outpatient antibiotic prescribing for acute respiratory...

Original Investigation | Infectious Diseases

Outpatient Antibiotic Prescribing for Acute Respiratory InfectionsDuring Influenza SeasonsFiona P. Havers, MD; Lauri A. Hicks, DO; Jessie R. Chung, MPH; Manjusha Gaglani, MD; Kempapura Murthy, MPH; Richard K. Zimmerman, MD; Lisa A. Jackson, MD;Joshua G. Petrie, PhD; Huong Q. McLean, PhD; Mary Patricia Nowalk, PhD; Michael L. Jackson, PhD; Arnold S. Monto, MD; Edward A. Belongia, MD;Brendan Flannery, PhD; Alicia M. Fry, MD

Abstract

IMPORTANCE Acute respiratory infections (ARIs) are the syndrome for which antibiotics are mostcommonly prescribed; viruses for which antibiotics are ineffective cause most ARIs.

OBJECTIVES To characterize antibiotic prescribing among outpatients with ARI during influenzaseason and to identify targets for reducing inappropriate antibiotic prescribing for common ARIdiagnoses, including among outpatients with laboratory-confirmed influenza.

DESIGN, SETTING, AND PARTICIPANTS Cohort study enrolling outpatients aged 6 months or olderwith ARI evaluated at outpatient clinics associated with 5 US Influenza Vaccine EffectivenessNetwork sites during the 2013-2014 and 2014-2015 influenza seasons. All patients received influenzatesting by real-time reverse transcriptase–polymerase chain reaction for research purposes only.Antibiotic prescriptions, medical history, and International Classification of Diseases, Ninth Revisiondiagnosis codes were collected from medical and pharmacy records, as were group A streptococcal(GAS) testing results in a patient subset.

EXPOSURE Visit for ARI, defined by a new cough of 7 days’ duration or less.

MAIN OUTCOMES AND MEASURES Antibiotic prescription within 7 days of enrollment.Appropriateness of antibiotic prescribing was based on diagnosis codes, clinical information, andinfluenza and GAS testing results.

RESULTS Of 14 987 patients with ARI (mean [SD] age, 32 [24] years; 8638 [58%] women; 11 892[80%] white), 6136 (41%) were prescribed an antibiotic. Among these 6136 patients, 2522 (41%) haddiagnoses for which antibiotics are not indicated; 2106 (84%) of these patients were diagnosed ashaving a viral upper respiratory tract infection or bronchitis (acute or not otherwise specified).Among the 3306 patients (22%) not diagnosed as having pneumonia and who had laboratory-confirmed influenza, 945 (29%) were prescribed an antibiotic, accounting for 17% of all antibioticprescriptions among patients with nonpneumonia ARI. Among 1248 patients with pharyngitis, 1137(91%) had GAS testing; 440 of the 1248 patients (35%) were prescribed antibiotics, among whom168 (38%) had negative results on GAS testing. Of 1200 patients with sinusitis and no otherindication for antibiotic treatment who received an antibiotic, 454 (38%) had symptoms for 3 daysor less prior to the outpatient visit, suggesting acute viral sinusitis not requiring antibiotics.

CONCLUSIONS AND RELEVANCE Antibiotic overuse remains widespread in the treatment ofoutpatient ARIs, including among patients with laboratory-confirmed influenza, although study sitesmay not be representative of other outpatient settings. Identified targets for improved outpatient

(continued)

Key PointsQuestion What are targets for

improving antibiotic stewardship for

outpatient acute respiratory infections?

Findings Among 14 987 outpatients

with acute respiratory infections

enrolled in this cohort study during

influenza seasons, 41% were prescribed

antibiotics, 41% of whom had diagnoses

for which antibiotics are not indicated,

primarily viral upper respiratory tract

infections and bronchitis; 29% of

patients with influenza confirmed

through research testing were

prescribed antibiotics. Among patients

prescribed antibiotics, 38% with

pharyngitis tested negative for group A

streptococcus and 38% with sinusitis

had symptoms for 3 days or less before

the visit, suggesting antibiotic therapy

was not required.

Meaning Eliminating antibiotic

treatment of viral upper respiratory tract

infections and bronchitis, improving

influenza diagnosis and treatment, and

reinforcing prescription guidelines for

pharyngitis and sinusitis could improve

outpatient antibiotic stewardship.

+ Supplemental content

Author affiliations and article information arelisted at the end of this article.

Open Access. This is an open access article distributed under the terms of the CC-BY License.

JAMA Network Open. 2018;1(2):e180243. doi:10.1001/jamanetworkopen.2018.0243 June 8, 2018 1/13

Downloaded From: on 06/28/2018

https://jama.jamanetwork.com/article.aspx?doi=10.1001/jamanetworkopen.2018.0243&utm_campaign=articlePDF%26utm_medium=articlePDFlink%26utm_source=articlePDF%26utm_content=jamanetworkopen.2018.0243

Abstract (continued)

antibiotic stewardship include eliminating antibiotic treatment of viral upper respiratory tractinfections and bronchitis and improving adherence to prescribing guidelines for pharyngitis andsinusitis. Increased access to sensitive and timely virus diagnostic tests, particularly for influenza,may reduce unnecessary antibiotic use for these syndromes.

JAMA Network Open. 2018;1(2):e180243. doi:10.1001/jamanetworkopen.2018.0243

Introduction

Acute respiratory infections (ARIs) remain the clinical category for which antibiotics are mostcommonly prescribed.1,2 However, most ARIs are caused by viruses for which antibiotics have no rolein treatment. Inappropriate antibiotic use contributes to the development of antibiotic-resistantorganisms, which cause an estimated 2 million illnesses and 23 000 deaths annually in the UnitedStates.3 Understanding antibiotic prescribing practices for ARIs in outpatient settings is critical todesigning strategies for reducing inappropriate antibiotic use.

Previous studies examining antibiotic overuse have relied on national survey data, and manylack clinical and laboratory testing results.2,4-7 In this study, we include data on illness onset date,laboratory testing for influenza in all patients, and, for many patients, self-reported fever and theresults of clinician-ordered group A streptococcal (GAS) testing. We used 3 approaches to assessantibiotic prescribing among outpatients with ARIs during the influenza season. First, we examinedantibiotic prescribing by age group, antibiotic type, and diagnosis. Second, because influenza is afrequent cause of ARIs, sensitive point-of-care influenza testing is not widely available,8,9 andindividuals with influenza are often inappropriately prescribed antibiotics,10 we characterizedantibiotic prescribing among those with laboratory-confirmed influenza, excluding syndromes forwhich antibiotics are indicated (eg, pneumonia). Third, we examined appropriateness of antibioticprescribing for those with pharyngitis, sinusitis, and otitis media (OM), diagnoses that account fornearly one-third of prescribed outpatient antibiotics,2 based on symptom duration, presence offever, prescription of recommended first-line antibiotics,4 and GAS testing results.

