[Skip to Navigation]
Sign In
Figure 1. Antibiotic Prescriptions for Adults With Nonpneumonic Acute Respiratory Tract Infections, by Diagnosis
Antibiotic Prescriptions for Adults With Nonpneumonic Acute Respiratory Tract Infections, by Diagnosis

Overall, 63% of all patients with an acute respiratory tract infection (ARTI) received an antibiotic, including 29% who received narrow-spectrum therapy and 34% who received broad-spectrum therapy. Broad-spectrum antibiotics included the quinolones, amoxicillin/clavulanate, the second- and third-generation cephalosporins, and azithromycin and clarithromycin. Other upper respiratory tract infection (URTI) includes pharyngitis, laryngitis, and tracheitis.

Figure 2. Type of Broad-Spectrum Antibiotic Prescribed for Adults With Nonpneumonic Acute Respiratory Tract Infections, by Diagnosis
Type of Broad-Spectrum Antibiotic Prescribed for Adults With Nonpneumonic Acute Respiratory Tract Infections, by Diagnosis

The total number of antibiotic prescriptions by diagnostic category is common cold, 228; sinusitis, 332; otitis media, 67; acute bronchitis, 286; and other URTI, 157. Prescriptions for narrow-spectrum antibiotics are not shown. URTI indicates upper respiratory tract infection.

