a b s t r a c t

Background: The rise of antibiotic-resistant pathogens is believed to have influenced the emergency department (ED) epidemiology and management of infectious diseases (IDs) since 2000.

Methods: Data from the National Hospital Ambulatory Medical Care Survey from 2000 to 2010 were used to ex- amine temporal trends in the incidence of IDs presenting to EDs. Outcome measures included national visit rates, visit proportions, and antimicrobial prescriptions for all ID primary diagnoses, as well as for specific organ sys- tems of interest (respiratory tract, skin/soft tissue, and urinary tract).

Results: An ID-related primary diagnosis was given in 18.3% (95% confidence interval, 17.9%-18.8%) of all ED visits during the study period. The hospitalization rate for these conditions is 7.8% (95% confidence interval, 7.3%-8.3%). The share of macrolide prescriptions for Upper respiratory tract infections and lower respiratory tract infections increased by 34% and 46%, respectively, and that of quinolone prescription for lower respiratory tract infections doubled from 9% to 18.4% during the study period. Management of skin and soft tissue infections shifted from predominant use of cephalosporins to sulfonamides. For UTIs, quinolones were most commonly prescribed, with an increasing use of third-generation cephalosporins. Antibiotics were more frequently prescribed to pa- tients who are white compared with (white: 60%, black: 57%, other races: 52%, P b .05).

Conclusion: The changing epidemiology of IDs diagnosed in US EDs reflects national trends in emerging patho- gens and drug resistance. broad-spectrum antibiotics are being prescribed at increasing rates. There are signifi- cant demographic disparities in nationwide antibiotic prescription practices.

(C) 2016

Introduction

Emergency departments (EDs) are an important venue to study trends in infectious disease (ID) epidemiology and management [1,2]. EDs have previously been shown to carry a disproportionate burden of infectious complaints and are often the first venue to identify and man- age emerging outbreaks [1]. Among ID complaints in US EDs, the most common categories are upper respiratory tract infections (URIs), lower respiratory tract infections (LRIs), Skin and soft tissue infections , and Urinary tract infections [3].

The management of these infections has substantially changed over the past 15 years. For example, bacterial causes of Respiratory infections have shown emerging resistance to first-line antibiotics [4]; meanwhile,

? Authors’ contributions: Y-HH designed the study. Y-HH performed data analyses. Y- HH, AMM, and AFD interpreted results. YH-H primarily designed tables and figures. AMM primarily drafted the manuscript. Y-HH and AFD assisted with literature review and critical editing of the manuscript.

* Corresponding author at: Johns Hopkins University Department of Emergency Medi- cine 5801 Smith Ave, Suite 3220 Davis Bldg, Baltimore, MD 21209. Tel.: +1 410 735 6413; fax: +1 410 735 6425.

E-mail address: [email protected] (Y.-H. Hsieh).

viral respiratory infections have resulted in numerous major outbreaks, including the 2003 SARS epidemic and the 2009 H1N1 influenza pandem- ic. Similarly, the epidemiology and management of SSTIs are believed to have changed since the rise of community-associated methicillin- resistant Staphylococcus aureus (CA-MRSA) in the first decade of this cen- tury [5-8]. New resistance patterns have also been noted in UTIs, particu- larly with regards to ?-lactams and fluoroquinolones [9,10].

In the wake of increasing resistance, health care settings, including

EDs, have adopted strategies of antibiotic stewardship, such as formu- lary restrictions, policies of prior approval, and provider education [11]. These efforts take into account local resistance patterns because, both globally and within the United States, there are significant Regional variations in antibiotic resistance [12]. There are also regional variations in provider prescription practices for common infections [3]. However, few studies have examined individual demographic disparities in pre- scription practices for commonly treated infections.

The primary objective of this study is to characterize the temporal trend of ED visits from 2000 to 2010 for all ID-related diagnoses, with an added focus on the most common organ systems of infections (URIs, LRIs, SSTIs, and UTIs). A secondary objective is to characterize an- timicrobial prescription practices for each of these classes and relate them to the described epidemiology. Finally, we aimed to explore

http://dx.doi.org/10.1016/j.ajem.2016.03.024

0735-6757/(C) 2016

whether demographic and regional disparities accounted for variability in Antibiotic prescriptions.

Methods

Study design

This study examines a cohort of ED visits from 2000 to 2010 using the National Hospital Medical Care Survey (NHAMCS). For each year, we evaluated the total number of ED visits, the number of infection- related ED visits, as well as the basic demographics and antibiotic pre- scribing practices associated with each visit. This study was granted ex- empt status for human subjects research by Johns Hopkins University School of Medicine Institutional Review Board.

Study setting

The National Hospital Medical Care Survey is a nationally represen- tative survey conducted by the Centers for Disease Control and Preven- tion and the National Center for Health Statistics (NCHS) [13]. It uses probability sampling of ED visits in 50 states and District of Columbia, excluding federal, military, and Veterans Administration hospitals. Basic ED visit information, including demographic (eg, age, sex, race) and clinical information (eg, reason for visit, prescriptions, ED diagno- ses, and disposition) of sampled ED visits, was recorded in a standard- ized patient record form by trained staff.