Methods

EnrollmentIn this cohort study, we used data collected by the US Influenza Vaccine Effectiveness Networkduring the 2013-2014 (defined here as December 2, 2013, through April 16, 2014) and 2014-2015(defined here as November 10, 2014, through April 10, 2015) influenza seasons, as describedelsewhere.11 The network, which included 57 (2013-2014) and 66 (2014-2015) clinics associated with5 sites that are geographically diverse academic medical centers and health care organizations, wasdesigned to assess influenza vaccine effectiveness annually; additional data were collected for thissecondary analysis. Once sites confirmed local influenza circulation, patients aged 6 months or olderwith ARI, defined by a new cough of 7 days’ duration or less, were eligible for enrollment. Respiratoryswab specimens were obtained and tested for influenza with real-time reverse transcriptase–polymerase chain reaction for research purposes only. Patients could enroll more than once perseason more than 14 days apart; we analyzed each patient visit separately. Eligible patients werescreened and enrolled if they or their legal guardians provided written or oral informed consent,depending on requirements of local institutional review boards. Study procedures, forms, andconsent documents were approved by site institutional review boards.

Patient characteristics, including self-reported race and ethnicity selected from options definedby the investigator, and symptom onset date were ascertained by interview. In addition, patientswere asked about self-reported fever at 2 sites in 2013-2014 and all sites in 2014-2015. We obtainedpatients’ history of underlying medical conditions by examining International Classification of

JAMA Network Open | Infectious Diseases Outpatient Antibiotic Prescribing for Acute Respiratory Infections




Diseases, Ninth Revision (ICD-9) codes assigned to medical encounters during the year beforeenrollment. We extracted dates of hospitalizations occurring less than 30 days after enrollment frommedical records. Excluding topical antibiotics, we obtained antibiotic and influenza antiviralprescriptions within 7 days of enrollment from pharmacy, insurance, or electronic medical records.Patients were excluded if dates indicated antibiotics were first prescribed during a hospitalization. Ifmore than 1 antibiotic was prescribed, the antibiotic with the earliest prescription date was included.

Diagnostic CategoriesWe classified the first 4 ICD-9 codes assigned to the enrollment visit into 3 tiers based on previouslypublished classifications (eTable 1 in the Supplement). Tier 1 includes diagnoses for which antibioticsare almost always indicated, including pneumonia; per Infectious Diseases Society of America/American Thoracic Society guidelines, empirical antibiotic treatment is appropriate for persons withpneumonia, regardless of the etiology.12 Tier 2 includes diagnoses for which antibiotics may beindicated, eg, sinusitis, pharyngitis, and suppurative OM. Tier 3 includes all other diagnoses for whichantibiotics are not indicated or the indication was unclear; these include viral upper respiratory tractinfection (URI), influenza, allergy, asthma, and acute bronchitis.2 Because current guidancerecommends antibiotic treatment for some patients with chronic obstructive pulmonary disease(COPD) exacerbations,13 adults with a history of COPD or COPD diagnosis code at enrollment wereexcluded from bronchitis analyses. We excluded patients missing ICD-9 codes or assigned diagnosticcodes compatible with nonrespiratory diagnoses for which antibiotics are potentially indicated (eg,urinary tract infection). If a patient had multiple diagnoses, priority was given to tier 1 diagnoses, thentier 2 diagnoses, then tier 3 diagnoses.

Antibiotic ClassificationWe classified antibiotics based on previously established categories.14 Narrow-spectrum antibioticsincluded narrow-spectrum penicillins, tetracyclines, first-generation cephalosporins, andsulfonamides. Broad-spectrum antibiotics included macrolides (eg, azithromycin), broad-spectrumpenicillins (eg, amoxicillin-clavulanate), advanced-generation cephalosporins, quinolones, andlincomycin derivatives (clindamycin). We examined whether clinicians prescribed first-line antibioticsfor the most common tier 2 diagnoses, based on national guidelines: penicillin or amoxicillin forpharyngitis, amoxicillin or amoxicillin-clavulanate for sinusitis, and amoxicillin or amoxicillin-clavulanate (alternative) for suppurative OM.4,15-19

Influenza TestingAt 1 site, clinicians were provided study real-time reverse transcriptase–polymerase chain reactionresults for influenza within 48 hours of enrollment. Results were not available to clinicians at othersites. At 4 sites, we examined data on rapid influenza diagnostic tests or other diagnostic testsordered the day of enrollment and not performed as part of the study protocol.

Pharyngitis and GAS TestingFor pharyngitis, guidelines recommend antibiotic therapy only for GAS pharyngitis and diagnostictesting if GAS is suspected.20 We examined a GAS testing subset comprising patients with apharyngitis diagnosis with no other diagnosis for which antibiotics may be warranted at 4 sites duringthe 2014-2015 season for whom we obtained information from medical records on clinician-orderedGAS testing. Rapid antigen detection test results were available at all sites, and some sites performedcultures when rapid antigen detection test results were negative. If a patient had positive results oneither the GAS rapid antigen detection test or culture, the patient was considered GAS positive. Theproportion of patients in the entire study population with pharyngitis who were prescribed anantibiotic despite a negative result on GAS testing was estimated based on results from the GAStesting subset.





Sinusitis and OMFor sinusitis, national guidelines recommend antibiotics for adults and children with severesymptoms, defined by a temperature of 39°C or higher, purulent nasal discharge, or facial pain, for atleast 3 consecutive days; worsening course after initial improvement; or persistent illness (symptomslasting �10 days).16-18 Among patients with sinusitis and suppurative OM,15 we examined antibioticprescribing and its relationship to the timing of symptom onset and self-reported fever, whenavailable.

Statistical AnalysisCategorical data were analyzed with a χ2 test. We used logistic regression to develop 2 models withpredictors of antibiotic prescribing. A model examining antibiotic prescribing among those with onlytier 3 diagnoses adjusted for sex, site, age group, race, number of chronic conditions, time fromsymptom onset to presentation for care, and clinical diagnosis. A model that examined antibioticprescribing among those with laboratory-confirmed influenza adjusted for sex, site, age group,number of chronic medical conditions, time from symptom onset, clinical diagnosis, and presence offever. Variables with 2-sided P < .20 on univariate analysis were included in multivariable analysesand subjected to model-fitting procedures. A 2-sided P value less than .05 was consideredstatistically significant. Statistical analyses were conducted using SAS version 9.4 statistical software(SAS Institute Inc).

