Original Contribution

Antibiotic prescribing for presumed nonbacterial acute respiratory tract infections^B

Sherrie L. Aspinall PharmD, MSc^a,b,c,?, Chester B. Good MD, MPH^a,b,d,

Joshua P. Metlay MD, PhD^e,f, Maria K. Mor PhD^a,g, Michael J. Fine MD, MSc^a,d

^aVA Pittsburgh Healthcare System, Center for Health Equity Research and Promotion, Pittsburgh, PA 15206, USA

^bVA Center for Medication Safety, Hines VA, Hines, IL 60141, USA ^cUniversity of Pittsburgh, School of Pharmacy, Pittsburgh, PA 15261, USA ^dUniversity of Pittsburgh, School of Medicine, Pittsburgh, PA 15261, USA

^ePhiladelphia VA Medical Center, Center for Health Equity Research and Promotion, Philadelphia, PA 19104, USA

^fUniversity of Pennsylvania School of Medicine, Philadelphia, PA 19104, USA

^gUniversity of Pittsburgh, Graduate School of Public Health, Pittsburgh, PA 15261, USA

Received 16 July 2007; revised 31 March 2008; accepted 10 April 2008

Abstract

Objective: The objective of the study was to identify patient and provider factors associated with prescribing antibiotics for outpatients with acute respiratory tract infections of likely nonbacterial etiology.

Methods: We identified outpatients who were diagnosed in the emergency department with nonspecific upper respiratory tract infections (URIs) and acute bronchitis at the VA Pittsburgh Healthcare System from June 15, 2003, to June 14, 2004, and the Philadelphia VA Medical Center from November 30, 2003, to March 31, 2004. Stepwise logistic regression was used to identify factors independently associated with antibiotic prescribing.

Results: Overall, 26% of the 667 eligible patients with URIs and/or acute bronchitis received antibiotics. Antibiotics were prescribed significantly more frequently for acute bronchitis at one site (97% vs 65%, P b .001). Using multivariable analysis, the following factors were independently associated with antibiotic prescribing (odds ratio, 95% confidence interval): presence of 1 or more comorbidities (2.1, 1.2-3.5), fever (2.5, 1.4-4.4), purulent sputum (2.5, 1.5-4.4), shortness of breath (2.8, 1.4-5.4), altered breath sounds (4.6, 2.4-8.6), diagnosis of acute bronchitis (15.9, 8.0-31.8), provider age

>=30 years (2.6, 1.1-6.3), and noninternal medicine specialty (2.7, 1.2-6.0).

Conclusions: Antibiotic use was high and varied substantially for URIs and acute bronchitis. Specific signs and symptoms, a diagnosis of acute bronchitis, and provider age and specialty were associated

^? Source of support: Dr Aspinall was a VA HSR&D Research Career Development Awardee (RCD 03-035) at the time of the study. Dr Fine was supported in part by a K-24 midcareer development award from the National Institute of Allergy and Infectious Diseases (5K24 A101769). Financial support was provided by a pilot grant through the Center for Health Equity Research and Promotion and an unrestricted educational grant from Pfizer, Inc. The Center for Health Equity Research and Promotion Intramural Review Committee and Pfizer, Inc, had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; and preparation, review, or approval of the manuscript.

Results were presented as a poster at the Health Services research and development National Meeting in Baltimore, MD, on February 17, 2005.

* Corresponding author. VA Pittsburgh Healthcare System, Center for Health Equity Research and Promotion (151C-H), Pittsburgh, PA 15206, USA. Tel.: +1 412 365 4969; fax: +1 412 365 4386.

E-mail address: [email protected] (S.L. Aspinall).

0735-6757/$ - see front matter. doi:10.1016/j.ajem.2008.04.015

with antibiotic prescribing. Interventions to decrease Inappropriate prescribing should address the perceived utility of antibiotics in acute bronchitis and the accuracy of signs and symptoms in diagnosing a bacterial respiratory infection.