Before 2006, antimicrobials were coded based on the drug “as en- tered” on the NHAMCS patient record form using NCHS-developed “drug entry code” and “generic code” classification as well as the FDA’s National Drug Code Directory (NDCD) [14]. After 2006, drugs were coded using the “drug entry code” described above and coded based on their generic components and therapeutic classes using Multum Lexicon Drug Database. Multum classification categorized drugs into 3 therapeutic class schemes. Antimicrobials were coded as “anti-infectives” at level 1 with 20 second-level codes for major antimi- crobial drug class (eg, cephalosporins) and 36 third-level codes for sub- category class (eg, third-generation cephalosporins). NCHS has provided guidance to transform NDCD before 2006 to the correspond- ing codes in the Multum Database to compare the drug prescription trend before and after the drug Coding system change [15]. Data regard- ing ED diagnoses are coded according to the International Classification of Disease, Ninth Revision (ICD-9). We classified the primary ED diagno- sis of each ED visit as infectious or noninfectious based on a previously described protocol (eTable 1 in the Supplement) which was restricted to those diagnoses that are always or almost always caused by infectious agents [16]. Thus, codes for sepsis and severe sepsis are not included be- cause they have less specificity as to the cause or organ system involved and can include terms of noninfectious diagnoses [17].

Statistical analysis

Adjusted Sample weights, strata, and primary sampling unit design variables provided by NHAMCS were included in all analyses using the SAS 9.3 SURVEYFREQ procedure (SAS Institute Inc, Cary, NC) which accounted for the 4-stage probability sampling scheme. These provided an unbiased national estimate of US ED visits if it was based on more than 30 cases in the sample data or if the relative standard error was greater than 30%. Results are reported as weighted frequencies, percent- ages, and 95% confidence intervals (CIs) for individual characteristics of interest. The visit rate per 1000 US civilian noninstitutionalized persons, which describes the amount of infectious diagnoses in the general pop- ulation, could serve as a surrogate for the incidence of ID complaints among EDs throughout the United States. US population estimates of the civilian noninstitutionalized population used in computing annual visit rates for NHAMCS by selected demographics were provided by NCHS based on US Bureau of the Census estimates. We also reported

the visit proportion, which reflects the contribution of these visits rela- tive to the other diseases. The visit proportion is valuable because it is independent of global changes in ED volume and utilization.

The national annual overall visit rates as well as sex-, age-, race-, and region-specific rates for IDs per 1000 US civilian, noninstitutionalized population were estimated using procedures for multistage survey data and US census data provided in NHAMCS documentation for public users of the NHMACS. The significance of differences in the ID-related ED visit rates or antibiotics prescription rate by specific demographic group in the data set was assessed using the ?² test. All percentages pre- sented are weighted percentages. A multivariate logistic regression was performed to determine whether demographic characteristics (age group, sex, race, ethnicity, payment type, year of visit) of an ED visit or ED characteristics (geographical region, metropolitan statistical area, and type of hospital ownership) were associated with receipt of an an- tibiotic prescription for each ED diagnosis of interest after controlling for covariates. P b .05 was considered statistically significant.

Results

Among 1,290,774,445 total ED visits from 2000 to 2010 (95% CI, 1,174,165,294-1,407,383,596), IDs accounted for the primary ED diag- nosis in 236,291,402 visits (95% CI, 213,269,698-259,313,106), corre-

sponding to 18.3% (95% CI, 17.9%-18.8%) of all ED visits or 74 (95% CI, 67-81) ED visits per 1000 US civilian persons. As a group, 18.5 million (95% CI, 16.7-20.2 million) of these patient visits were admitted to hos- pital, which correspond to a hospitalization rate of 7.8% (95% CI, 7.3%- 8.3%). Hospitalizations differed based on Infection type: 19.1% of LRIs, 12.6% of UTIs, 11.1% of SSTIs, and 1.1% of URIs (P b .001). As compared with the Northeast, patients presenting to the South region were more likely to be hospitalized for URIs (OR, 2.02; 95% CI, 1.31-3.12) and LRIs (OR, 1.63; 95% CI, 1.34-1.99). white race was found to be a predictor in hospitalizations only for SSTIs (OR, 0.74; 95% CI, 0.60-0.93) when compared with other races.

Upper respiratory tract infections, LRIs, SSTIs, and UTIs together accounted 71.8% of all IDs visits (95% CI, 71.1%-72.5%). Visit rates per 1000 US civilian persons stratified by age, sex, race, ethnicity, and region are shown in Table 1, and proportions of ID-related US ED visits per 1000 ED visits by these demographics are shown in eTable 2 in the Sup- plement. Female patients had a significantly higher rate and proportion of ID-related visits than male patients (both P b .05). black patients ex- perienced a significantly higher share of diagnoses in US civilian popu- lation and in all ED visits compared with white patients (both P b .05). Visit rates also significantly varied by geographical region.

Overall ED visits increased from 108 million in 2000 to 130 million in

2010. The annual visit rate for all ID-related diagnoses remained stable between 71 and 77 per 1000 civilian population from 2000 to 2008; however, the rate increased by almost 20% to 88 per 1000 persons in 2009 before decreasing back to 77 per 1000 persons during 2010. As a proportion of all ED visits, ID-related complaints were stable between 17.2% and 18.7% for most years of the study period. The 2 years with higher ID-related visit proportions were in 2003 (19.2%) and 2009 (19.5%). Temporal trends are shown for each of the ID categories of in- terest in Fig. 1. During 2009, URIs and LRIs increased by 17% and 21%, re- spectively. The rate of ED visits for SSTIs also shows an increase during the past decade, with a higher rate of increase between 2000 and 2004 (80%; from 6.4 to 11.5 per 1000 persons) compared with 2005-

2010 (18%; from 11.4 to 13.5 per 1000 persons).