Results

During the 2013-2014 and 2014-2015 influenza seasons, we enrolled 15 714 patients. Among these,727 were excluded: 126 had missing ICD-9 codes, 118 were prescribed an antibiotic while hospitalized,and 483 were diagnosed with a nonrespiratory condition for which antibiotics are potentiallyindicated. Among the 14 987 patients included, the mean (SD) age was 32 (24) years, 8638 (58%)were women, and 11 892 (80%) were white (Table 1). The 14 987 included patient visits had a mean(SD) of 2.0 (1.2) ICD-9 codes; 338 (2%) had at least 5 codes, less than 1% of which had additionalcodes indicating tier 1 or tier 2 diagnoses. Results were unchanged when we analyzed alldiagnosis codes.

Among 14 987 patient visits analyzed, 6136 (41%) were associated with an antibioticprescription, including 3423 (56%) with broad-spectrum antibiotics. Five drugs accounted for 90%of antibiotic prescriptions: azithromycin (37%), amoxicillin (35%), amoxicillin-clavulanate (10%),doxycycline (5%), and levofloxacin (3%) (Figure 1). Macrolides were most frequently prescribed inadults; 960 of 4339 adults (22%) aged 50 years or older were prescribed azithromycin, regardless ofclinical diagnosis, symptoms, symptom duration, or laboratory findings.

Influenza and Antibiotic PrescribingInfluenza was confirmed through research testing in 3381 patients (23%), 902 (27%) of whomreceived a clinical diagnosis code for influenza. Excluding those with a tier 1 diagnosis, among 3306patients (22%) with laboratory-confirmed influenza, 945 (29%) were prescribed an antibiotic,accounting for 17% of all antibiotic prescriptions among patients with nonpneumonia ARI; 656(20%) were prescribed an influenza antiviral medication, including 89 prescribed both antibiotic andantiviral medications. Laboratory-confirmed influenza was detected in 72 of 375 patients (16%)diagnosed as having pneumonia, 446 of 2065 patients (18%) with pharyngitis, 129 of 1099 patients(11%) with suppurative OM, and 227 of 1486 patients (13%) with sinusitis; the proportion withlaboratory-confirmed influenza among those prescribed antibiotics is shown in Figure 2. In amultivariable model among those with laboratory-confirmed influenza, antibiotic prescribing wassignificantly associated with a clinical diagnosis of pneumonia (adjusted odds ratio [AOR] = 145.74;95% CI, 34.37-617.82), sinusitis (AOR = 15.34; 95% CI, 9.83-23.94), OM (AOR = 71.77; 95% CI,30.00-171.71), and bronchitis (AOR = 4.60; 95% CI, 3.27-6.47), but not pharyngitis (AOR = 1.23; 95%




CI, 0.90-1.69). Other factors associated with antibiotic prescribing include site, longer symptomduration, and older age. Those with a clinical diagnosis code for influenza (AOR = 0.30; 95% CI,

Table 1. Characteristics of Patients With Acute Respiratory Infections in the 2013-2014 and 2014-2015 InfluenzaSeasons at Ambulatory Care Settings Affiliated With the US Influenza Vaccine Effectiveness Networka

Characteristic

Patients, No. (%)

Total (N = 14 987)b Prescribed Antibiotic (n = 6136)c

Sex

Male 6349 (42) 2567 (40)

Female 8638 (58) 3569 (41)

Age group, y

0.5-2 918 (6) 398 (43)

2 to <5 1488 (10) 589 (40)

5 to <18 3294 (22) 1186 (36)

18 to <50 4948 (33) 1938 (39)

50 to <65 2563 (17) 1183 (46)

≥65 1776 (12) 842 (47)

Race

White 11 892 (80) 5150 (43)

Black 1407 (9) 424 (30)

Other 1619 (11) 534 (33)

Study site

A 4286 (29) 1353 (32)

B 3112 (21) 1629 (52)

C 2699 (18) 1166 (43)

D 2315 (15) 858 (37)

E 2575 (17) 1130 (44)

Time from symptom onset to presentationfor care, d

≤2 4678 (31) 1610 (34)

3-4 5719 (38) 2406 (42)

5-7 4590 (31) 2120 (46)

≥1 Chronic medical condition 5331 (36) 2315 (43)

Laboratory-confirmed influenzad 3381 (22) 1017 (30)

a Acute respiratory infection is defined by cough with7 days’ duration or less.

b Percentage indicates the percentage in the column.c Percentage indicates the percentage in the row.d Tested by real-time reverse transcriptase–

polymerase chain reaction in all enrolled patients.Testing was performed for research purposes.Clinicians were unaware of results, except for 1 site atwhich clinicians were notified of positive real-timereverse transcriptase–polymerase chain reactionresearch results within 48 hours of enrollment.

Figure 1. Antibiotic Classes Prescribed in Pediatric and Adult Outpatients With Acute Respiratory Infectionsin the 2013-2014 and 2014-2015 Influenza Seasons in the US Influenza Vaccine Effectiveness Network

70

50

60

40

30

20

10

0Narrow-Spectrum

PenicillinsOthera Macrolides Broad-Spectrum

PenicillinsBroad-SpectrumCephalosporins

Quinolones

Prop

ortio

n of

Ant

ibio

tics P

resc

ribed

, %

Narrow-Spectrum Antibiotics Broad-Spectrum Antibiotics

Tetracyclines

Pediatric outpatients

Adult outpatients

Acute respiratory infections are defined by cough with7 days’ duration or less. Pediatric indicates patientsaged 6 months to younger than 18 years.a Includes first-generation cephalosporins and

sulfonamides.




0.21-0.41), URI (AOR = 0.75; 95% CI, 0.57-0.99), or other tier 3 diagnosis (AOR = 0.50; 95% CI, 0.33-0.76) had decreased odds of receiving an antibiotic prescription (eFigure 1 in the Supplement). At 4sites, clinical nonresearch influenza testing was performed on the day of enrollment for 198 of 12 412patients (2%).

Antibiotic Prescribing Without an Appropriate Indication (Tier 3 Diagnoses)Among 6136 patients who received antibiotic prescriptions, 2522 (41%) had only diagnoses for whichantibiotics are not indicated (tier 3) (Table 2; eFigure 2 in the Supplement). The proportion ofpatients with only tier 3 diagnoses who were prescribed antibiotics increased with increasing age:25%, 46%, and 65% of antibiotic prescriptions among those younger than 18 years, 18 to 64 years,and 65 years or older, respectively (P < .01). Azithromycin was the most commonly prescribedantibiotic given to adults who lacked a diagnosis for which antibiotics are indicated; among the 3186adults aged 50 years and older with no clinical indication for antibiotics (tier 3 diagnoses only), 1183(37%) received an antibiotic prescription; 688 (58%) of these prescriptions were for azithromycin.