Introduction

Although antibiotic prescribing for adults with acute respiratory tract infections in the outpatient setting has declined over the past 10 years, antibiotics are still overused for many infections with a predominantly viral etiology such as nasopharyngitis, acute bronchitis, and nonspecific upper respiratory tract infections (URIs) [1]. In a cross-sectional study of data from the National Ambulatory Medical Care Survey and the National Hospital Ambulatory Medical Care Survey , an antibiotic was prescribed in approximately 49% of the visits for these types of acute respiratory tract infections in 2001 to 2002. In addition, broad-spectrum antibiotics accounted for more than three quarters of all Antibiotic prescriptions [1]. Antibiotic over- utilization is problematic because previous antibiotic use is a risk factor for infection with several antibiotic-resistant pathogens, particularly drug-resistant Streptococcus pneu- monia, an organism that is the most common cause of community-acquired pneumonia, Otitis media, sinusitis, and Bacterial meningitis [2-4].

Numerous factors have been associated with antibiotic prescribing for nonbacterial acute respiratory tract infections in outpatient and emergency department (ED) settings, including patient (eg, age, race, expectations), provider (eg, type, specialty), and system (eg, time pressures, facility size) characteristics [5-17]. However, most prior studies involved outpatient visits to physicians’ offices. Of those studies evaluating antibiotic use in the ED, all used data from NHAMCS [5,9,17]. Therefore, a limited number of patient (age, race, ethnicity, sex), provider (type, geographic region), and visit (Insurance type, primary and secondary diagnoses, medications) characteristics with the potential to influence prescribing were available for the analyses using this data set. For example, clinical findings such as purulent Nasal discharge and green sputum have been associated with antibiotic prescribing for URIs in the ambulatory care practice setting; but they are not included in the NHAMCS data [11]. It is essential to understand these potentially erroneous clinical practice patterns (eg, purulence is synonymous with bacterial infection) to address them. In addition, the factors that influence antibiotic prescribing in the ED may vary from those identified in physicians’ offices given potential differences in the 2 settings and the type of patients who use these services. Staffing in the ED and the availability of ancillary services are frequently limited during non-daylight/weekday hours. Therefore, time pressures and a fear of missed diagnoses may contribute to more frequent antibiotic prescribing. In addition, patients with poor

physical and mental health use the ED more frequently [18]. Identification of the most salient factors from a broad range of variables is an essential first step in the development and implementation of interventions to improve the appro- priate and cost-effective use of antibiotics in the ED.

The objective of this study was to identify patient and provider factors associated with prescribing antibiotics for outpatients diagnosed with nonspecific URIs or acute bronchitis in the ED. We hypothesized that increased age (>=65 years); a greater number of comorbidities; and presentation to the ED during evening, night, and weekend hours would be associated with a higher frequency of antibiotic prescribing for URIs and acute bronchitis.

Methods

Study design and patient eligibility

Using the nursing assessment in the medical record, we prospectively identified consecutive patients who presented to the ED 24 hours a day, 7 days a week with a chief complaint consistent with an acute respiratory tract infection (eg, cough, sore throat, runny nose) at the VA Pittsburgh Healthcare System from June 15, 2003, to June 14, 2004, and at the Philadelphia VA Medical Center from November 30, 2003, to March 31, 2004. Patients with a provider-assigned diagnosis in the medical record of a nonspecific URI, nasopharyngitis, acute bronchitis, pharyngitis, sinusitis, acute exacerbation of chronic bronchitis, or community- acquired pneumonia were eligible for inclusion. We excluded patients who were admitted to the hospital, were immunosuppressed, had tuberculosis, were previously diag- nosed with an acute respiratory tract infection or discharged from the hospital within the preceding 2 weeks, were transferred to the ED from a nursing home or acute care hospital, were previously enrolled in the study, or were evaluated in the ED by one of the study investigators (CBG). For this analysis, we were interested in the subset of patients with acute respiratory tract infections of a presumed viral etiology (ie, URIs, nasopharyngitis, and acute bronchitis) to assess factors associated with inappropriate antibiotic use; so we subsequently excluded subjects diagnosed with a potential bacterial infection (ie, pharyngitis, sinusitis, acute exacerbation of chronic bronchitis, and community-acquired pneumonia). To eliminate any confusion between an acute exacerbation of chronic bronchitis and acute bronchitis, we also excluded patients with chronic obstructive pulmonary

disease (COPD) who were diagnosed with acute bronchitis. Finally, the diagnoses of nonspecific URI and nasophar- yngitis were combined into one group and termed URIs. The study was approved by the Institutional Review Boards at both hospitals.