Trends of antimicrobial prescriptions in EDs for each of the infectious categories of interest are shown in Fig. 2A-D. Patients with URIs were most commonly prescribed penicillins, and patients with LRIs were most commonly prescribed macrolide antibiotics. The use of macrolides for URIs and LRIs increased by 34% and 46%, respectively, during the study period. The use of quinolone antibiotics for LRIs doubled during the study period (from 9.0% to 18.4%). A total of 1.58 million (95% CI, 1.08-2.09 million) ED visits for LRIs were prescribed antivirals, which

Table 1

Rates of ID-related US ED visits per 1000 US civilian, noninstitutional population, 2000-2010

ED visit rates per 1000 population (95% CI)

Characteristics	Categories	Total ID	URI	LRI	SSTI	UTI
Overall		74 (67-81)	21 (19-23)	14 (13-16)	10 (9-11)	8 (7-9)
Age	0-4	264 (215-313)?	90 (79-101)	39 (34-43)	10 (8-12)	6 (5-7)
	5-14	71 (58-83)	28 (25-31)	8 (7-10)	6 (5-7)	3 (3-4)
	15-24	85 (72-99)	27 (24-30)	11 (9-12)	13 (11-15)	14 (12-15)
	25-34	73 (62-84)	21 (19-23)	12 (11-13)	13 (11-15)	10 (9-11)
	35-44	51 (43-59)	12 (11-13)	12 (10-13)	11 (9-12)	6 (6-7)
	45-54	41 (35-48)	8 (7-9)	11 (10-12)	10 (9-11)	5 (5-6)
	55-64	35 (28-41)	6 (5-7)	11 (9-12)	8 (7-9)	5 (4-6)
	65-74	41 (32-49)	4 (3-5)	15 (13-17)	8 (7-10)	9 (7-10)
	>= 75	74 (61-87)	5 (4-6)	31 (27-34)	10 (9-12)	21 (19-24)
Sex	Female	82 (64-90)?	23 (20-25)	15 (14-17)	9 (8-11)	13 (11-14)
	Male	65 (59-72)	19 (17-21)	13 (12-14)	11 (10-12)	3 (3-4)
Race	White	65 (58-71)	18 (16-20)	13 (11-14)	9 (8-10)	7 (7-8)
	Black	152 (133-172)?	47 (40-53)	26 (22-29)	20 (16-23)	15 (13-17)
	Other	39 (32-46)	11 (9-13)	7 (6-9)	4 (3-5)	4 (3-5)
Ethnicity	Hispanic	80 (67-93)	24 (20-28)	13 (11-15)	9 (7-11)	8 (6-9)
	Non-Hispanic	73 (65-80)	20 (18-23)	14 (13-16)	10 (9-12)	8 (7-9)
Region	Northeast	69 (59-80)	20 (17-23)	12 (10-14)	9 (7-11)	7 (6-8)
	Midwest	74 (60-88)	21 (17-25)	14 (11-17)	9 (6-11)	8 (6-10)
	South	88 (73-103)?	25 (21-30)	17 (14-20)	13 (10-15)	10 (8-11)
	West	56 (43-69)	14 (11-18)	11 (8-14)	8 (6-10)	6 (5-8)
* P b .05.

correspond to 3.5% (95% CI, 2.5-4.5%) of all visits for an LRI diagnosis. The proportion of antiviral prescription for LRIs more than doubled in 2009 (9.3%) as compared with 2008 (4.2%). Patients with SSTIs were most commonly prescribed cephalosporins at the start of the decade, but by 2007, these were overtaken by sulfonamides. A total of 4.9% of patients received treatment with both cephalosporins and sulfonamides for SSTIs. The Prescription rates for UTIs remained relatively stable, with fluoroquinolones continuing to be the most commonly prescribed class of antimicrobials.

Table 2 shows the antibiotics prescription rates for each category of infection stratified by demographic characteristics. There were signifi- cant differences in the antibiotic prescription rates based on age, race, ethnicity, insurance status, and region for specific categories of infec- tions after adjustment for covariates, as shown in Table 3. For URIs, white patients were more likely to be prescribed antibiotics than other races (OR, 1.38; 95% CI, 1.24-1.53), and non-Hispanic patients were more likely to receive antibiotics than Hispanic patients (OR, 1.27; 95% CI, 1.12-1.43). Antibiotics for SSTI visits were also more fre- quently prescribed to patients of white and non-Hispanic race. Female patients were less likely to be prescribed antibiotics when they had URIs or LRIs as compared with male patients. prescription patterns by region varied, with the South and Midwest regions having higher rates of prescriptions for URIs and the South and West regions having higher rates of prescriptions for SSTIs. Regarding medical insurance

Fig. 1. ED visit rate (per 1000 US population) of IDs: 2000-2010.

type, rate of antibiotic prescription for patients with private, public, or self-pay did not differ from each other in all 4 IDs of interest but was sig- nificantly higher than those with other or unknown insurance payment type for UTIs (private, public, and self-pay), SSTIs (private and public), and LRIs (public and self-pay).