Among the 2522 patients with tier 3 diagnoses prescribed an antibiotic, 2106 (84%) werediagnosed as having either a viral URI or bronchitis (acute or not otherwise specified). Among 5553

Figure 2. Proportion of Pediatric and Adult Outpatients With Acute Respiratory Infections Who Were PrescribedAntibiotics, by Selected International Classification of Diseases, Ninth Revision Diagnostic Codesand Laboratory-Confirmed Influenza Status

100

80

60

40

90

70

50

30

20

0Pneumonia Otitis Media Viral Upper

RespiratoryTract Infection

Bronchitis Allergyor Asthma

ClinicalInfluenzab

OtherSinusitis

Patie

nts P

resc

ribed

Ant

ibio

tics,

%

Tier 2: AntibioticsPotentially Indicated

Tier 1: Antibiotics

Almost AlwaysIndicated

Tier 3: Antibiotics Not Indicated

Pharyngitis

Pediatric outpatientsA

Influenza negativea

Laboratory-confirmedinfluenzaa

100

80

60

40

90

70

50

30

20

0Pneumonia Otitis Media Viral Upper

RespiratoryTract Infection

Bronchitisc Allergyor Asthma

ClinicalInfluenzab

OtherSinusitis

Patie

nts P

resc

ribed

Ant

ibio

tics,

%

Tier 2: AntibioticsPotentially Indicated

Tier 1:Antibiotics

Almost AlwaysIndicated

Tier 3: Antibiotics Not Indicated

Pharyngitis

Adult outpatientsB

Acute respiratory infections are defined by cough with7 days’ duration or less. Pediatric indicates patientsaged 6 months to younger than 18 years.a All enrollees received influenza testing by real-time

reverse transcriptase–polymerase chain reaction forresearch purposes only. At 1 site, clinicians wereprovided study real-time reverse transcriptase–polymerase chain reaction results for influenzawithin 48 hours of enrollment. Results were notavailable to clinicians at other sites.

b Indicates an International Classification of Diseases,Ninth Revision diagnosis code of influenza assignedby the clinician who saw the patient.

c Adults with a history of chronic obstructivepulmonary disease or a visit diagnosis code forchronic obstructive pulmonary disease wereexcluded from the analysis of bronchitis.






patients with a viral URI diagnosis, 1376 (25%) received an antibiotic, accounting for 22% of allantibiotic prescriptions (Figure 2). Excluding patients with COPD, 1157 persons were assigned adiagnosis for bronchitis; 730 (63%) received an antibiotic, accounting for 12% of all antibioticprescriptions. In a multivariable model among those with only tier 3 diagnoses, antibiotic prescribingwas most strongly associated with a diagnosis of bronchitis (AOR = 4.71; 95% CI, 3.82-5.80)(Figure 3). Other factors associated with antibiotic prescribing include site, longer symptomduration, older age, at least 1 chronic medical condition, worse self-rated health, and fever. Blackpatients were significantly less likely to be prescribed antibiotics than white patients (AOR= 0.60;95% CI, 0.48-0.76; P < .001). Both a clinical diagnosis of influenza and laboratory-confirmedinfluenza through research testing were significantly associated with decreased odds of an antibioticprescription (clinical diagnosis: AOR = 0.32; 95% CI, 0.25-0.42; laboratory-confirmed diagnosis:AOR = 0.62; 0.54-0.71).

Pharyngitis, Sinusitis, and OM (Tier 2 Diagnoses)The 3 most frequent diagnoses for which antibiotics are potentially indicated (tier 2 diagnoses) werepharyngitis (diagnosed in 2494 patients [17%]), sinusitis (1707 patients [11%]), and suppurative OM(1212 patients [8%]), which together accounted for 52% of all antibiotics prescribed. Among 3198persons who received antibiotics for these diagnoses, 2050 (64%) received first-line antibiotics(eFigure 2 and eTable 2 in the Supplement).

Pharyngitis accounted for 16% of antibiotic prescriptions. Among 1000 of 2494 patients (40%)with a pharyngitis diagnosis who received an antibiotic, 122 (12%) had laboratory-confirmedinfluenza (Figure 2; eTable 2 in the Supplement). Among the 1248 patients assigned a pharyngitis

Table 2. Antibiotic Prescribing by Diagnostic Category Tier Among Patients With Acute Respiratory Infections in the 2013-2014 and 2014-2015 Influenza Seasonsat Ambulatory Care Settings Affiliated With the US Influenza Vaccine Effectiveness Networka

Diagnosis

Patients, No. (%)

All Ages Pediatric Adult

Total (N = 14 987)bPrescribed Antibiotic(n = 6136)c Total (n = 5700)b

Prescribed Antibiotic(n = 2173)c Total (n = 9287)b

Prescribed Antibiotic(n = 3963)c

Tier 1: antibiotics almostalways indicated

467 (3) 416 (89) 218 (4) 187 (86) 249 (3) 229 (92)

Pneumonia 447 (3) 400 (89) 202 (4) 174 (86) 245 (3) 226 (92)

Miscellaneous bacterialinfections

20 (0.1) 16 (80) 16 (0.3) 13 (81) 4 (0.04) 3 (75)

Tier 2: antibiotics potentiallyindicatedd

5129 (34) 3198 (62) 2322 (41) 1445 (62) 2807 (30) 1753 (62)

Pharyngitis 2494 (17) 1000 (40) 1167 (20) 439 (38) 1327 (14) 561 (42)

Sinusitis 1707 (11) 1385 (81) 243 (4) 192 (79) 1464 (16) 1193 (81)

Suppurative otitis media 1212 (8) 1055 (87) 993 (17) 881 (89) 219 (2) 174 (79)

Tier 3: antibiotics notindicatede

9391 (63) 2522 (27) 3160 (55) 541 (17) 6231 (67) 1981 (32)

Viral upper respiratory tractinfection

5553 (37) 1376 (25) 1964 (34) 299 (15) 3589 (39) 1077 (30)

Bronchitisf 1157 (8) 730 (63) 107 (2) 74 (7) 1050 (11) 656 (62)

Allergy or asthma 1217 (8) 289 (24) 451 (8) 61 (14) 766 (8) 228 (30)

Influenza diagnosis codeg 1257 (8) 167 (13) 317 (6) 43 (14) 940 (10) 124 (13)

Viral pneumonia 3 (0.02) 2 (67) 2 (0.04) 1 (50) 1 (0.01) 0

Remaining codes not listedelsewhere

1463 (10) 373 (25) 602 (11) 113 (19) 861 (9) 260 (30)

a The diagnostic category tiers are adapted from Fleming-Dutra et al2 and Shapiro et al.14