Table. Predictors of Broad-Spectrum Antibiotic Choice Among Patients Prescribed an Antibiotic*
Predictors of Broad-Spectrum Antibiotic Choice Among Patients Prescribed an Antibiotic*
1.
Gonzales  R, Steiner  JF, Sande  MA.  Antibiotic prescribing for adults with colds, upper respiratory tract infections, and bronchitis by ambulatory care physicians.  JAMA. 1997;278:901-904. PubMedGoogle ScholarCrossref
2.
Linder  JA, Stafford  RS.  Antibiotic treatment of adults with sore throat by community primary care physicians: a national survey, 1989-1999.  JAMA. 2001;286:1181-1186. PubMedGoogle ScholarCrossref
3.
Piccirillo  JF, Mager  DE, Frisse  ME, Brophy  RH, Goggin  A.  Impact of first-line vs second-line antibiotics for the treatment of acute uncomplicated sinusitis.  JAMA. 2001;286:1849-1856. PubMedGoogle ScholarCrossref
4.
McCaig  LF, Besser  RE, Hughes  JM.  Trends in antimicrobial prescribing rates for children and adolescents.  JAMA. 2002;287:3096-3102. PubMedGoogle ScholarCrossref
5.
Snow  V, Mottur-Pilson  C, Gonzales  R.  Principles of appropriate antibiotic use for treatment of nonspecific upper respiratory tract infections in adults.  Ann Intern Med. 2001;134:487-489. PubMedGoogle ScholarCrossref
6.
Snow  V, Mottur-Pilson  C, Gonzales  R.  Principles of appropriate antibiotic use for treatment of acute bronchitis in adults.  Ann Intern Med. 2001;134:518-520. PubMedGoogle ScholarCrossref
7.
Takata  GS, Chan  LS, Shekelle  P, Morton  SC, Mason  W, Marcy  SM.  Evidence assessment of management of acute otitis media, I: the role of antibiotics in treatment of uncomplicated acute otitis media.  Pediatrics. 2001;108:239-247. PubMedGoogle ScholarCrossref
8.
Austin  DJ, Kristinsson  KG, Anderson  RM.  The relationship between the volume of antimicrobial consumption in human communities and the frequency of resistance.  Proc Natl Acad Sci U S A. 1999;96:1152-1156. PubMedGoogle ScholarCrossref
9.
Not Available.  Resistance of Streptococcus pneumoniae to fluoroquinolones—United States, 1995-1999  MMWR Morb Mortal Wkly Rep. 2001;50:800-804. PubMedGoogle Scholar
10.
Hyde  TB, Gay  K, Stephens  DS,  et al.  Macrolide resistance among invasive Streptococcus pneumoniae isolates.  JAMA. 2001;286:1857-1862. PubMedGoogle ScholarCrossref
11.
Gonzales  R, Barrett Jr  PH, Crane  LA, Steiner  JF.  Factors associated with antibiotic use for acute bronchitis.  J Gen Intern Med. 1998;13:541-548. PubMedGoogle ScholarCrossref
12.
Gonzales  R, Barrett Jr  PH, Steiner  JF.  The relation between purulent manifestations and antibiotic treatment of upper respiratory tract infections.  J Gen Intern Med. 1999;14:151-156. PubMedGoogle ScholarCrossref
13.
Hemminki  E.  Review of literature on the factors affecting drug prescribing.  Soc Sci Med. 1975;9:111-116. PubMedGoogle ScholarCrossref
14.
Soumerai  SB, Ross-Degnan  D, Avorn  J, McLaughlin  T, Choodnovskiy  I.  Effects of Medicaid drug-payment limits on admission to hospitals and nursing homes.  N Engl J Med. 1991;325:1072-1077. PubMedGoogle ScholarCrossref
15.
Wazana  A.  Physicians and the pharmaceutical industry: is a gift ever just a gift?  JAMA. 2000;283:373-380. PubMedGoogle ScholarCrossref
16.
Avorn  J, Soumerai  SB.  Improving drug-therapy decisions through educational outreach: a randomized controlled trial of academically based "detailing."  N Engl J Med. 1983;308:1457-1463. PubMedGoogle ScholarCrossref
17.
Avorn  J, Solomon  DH.  Cultural and economic factors that (mis)shape antibiotic use: the nonpharmacologic basis of therapeutics.  Ann Intern Med. 2000;133:128-135. PubMedGoogle ScholarCrossref
18.
National Center for Health Statistics.  Public Use Microdata File Documentation, National Ambulatory Medical Care Survey, 1997 Hyattsville, Md: National Technical Information Service; 1999.
19.
National Center for Health Statistics.  Public Use Microdata File Documentation, National Ambulatory Medical Care Survey, 1998 Hyattsville, Md: National Technical Information Service; 2000.
20.
National Center for Health Statistics.  Public Use Microdata File Documentation, National Ambulatory Medical Care Survey, 1999 Hyattsville, Md: National Technical Information Service; 2001.
21.
Not Available.  International Classification of Diseases, Ninth Revision, Clinical Modification. Washington, DC: Public Health Service, US Dept of Health and Human Services; 1988.
22.
Gonzales  R, Malone  DC, Maselli  JH, Sande  MA.  Excessive antibiotic use for acute respiratory infections in the United States.  Clin Infect Dis. 2001;33:757-762. PubMedGoogle ScholarCrossref
23.
Ioannidis  JP, Contopoulos-Ioannidis  DG, Chew  P, Lau  J.  Meta-analysis of randomized controlled trials on the comparative efficacy and safety of azithromycin against other antibiotics for upper respiratory tract infections.  J Antimicrob Chemother. 2001;48:677-689. PubMedGoogle ScholarCrossref
24.
Cabana  MD, Rand  CS, Powe  NR,  et al.  Why don't physicians follow clinical practice guidelines? a framework for improvement.  JAMA. 1999;282:1458-1465. PubMedGoogle ScholarCrossref
25.
Schwartz  RK, Soumerai  SB, Avorn  J.  Physician motivations for nonscientific drug prescribing.  Soc Sci Med. 1989;28:577-582. PubMedGoogle ScholarCrossref
26.
Metlay  JP, Shea  JA, Crossette  LB, Asch  DA.  Tensions in antibiotic prescribing: pitting social concerns against the interests of individual patients.  J Gen Intern Med. 2002;17:87-94. PubMedGoogle ScholarCrossref
27.
Epstein  AM, Read  JL, Winickoff  R.  Physician beliefs, attitudes, and prescribing behavior for anti-inflammatory drugs.  Am J Med. 1984;77:313-318. PubMedGoogle ScholarCrossref
28.
Carrin  G.  Drug prescribing: a discussion of its variability and (ir)rationality.  Health Policy. 1987;7:73-94. PubMedGoogle ScholarCrossref
29.
Cockburn  J, Pit  S.  Prescribing behaviour in clinical practice: patients' expectations and doctors' perceptions of patients' expectations—a questionnaire study.  BMJ. 1997;315:520-523. PubMedGoogle ScholarCrossref
