8 20.0

In hospital 24.0 25.2 27.9 29.9 29.6 32.9 36.8

Visits untreated 43.3 40.0 44.2 45.3 44.8 43.5 43.8 38.4

No Pain

Fixed-dose combination uses

At discharge 23.1 23.6 24.3 24.7 25.2 25.5 10.3

Visits treated with opioids

Visits treated with

non-opioids

15.9 13.7 10.9 13.5 15.0 15.2 13.7 14.1

26.7 27.9 26.3 23.1 24.6 28.1 24.4 28.7

In hospital 10.7 10.2 10.1 11.1 11.4 11.6 8.6

Opioid only uses

At discharge 2.8 3.7 3.5 4.3 5.1 5.6 96.8

In hospital 14.5 16.0 19.1 20.0 19.9 23.1 59.5

Visits untreated 57.5 58.4 62.8 63.4 60.4 56.7 61.9 57.3

Mild Pain

NSAIDs uses

At discharge 5.2 5.4 5.3 6.2 6.1 7.0 34.1

Visits treated with opioids

Visits treated with

non-opioids

29.2 27.7 28.3 29.1 27.7 27.1 28.8 28.4

27.4 29.6 28.0 30.1 25.6 29.5 31.5 25.2

In hospital 6.3 5.9 6.7 7.0 7.3 8.4 33.5

acetaminophen uses

At discharge 0.6 0.3 0.5 0.6 0.5 0.5 -21.3

In hospital 0.6 0.6 0.8 0.7 0.7 0.9 38.4

Visits untreated 43.5 42.7 43.8 40.8 46.7 43.4 39.7 46.4

Moderate Pain

Adjunctive treatment uses

At discharge 3.6 4.1 4.2 4.0 4.4 4.7 31.8

Visits treated with opioids

Visits treated with non-opioids

35.4 40.0 41.2 41.2 40.7 43.2 36.5 38.5

30.1 27.7 27.1 25.6 27.3 25.4 27.5 27.2

In hospital 2.0 2.4 2.9 3.1 3.3 3.8 92.5

Non-opioid Visits Visits treated with

non-opioids only

Visits untreated 34.5 32.4 31.7 33.3 32.0 31.4 36.1 34.3

At discharge	13.9	14.2	13.3	13.2	12.9	12.1	-12.7
In hospital	20.4	20.0	19.1	19.9	19.4	19.8	-2.8
NSAIDs uses At discharge	10.6	10.6	9.8	9.9	9.8	8.8	-16.8
In hospital Acetaminophen uses	16.3	15.9	15.6	15.9	15.4	15.7	-4.0
At discharge	2.9	2.9	3.1	3.0	2.4	2.4	-18.0

Severe Pain

Visits treated with opioids

Visits treated with non-opioids

50.5 50.9 52.5 54.7 58.0 60.5 59.1 60.3

24.0 25.4 22.1 21.5 20.2 19.8 19.5 18.5

Visits untreated 25.4 23.7 25.4 23.8 21.8 19.7 21.4 21.2

Source: National Hospital Ambulatory Medical Care Survey, 2000-2010. * Pain visits include visits with a primary patient-reported symptom of pain or a primary physician diagnosis of pain.	In hospital Adjunctive treatment uses At discharge	2.7 3.3	3.0 2.9	2.6 3.0	3.0 2.9	3.1 3.1	2.9 3.3	8.2 0.8
	In hospital	3.8	3.3	3.4	3.2	3.3	4.2	9.7
	All pain visits (millions), N??	38.5	41.7	41.5	42.9	47.6	48.2	—

Source: National Hospital Ambulatory Medical Care Survey, 2000-2010.

* Except where otherwise indicated values represent column percents.

?? Pain visits include visits with a primary patient-reported symptom of pain or a primary physician diagnosis of pain.

??? Pharmacotherapies include opioids, non-steroidal anti-inflammatory agents (NSAIDs),

acetaminophen, and adjunctive treatments (gabapentin, pregabalin, tricyclic antidepressants, local injectable anesthetics, muscle relaxants and topical therapies).