Data collection

We recorded patient demographics, antibiotic allergies, and the following clinical information documented during the ED visit only from the electronic medical record: chief complaint, vital signs (ie, temperature, blood pressure, pulse, respiratory rate), smoking status (ie, current cigarette smoker vs former or nonsmoker), alcohol abuse (ie, defined by the provider as alcoholism or alcohol abuse, tolerance, or dependence), symptoms consistent with an acute respiratory tract infection (eg, purulent nasal discharge, sputum production), duration of symptoms, relevant physical examination findings (eg, cervical lymphadenopathy), pertinent laboratory tests (eg, white blood cell count) and microbiology results (ie, throat and sputum cultures), and the presence or absence of chest radiography and the results. Comorbid conditions (eg, coronary artery disease, diabetes) were also abstracted from the electronic progress note for the ED visit and the most recent primary care note within the past 6 months.

We surveyed ED providers at each site to collect personal demographic and professional information, including age, sex, race, year of graduation from medical school, Medical specialty, and provider type (ie, housestaff, attending physician, nurse practitioner/physician assistant). Although attending physicians supervised all housestaff, nurse practi- tioners, and physician assistants, the individual who documented most of the clinical encounter information was identified as the provider of record for our analyses.

We also recorded the day of the week and the time of day that patients presented to the ED. Data on any antibiotics prescribed, including dose and duration of therapy, were obtained from the patient record. Only those antibiotics prescribed on the day of the visit were included. Although antibiotic prescribing was the primary outcome measure, we also followed patients for 30 days from the initial ED visit and gathered the following clinical outcomes for patients who returned to either study site: repeated ED visits with an acute respiratory tract infection, prescription of antibiotics at the repeated visit, admission to the hospital with an acute respiratory tract infection, or death.

Analysis

To mask identity, study sites were arbitrarily designated as site 1 and site 2. Patient and Provider characteristics, diagnoses, and antibiotic prescribing practices were com- pared across sites. Antibiotic prescribing by patient and provider characteristics was also compared. Fisher exact tests

Fig. 1 Flow diagram for patient enrollment.

for discrete variables and t tests for continuous variables were used to check for differences by site or in antibiotic prescribing. To reduce issues with confounding and collinearity, only those factors associated with antibiotic prescribing (P <= .25) after adjustment for site and diagnosis were considered for inclusion in the final multivariable regression model. Stepwise backward logistic regression was used for these exploratory analyses to identify potential factors independently associated with antibiotic prescribing. Generalized Estimating Equations were used to account for the clustering of patients by provider [19]. Those factors that were significant at P <= .05 were retained in the final model. Based on the design of the study, site was included in all models, regardless of statistical significance. The sensitivity of estimates from the final model was checked by analyses that excluded patients with COPD and a diagnosis of URI (patients with COPD and a diagnosis of acute bronchitis were excluded from all analyses as described previously), excluded patients with a diagnosis of acute bronchitis, and included interaction terms of interest. All analyses were conducted using SAS (version 9.1; Cary NC) and STATA (10th version; College Station, TX).

Results

Of the 2600 patients who presented with a chief complaint consistent with a potential acute respiratory tract infection, 1184 patients met the eligibility criteria for an acute respiratory tract infection. Of these, 721 were diagnosed with a URI and/or acute bronchitis; and 667 are included in these subanalyses after the exclusion of 54 patients with both a history of COPD and a diagnosis of acute bronchitis

(Fig. 1). Study patients were predominantly male (91%), with a mean age of 55 (+-17) years (Table 1). A larger proportion of patients at site 1 were black (58% vs 28%, P b

.001), were current smokers (43% vs 32%, P = .02), and had a diagnosis of alcohol abuse (27% vs 9%, P b .001).