30.
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;315:1211-1214. PubMedGoogle ScholarCrossref
31.
Wilkes  MS, Bell  RA, Kravitz  RL.  Direct-to-consumer prescription drug advertising: trends, impact, and implications.  Health Aff (Millwood). 2000;19:110-128. PubMedGoogle ScholarCrossref
32.
Hamm  RM, Hicks  RJ, Bemben  DA.  Antibiotics and respiratory infections: are patients more satisfied when expectations are met?  J Fam Pract. 1996;43:56-62. PubMedGoogle Scholar
33.
Butler  CC, Rollnick  S, Pill  R, Maggs-Rapport  F, Stott  N.  Understanding the culture of prescribing: qualitative study of general practitioners' and patients' perceptions of antibiotics for sore throats.  BMJ. 1998;317:637-642. PubMedGoogle ScholarCrossref
34.
Chew  LD, O'Young  TS, Hazlet  TK, Bradley  KA, Maynard  C, Lessler  DS.  A physician survey of the effect of drug sample availability on physicians' behavior.  J Gen Intern Med. 2000;15:478-483. PubMedGoogle ScholarCrossref
35.
Bell  RA, Wilkes  MS, Kravitz  RL.  Advertisement-induced prescription drug requests: patients' anticipated reactions to a physician who refuses.  J Fam Pract. 1999;48:446-452. PubMedGoogle Scholar
36.
Rosenthal  MB, Berndt  ER, Donohue  JM, Frank  RG, Epstein  AM.  Promotion of prescription drugs to consumers.  N Engl J Med. 2002;346:498-505. PubMedGoogle ScholarCrossref
37.
Chren  MM, Landefeld  CS, Murray  TH.  Doctors, drug companies, and gifts.  JAMA. 1989;262:3448-3451. PubMedGoogle ScholarCrossref
38.
Chren  MM, Landefeld  CS.  Physicians' behavior and their interactions with drug companies: a controlled study of physicians who requested additions to a hospital drug formulary.  JAMA. 1994;271:684-689. PubMedGoogle ScholarCrossref
39.
Lurie  N, Rich  EC, Simpson  DE,  et al.  Pharmaceutical representatives in academic medical centers: interaction with faculty and housestaff.  J Gen Intern Med. 1990;5:240-243. PubMedGoogle ScholarCrossref
40.
Ziegler  MG, Lew  P, Singer  BC.  The accuracy of drug information from pharmaceutical sales representatives.  JAMA. 1995;273:1296-1298. PubMedGoogle ScholarCrossref
41.
Lambert  BL, Salmon  JW, Stubbings  J, Gilomen-Study  G, Valuck  RJ, Kezlarian  K.  Factors associated with antibiotic prescribing in a managed care setting: an exploratory investigation.  Soc Sci Med. 1997;45:1767-1779. PubMedGoogle ScholarCrossref
42.
Avery  AJ, Walker  B, Heron  T, Teasdale  SJ.  Do prescribing formularies help GPs prescribe from a narrower range of drugs? a controlled trial of the introduction of prescribing formularies for NSAIDs.  Br J Gen Pract. 1997;47:810-814. PubMedGoogle Scholar
43.
Mather  JL, Bayliff  CD, Reider  MJ, Hussain  Z, Colby  WD.  The impact of formulary reservations on drug utilization: a controlled trial.  Can J Hosp Pharm. 1994;47:111-116. PubMedGoogle Scholar
44.
Harrold  LR, Field  TS, Gurwitz  JH.  Knowledge, patterns of care, and outcomes of care for generalists and specialists.  J Gen Intern Med. 1999;14:499-511. PubMedGoogle ScholarCrossref
45.
Walker  S, McGeer  A, Simor  AE, Armstrong-Evans  M, Loeb  M.  Why are antibiotics prescribed for asymptomatic bacteriuria in institutionalized elderly people? a qualitative study of physicians' and nurses' perceptions.  CMAJ. 2000;163:273-277. PubMedGoogle Scholar
46.
Safavi  KT, Hayward  RA.  Choosing between apples and apples: physicians' choices of prescription drugs that have similar side effects and efficacies.  J Gen Intern Med. 1992;7:32-37. PubMedGoogle ScholarCrossref
47.
Hasty  M, Schrager  J, Wrenn  K.  Physicians' perceptions about managed care restrictions on antibiotic prescribing.  J Gen Intern Med. 1999;14:756-758. PubMedGoogle ScholarCrossref
48.
Hadley  J, Mitchell  JM, Sulmasy  DP, Bloche  MG.  Perceived financial incentives, HMO market penetration, and physicians' practice styles and satisfaction.  Health Serv Res. 1999;34:307-321. PubMedGoogle Scholar
49.
Reschovsky  J, Reed  M, Blumenthal  D, Landon  B.  Physicians' assessments of their ability to provide high-quality care in a changing health care system.  Med Care. 2001;39:254-269. PubMedGoogle ScholarCrossref
50.
Mainous 3rd  AG, Hueston  WJ, Eberlein  C.  Colour of respiratory discharge and antibiotic use.  Lancet. 1997;350:1077. PubMedGoogle ScholarCrossref
51.
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;158:1813-1818. PubMedGoogle ScholarCrossref
52.
Mainous III  AG, Hueston  WJ, Love  MM.  Antibiotics for colds in children: who are the high prescribers?  Arch Pediatr Adolesc Med. 1998;152:349-352. PubMedGoogle ScholarCrossref
53.
Hutchinson  JM, Foley  RN.  Method of physician remuneration and rates of antibiotic prescription.  CMAJ. 1999;160:1013-1017. PubMedGoogle Scholar
54.
Oxman  AD, Thomson  MA, Davis  DA, Haynes  RB.  No magic bullets: a systematic review of 102 trials of interventions to improve professional practice.  CMAJ. 1995;153:1423-1431. PubMedGoogle Scholar
55.
Bero  LA, Grilli  R, Grimshaw  JM, Harvey  E, Oxman  AD, Thomson  MA, for the Cochrane Effective Practice and Organization of Care Review Group.  Closing the gap between research and practice: an overview of systematic reviews of interventions to promote the implementation of research findings.  BMJ. 1998;317:465-468. PubMedGoogle ScholarCrossref
56.
Soumerai  SB, Majdumar  S, Lipton  HL.  Evaluating and improving physician prescribing.  In: Strom  BL. ed.  Pharmacoepidemiology. New York, NY: John Wiley & Sons Inc; 2000:483-503.Google Scholar
57.
Soumerai  SB, Avorn  J.  Principles of educational outreach ("academic detailing") to improve clinical decision making.  JAMA. 1990;263:549-556. PubMedGoogle ScholarCrossref
58.
Gonzales  R, Steiner  JF, Lum  A, Barrett Jr  PH.  Decreasing antibiotic use in ambulatory practice: impact of a multidimensional intervention on the treatment of uncomplicated acute bronchitis in adults.  JAMA. 1999;281:1512-1519. PubMedGoogle ScholarCrossref
Original Contribution
February 12, 2003