Appendix Table 4

Regression estimates for non-opioid receipt from sensitivity analyses including and excluding pain severity

Variable	2003 - 2010				2000 - 2010				2003 - 2008
	Odds ratio	Upper 95% CI	Lower 95% CI		Odds ratio	Upper 95% CI	Lower 95% CI		Odds ratio	Upper 95% CI	Lower 95% CI
Intercept Age 18-24	0.25 1.35	0.21 1.24	0.28 1.46		0.38 1.28	0.34 1.20	0.44 1.37		0.29 1.32	0.25 1.22	0.34 1.44
25-44	1.06	0.99	1.13		1.04	0.98	1.09		1.05	0.97	1.14
45-64	REF	REF	REF		REF	REF	REF		REF	REF	REF
65-74	1.08	0.96	1.21		1.16	1.05	1.29		1.02	0.90	1.15
75 or more Sex Male	1.03 REF	0.93 REF	1.15 REF		1.14 REF	1.04 REF	1.24 REF		1.07 REF	0.95 REF	1.20 REF
Female Race	1.06	1.01	1.12		1.05	1.01	1.10		1.04	0.98	1.10
White	REF	REF	REF		REF	REF	REF		REF	REF	REF
Black	1.61	1.48	1.74		1.63	1.51	1.74		1.60	1.45	1.77
Other Ethnicity Hispanic	1.45 1.43	1.22 1.31	1.72 1.55		1.39 1.44	1.19 1.34	1.62 1.55		1.39 1.38	1.17 1.26	1.64 1.52
Non-Hispanic Pain Severity None	REF 4.77	REF 4.09	REF 5.57		REF —	REF —	REF —		REF 4.36	REF 3.66	REF 5.20
Mild	2.38	2.17	2.61		—	—	—		2.21	2.00	2.45
Moderate	1.71	1.61	1.80		—	—	—		1.58	1.47	1.70
Severe	REF	REF	REF		—	—	—		REF	REF	REF
Unknown Mental Illness No	1.73 REF	1.60 REF	1.87 REF		— REF	— REF	— REF		1.72 REF	1.56 REF	1.89 REF
Yes Number of medications	2.24	1.99	2.52		2.27	2.05	2.53		2.23	1.95	2.55
1	2.91	2.69	3.14		2.98	2.80	3.16		2.64	2.43	2.86
2	1.67	1.56	1.79		1.63	1.54	1.72		1.55	1.43	1.67
3	1.30	1.21	1.39		1.26	1.19	1.33		1.20	1.10	1.31
4-6 Provider reported pain No	REF REF	REF REF	REF REF		REF REF	REF REF	REF REF		REF REF	REF REF	REF REF
Yes	1.15	1.09	1.21		1.14	1.09	1.19		1.13	1.06	1.19
Metropolitan area	REF	REF	REF		REF	REF	REF		REF	REF	REF
No Hospital ownership Voluntary	1.32 REF	1.11 REF	1.56 REF		1.24 REF	1.08 REF	1.42 REF		1.29 REF	1.06 REF	1.56 REF
Government	1.34	1.14	1.58		1.33	1.17	1.50		1.38	1.14	1.68
Proprietary Geographic location Northeast	0.99 1.61	0.86 1.45	1.13 1.79		0.96 1.62	0.86 1.46	1.08 1.79		0.96 1.58	0.82 1.42	1.11 1.76
Midwest	1.08	0.94	1.25		1.11	0.98	1.26		1.04	0.89	1.20
South	REF	REF	REF		REF	REF	REF		REF	REF	REF
West Immediacy/triage level No Triage/ Unknown	0.79 0.71	0.67 0.64	0.94 0.79		0.77 0.82	0.67 0.75	0.88 0.90		0.79 0.73	0.67 0.65	0.93 0.82
Less than 15 min	REF	REF	REF		REF	REF	REF		REF	REF	REF
15-60 min	0.70	0.64	0.76		0.74	0.68	0.80		0.72	0.65	0.80
1-2 h	0.70	0.63	0.78		0.72	0.65	0.78		0.72	0.64	0.82
2-24 h Admission to hospital No	0.62 REF	0.54 REF	0.70 REF		0.67 REF	0.60 REF	0.75 REF		0.65 REF	0.57 REF	0.74 REF
Yes Length of visit Short	1.41 REF	1.31 REF	1.53 REF		1.47 REF	1.37 REF	1.58 REF		1.39 REF	1.27 REF	1.52 REF
Medium	1.00	0.94	1.07		1.02	0.97	1.09		1.01	0.94	1.09
Long	0.86	0.79	0.93		0.89	0.83	0.96		0.86	0.79	0.94
Missing Procedure provided No	0.96 REF	0.84 REF	1.09 REF		1.02 REF	0.93 REF	1.11 REF		0.95 REF	0.82 REF	1.09 REF
Yes Year 2009	0.87 REF	0.82 REF	0.93 REF		0.87 REF	0.82 REF	0.92 REF		0.85 REF	0.79 REF	0.91 REF
2010	0.96	0.95	0.97		0.93	0.92	0.94		0.94	0.92	0.96