Providers at site 1 were older (mean age of 42 vs 32 years, P b .001), and a larger proportion were nurse practitioners or physician assistants (30% vs 1%, P b .001) or were Hispanic

Characteristics	Total (%)	Site 1 (%)	Site 2 (%)	P value
Patient characteristics Age (mean +- SD) Sex (male) Race White Black Other Unknown Ethnicity (Hispanic) ? Comorbid conditions Coronary artery disease Congestive heart failure Peripheral vascular disease Cerebrovascular disease Dementia COPD Diabetes Liver disease Renal disease Smoker Alcohol abuse N1 Comorbid condition + Provider characteristics Age (mean +- SD) Sex (male) Race White Black Asian Unknown Ethnicity (Hispanic) ? Provider type House officer Attending physician Nurse practitioner/ physician assistant Physician specialty Internal medicine Emergency medicine Other	n = 667	n = 118	n = 549
	55 +- 17	54 +- 16	55 +- 17	.99
	91	92	90	.86
	58	26	65
	33	58	28
	1	0	1
	8	16	6	b.001
	1	3	1	.21
	17	11	18	.08
	5	5	5	.81
	3	2	3	.55
	3	2	4	.40
	1	3	1	.11
	10	10	11	1.00
	18	14	19	.29
	5	8	4	.16
	4	5	4	.62
	34	43	32	.02
	12	27	9	b.001
	30	28	31	.66
	n = 126	n = 30	n = 96
	34 +- 8	42 +- 10	32 +- 6	b.001
	62	45	67	.049
	64	67	64
	3	7	2
	31	20	34
	2	7	0	.03
	2	8	0	.04
	58	0	76
	34	70	23
	8	30	1	b.001
	87	67	92
	3	19	0
	9	14	8	b.001
* Ethnicity was unknown in 213 patients and 5 providers. + Does not include COPD.

Diagnosis ?				.003
URI 547 (82)	84 (71)	463	(84)
Acute bronchitis 97 (15)	26(22)	71	(13)
URI and acute 23 (3)	8 (7)	15	(3)
bronchitis +
Antibiotics prescribed
URI 85 (16)	18 (21)	67	(14)	.14
Acute bronchitis 76 (78)	26 (100)	50	(70)	b.001
URI and acute 13 (57)	7 (88)	6	(40)	.07
bronchitis +
Total 174 (26)	51 (43)	123	(22)	b.001
* The Diagnostic categories are mutually exclusive. + Denotes patients diagnosed with both a nonspecific URI and acute bronchitis.

(8% vs 0%, P = .04) (Table 1). Although most physicians at both hospitals were specialists in internal medicine, none at site 1 were housestaff (ie, residents and interns) and none at site 2 were in the specialty of emergency medicine.

Table 2 Diagnosis and antibiotic prescribing by study site

Total, Site 1,

n (%) n (%)

n = 667 n = 118

Site 2, n (%)

n = 549

P value

Table 1 Baseline patient and provider characteristics by study site

Diagnosis and antibiotic treatment of likely nonbacterial respiratory tract infections

Although more patients were diagnosed with URIs than acute bronchitis at both sites, acute bronchitis (either alone or with a URI) was diagnosed more often at site 1 (29% vs 16%, P = .001) (Table 2). Overall, 26% of the patients with URIs and/or acute bronchitis received antibiotics (16% for URIs only, 78% for acute bronchitis only, and 57% for those with both infections). The frequency of antibiotic prescribing at site 1 was significantly higher for patients with acute bronchitis (either alone or with a URI) (97% vs 65%, P b .001).

Factors associated with antibiotic prescribing

Overall, multiple patient-related factors, such as smoking, diagnosis, day and time of visit, and certain signs and symptoms, as well as several provider factors and site were significantly associated with antibiotic prescribing in the unadjusted analyses (Table 3). In the final model, the presence of comorbid conditions, self-reported fever, purulent sputum, shortness of breath, altered breath sounds, a diagnosis of acute bronchitis, and advanced provider age and noninternal medicine specialty were independently associated with antibiotic prescribing for outpatients diag- nosed with URIs or acute bronchitis in the ED (Table 4). The odds ratios ranged from 2.1 for patients with >=1 comorbidity to 15.9 for a diagnosis of acute bronchitis. The odds ratio for site did not reach statistical significance, but was included in

Table 3 (continued)

Factor Antibiotic,

n (%)

P value

Sore throat

Table 3 Unadjusted associations between patient and provider factors and antibiotic prescribing in patients with URIs and/or acute bronchitis