Discussion

ID epidemiology

Acute infections continue to be one of the most common reasons for Americans to present to the ED. Our results of the ED visit and hospital- ization rates for infections are most likely conservative estimates be- cause we excluded primary ICD-9 codes for sepsis and severe sepsis. Using an inpatient database, one study noted increasing rate of ID- related hospitalizations between 1998 and 2006 [18,19]. That study demonstrated demographic variations in ID diagnoses similar to what we report here: specifically, a higher burden of disease in females, blacks, and Pediatric populations. The vast majority (N 70%) of ID diag- noses in our data set were comprised of URIs, LRIs, SSTIs, and UTIs.

Respiratory infections

Respiratory infections continue to have a high burden of disease in the ED population. Our analysis suggests that a surge in URIs and LRIs in 2009 was the primary driver of an overall increase in ID-related diag- noses during those years. This surge may be attributed to the 2009 H1N1 influenza A pandemic. A smaller increase in ID-related complaints as a proportion of all ED visits was also evident in 2003, the year of the SARS epidemic. Recent studies have shown that antibacterial agents are generally overused in the ambulatory setting for specific URI diagnoses and that their use is higher in EDs compared with outpatient clinics [20-22]. Our results show that 53% of URIs were prescribed antibiotics, with penicillins being the most common, followed by macrolides. This suggests a high frequency of prescriptions for a set of conditions thought to be predominately viral in etiology [23]. ED management of URIs thus represents a potential point of intervention for ED-based an- tibiotic stewardship programs to minimize unnecessary and inappro- priate antibiotic use to curb emerging resistance per choosing wisely campaign [24,25].

Our study also examined the prescription practices for LRIs, which include viral and Bacterial pneumonia. A prior study found a more than 5-fold increase in the use of quinolones and macrolides for respira- tory complaints from 1995 to 2006 [26]. These classes of antibiotics con- tinued to be prescribed at high frequencies by ED providers for LRIs, with quinolone use doubling during our study period. Our results also show that the rate of antiviral prescriptions increased from 2004 until 2009, its peak year, which corresponds to the H1N1 influenza A pan- demic. In 2009, antiviral prescriptions comprised almost 9.3% of all an- timicrobial prescriptions for LRIs.

Skin and soft tissue infections

The microbial epidemiology of SSTIs has seen a dramatic change over the past decade because of increased prevalence of CA-MRSA [7,27]. Controlling the rising rates of CA-MRSA SSTIs has been extraordi- narily difficult on a national level, especially as it appears that its spread has converged with epidemics of other infections, including MRSA bac- teremia and HIV [28,29]. However, our results show that the rate of ED visits for SSTIs, although still increasing, was rising at a slower rate be- tween 2005 and 2010 when compared with 2000-2004. Management of SSTIs in the ED depends on knowledge of local and national resistance patterns because ED providers often have to provide targeted Empiric therapy without the availability of culture results or certitude of patient follow-up [7]. As the decade progressed, ED providers transitioned from most commonly prescribing early-generation cephalosporins-which

have activity against Streptococcus and methicillin-susceptible S aureus-to sulfa antibiotics, such as trimethoprim-sulfamethoxazole, which are effective against CA-MRSA. A total of 4.9% of patients with SSTIs received treatment with both cephalosporins and sulfonamides. Presumably, this represents empiric coverage of both streptococcal and MRSA species in patients with SSTIs.

Urinary tract infections

Another commonly diagnosed and treated infection in EDs con- tinues to be UTIs. Eighty-four percent of ED patients with UTIs received an antibiotic prescription, a proportion that was relatively stable across demographic and regional variables. Fluoroquinolones continued to be the most commonly prescribed antibiotic class, with increasing use through the first half of the study period followed by a decline between 2007 and 2010 from 45.8% to 36.8% as a share of all antibiotics. Sulfon- amides represented the second most commonly prescribed antibiotics for UTIs until the latter 3 years of the decade when they were overtaken by cephalosporins, which included a sharp increase in the use of third- generation cephalosporins in 2010.

The high proportion of fluoroquinolone use for UTIs is a trend con- tinued from decades prior, as other studies have shown [30]. A recent analysis of Escherichia coli isolates collected nationwide showed that an- tibiotic resistance to quinolones increased substantially during our study period, reaching 17.1% for ciprofloxacin by 2010. [31] It could be for this reason that the empiric use of fluoroquinolones begins to

A CEPHALOSPORINS MACROLIDES PENICILLINS QUINOLONES 3rd Gen CEPHALOSPORINS

35

30

25

Percentage (%)

20

15

10

5

0

2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010

Year

CEPHALOSPORINS MACROLIDES PENICILLINS QUINOLONES ANTIVIRALS 3rd Gen CEPHALOSPORINS

B

40

35

30

25

Percentage (%)

20

15

10

5

0

2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010

Year

Fig. 2. Trends of antimicrobial prescriptions in EDs for each of the infectious categories of interest.