References

1. Pletcher MJ, Kertesz SG, Kohn MA, Gonzales R. Trends in opioid prescribing by race/ethnicity for patients seeking care in US emergency departments. JAMA 2008;299:70-8.
2. Cordell WH, Keene KK, Giles BK, Jones JB, Jones JH, Brizendine EJ. The high prevalence of pain in Emergency medical care. Am J Emerg Med 2002;20:165-9.
3. Johnston CC, Gagnon AJ, Fullerton L, Common C, Ladores M, Forlini S. One-week survey of pain intensity on admission to and discharge from the emergency department: a pilot study. J Emerg Med 1998;16:377-82.
4. Tanabe P, Buschmann M. A prospective study of ED pain management practices and the patient’s perspective. J Emerg Nurs 1999;25:171-7.
5. Singer AJ, Garra G, Chohan JK, Dalmedo C, Thode Jr HC. Triage pain scores and the desire for and use of analgesics. Ann Emerg Med 2008;52:689-95.
6. Rupp T, Delaney KA. Inadequate analgesia in emergency medicine. Ann Emerg Med 2004;43:494-503.
7. Todd HK, Ducharme J, Choiniere M, Crandall CS, Fosnocht DE, Homel P, et al. for the PEMI Study Group. Pain in the emergency department: results of the pain and emergency medicine initiative (PEMI) multicenter study. J Pain 2007;8: 460-6.
8. Platts-Mills TF, Hunold KM, Bortsov AV, Soward AC, Peak DA, Jones JS, et al. More educated emergency department patients are less likely to receive opioids for acute pain. Pain 2012;153(5):967-73.
9. Ritsema TS, Kelen GD, Pronovost PJ, Pham JC. The national trend in quality of emergency department pain management for long bone fractures. Acad Emerg Med 2007;14(2):163-9.
10. Brown JC, Klein EJ, Lewis CW, Johnston BD, Cummings P. Emergency department analgesia for fracture pain. Ann Emerg Med 2003;42(2):197-205.
11. IOM (Institute of Medicine). Relieving pain in America: a blueprint for transforming prevention, care, education, and research. Washington, DC: The National Academies Press; 2011.
12. Phillips DM. Joint commission on accreditation of healthcare organizations pain management standards are unveiled. J Am Med Assoc 2000;284:428-9.
13. American Pain Society. Principles of analgesic Use in the treatment of acute pain and cancer pain. 4. Glenview, IL: American Pain Society; 1999.
14. World Health Organization. WHO’s pain relief ladder. www.who.int/cancer/ palliative/painladder/en/; 2009. (Accessed April 2, 2013).
15. American Pain Society Quality of Care Committee. Quality improvement guidelines for the treatment of acute pain and cancer pain. JAMA 1995;274: 1874-80.
16. American College of Emergency Physicians. Optimizing the treatment of pain in patients with acute presentations. http://www.acep.org/Clinical-Practice- Management/Optimizing-the-Treatment-of-Pain-in-Patients-with-Acute- Presentations/; 2009.
17. Phillips DM. JCAHO pain management standards are unveiled. Joint Commission on Accreditation of Healthcare Organizations. JAMA 2000;284:428-42.
18. Van Zee Art. The promotion and marketing of OxyContin: commercial triumph, public health tragedy. Am J Public Health 2009;99(2):221.
19. Kuehn BM. Opioid prescriptions soar: increase in legitimate use as well as abuse. JAMA 2007;297:249-51.
20. Alexander GC, Kruszewski SP, Webster DW. Rethinking opioid prescribing to protect patient safety and public health. JAMA 2012;308:1865-6.
21. Manchikanti L, Fellows B, Ailinani H, Pampati V. therapeutic use, abuse, and nonmedical use of opioids: a ten-year perspective. Pain Phys 2010;13(5): 401-35.
22. CDC. Vital signs: prescription painkiller overdoses in the US. http://www.cdc.gov/ vitalsigns/PainkillerOverdoses/index.html. November 2011. (Accessed April 2,

2013).