Factor	Antibiotic, n (%)	P value
Patient factors
Age ?
b65 y (n = 486)	123 (25)
>=65 y (n = 181)	51 (28)	.49
Sex
Male (n = 604)	157 (26)
Female (n = 63)	17 (27)	.88
Race
White (n = 392)	98 (25)
Nonwhite (n = 275)	76 (28)	.47
Smoker
Yes (n = 227)	78 (34)
No (n = 440)	96 (22)	b.001
Alcohol abuse
Yes (n = 81)	28 (35)
No (n = 586)	146 (25)	.08
>=1 Comorbid condition ?, +
Yes (n = 201)	60 (30)
No (n = 466)	114 (24)	.15
COPD ?
Yes (n = 70)	22 (31)
No (n = 597)	152 (25)	.31
URI or nasopharyngitis
Yes (n = 570)	98 (17)
No (n = 97)	76 (78)	b.001
Acute bronchitis ?
Yes (n = 120)	89 (74)
No (n = 547)	85 (16)	b.001
Day and time of visit ?
Evening or weekend visit (n = 262)	90 (34)
Daytime (n = 405)	84 (21)	b.001
Patient signs and symptoms
Unspecified cough
Yes (n = 108)	27 (25)
No (n = 559)	147 (26)	.81
Dry cough ?
Yes (n = 132)	19 (14)
No (n = 535)	155 (29)	b.001
Productive cough ?
Yes (n = 322)	108 (34)
No (n = 345)	66 (19)	b.001
Fever, self-report ?
Yes (n = 176)	70 (40)
No (n = 491)	104 (21)	b.001
Chills ?
Yes (n = 135)	42 (31)
No (n = 532)	132 (25)	.15
Rhinorrhea
Yes (n = 225)	47 (21)
No (n = 442)	127 (29)	.03
Postnasal drip
Yes (n = 50)	8 (16)
No (n = 617)	166 (27)	.10

Yes (n = 198) 44 (22)

No (n = 469) 130 (28) .15

Weakness

Yes (n = 46) 14 (30)

No (n = 621) 160 (26) .49

Muscle pain

Yes (n = 94) 27 (29)

No (n = 573) 147 (26) .53

Shortness of breath ?

Yes (n = 106) 47 (44)

No (n = 561) 127 (23) b.001

Sputum production ?

Yes (n = 283) 95 (34)

No (n = 384) 79 (21) b.001

Purulent sputum ?

Yes (n = 194) 81 (42)

No (n = 473) 93 (20) b.001

Purulent nasal discharge ?

Yes (n = 16) 5 (31)

No (n = 651) 169 (26) .58

Temperature N99?F, on examination ?

Yes (n = 143) 48 (34)

No (n = 513) 121 (24) .02

Altered breath sounds ?

Yes (n = 126) 66 (52)

No (n = 541) 108 (20) b.001

Wheezing ?

Yes (n = 43) 20 (47)

No (n = 624) 154 (25) .003

Cervical lymphadenopathy

Yes (n = 14) 4 (29)

No (n = 653) 170 (26) .77

Duration of symptoms

<=7 days (n = 488) 116 (24)

N7 days (n = 179) 58 (32) .03

Provider factors

Age ?

b30 (n = 134) 22 (16)

>= 30 (n = 500) 141 (28) .005

Specialty ?

Internal medicine (n = 404) 81 (20)

Not internal medicine (n = 263) 93 (35) b.001 Sex

Male (n = 401) 99 (25)

Female (n = 238) 65 (27) .51

Race

White (n = 447) 106 (24)

Nonwhite (n = 195) 60 (31) .06

* Variables considered in the final multivariable logistic regression model (P <= .25 after adjustment for site, a diagnosis of acute bronchitis, and provider-level clustering).

⁺ >=Does not include COPD.

Factor	Odds ratio (95% confidence interval)
Patient
>=1 Comorbid condition ?	2.1	(1.2, 3.5)
Fever, self-report	2.5	(1.4, 4.4)
Purulent sputum	2.5	(1.5, 4.4)
Shortness of breath	2.8	(1.4, 5.4)
Altered breath sounds	4.6	(2.4, 8.6)
Acute bronchitis	15.9	(8.0, 31.8)
Provider
Provider age >=30 y	2.6	(1.1, 6.3)
Provider specialty,	2.7	(1.2, 6.0)
not internal medicine
Site 1	1.9	(0.8, 4.5)
Adjusted for provider-level clustering and site. * Does not include COPD.

the model because of the study design. Estimates from this model were similar when patients with a history of COPD and a diagnosis of URI were excluded or interactions of interest with site were included. When patients with a diagnosis of acute bronchitis were excluded, the impact of shortness of breath on antibiotic prescribing was reduced. Given the different data collection periods at the 2 hospitals, we evaluated a variable for enrollment time; but it had no effect on the modeling results.