Predictors of Broad-Spectrum Antibiotic Prescribing for Acute Respiratory Tract Infections in Adult Primary Care

Author Affiliations

Author Affiliations: Division of Geriatrics and VA National Quality Scholars Program, San Francisco VA Medical Center, San Francisco, Calif (Drs Steinman and Landefeld), and Department of Medicine, University of California, San Francisco (Drs Steinman, Landefeld, and Gonzales).

JAMA. 2003;289(6):719-725. doi:10.1001/jama.289.6.719
Abstract

Context  Broad-spectrum antibiotics are commonly prescribed, but little is known about the physicians who prescribe and the patients who take these agents.

Objective  To identify factors associated with prescribing of broad-spectrum antibiotics by physicians caring for patients with nonpneumonic acute respiratory tract infections (ARTIs).

Design, Setting, and Patients  Cross-sectional study using data from the National Ambulatory Medical Care Survey between 1997 and 1999. Information was collected on a national sample of 1981 adults seen by physicians for the common cold and nonspecific upper respiratory tract infections (URTIs) (24%), acute sinusitis (24%), acute bronchitis (23%), otitis media (5%), pharyngitis, laryngitis, and tracheitis (11%), or more than 1 of the above diagnoses (13%).

Main Outcome Measure  Prescription of broad-spectrum antibiotics, defined for this study as quinolones, amoxicillin/clavulanate, second- and third-generation cephalosporins, and azithromycin and clarithromycin.

Results  Antibiotics were prescribed to 63% of patients with an ARTI, ranging from 46% of patients with the common cold or nonspecific URTIs to 69% of patients with acute sinusitis. Broad-spectrum agents were chosen in 54% of patients prescribed an antibiotic, including 51% of patients with the common cold and nonspecific URTIs, 53% with acute sinusitis, 62% with acute bronchitis, and 65% with otitis media. Multivariable analysis identified several clinical and nonclinical factors associated with choice of a broad-spectrum agent. After adjusting for diagnosis and chronic comorbid illnesses, the strongest independent predictors of broad-spectrum antibiotic prescribing were physician specialty (odds ratio [OR], 2.4; 95% confidence interval [CI], 1.6-3.5 for internal medicine physicians compared with general and family physicians) and geographic region (OR, 2.6; 95% CI, 1.4-4.8 for Northeast and OR, 2.4; 95% CI, 1.4-4.2 for South [both compared with West]). Other independent predictors of choosing a broad-spectrum agent included black race, lack of health insurance, and health maintenance organization membership, each of which was associated with lower rates of broad-spectrum prescribing. Patient age, sex, and urban vs rural location were not significantly associated with prescribing choice.

Conclusions  Broad-spectrum antibiotics are commonly prescribed for the treatment of ARTIs, especially by internists and physicians in the Northeast and South. These high rates of prescribing, wide variations in practice patterns, and the strong association of nonclinical factors with antibiotic choice suggest opportunities to improve prescribing patterns.