The first 4 International Classification of Diseases, Ninth Revision diagnostic codes wereexamined. Each patient was classified in a tier; priority was given to tier 1 diagnoses,then tier 2 diagnoses, then tier 3 diagnoses. However, individuals could be in more than1 diagnosis category.

b Percentage indicates the percentage in the column.c Percentage indicates the percentage in the row.d Excludes patients with any tier 1 diagnosis.

e Excludes patients with any tier 1 or tier 2 diagnosis.f Excludes 31 adults with a history of chronic obstructive pulmonary disease per medical

record or who had a chronic obstructive pulmonary disease diagnosis code. If theseare included, 684 of 1081 adults (63%) with a diagnosis of bronchitis receivedantibiotics.

g Indicates an International Classification of Diseases, Ninth Revision code diagnosis ofinfluenza, not laboratory confirmation of influenza, which was performed for researchpurposes only.






diagnosis code who had information on GAS testing available, 1137 (91%) had GAS testing performed.A total of 440 patients with pharyngitis (35%) were prescribed an antibiotic, among whom 208(47%) lacked a positive GAS test, either because they did not have testing performed (40 patients[9%]) or their test result was negative (168 patients [38%]) (eFigure 3 in the Supplement). Anestimated 327 of 860 patients (38%) in the entire study population with only a pharyngitis diagnosiswere given an antibiotic prescription despite a negative GAS test result.

Among 1707 patients assigned a sinusitis diagnosis, 1385 (81%) were prescribed antibiotics, 188(14%) of whom had laboratory-confirmed influenza (Figure 2; eTable 2 in the Supplement). Amongthe 1295 patients diagnosed as having sinusitis for whom fever information was collected, 1087(84%) were prescribed an antibiotic; 558 (51%) reported a fever, whereas 203 (19%) had neitherfever nor symptom duration longer than 3 days. Among patients with sinusitis and no otherindication for antibiotic treatment who received an antibiotic, the median time from symptom onsetto prescription date was 4 days (interquartile range, 3-6 days); 454 of 1200 patients (38%) were

Figure 3. Multivariable Analysis of Predictors of Antibiotic Prescribing Among Persons With Acute Respiratory Infections and Assigned Diagnosis Codesfor Which Antibiotics Are Not Indicated (Tier 3 Diagnoses Only)

0.10 1.00 2.50 10.000.50 5.00Adjusted Odds Ratio (95% CI)

0.25

Decreased Odds ofAntibiotic Prescribing

Increased Odds ofAntibiotic PrescribingVariable

Patients, No.(n = 9391)

Adjusted OddsRatio (95% CI)

SexMale 4059 1.02 (0.90-1.15)Female 5332 1 [Reference]

SiteA 2966 0.42 (0.34-0.52)B 1939 1.85 (1.51-2.26)C 1496 1.16 (0.92-1.46)D 1597 1.16 (0.93-1.43)E 1393 1 [Reference]

Age group, y<5 1446 0.45 (0.36-0.55)5 to <18 1714 0.70 (0.58-0.83)18 to <50 3045 1 [Reference]50 to <65 1796 1.39 (1.18-1.64)65 to <80 1122 1.51 (1.24-1.83)≥80 268 2.21 (1.53-2.94)

RaceWhite 7387 1 [Reference]Black 940 0.60 (0.48-0.76)Other 1024 0.84 (0.67-1.04)

Time from symptom onset to presentation for care, d

≤2 2954 1 [Reference]3-4 3559 1.22 (1.06-1.41)5-7 2878 1.31 (1.12-1.52)

Clinical diagnosisa

Fever 4235 1.12 (1.01-1.24)

Self-rated health status on day of visitd 9375 0.94 (0.91-0.96)

Bronchitis 1250 4.71 (3.82-5.80)Viral upper tract respiratory infection 5553 0.87 (0.70-1.07)Allergy or asthma 1217 0.92 (0.74-1.14)Influenzab 1257 0.32 (0.25-0.42)Other tier 3 diagnosis 1463 0.89 (0.68-1.15)Laboratory-confirmed influenzac 3381 0.62 (0.54-0.71)

≥1 Chronic medical condition 3737 1.33 (1.16-1.51)

a By International Classification of Diseases, Ninth Revision diagnosis code.b Indicates an International Classification of Diseases, Ninth Revision code diagnosis of

influenza, not real-time reverse transcriptase–polymerase chain reaction confirmationof influenza.

c Indicates real-time reverse transcriptase–polymerase chain reaction confirmation ofinfluenza, which was performed for research purposes only.

d Self-rated health on a scale of 0 (worst) to 100 (best), analyzed as a continuousvariable.






prescribed an antibiotic within 3 days following symptom onset, a proportion that was unchangedwhen these patients were stratified by fever status.

Suppurative OM accounted for 17% of antibiotic prescriptions, including 41% of those inchildren and 4% in adults. Among 1055 persons with this diagnosis prescribed an antibiotic, 121 (11%)had laboratory-confirmed influenza (Figure 2; eTable 2 in the Supplement).

Among all 6136 patients with antibiotic prescriptions, we estimated 3303 (54%) likely did notmeet clinical criteria or laboratory criteria for GAS pharyngitis for antibiotic treatment. Of these 6136patients, 2522 (41%) had only clinical diagnoses that do not warrant antibiotics (tier 3 diagnosesonly), 454 (7%) had sinusitis that did not meet clinical criteria, and an estimated 327 (5%) hadpharyngitis with a negative GAS test result. Among the remaining 2833 patients given antibioticprescriptions who had a clinical diagnosis for which antibiotic therapy may be indicated (eg,suppurative OM, pharyngitis, or sinusitis), an additional 369 (13%) had influenza virus infectionconfirmed through research testing and likely did not benefit from antibiotic therapy.

Discussion

Antibiotics were likely prescribed inappropriately to a majority of the nearly 15 000 outpatients inthis study who presented during influenza season with symptoms of a broadly defined ARIcharacterized by cough. Among all patients prescribed antibiotics, 41% lacked a diagnosis code forwhich antibiotic therapy is potentially indicated. Those with influenza confirmed through researchtesting accounted for a substantial proportion (17%) of all antibiotics prescribed; fewer thanone-third of patients with laboratory-confirmed influenza were given a clinical diagnosis of influenza.Patients with influenza virus infection accounted for a substantial proportion of those givenantibiotics for diagnoses for which antibiotics may be appropriate, including pharyngitis (12%),sinusitis (14%), and suppurative OM (11%), although most patients with influenza are unlikely tobenefit from antibiotic treatment. In addition, among those patients given diagnoses for whichantibiotic therapy may be appropriate, many patients prescribed antibiotics likely did not meetcriteria for antibiotic therapy based on clinician-ordered laboratory testing and clinical criteria,including 47% of those with a pharyngitis diagnosis and 38% with a sinusitis diagnosis.