Clinical outcomes

A minority of the patients had repeated ED visits for an acute respiratory tract infection (9%), prescriptions for antibiotics at the repeated visit (5%), or admissions to the hospital with an acute respiratory tract infection (1%). No patient died within 30 days of initial presentation. These outcomes were not significantly different if analyzed by study site or receipt of antibiotic therapy.

Discussion

Our study of antibiotic prescribing for URIs and acute bronchitis diagnosed in EDs demonstrated that prescribing rates varied widely between 2 VA Medical Centers in the same Geographical area and was especially high for patients with acute bronchitis. Signs and symptoms of potential chest consolidation, advanced provider age, and noninternal medicine provider specialty were independently associated with antibiotic prescribing.

Consistent with a report on antibiotic prescribing for URIs and acute bronchitis at outpatient clinic visits across the VA,

the rate of antibiotic prescribing varied significantly between the 2 sites in our study, although they are both VA facilities and located in the Northeast [16]. Patients at both of the medical centers were fairly similar in terms of age, sex, and comorbid conditions; so this does not explain the observed difference. More people were current smokers at site 1, and cigarette smoking was significantly associated with anti- biotic prescribing when no other variables were included in the model (P b .001). However, once the model was adjusted for a diagnosis of acute bronchitis, site, and provider-level clustering, smoking was no longer significant. The literature suggests that antibiotics offer no greater benefit for smokers than nonsmokers with acute bronchitis, but they are still prescribed more frequently for smokers [20]. In addition, a higher percentage of the patients at site 1 had a history of alcohol abuse. However, alcohol abuse was not associated with antibiotic prescribing in the univariable or multivariable analyses. It is possible that providers at site 1 were concerned about the ability of these patients to return for care if their condition worsened. Although the patients were relatively similar, the characteristics of the providers differed by site. On average, providers at site 2 were significantly younger because residents and interns evaluated patients. The participation of housestaff in the management of patients with acute respiratory tract infections may lead to reduced rates of antibiotic prescribing because of genera- tional effects that have led to increased awareness of the problem of antibiotic overuse [21]. In addition, there could be differences in the practice styles of providers and exposure to opinion leaders that create distinctive local medical cultures.

Although antibiotics were overprescribed for both infec- tions, it seemed to be especially problematic in acute bronchitis. Overall, 78% of patients with a diagnosis of acute bronchitis alone and 57% of those with diagnoses of both acute bronchitis and nonspecific URI received anti- biotics (vs 16% for URIs). This is comparable with the rate reported in the previously mentioned VA study where the median proportion of acute bronchitis visits with an antibiotic dispensed was 83% [16]. This seems to suggest that either physicians equate a diagnosis of acute bronchitis with the need for antibiotics or the diagnosis documented in the medical record is inaccurate [22].

Table 4 Factors independently associated with antibiotic prescribing in patients with URIs and/or acute bronchitis

Interventions have been designed to decrease antibiotic prescribing for patients with either of these acute respiratory tract infections; but as mentioned previously, our data suggest that physicians treat URIs and acute bronchitis differently [23-25]. This practice distinction would seem to be supported in the literature, where a recent Cochrane review of 9 randomized placebo-controlled trials of anti- biotic therapy for acute bronchitis reported that patients receiving antibiotics were less likely to have a cough, show no improvement on physician assessment, or have abnormal lung findings at a follow-up visit [26]. However, meta- analyses have found no impact of antibiotic treatment on overall duration of illness [27,28]. In addition, fewer than

10% of patients with uncomplicated acute bronchitis have a bacterial etiology; and the routine use of antibiotics is not recommended in other reviews of acute bronchitis [29,30]. Future interventions may need to speak to potential practice differences in the management of patients with acute bronchitis vs URIs.