Antibiotics are commonly prescribed for acute respiratory tract infections (ARTIs).1-4 Some of this use, such as for the common cold and uncomplicated acute bronchitis, is almost always unnecessary.5,6 For other infections, such as otitis media and acute sinusitis, antibiotics provide some benefit, but the value of their use as first-line treatment has been debated.6,7

Even for conditions in which antibiotic use might be justified, many experts have expressed concern about substantial overuse of newer broad-spectrum antibiotics such as the quinolones, amoxicillin/clavulanate, second- and third-generation cephalosporins, and second-generation macrolides. Recent studies have documented frequent use of nonrecommended and second-line antibiotics for patients with sore throat and sinusitis, and found that nonrecommended antibiotic use increased over the last decade among patients with sore throat.2,3 These agents, while providing little therapeutic benefit for most patients, add substantially to health care costs. More importantly, frequent use of these agents promotes bacterial resistance.8-10 As a result, future patients may face high levels of resistance to some of today's most powerful therapies.

A variety of factors affect prescribing behavior, including the clinical characteristics of patients, physician training and local patterns of practice, drug cost and formulary restrictions, and pharmaceutical detailing and marketing.11-17 However, most studies of outpatient antibiotic prescribing have focused on the decision to treat or not to treat, rather than on the choice of a particular antibiotic. Information about which factors affect the choice of a particular drug could help efforts to improve both the quality and the quantity of antibiotic prescribing.

In this study, we sought to identify factors associated with the prescribing of broad-spectrum antibiotics by primary care physicians caring for patients with nonpneumonic ARTIs.

METHODS
Sample and Data Set

We used data from the National Ambulatory Medical Care Survey (NAMCS) between 1997 and 1999. Conducted annually by the National Center for Health Statistics, the NAMCS is a nationally representative probability sample of patient visits to community-based outpatient physicians.18-20 Upon selection for participation in this survey, physicians were asked to complete a standardized encounter form for each eligible patient visit during a 1-week period. Eligible encounters ranged from each patient encounter in small practices to every fifth encounter in large practices. Visits outside the physician office (eg, in nursing homes), visits to federal facilities, and visits to hospital-based outpatient or emergency departments were not included. Over the study period, approximately 1800 physicians were recruited each year, with approximately 63% to 69% agreeing to participate.18-20

All information was recorded by the physician or office staff on a standardized form. This form had space for up to 3 diagnoses and 6 medications, which were converted by NAMCS data entry staff into International Classification of Diseases, Ninth Revision, Clinical Modification (ICD-9-CM) codes and a standardized drug coding schema.21 Using a 4-step process, each patient visit was weighted by the inverse probability of its selection (ie, the number of visits in the population that the sampled visit was taken to represent).20 Using these weights, data from this survey can be extrapolated to the universe of community-based outpatient visits that occur in the United States each year.

Inclusion Criteria and Definitions

The study sample was drawn from 3151 patient visits made by adults aged 18 years or older who received a primary, secondary, or tertiary diagnosis of a nonpneumonic ARTI. These diagnoses included nasopharyngitis (ie, the common cold) or upper respiratory tract infection (URTI) not otherwise specified (ICD-9-CM codes 460 and 465), acute or unspecified sinusitis (ICD-9-CM codes 461 and 473.9), suppurative or nonsuppurative otitis media (ICD-9-CM codes 381.0-381.4 and 382), acute or unspecified bronchitis and bronchiolitis (ICD-9-CM codes 466 and 490), and other URTI including pharyngitis, laryngitis, tracheitis, and streptococcal sore throat (ICD-9-CM codes 462, 464, and 34.0). During the study period, these diagnoses accounted for 7% of all adult outpatient visits to community-based physicians in the United States, and for 44% of all visits involving an antibiotic prescription.

We excluded 66 visits that involved a concomitant diagnosis with other common outpatient infections that might be treated with an antibiotic, including chronic sinusitis (ICD-9-CM codes 473.0-473.8), influenza (ICD-9-CM code 487), bacterial or unspecified pneumonia (ICD-9-CM codes 481-483 and 485-486), urinary tract infection or acute/unspecified cystitis (ICD-9-CM codes 599.0, 595.0, and 595.9), cellulitis, carbuncle, or furuncle (ICD-9-CM codes 680-682), prostatitis or pelvic inflammatory disease (ICD-9-CM codes 601 and 614), and sexually transmitted diseases including syphilis, gonococcal infections, and other venereal infections (ICD-9-CM codes 90-99 and 647.0-647.2).

We subsequently excluded 1054 visits to pediatricians and nonprimary care physicians (ie, physicians with specialties other than general practice, family practice, or internal medicine). In the NAMCS database, most internal medicine subspecialists were coded separately from generalists and were excluded from this analysis. Finally, we excluded an additional 50 visits for which the patient was referred to the treating physician.