Our study is consistent with previous studies that indicate clinicians overprescribe antibioticsfor outpatient ARIs. These studies used national survey data, which lack detailed clinical data,including laboratory testing results.2,4-7 We used a sensitive clinical definition of ARI and had dataincluding illness onset date, laboratory confirmation of influenza, and, for many patients, self-reported fever and the results of clinician-ordered GAS testing. As noted in previous studies,2,6 wefound that a substantial proportion of antibiotic overuse was driven by prescribing for conditions forwhich antibiotics were not indicated, including viral URIs and acute bronchitis.19 Our results alsosuggest antibiotics were frequently overprescribed for patients with 2 common diagnoses for whichantibiotics are potentially indicated, pharyngitis and sinusitis, underscoring that these diagnoses areimportant targets for improving antibiotic use. For pharyngitis, national guidelines recommendantibiotic therapy only for laboratory-confirmed GAS19,20; in our study, nearly half of patientsdiagnosed as having pharyngitis who received antibiotics either tested negative for GAS or were nottested. For patients with sinusitis, among those prescribed an antibiotic, 38% had symptoms for 3days or less and most did not report fever, indicating that many patients likely did not meet clinicalcriteria for antibiotic treatment as defined by national guidelines.

Older adults were more likely than younger adults and children to receive antibiotics withoutdiagnostic documentation of an appropriate indication and were far more likely to receive broad-spectrum antibiotics, particularly macrolides. Azithromycin was prescribed to nearly one-quarter ofadults aged 50 years or older with an ARI, regardless of diagnosis, laboratory testing, symptomduration, or medical history, including to 58% of those older adults given a prescription who lackedan indication for antibiotic treatment. Azithromycin also accounted for more than one-quarter ofprescriptions among those adults diagnosed as having pharyngitis and sinusitis, even though it is not





the first-line agent for either condition. A concerning increase in macrolide use for these conditionshas been previously noted in both adults and children.5,7 Choosing a macrolide when amoxicillin oramoxicillin-clavulanate is the recommended first-line antibiotic is a potential patient safety issue;GAS and Streptococcus pneumoniae infections are more likely to be resistant to macrolides than toamoxicillin or amoxicillin-clavulanate. While all antibiotics have potential risks, macrolides have alsobeen associated with an increased risk of cardiovascular events.21

This study was conducted during influenza seasons, and we found that 17% of patients with ARIprescribed antibiotics had laboratory-confirmed influenza. Influenza can predispose patients toincreased risk of bacterial coinfections. In some instances of influenza infection, antibiotics may beappropriately prescribed if an associated bacterial infection is suspected. However, bacterialcoinfection was likely only in a minority of influenza cases.22 As in previous studies,10 we found thatpatients with influenza confirmed through research testing were more likely to receive an antibioticthan an antiviral medication. In addition, a substantial proportion of patients with diagnoses forwhich antibiotics may be considered appropriate—pharyngitis, suppurative OM, and sinusitis—actually had influenza virus infection, and depending on the clinical situation, many would likely notbenefit from antibiotic treatment. Current widely available point-of-care influenza diagnostic testshave highly variable sensitivity8 and are not recommended for ruling out influenza infection for thepurposes of treatment decisions9; clinician-ordered testing was infrequent in this study. Cliniciansassigned an influenza diagnosis code to only one-quarter of patients with influenza confirmedthrough research testing, indicating that influenza virus infections were underdiagnosed. However,among all patients with ARI, those patients to whom clinicians assigned an influenza diagnosis weresignificantly less likely to receive an antibiotic prescription. The development of sensitive and specificpoint-of-care testing for influenza may assist clinicians in making treatment decisions for patientswith ARI during influenza season and may help to reduce unnecessary antibiotic use for influenza inoutpatient settings.23 Inappropriate antibiotic prescribing exposes patients to the risks ofunnecessary antibiotics and represents a potential missed opportunity for patients to benefit frominfluenza antiviral medications.9 Clinicians should be encouraged to consider influenza as a clinicaldiagnosis during the influenza season, refrain from prescribing antibiotics in situations in which theyare not recommended by guidelines, and prescribe influenza antiviral medications when indicated.9

LimitationsThis study has several limitations. We used up to 4 ICD-9 codes to determine whether antibioticswere prescribed appropriately, but these may inaccurately reflect clinical decision making. Medicalhistory information may have been incomplete in some instances, and other factors may haveinformed prescribing decisions. Our study sites may not be representative of other outpatientsettings; some sites are health care organizations that may have institutional prescribing policies ormostly serve insured patients. In addition, sites have been engaged in influenza research for multipleyears, which may have influenced clinician awareness of influenza and antibiotic prescribingpractices. Of note, the proportion of patients prescribed antibiotics for conditions such as pharyngitisand bronchitis was lower in this study compared with other studies, and assessment of antibioticoveruse may be underestimated.2,6,7 We also assumed prescriptions were related to the study visit,but this could not always be verified. We lacked dispensing information for several sites, limiting ourability to know whether clinicians used delayed prescribing strategies in which patients are instructedto fill prescriptions only if symptoms failed to improve. This could have led to an overestimation ofantibiotic use. Previous antibiotic treatment of prior infections and allergy history were not available,both of which may have affected antibiotic choice.

Conclusions

Our study adds to evidence that misuse of antibiotics, characterized by antibiotic overuse andinappropriate antibiotic selection, is widespread in the treatment of outpatient ARIs. The study




indicates a number of potential targets to achieve the goal of the National Action Plan for CombatingAntibiotic-Resistant Bacteria of reducing inappropriate outpatient antibiotic use by 50% by 2020.24

We must strengthen outpatient antibiotic stewardship efforts to eliminate antibiotic treatment forviral URIs and acute bronchitis, which our study indicates would make the largest contribution todecreasing unnecessary antibiotic prescriptions. Increased efforts are needed to support improvedadherence to guidelines for antibiotic prescribing for common diagnoses, including more stringentadherence to GAS pharyngitis testing guidelines and clinical criteria for antibiotic treatment ofsinusitis, as well as interventions focused on appropriate selection of first-line antibiotics for theseconditions if treatment is indicated. In addition, our findings indicate that improved point-of-careinfluenza diagnostics and increased recognition and appropriate treatment of influenza virusinfection may also aid in decreasing unnecessary antibiotic use for ARIs. The Centers for DiseaseControl and Prevention published “Core Elements of Outpatient Antibiotic Stewardship,”25 whichprovides guidance to clinicians and facility leadership to implement activities to improve antibioticuse. Improving antibiotic prescribing for ARIs during the influenza season represents an importantopportunity to improve the long-term quality of patient care.