Finally, unlike previous studies of patients in the ED, we collected detailed clinical data from the visit; and several signs and symptoms were independently associated with the receipt of antibiotics. One of these clinical findings was purulent sputum production. Our results confirm the need to address this perceived association between purulent mani- festations and the presence of a bacterial infection, as well as other clinical signs and symptoms that may be influencing antibiotic prescribing [11,31]. We had postu- lated that increased age; a greater number of comorbidities; and presentation to the ED during evening, night, and weekend hours would be independently associated with a higher frequency of antibiotic prescribing for URIs and acute bronchitis. However, neither age nor the time of presentation was in the final model. Instead, antibiotic prescribing was associated with the diagnosis, certain signs and symptoms, and provider characteristics. Adequate staffing during evenings and weekends may not have been an issue for these less severe illnesses. Providers may have considered comorbidities and symptoms more important than age, although neither should be important in these likely viral infections.

There are several limitations to our study. It involved 2 VA Medical Centers with EDs primarily staffed by internists, which limits the generalizability of our findings to other clinical settings. However, many of the issues we identified are likely to occur throughout EDs in both VA and non-VA institutions. Given the variability in antibiotic prescribing across sites, it is likely that organizational factors are involved; and we could not address site-level variables because our study involved 2 medical centers. In addition, data were collected for a shorter period at one of the sites because of budget constraints. However, the same respiratory season was included at both sites; and we evaluated a variable for enrollment time, but it had no effect on the modeling results. We also do not know if patient and provider expectations influenced the prescribing decision. Finally, most of the predictors of antibiotic prescribing are based on clinician report in the medical record. It is unknown how often signs or symptoms are emphasized to justify an antibiotic prescription.

Conclusions

Antibiotic prescribing for URIs and acute bronchitis was high and varied substantially between 2 VA Medical Centers located in the same part of the country and with similar types of patients. Specific signs and symptoms, a diagnosis of

acute bronchitis, and provider age and specialty were independently associated with antibiotic prescribing for infections of likely nonbacterial etiology. Multifaceted interventions to decrease prescribing in nonspecific URIs and acute bronchitis will likely be needed because a variety of patient and provider factors appears to be involved. In addition, our data suggest that physicians think URIs and acute bronchitis are managed differently. Therefore, inter- ventions may need to address the perceived utility of antibiotics in acute bronchitis, as well as the accuracy of signs and symptoms in diagnosing a bacterial respiratory tract infection.