LINCOMYCIN CEPHALOSPORINS PENICILLINS

SULFONAMIDES 1st Gen CEPHALOSPORINS MISCELLANEOUS ANTIMICROBIALS

60

50

40

Percentage (%)

30

20

10

0

2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010

Year

URINARY ANTI-INFECTIVES CEPHALOSPORINS QUINOLONES

50

SULFONAMIDES

3rd Gen CEPHALOSPORINS

45

40

35

Percentage (%)

30

25

20

15

10

5

0

2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010

Year

Fig. 2 (continued).

decline in the latter years of the decade in favor of broader antibiotics, such as third-generation cephalosporins. As a percentage of all antibi- otics used for UTIs, third-generation cephalosporins increased from 14.3% to 23.3% during the end of the study period.

Demographic disparities

Prior studies have shown that infections seem to disproportionately affect racial and Ethnic minorities [32,33]. Our study demonstrates that this was true for the US ED population as well. However, we also ob- serve regional and demographic disparities that independently underlie antibiotic prescription practices in the multivariate regression analysis. ED clinical providers in the South region were more likely to prescribe antibiotics for URIs and SSTIs. White patients were more likely to re- ceive antibiotic prescriptions for all URIs and SSTIs when compared with patients of other races. Together, for those with URIs, white ED pa- tients in the South had the highest odds (OR, 2.61) to receive antibiotics, followed by other race patients in the South (OR, 1.89) and white pa- tients in the Northeast (OR, 1.38), as compared with other race patients in the Northeast (the reference group). However, we did not have enough information to explain the finding that EDs in the South are more likely to prescribe antibiotics because we used secondary data

from NHAMCS and did not collect direct relevant information to test this hypothesis.

Racial and Ethnic disparities have been a topic of heightened nation- al interest for many years [34]. In the ED setting, prior research has shown that minority race patients are less likely to receive analgesic prescriptions for pain control [35,36]. However, the literature is more limited with regard to demographic differences in treatment of ID com- plaints. Studies investigating the receipt of antibiotics in pediatric ED patients presenting with acute otitis media have found varying results [37,38]. A study analyzing inpatients with pneumonia did not find racial disparities in antibiotic use [39]. Our study demonstrates racial and eth- nicity disparities in antibiotic prescription practices for URI- and SSTI- related ED visits, independently from other confounders, for example, age, sex, region, and medical insurance payment type. It is possible that disparities can be explained by racial disparities in other aspects of medical care. For example, inferior access to primary care among mi- nority populations [40] may increase their likelihood of having a low- acuity infectious complaint present to the ED. Similarly, they are more likely to present to the ED with asthma exacerbations [41], which may be concomitantly diagnosed as viral URIs and not prescribed antibiotics. Using the available data, it is not possible to determine whether our finding represents an undertreatment of minority populations, over- treatment of white patients, or a combination thereof.

Table 2

Comparison of demographic characteristics of patients receiving antibiotic prescriptions in ED visits, 2000-2010

Percentage of ED ID-related visits of receiving antibiotic prescriptions (95% CI)

Characteristics	Categories	Total ID	URI	LRI	SSTI	UTI
Overall		59 (58-60)	53 (51-55)	66 (64-68)	71 (69-72)	84 (83-85)
Age	0-4	49 (48-51)	35 (32-39)	57 (54-60)	65 (59-70)	84 (79-89)
	5-14	55 (53-56)	55 (52-58)	58 (53-63)	69 (64-74)	86 (82-90)
	15-24	63 (62-65)	64 (62-67)	63 (59-67)	74 (71-77)	87 (85-89)
	25-34	63 (62-65)	63 (60-65)	67 (65-71)	73 (70-76)	86 (83-89)
	35-44	64 (63-66)	63 (59-67)	67 (65-70)	73 (70-75)	86 (84-89)
	45-54	65 (63-67)	61 (57-65)	71 (67-75)	72 (68-75)	83 (80-87)
	55-64	65 (63-68)	64 (58-69)	72 (68-76)	71 (66-75)	80 (75-85)
	65-74	68 (65-71)	62 (54-70)	72 (68-76)	69 (64-74)	82 (77-87)
	>= 75	67 (65-69)	55 (47-63)	71 (67-74)	63 (58-67)	78 (75-82)
Sex	Female	60 (59-61)?	54 (51-56)	65 (63-67)	70 (68-72)	85 (83-86)
	Male	59 (57-60)	53 (50-55)	66 (64-68)	72 (70-74)	82 (79-84)
Race	White	60 (59-62)	55 (53-57)	66 (64-68)	72 (70-74)	84 (82-85)
	Black	57 (56-59)?	49 (47-52)	66 (63-68)	69 (66-72)	85 (83-88)
	Other	52 (49-55)	44 (39-50)	65 (59-71)	58 (49-66)	87 (81-92)
Ethnicity	Hispanic	54 (52-56)	43 (40-46)	64 (61-67)	67 (63-71)	85 (82-87)
Region	Northeast	54 (52-56)	44 (41-46)	66 (61-71)	67 (64-70)	84 (82-87)
	Midwest	59 (56-61)	53 (49-56)	66 (62-70)	68 (65-71)	81 (79-84)
	South	63 (61-64)?	60 (57-63)	65 (62-68)	73 (71-76)	85 (83-87)
	West	58 (56-60)	46 (43-49)	66 (63-70)	72 (69-75)	85 (82-88)
Payment type	Private	60 (58-61)	55 (53-58)	65 (62-68)	73 (71-75)	86 (84-87)
	Public	59 (58-60)	50 (47-52)	67 (65-69)	70 (68-72)	83 (81-85)
	Self-pay	63 (61-65)	59 (56-66)	67 (64-70)	72 (69-75)	88 (85-90)
	Other/unknown	55 (51-58)	49 (44-54)	61 (57-66)	66 (62-70)	78 (73-84)
Hospital Ownership	Voluntary nonprofit	59 (58-60)	51 (49-53)	66 (64-68)	71 (69-73)	84 (83-86)
	Government, nonfederal	60 (56-63)	58 (53-64)	63 (60-67)	69 (66-73)	83 (79-87)
	Proprietary	63 (61-65)	58 (55-62)	68 (62-75)	72 (69-76)	86 (83-89)
* P b .05.