ARTICLE INFORMATIONAccepted for Publication: March 5, 2018.

Open Access: This is an open access article distributed under the terms of the CC-BY License. © 2018 Havers FPet al. JAMA Network Open.

Corresponding Author: Fiona P. Havers, MD, Division of Bacterial Diseases, Centers for Disease Control andPrevention, 1600 Clifton Rd NE, Mailstop C-25, Atlanta, GA 30333 ([email protected]).

Author Affiliations: Centers for Disease Control and Prevention, Atlanta, Georgia (Havers, Hicks, Chung, Flannery,Fry); Baylor Scott & White Health, Texas A&M University Health Science Center College of Medicine, Temple(Gaglani, Murthy); University of Pittsburgh, Pittsburgh, Pennsylvania (Zimmerman, Nowalk); Kaiser PermanenteWashington Health Research Institute, Seattle (L. A. Jackson, M. L. Jackson); Department of Epidemiology,University of Michigan School of Public Health, Ann Arbor (Petrie, Monto); Marshfield Clinic Research Foundation,Marshfield, Wisconsin (McLean, Belongia).

Author Contributions: Dr Havers had full access to all the data in the study and takes responsibility for theintegrity of the data and the accuracy of the data analysis.

Concept and design: Havers, Hicks, McLean, M. L. Jackson, Flannery, Fry.

Acquisition, analysis, or interpretation of data: Havers, Hicks, Chung, Gaglani, Murthy, Zimmerman, L. A. Jackson,Petrie, McLean, Nowalk, M. L. Jackson, Monto, Belongia, Fry.

Drafting of the manuscript: Havers.

Critical revision of the manuscript for important intellectual content: All authors.

Statistical analysis: Havers, Chung.

Obtained funding: Flannery.

Administrative, technical, or material support: Hicks, Zimmerman, McLean, M. L. Jackson, Belongia.

Supervision: Havers, Hicks, Monto, Fry.

Conflict of Interest Disclosures: Drs Gaglani, McLean, and Belongia reported receiving research support fromMedimmune. Mr Murthy reported serving as an institutional research study data manager for Medimmune/AstraZeneca. Dr Zimmerman reported receiving research grant support from Medimmune, Sanofi Pasteur, PfizerInc, and Merck and Co and consulting fees from Medimmune. Dr L. A. Jackson reported receiving research grantsupport from Novartis and Takeda. Dr Nowalk reported receiving research funding from Pfizer Inc and Merck andCo. Dr Monto reported receiving consulting fees from Roche, Novartis, and Sanofi. No other disclosures werereported.

Funding/Support: This work was supported by the Centers for Disease Control and Prevention throughcooperative agreements with the University of Michigan (grant U01 IP000474), Group Health Research Institute(now named Kaiser Permanente Washington Health Research Institute; grant U01 IP000466), Marshfield ClinicResearch Foundation (grant U01 IP000471), University of Pittsburgh (grant U01 IP000467), and Baylor Scott &




https://jamanetwork.com/journals/jamanetworkopen/pages/instructions-for-authors#SecOpenAccess/?utm_campaign=articlePDF%26utm_medium=articlePDFlink%26utm_source=articlePDF%26utm_content=jamanetworkopen.2018.0243

mailto:[email protected]

White Health (grant U01 IP000473). At the University of Pittsburgh, the project was also supported by grants UL1RR024153 and UL1TR000005 from the National Institutes of Health.

Role of the Funder/Sponsor: The Centers for Disease Control and Prevention Health designed and conducted thestudy; collected, managed, analyzed, and interpreted the data; prepared, reviewed, and approved the manuscript;and had a role in the decision to submit the manuscript for publication. The National Institutes of Health had norole in the design and conduct of the study; collection, management, analysis, and interpretation of the data;preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication.

Disclaimer: The findings and conclusions in this article are those of the authors and do not necessarily representthe official views of the Centers for Disease Control and Prevention.

Additional Contributions: We thank the US Influenza Vaccine Effectiveness Network study participants and studystaff. The following persons contributed to study support, data collection, and management and received studyfunds: Kaiser Permanente Washington Health Research Institute, Seattle: C. Hallie Phillips, MS, Joyce Benoit, BSN,Lawrence Madziwa, MS, and Matt Nguyen, MS; Marshfield Clinic Research Foundation, Marshfield, Wisconsin:Jennifer King, MPH, Sarah Kopitzke, and Carla Rottscheit; University of Michigan, Ann Arbor: EJ McSpadden, MPH,Rachel Truscon, MPH, Emileigh Johnson, Caroline Cheng, MS, Anne Kaniclides, Emily T. Martin, PhD, HeatherLipkovich, MPH, and Lois Lamerato, PhD; University of Pittsburgh, Pittsburgh, Pennsylvania: GK Balasubramani,PhD, Heather Eng, BA, Theresa Sax, BS, Krissy K. Moehling, PhD, Michael Susick, MPH, Samantha Ford, BS, andStephen Wisniewski, PhD; Baylor Scott & White Health, Texas A&M University Health Science Center College ofMedicine, Temple: Michael Reis, MD, Lydia Clipper, RN, Archana Nangrani, MBBS, Anne Robertson, JessicaPruszynski, PhD, Virginia Gandy, RN, Teresa Ponder, Hope Gonzales, Martha Zayed, and Deborah Furze; and BaylorCollege of Medicine, Houston, Texas: Pedro A. Piedra, MD, W. Paul Glezen, MD, Vasanthi Avadhanula, PhD, AlanJewell, Kirtida Patel, and Sneha Thaker. The following persons contributed to study support and did not receivestudy funds: Baylor Scott & White Health, Texas A&M University Health Science Center College of Medicine,Temple: Madhava Beeram, MD, Alejandro Arroliga, Donald Wesson, MD, Richard Beswick, PhD, Monica Weir, MaryKylberg, Jeremy Ray, Jessica Rostocykj, and Glen Cryer.

REFERENCES1. Centers for Disease Control and Prevention. Office-related antibiotic prescribing for persons aged � 14 years—United States, 1993-1994 to 2007-2008. MMWR Morb Mortal Wkly Rep. 2011;60(34):1153-1156.

2. Fleming-Dutra KE, Hersh AL, Shapiro DJ, et al. Prevalence of inappropriate antibiotic prescriptions among USambulatory care visits, 2010-2011. JAMA. 2016;315(17):1864-1873.