References

1. Roumie CL, Halasa NB, Grijalva CG, Edwards KM, Zhu Y, Dittus RS, et al. Trends in antibiotic prescribing for adults in the United States- 1995 to 2002. J Gen Intern Med 2005;20:697-702.
2. Fuller JD, Low DE. A review of Streptococcus pneumoniae infection treatment failures associated with fluoroquinolone resistance. Clin Infect Dis 2005;41:118-21.
3. Ho PL, Tse WS, Tsang KWT, Kwok TK, Ng TK, Cheng VCC, et al. Risk factors for acquisition of levofloxacin-resistant Strepto- coccus pneumoniae: a case-control study. Clin Infect Dis 2001;32: 701-7.
4. Nuorti JP, Butler JC, Crutcher JM, Guevara R, Welch D, Holder P, et al. An outbreak of multidrug-resistant pneumococcal pneumonia and bacteremia among unvaccinated nursing home residents. N Engl J Med 1998;338:1861-8.
5. Roumie CL, Halasa NB, Edwards KM, Zhu Y, Dittus RS, Griffin MR. Differences in antibiotic prescribing among physicians, residents, and nonphysician clinicians. Am J Med 2005;118:641-8.
6. Rutschmann OT, Domino ME. Antibiotics for upper respiratory tract infections in ambulatory practice in the United States, 1997-1999: does physician specialty matter? J Am Board Fam Pract 2004;17: 196-200.
7. Linder JA, Singer D, Stafford RS. Association between antibiotic prescribing and Visit duration in adults with upper respiratory tract infections. Clin Ther 2003;25:2419-30.
8. Gonzales R, Steiner JF, Sande MA. Antibiotic prescribing for adults with colds, upper respiratory tract infections, and bronchitis by ambulatory care physicians. JAMA 1997;278:901-4.
9. Stone S, Gonzales R, Maselli J, Lowenstein SR. Antibiotic prescribing for patients with colds, upper respiratory tract infections, and bronchitis: a national study of hospital-based emergency departments. Ann Emerg Med 2000;36:320-7.
10. Dosh SA, Hickner JM, Mainous AG, Ebell MH. Predictors of antibiotic prescribing for nonspecific upper Respiratory infections, acute bronchitis, and acute sinusitis. J Fam Pract 2000;49:407-14.
11. Gonzales R, Barrett PH, Steiner JF. The relation between purulent manifestations and antibiotic treatment of upper respiratory tract infections. J Gen Intern Med 1999;14:151-6.
12. Metlay JP, Stafford RS, Singer DE. National trends in the use of antibiotics by primary care physicians for adult patients with cough. Arch Intern Med 1998;158:1813-8.
13. Macfarlane J, Holmes W, Macfarlane R, Britten N. Influence of patients’ expectations on antibiotic management of acute lower respiratory tract illness in General practice: questionnaire study. BMJ 1997;315:1211-4.
14. Steinman MA, Sauaia A, Maselli J, Houck PM, Gonzales R. Office evaluation and treatment of elderly patients with acute bronchitis. J Am Geriatr Soc 2004;52:875-9.
15. Avorn J, Solomon DH. Cultural and economic factors that (mis)shape antibiotic use: the nonpharmacologic basis of therapeutics. Ann Intern Med 2000;133:128-35.
16. Aspinall SL, Berlin JA, Zhang Y, Metlay JP. Facility-level variation in antibiotic prescriptions for veterans with upper respiratory infections. Clin Ther 2005;27:258-62.
17. Thorpe JM, Smith SR, Trygstad TK. Trends in emergency department antibiotic prescribing for acute respiratory tract infections. Ann Pharmacother 2004;38:928-35.
18. Hunt KA, Weber EJ, Showstack JA, Colby DC, Callaham ML. Characteristics of frequent users of emergency departments. Ann Emerg Med 2006;48:1-8.
19. Zeger SL, Liang KY. Longitudinal data analysis for discrete and continuous outcomes. Biometrics 1986;42:121-30.
20. Linder JA, Sim I. Antibiotic treatment of acute bronchitis in smokers: a systematic review. J Gen Intern Med 2002;17:230-4.
21. Gonzales R, Camargo Jr CA, MacKenzie T, Kersey AS, Maselli JH, Levin SK, et al. Antibiotic treatment of Acute respiratory infections in Acute care settings. Acad Emerg Med 2006;13:288-94.
22. Sethi S, Murphy TF. acute exacerbations of chronic bronchitis: new developments concerning microbiology and pathophysiology- impact on approaches to risk stratification and therapy. Infect Dis Clin N Am 2004;18:861-82.
23. Gonzales R, Steiner JF, Lum A, Barrett PH. Decreasing antibiotic use in ambulatory practice. Impact of a multidimensional intervention on

the treatment of uncomplicated acute bronchitis in adults. JAMA 1999; 281:1512-9.

1. Finkelstein JA, Davis RL, Dowell SF, Metlay JP, Soumerai SB, Rifas- Shiman SL, et al. Reducing antibiotic use in children: a randomized trial in 12 practices. Pediatrics 2001;108:1-7.
2. Harris RH, MacKenzie TED, Leeman-Castillo B, et al. Opti- mizing antibiotic prescribing for acute respiratory tract infections in an urban urgent care clinic. J Gen Intern Med 2003;18: 326-34.
3. Smucny J, Fahey T, Becker L, Glazier R. Antibiotics for acute bronchitis. Cochrane Database Syst Rev 2004;4:CD000245.
4. Fahey T, Stocks N, Thomas T. Quantitative systematic review of Randomised controlled trials comparing antibiotic with placebo for acute cough in adults. BMJ 1998;316:906-10.
5. Bent S, Saint S, Vittinghoff E, Grady D. Antibiotics in acute bronchitis: a meta-analysis. Am J Med 1999;107:62-7.
6. Gonzales R, Bartlett JG, Besser RE, Hickner JM, Hoffman JR, Sande MA. Principles of appropriate antibiotic use for treatment of nonspecific upper respiratory tract infections in adults: background. Ann Intern Med 2001;134:490-4.
7. Aagaard E, Gonzales R. Management of acute bronchitis in Healthy adults. Infect Dis Clin N Am 2004;18:919-37.
8. Murray S, Del Mar C, O’Rourke P. Predictors of an antibiotic prescription by GPs for respiratory tract infections: a pilot. Fam Pract 2000;17:386-8.