Limitations

This study has several limitations. Consistent with previous studies, we chose not to use primary ICD-9 codes related to sepsis or septic shock. Although this likely increased the specificity of our cases, we be- lieve that our findings underestimate the prevalence and severity of ID

complaints [42]. Because we use the NHAMCS database, we depend upon the integrity of codified diagnoses given by ED providers at the time of the ED visit, which can have variable reliability [43,44]. There are no uniform case definitions for the diagnoses between providers, and there is no assurance that the diagnoses were confirmed with any diagnostic testing. As such, we are not able to determine which, if any,

Table 3

Factors associated with increasing odds of receiving antibiotic prescription during an ED visit with an ED ID diagnosis, United States, 2000-2010

		Odds ratios (95% CI)
Characteristics	Categories	URI	LRI	SSTI	UTI
Age	0-4	0.48 (0.33-0.69)?	0.54 (0.44-0.66)?	1.17 (0.86-1.60)	1.45 (0.92-2.28)
	5-14	1.05 (0.74-1.50)	0.57 (0.44-0.74)?	1.38 (1.00-1.90)	1.69 (1.10-2.62)?
	15-24	1.64 (1.13-2.37)?	0.73 (0.57-0.93)?	1.82 (1.35-2.45)?	1.75 (1.31-2.34)?
	25-34	1.55 (1.06-2.26)?	0.92 (0.74-1.15)	1.68 (1.25-2.28)?	1.68 (1.21-2.35)?
	35-44	1.52 (1.02-2.26)?	0.88 (0.72-1.09)	1.72 (1.30-2.26)?	1.74 (1.25-2.41)?
	45-54	1.45 (0.99-2.12)	1.02 (0.79-1.32)	1.58 (1.18-2.13)?	1.25 (0.91-1.73)
	55-64	1.55 (1.07-2.26)?	1.12 (0.88-1.42)	1.49 (1.09-2.05)?	1.04 (0.70-1.54)
	65-74	1.43 (0.92-2.22)	1.06 (0.85-1.33)	1.33 (0.96-1.83)	1.27 (0.90-1.79)
	>= 75	1.00	1.00	1.00	1.00
Sex	Female	0.88 (0.81-0.96)?	0.90 (0.83-0.98)?	0.91 (0.81-1.02)	1.12 (0.92-2.28)
	Male	1.00	1.00	1.00	1.00
Race	White	1.38 (1.24-1.53)?	0.96 (0.85-1.09)	1.30 (1.12-1.50)?	0.92 (0.74-1.14)
	Nonwhite	1.00	1.00	1.00	1.00
Ethnicity	Non-Hispanic	1.27 (1.12-1.43)?	1.01 (0.86-1.18)	1.29 (1.07-1.56)?	1.11 (0.85-1.44)
	Hispanic	1.00	1.00	1.00	1.00
Region	Northeast	1.00	1.00	1.00	1.00
	Midwest	1.40 (1.19-1.65)?	1.04 (0.80-1.35)	1.04 (0.87-1.25)	0.80 (0.62-1.04)
	South	1.89 (1.60-2.22)?	1.01 (0.80-1.28)	1.42 (1.18-1.71)?	0.97 (0.74-1.27)
	West	1.11 (0.92-1.34)	1.05 (0.82-1.34)	1.31 (1.09-1.57)?	1.07 (0.78-1.45)
Payment type	Private	1.17 (0.99-1.38)	1.20 (0.98-1.46)	1.40 (1.14-1.73)?	1.75 (1.21-2.53)?
	Public	1.15 (0.98-1.36)	1.25 (1.04-1.51)?	1.37 (1.08-1.74)?	1.59 (1.13-2.22)?
	Self-pay	1.18 (0.98-1.43)	1.31 (1.05-1.62)?	1.25 (1.00-1.56)	1.88 (1.24-2.84)?
	Other/unknown	1.00	1.00	1.00	1.00
Hospital Ownership	Voluntary nonprofit	0.80 (0.68-0.94)?	0.86 (0.66-1.12)	1.07 (0.87-1.31)	0.95 (0.69-1.30)
	Government, nonfederal	0.96 (0.98-1.43)	0.77 (0.58-1.03)	0.95 (0.75-1.20)	0.85 (0.59-1.22)
	Proprietary	1.00	1.00	1.00	1.00

multivariate logistic regression model which also adjusted for variables metropolitan statistical area and calendar year.