3. Centers for Disease Control and Prevention. Antibiotic Resistance Threats in the United States, 2013. Atlanta,GA: Centers for Disease Control & Prevention; 2013.

4. Hersh AL, Fleming-Dutra KE, Shapiro DJ, Hyun DY, Hicks LA; Outpatient Antibiotic Use Target-SettingWorkgroup. Frequency of first-line antibiotic selection among US ambulatory care visits for otitis media, sinusitis,and pharyngitis. JAMA Intern Med. 2016;176(12):1870-1872.

5. Dooling KL, Shapiro DJ, Van Beneden C, Hersh AL, Hicks LA. Overprescribing and inappropriate antibioticselection for children with pharyngitis in the United States, 1997-2010. JAMA Pediatr. 2014;168(11):1073-1074.

6. Barnett ML, Linder JA. Antibiotic prescribing for adults with acute bronchitis in the United States, 1996-2010.JAMA. 2014;311(19):2020-2022.

7. Barnett ML, Linder JA. Antibiotic prescribing to adults with sore throat in the United States, 1997-2010.JAMA Intern Med. 2014;174(1):138-140.

8. Chartrand C, Leeflang MM, Minion J, Brewer T, Pai M. Accuracy of rapid influenza diagnostic tests: a meta-analysis. Ann Intern Med. 2012;156(7):500-511.

9. Fiore AE, Fry A, Shay D, Gubareva L, Bresee JS, Uyeki TM; Centers for Disease Control and Prevention. Antiviralagents for the treatment and chemoprophylaxis of influenza—recommendations of the Advisory Committee onImmunization Practices (ACIP). MMWR Recomm Rep. 2011;60(1):1-24.

10. Havers F, Thaker S, Clippard JR, et al. Use of influenza antiviral agents by ambulatory care clinicians during the2012-2013 influenza season. Clin Infect Dis. 2014;59(6):774-782.

11. Ohmit SE, Thompson M, Petrie JG, et al. Influenza vaccine effectiveness in the 2011-2012 season: protectionagainst each circulating virus and the effect of prior vaccination on estimates. Clin Infect Dis. 2014;58(3):319-327.

12. Mandell LA, Wunderink RG, Anzueto A, et al; Infectious Diseases Society of America; American ThoracicSociety. Infectious Diseases Society of America/American Thoracic Society consensus guidelines on themanagement of community-acquired pneumonia in adults. Clin Infect Dis. 2007;44(suppl 2):S27-S72.

13. Vestbo J, Hurd SS, Agustí AG, et al. Global strategy for the diagnosis, management, and prevention of chronicobstructive pulmonary disease: GOLD executive summary. Am J Respir Crit Care Med. 2013;187(4):347-365.




https://www.ncbi.nlm.nih.gov/pubmed/21881545












14. Shapiro DJ, Hicks LA, Pavia AT, Hersh AL. Antibiotic prescribing for adults in ambulatory care in the USA,2007-09. J Antimicrob Chemother. 2014;69(1):234-240.

15. Lieberthal AS, Carroll AE, Chonmaitree T, et al. The diagnosis and management of acute otitis media [publishedcorrection appears in Pediatrics. 2014;133(2):346]. Pediatrics. 2013;131(3):e964-e999.

16. Wald ER, Applegate KE, Bordley C, et al; American Academy of Pediatrics. Clinical practice guideline for thediagnosis and management of acute bacterial sinusitis in children aged 1 to 18 years. Pediatrics. 2013;132(1):e262-e280.

17. Chow AW, Benninger MS, Brook I, et al; Infectious Diseases Society of America. IDSA clinical practice guidelinefor acute bacterial rhinosinusitis in children and adults. Clin Infect Dis. 2012;54(8):e72-e112.

18. Rosenfeld RM, Piccirillo JF, Chandrasekhar SS, et al. Clinical practice guideline (update): adult sinusitis.Otolaryngol Head Neck Surg. 2015;152(2)(suppl):S1-S39.

19. Hersh AL, Jackson MA, Hicks LA; American Academy of Pediatrics Committee on Infectious Diseases.Principles of judicious antibiotic prescribing for upper respiratory tract infections in pediatrics. Pediatrics. 2013;132(6):1146-1154.

20. Shulman ST, Bisno AL, Clegg HW, et al. Clinical practice guideline for the diagnosis and management of groupA streptococcal pharyngitis: 2012 update by the Infectious Diseases Society of America [published correctionappears in Clin Infect Dis. 2014;58(10):1496]. Clin Infect Dis. 2012;55(10):1279-1282.

21. Ray WA, Murray KT, Hall K, Arbogast PG, Stein CM. Azithromycin and the risk of cardiovascular death.N Engl J Med. 2012;366(20):1881-1890.

22. Klein EY, Monteforte B, Gupta A, et al. The frequency of influenza and bacterial coinfection: a systematicreview and meta-analysis. Influenza Other Respir Viruses. 2016;10(5):394-403.

23. Merckx J, Wali R, Schiller I, et al. Diagnostic accuracy of novel and traditional rapid tests for influenza infectioncompared with reverse transcriptase polymerase chain reaction: a systematic review and meta-analysis. Ann InternMed. 2017;167(6):394-409.

24. Task Force for Combating Antibiotic-Resistant Bacteria. National Action Plan for Combating Antibiotic-Resistant Bacteria. Washington, DC: The White House; 2015.

25. Sanchez GV, Fleming-Dutra KE, Roberts RM, Hicks LA. Core elements of outpatient antibiotic stewardship.MMWR Recomm Rep. 2016;65(6):1-12.

SUPPLEMENT.eTable 1. Diagnostic Categories by Tier With Corresponding International Classification of Diseases, Ninth Revision(ICD-9) Diagnosis CodeseTable 2. Antibiotic Selection Among Those With an International Classification of Diseases, Ninth Revision (ICD-9)Diagnosis Code for Pharyngitis, Sinusitis, and Suppurative Otitis Media Among Those Prescribed AntibioticseFigure 1. Predictors of Antibiotic Prescribing Among Persons With Laboratory-Confirmed Influenza and WhoSought Care at Ambulatory Care Settings Associated With 5 Institutions, US Influenza Vaccine EffectivenessNetwork, 2013-2014 and 2014-2015 Influenza SeasonseFigure 2. Number of Enrollees With Acute Respiratory Infections Defined by Cough �7 Days’ Duration, SelectedICD-9 Diagnostic Codes, by Antibiotic Prescribing StatuseFigure 3. Antibiotic Prescribing Among Those With an International Classification of Diseases, Ninth Revision(ICD-9) Diagnosis Code for Pharyngitis, by Group A Streptococcus Testing Status, 2014-2015 Influenza Season


















outpatient antibiotic prescribing for acute respiratory...

Documents