Of the remaining 1981 patient visits, 1257 involved an antibiotic. Antibiotics were defined as antibacterial agents commonly available in oral or intramuscular form for outpatient use. We did not count polymyxins and aminoglycosides in our tally of antibiotics, as these medications are used almost exclusively in topical form for outpatients. Similarly, we did not count antimycobacterial medications as antibiotics, as they are uncommonly used in the treatment of typical ARTIs. Of the remaining agents, we classified the quinolones, amoxicillin/clavulanate, second- and third-generation cephalosporins, and azithromycin and clarithromycin as broad-spectrum antibiotics. All other antibiotics were classified as narrow-spectrum.

Thirty-seven (3%) of 1257 patients received more than 1 antibiotic, for a total of 1298 antibiotic prescriptions. For most of our analyses, the patient was the unit of analysis (ie, patients who received at least 1 broad-spectrum antibiotic were counted in the broad-spectrum antibiotic category). For our analysis evaluating use of specific agents, we made the antibiotic the unit of analysis, such that each antibiotic was counted separately.

Analysis

To identify patients with active comorbid conditions, we searched the 3 diagnosis fields for common chronic diseases that were listed alongside the ARTI diagnosis. These included chronic obstructive pulmonary disease or asthma (ICD-9-CM codes 491-493 and 496), congestive heart failure (ICD-9-CM code 428), ischemic or hypertensive heart disease (ICD-9-CM codes 410-414, 402, and 404), diabetes mellitus (ICD-9-CM code 250), malignant neoplasms and neoplasms of uncertain or unspecified behavior (ICD-9-CM codes 140-208 and 235-239), cerebrovascular disease (ICD-9-CM codes 430-438), acute or chronic renal disease (ICD-9-CM codes 403 and 580-588), and human immunodeficiency virus infection (ICD-9-CM code 042). Comorbid conditions were recorded in 96 (8%) of 1257 patients receiving an antibiotic.

In each analysis, we adjusted for patient weights to make our point estimates of antibiotic prescriptions nationally representative.20 These weights are assigned in inverse proportion to the probability of a visit being selected from the universe of patient visits. As such, they denote the number of visits in the US population that are represented by the sampled visit. To account for the sample's hierarchical design, we adjusted for clustering at the level of the individual physician, thus increasing the variance around our point estimates to account for nonindependence of observations within the sampling frame. The 1981 visits in our sample were conducted by 558 physicians, with a mean (SD) of 3.6 (2.8) visits per physician. The intracluster correlation coefficient for broad-spectrum antibiotic prescriptions among these physicians was 0.31. Additional cluster identifiers based on the local geographic sampling unit were not available in the public-use data set at the time of this analysis, and could not be used in our calculations.

Adjusting for patient weights and physician-level clustering, we performed bivariate analyses with the design-based F test, which is analogous to a χ2 test for complexly sampled data. Multivariable analyses were performed with generalized estimating equations, using forward stepwise regression with P<.10 to stay and P<.05 to report. We checked for potential interactions, none of which added significantly to the model.

All analyses were performed using STATA statistical software (Intercooled Version 6.0; STATA Corp, College Station, Tex). All results except raw numbers are presented after adjustment for patient weights. P<.05 was considered significant. This research was approved by the Committee on Human Research of the University of California, San Francisco.

RESULTS

Of 1981 adults with nonpneumonic ARTIs that participated in the NAMCS, the most common diagnoses were sinusitis (24% of patient visits), the common cold and nonspecific URTIs (24%), and acute bronchitis (23%). Less common diagnoses were pharyngitis, laryngitis, or tracheitis (11%), otitis media (5%), and combinations of more than 1 ARTI (13%).

Antibiotics were prescribed to 63% of patients with an ARTI, ranging from 46% of patients with the common cold or nonspecific URTI to 69% of patients with sinusitis and 83% of patients with more than 1 ARTI (Figure 1). Similarly, prescriptions for broad-spectrum antibiotics were common, but varied across conditions. Broad-spectrum agents were chosen in 54% of patients prescribed an antibiotic, ranging from 33% of patients with pharyngitis, laryngitis, or tracheitis to 65% of patients with otitis media. Overall, broad-spectrum therapy was ordered for more than half of patients prescribed an antibiotic for each type of ARTI except pharyngitis, laryngitis, or tracheitis (Figure 1).

Azithromycin and clarithromycin were the most commonly prescribed broad-spectrum agents, comprising 21% of all antibiotic prescriptions for patients with ARTIs. These were followed by second- and third-generation cephalosporins (17% of prescriptions), amoxicillin/clavulanate (8%), and the quinolones (7%). Prescriptions for these agents varied substantially across diagnoses (Figure 2).