* Statistically significant.

antibiotics are overused in specific Clinical contexts. Similarly, our data do not lend themselves to conclusions regarding clinical outcomes for the IDs included in sampling. Analysis of antibiotic prescriptions is lim- ited by the number of medications the data set coded for each encounter (up to 6 drugs before 2003 and up to 8 drugs thereafter). Finally, be- cause our results are confined to the ED, we cannot draw conclusions re- garding the outpatient setting in general, where there are known differences in the care for ID-related conditions [3].

Conclusions

This study demonstrates that ED visits for ID-related diagnoses remained stable between 2000 and 2010, with increasing proportions of UTIs and SSTIs in the latter years of the decade. URIs and LRIs together experienced a surge in ED visits in 2009 and incur high frequencies of broad-spectrum antibacterial and Antiviral use. Antibiotics most often prescribed for SSTIs seem to have changed in concert with the CA- MRSA epidemic. Meanwhile, UTIs are most commonly treated with quinolones and, more recently, broad-spectrum cephalosporins. Racial and ethnic disparities may contribute to variations in prescription prac- tices, as minority race patients were less likely to be prescribed antibi- otics during this study period.

Appendix A. Supplementary data

Supplementary data to this article can be found online at http://dx. doi.org/10.1016/j.ajem.2016.03.024.