1. CDC. CDC grand rounds: prescription Drug overdoses-a U.S. epidemic. MMWR 2012;61(01):10-3http://www.cdc.gov/mmwr/preview/mmwrhtml/mm6101a3. htm. (Accessed April 2, 2013).
2. Schappert SM, Rechtsteiner EA. Ambulatory medical care utilization estimates for 2007. Vital Health Stat 2011;13(169):1-31.
3. Rockett IRH, Putnam SL, Jia H, Smith GS. Assessing substance abuse treatment need: a statewide hospital emergency department study. Ann Emerg Med 2003; 41(6):802-13.
4. Rockett IR, Putnam SL, Jia H, Chang CF, Smith GS. Unmet substance abuse treatment need, health services utilization, and cost: a population-based emergency department study. Ann Emerg Med 2005;45(2):118-27.
5. SAMHSA. Highlights of the 2010 Drug Abuse Warning Network (DAWN) findings on drug-related emergency department visits: the DAWN report. July 2012. http:// www.samhsa.gov/data/2k12/DAWN096/SR096EDHighlights2010.htm; 2012.

(Accessed April 2, 2013).

1. Cantrill SV, Brown MD, Carlisle RJ, Delaney KA, Hays DP, Nelson LS, et al. Whitson RR; Clinical policy: critical issues in the prescribing of opioids for adult patients in the emergency department. Ann Emerg Med 2012;60:499-525.
2. American Society of Anesthesiologists. Practice guidelines for acute pain management in the perioperative setting. An updated report by the American Society of Anesthesiologists Task Force on acute pain management. Anesthesi- ology 2012;116:248-73.
3. McCaig LF, Burt CW. Understanding and interpreting the National Hospital Ambulatory Medical Care Survey: key questions and answers. Ann Emerg Med 2012;60(6):716-21.
4. NCHS. NHAMCS 2010 micro-data file documentation. ftp://ftp.cdc.gov/pub/ Health_Statistics/NCHS/Dataset_Documentation/NHAMCS/doc2010.pdf . (Accessed April 2, 2013).
5. Cooper RJ. NHAMCS: does it hold up to scrutiny? Ann Emerg Med 2012;60(6): 722-5.
6. Flash Code Medical Coding Software. http://flashcode.com (Accessed January 2, 2013).
7. Pain Intensity Instruments. National Institutes of Health-Warren Grant Magnuson Clinical Center. http://painconsortium.nih.gov/pain_scales/NumericRatingScale. pdf; 2003. (Accessed April 2, 2013).
8. Platts-Mills TF, Esserman DA, Brown DL, Bortsov AV, Sloane PD, McLean SA. Older US emergency department patients are less likely to receive pain medication than younger patients: results from a national survey. Ann Emerg Med 2012;60(2): 199-206.
9. Office of the Mayor. Mayor Bloomberg, Deputy Mayor Gibbs and Chief Policy Advisor Feinblatt announce new emergency room guidelines to prevent opioid prescription painkiller abuse. NYC.gov. http://www.nyc.gov/portal/site/nycgov/ menuitem.c0935b9a57bb4ef3daf2f1c701c789a0/index.jsp?pageID= mayor_press_release&catID=1194&doc_name=http%3A%2F%2Fwww.nyc. gov%2Fhtml%2Fom%2Fhtml%2F2013a%2Fpr015-13.html&cc=unused1978&rc= 1194&ndi=1 . January 10, 2013. Accessed July 17, 2013.
10. Bird J. Hospitals toughen ED limits on painkiller prescriptions. Fierce Healthcare. http://www.fiercehealthcare.com/story/hospitals-toughen-ed-limits-painkiller- prescriptions/2013-02-25; 2013.
11. Ciccone DS, Just N, Bandilla EB, Reimer E, Ilbeigi MS, Wu W. Psychological correlates of opioid use in patients with chronic nonmalignant pain: a preliminary test of the Downhill spiral hypothesis. J Pain Symptom Manage 2000;20:180-92.
12. Harden N. Chronic opioid therapy: another reappraisal. American Pain Society Bulletin: Pain and. public policy 2002;12(1).
13. Vogt MT, Kwoh CK MT, Cope DK, Osial TA, Culyba M, Starz TW. Analgesic usage for low back pain: impact on health care costs and service use. Spine 2005;30(9): 1075-81.
14. Iyer S, Davis KL, Candrilli S. Opioid use patterns and health care resource utilization in patients prescribed opioid therapy with and without constipation. Manag Care 2010;19(3):44-51.