Among patients prescribed antibiotics, several factors were associated with choice of a broad-spectrum agent (Table 1). On unadjusted analyses, there was substantial variation by diagnosis, comorbidity, sex, national region, and physician specialty (P<.05 for each). On multivariable analysis controlling for diagnosis and comorbidity, the strongest independent predictors of broad-spectrum antibiotic choice were national region and physician specialty, with physicians in the Northeast and South and internists prescribing at particularly high rates. Other independent predictors included black (non-Hispanic) patient race, patient health maintenance organization membership, and lack of patient health insurance for the visit, each of which was associated with lower rates of broad-spectrum antibiotic prescribing.

The association between prescribing behavior and physician specialty and region appeared additive. The percentage of patients whose antibiotic regimen included a broad-spectrum agent ranged from 41% for general and family physicians in the Midwest and West, to 50% for general and family physicians in the Northeast and South, to 56% for internists in the Midwest and West, to 76% for internists in the Northeast and South.

The total volume of broad-spectrum antibiotic prescriptions depends on overall rates of antibiotic prescribing and the proportion of those prescriptions that are for broad-spectrum agents. Our primary analysis measured only the proportion of broad-spectrum prescriptions among patients who received an antibiotic. To control for overall rates of antibiotic prescribing, we rebuilt our multivariable analysis to compare patients who received a broad-spectrum antibiotic with all other patients (ie, those who received either a narrow-spectrum agent or no antibiotic at all). After controlling for diagnosis and comorbidity, broad-spectrum antibiotic prescribing was still significantly predicted by internal medicine specialty (odds ratio [OR], 1.6; 95% confidence interval [CI], 1.2-2.3) compared with general and family physicians and region (Northeast: OR, 2.3; 95% CI, 1.4-3.8 and South: OR, 2.0; 95% CI, 1.3-3.2 [both compared with West]). Other variables from the original model remained generally stable as well.

COMMENT

Our study has 2 principal findings. First, when primary care physicians prescribe antibiotics for ARTIs, they choose a broad-spectrum agent more than half the time. Second, there is wide variation in prescribing of broad-spectrum agents among different groups of patients and physicians, even after controlling for diagnosis and comorbidities. Internists and physicians in the Northeast and South were particularly likely to prescribe broad-spectrum antibiotics. More than three quarters of antibiotic regimens among patients seen by internists in these 2 regions involved a broad-spectrum agent. Other factors including patient race, health maintenance organization membership, and insurance status were also significantly associated with choice of a broad-spectrum agent.

It is difficult to quantify the correct amount of broad-spectrum antibiotic prescribing, yet several features suggest that the rates we observed are too high. Many of the diseases we studied do not require any antibiotic therapy at all, such as the common cold and acute bronchitis.5,6 It has been estimated that 55% of all antibiotic prescriptions for ARTIs are unnecessary.22 Moreover, even for ARTIs that may benefit from antibiotics, there is little evidence of clinically meaningful differences in cure rates between broad- and narrow-spectrum agents.3,23

Overall rates of medication use and variation in prescribing between physicians stem from a wide array of factors that affect clinical decisions.13,17,24 Traditional biomedical factors (characteristics of diseases, drugs, and physician knowledge) form one important set of inputs to the prescribing decision.25 However, a host of other factors ranging from physician and patient attitudes to environmental constraints can play an equal or greater role in physician decision making.13,17,25 Physicians with a similar knowledge base may make different judgments about the relative importance of various characteristics of a drug, such as its efficacy, ease of use, adverse effect profile, cost, and potential effect on resistance in the community.26-28 Numerous studies document that patient expectations, driven by explanatory models of illness and increasingly by direct-to-consumer advertising, can be a major driver of prescribing behavior.17,25,29-31 For example, physicians are far more likely to prescribe an antibiotic when they think their patient expects a drug, even though these perceptions often do not match the reality of patients' wishes.32,33 Drug samples can also influence prescribing decisions for individual patients, as do the wide range of other marketing efforts such as advertisements, gifts, and promotional detailing.34-40 Finally, formulary restrictions and other external constraints can promote the use of certain drugs over others.41-43

Variation in prescribing behavior between different geographic regions and specialties may reflect an asymmetrical distribution of this wide range of factors. For example, knowledge of recommended treatment strategies for ARTIs may vary between groups of physicians. Perhaps more importantly, the medical culture in which physicians practice (ie, within a specialty or geographic region) may impart different values to therapeutic decisions.27,44-46 Highlighting this point, Metlay et al26 demonstrated that infectious disease specialists are more likely than generalists to incorporate public health concerns into their choice of antibiotic. In addition, the distribution of environmental factors may have substantial heterogeneity. For example, different degrees of managed care penetration (and attendant formulary restrictions) may help explain regional differences in patterns of care,47-49 and pharmaceutical marketing efforts may be greater in areas where the perceived reward is higher.