References

1. Talan D, Moran G, Mower W, et al. EMERGEncy ID NET: an emergency department- based emerging infections sentinel network. the EMERGEncy ID NET Study Group. Ann Emerg Med 1998;32(6):703-11.
2. Talan D, Moran G, Abrahamian F. Infectious disease emergencies, preface. Infect Dis Clin North Am 2008;22(1):ix-xi.
3. May L, Mullins P, Pines J. Demographic and treatment patterns for infections in am- bulatory settings in the United States, 2006-2010. Acad Emerg Med 2014;21(1): 17-24.
4. Hyde T, Gay K, Stephens D, Vugia D, Pass M, Johnson S, et al. Whitney CG, Active Bac- terial Core Surveillance/Emerging Infections Program Network. Macrolide resistance among invasive Streptococcus pneumoniae isolates. JAMA 2001;286(15):1857-62.
5. McCaig L, McDonald L, Mandal S, Jernigan D. Staphylococcus aureus-associated skin and soft tissue infections in ambulatory care. Emerg Infect Dis 2006;12(11): 1715-23.
6. Moran G, Krishnadasan A, Gorwitz R, et al. Methicillin-resistant S. aureus infections among patients in the emergency department. N Engl J Med 2006;355(7):666-74.
7. Pallin D, Egan D, Pelletier A, Espinola J, Hooper D, Camargo CJ. Increased US emer- gency department visits for skin and soft tissue infections, and changes in antibiotic choices, during the emergence of community-associated methicillin-resistant Staph- ylococcus aureus. Ann Emerg Med 2008;51(3):291-8.
8. Qualls M, Mooney M, Camargo CJ, Zucconi T, Hooper D, Pallin D. Emergency depart- ment visit rates for abscess versus other skin infections during the emergence of community-associated methicillin-resistant Staphylococcus aureus, 1997-2007. Clin Infect Dis 2012;55(1):103-5.
9. Karlowsky J, Kelly L, Thornsberry C, Jones M, Sahm D. Trends in antimicrobial resis- tance among urinary tract infection isolates of Escherichia coli from female outpa- tients in the United States. Antimicrob Agents Chemother 2002;46(8):2540-5.
10. Karlowsky J, Thornsberry C, Jones M, Sahm D. Susceptibility of antimicrobial- resistant urinary Escherichia coli isolates to fluoroquinolones and nitrofurantoin. Clin Infect Dis 2003;36(2):183-7.
11. Fishman N. Antimicrobial stewardship. Am J Infect Control 2006;34(5 Suppl 1): S55-63.
12. Thornsberry C, Sahm D, Kelly L, et al. Regional trends in antimicrobial resistance among clinical isolates of Streptococcus pneumoniae, Haemophilus influenzae, and Moraxella catarrhalis in the United States: results from the TRUST Surveillance Pro- gram, 1999-2000. Clin Infect Dis 2002;34(Suppl. 1):S4-S16.
13. National Center for Health Statistics. The National Ambulatory Medical Care Survey (NAMCS) description. http://www.cdc.gov/nchs/ahcd/ahcd_questionnaires.htm. [Accessed January 30, 2012].
14. Centers for Disease Control and Prevention. Ambulatory care drug database system. Ambul health care data 2011 [http://www.cdc.gov/nchs/ahcd/ahcd_database.htm. Accessed June 16, 2015].
15. Centers for Disease Control and Prevention. Trend analysis using NAMCS and NHAMCS drug data. Ambul Health Care Data 2010 [http://www.cdc.gov/nchs/ ahcd/trend_analysis.htm. Accessed June 16, 2015].
16. Armstrong G, Conn L, Pinner R. Trends in infectious disease mortality in the United States during the 20th century. JAMA 1999;281(1):61-6.
17. Pinner RW, Teutsch SM, Simonsen L, et al. Trends in infectious diseases mortality in the United States. JAMA 1996;275(3):189-93.
18. Christensen K, Holman R, Steiner C, Sejvar J, Stoll B, Schonberger L. Infectious disease hospitalizations in the United States. Clin Infect Dis 2009;49(7):1025-35.
19. Yorita K, Holman R, Sejvar J, Steiner C, Schonberger L. Infectious disease hospitaliza- tions among infants in the United States. Pediatrics 2008;121(2):244-52.
20. Donnelly J, Baddley J, Wang H. Antibiotic utilization for acute respiratory tract infections in U.S. emergency departments. Antimicrob Agents Chemother 2014;58(3):1451-7.
21. Neuman M, Shah S, Shapiro D, Hersh A. Emergency department management of childhood pneumonia in the United States prior to publication of national guide- lines. Acad Emerg Med 2013;20(3):240-6.
22. Shapiro D, Hicks L, Pavia A, Hersh A. Antibiotic prescribing for adults in ambulatory care in the USA, 2007-09. J Antimicrob Chemother 2014;69(1):234-40.
23. Gonzales R, Malone D, Maselli J, Sande M. Excessive antibiotic use for acute respira- tory infections in the United States. Clin Infect Dis 2001;33(6):757-62.
24. Jung N, Lehmann C, Fatkenheuer G. The “choosing wisely”: initiative in infectious diseases. Infection 2015.
25. Maughan B, Baren J, Shea J, Merchant R. Choosing wisely in emergency medicine: a national survey of emergency medicine academic chairs and division chiefs. Acad Emerg Med 2015;22(12):1506-10.
26. Grijalva C, Nuorti J, Griffin M. Antibiotic prescription rates for acute respiratory tract infections in US ambulatory settings. JAMA 2009;302(7):758-66.
27. Pallin D, Espinola J, Leung D, Hooper D, Camargo CJ. Epidemiology of dermatitis and skin infections in United States physicians’ offices, 1993-2005. Clin Infect Dis 2009; 49(6):901-7.
28. Popovich K, Weinstein R, Aroutcheva A, Rice T, Hota B. Community-associated methicillin-resistant Staphylococcus aureus and HIV: intersecting epidemics. Clin In- fect Dis 2010;50(7):979-87.
29. Tattevin P, Schwartz B, Graber C, et al. Concurrent epidemics of skin and soft tissue infection and bloodstream infection due to community-associated methicillin- resistant Staphylococcus aureus. Clin Infect Dis 2012;55(6):781-8.
30. Huang E, Stafford R. National patterns in the treatment of urinary tract infections in women by ambulatory care physicians. Arch Intern Med 2002;162(1):41-7.
31. Sanchez G, Master R, Karlowsky J, Bordon J. In vitro antimicrobial resistance of uri- nary Escherichia coli isolates among U.S. outpatients from 2000 to 2010. Antimicrob Agents Chemother 2012;56(4):2181-3.
32. Baker M, Barnard L, Kvalsvig A, et al. Increasing incidence of serious infectious dis- eases and inequalities in New Zealand: a national Epidemiological study. Lancet 2012;379(9821):1112-9.
33. Butler J, Crengle S, Cheek J, et al. Emerging infectious diseases among indigenous peoples. Emerg Infect Dis 2001;7(3 Suppl):554-5.
34. Institute of Medicine. Unequal treatment: confronting racial and ethnic disparities in Health Care (full printed version). Washington DC: The National Academies Press; 2003.
35. Hens J, Heins A, Grammas M, Costello M, Huang K, Mishra S. Disparities in analgesia and opioid prescribing practices for patients with Musculoskeletal pain in the emer- gency department. J Emerg Nurs 2006;32(3):219-24.
36. Tamayo-Sarver J, Hinze S, Cydulka R, Baker D. Racial and ethnic disparities in emer- gency department analgesic prescription. Am J Public Health 2003;93(12):2067-73.
37. Sidell D, Shapiro N, Bhattacharyya N. Demographic influences on antibiotic prescrib- ing for pediatric acute otitis media. Otolaryngol Head Neck Surg 2012;146(4):653-8.
38. Fleming-Dutra KE, Shapiro DJ, Hicks LA, Gerber JS, Hersh AL. Race, otitis media, and antibiotic selection. Pediatrics 2014;134(6):1059-66.
39. Frei C, Mortensen E, Copeland L, et al. Disparities of care for African-Americans and Caucasians with community-acquired pneumonia: a retrospective cohort study. BMC Health Serv Res 2010;10:143.
40. Bach P, Pham H, Schrag D, Tate R, Hargraves J. Primary care physicians who treat blacks and whites. N Engl J Med 2004;351(6):575-84.
41. Boudreaux E, Emond S, Clark S, Camargo CJ. acute asthma among adults presenting to the emergency department: the role of race/ethnicity and socioeconomic status. Chest 2003;124(3):803-12.
42. Filbin MR, Arias SA, Camargo Jr CA, Barche A, Pallin DJ. Sepsis visits and antibiotic uti- lization in U.S. emergency departments*. Crit Care Med 2014;42(3):528-35.
43. Aronsky D, Haug P, Lagor C, Dean N. Accuracy of administrative data for identifying patients with pneumonia. Am J Med Qual 2005;20(6):319-28.
44. Cadieux G, Tamblyn R. Accuracy of physician billing claims for identifying acute re- spiratory infections in primary care. Health Serv Res 2008;43(6):2223-38.