Other studies of antibiotic choice confirm the complex interplay of these influences and highlight the role of factors other than a standardized application of accurate clinical knowledge. In a recent study of nonrecommended antibiotic use among patients with pharyngitis, only nonclinical characteristics (ie, calendar year and type of health plan) were associated with medication choice.2 In contrast, another study of patients with sinusitis found that prescribing of second-line antibiotics was predicted by older age, comorbidity, and specialist care.3 Other research about whether patients receive an antibiotic has documented a similar mix of clinical and nonclinical inputs into the prescribing decision, including the patient's clinical presentation, age, and perceived patient expectation for antibiotics, as well as physician factors such as clinician age, specialty, method of renumeration, and practice volume.1,2,12,32,50-53

It is difficult to disentangle the web of factors that contribute to variation in broad-spectrum antibiotic prescribing. However, cultural factors shared within physician specialties and geographic regions that impact prescribing could serve as important targets for intervention efforts.13,17 The use of opinion leaders, clinical champions, and one-on-one education (termed academic detailing) have been among the most effective strategies for modifying physician prescribing.54,55 The success of these interventions has been attributed to their ability to address local cultures of practice by modeling approaches to drug use in a context that resonates with the audience.56,57 Similarly, other successful interventions have accounted for the unique patient, physician, and system influences that impact a discrete group of physicians, simultaneously addressing multiple targets to change the local culture of practice and thereby improve prescribing.58

Our findings should be interpreted in light of this study's limitations. Although better than most surveys of its kind, the response rate of the NAMCS was incomplete, and like all surveys there was potential for inaccurate data collection. Our data included little clinical information. For instance, there was no indication if the study visit was an initial or follow-up appointment. Also, our comorbidity variable likely missed many patients with quiescent comorbid conditions, although by preferentially capturing patients with active comorbidities, we likely overestimated the association between comorbidity and broad-spectrum antibiotic prescribing. Although we were unable to completely eliminate other clinical confounders (eg, specialty-based differences in illness severity and regional differences in microbiological resistance), these confounders are unlikely to fully account for the intergroup variation that we observed. Other important clinical variables such as age did not predict broad-spectrum antibiotic choice, even when comorbidity was not included in the multivariate model. Moreover, our findings persisted after expanding the analysis to include patients who did not receive an antibiotic, thus making it unlikely that physicians with the highest proportion of broad-spectrum agents were merely reserving antibiotic therapy for their sickest patients.

In summary, our findings show high rates of broad-spectrum antibiotic prescribing, even among conditions for which antibiotic therapy is not indicated at all. After controlling for clinical characteristics, physician specialty and region were strongly associated with choice of broad-spectrum agents. Further investigation, both globally and locally, is needed to more clearly delineate the forces that contribute to this variation. Whatever these reasons, past experience shows that interventions relying on passive education or a one-size-fits-all approach are not enough. Instead, interventions that focus on understanding and changing shared cultures of practice will be most likely to improve the quality of medication prescribing.

Back to top
Article Information

Corresponding Author and Reprints: Michael A. Steinman, MD, Division of Geriatrics, San Francisco VA Medical Center, 4150 Clement St, Box 181-G, San Francisco, CA 94121 (e-mail: mstein@itsa.ucsf.edu).

Author Contributions: The authors had full access to the data, which is contained in a publicly accessible data set available from the Centers for Disease Control and Prevention.

Study concept and design: Steinman, Gonzales.

Acquisition of data: Steinman.

Analysis and interpretation of data: Steinman, Landefeld, Gonzales.

Drafting of the manuscript: Steinman, Gonzales.

Critical revision of the manuscript for important intellectual content: Steinman, Landefeld, Gonzales.

Statistical expertise: Steinman.

Obtained funding: Steinman, Landefeld, Gonzales.

Administrative, technical, or material support: Steinman.

Study supervision: Landefeld, Gonzales.

Funding/Support: Research support was provided by the VA National Quality Scholars Program, the Robert Wood Johnson Foundation Minority Medical Faculty Development Program, and grants from the National Institute on Aging, the John A. Hartford Foundation, Dartmouth College, and a grant to Dartmouth College from the Pfizer Foundation.

Previous Presentation: Presented at the Society of General Internal Medicine Annual Meeting, May 2-4, 2002; Atlanta, Ga.

Acknowledgment: We thank Eric Vittinghoff, PhD, for his assistance with statistical issues.

Financial Disclosure: Dr Gonzales received honoraria from Abbott Pharmaceuticals for oral presentations, and unrestricted educational grants from Abbott Pharmaceuticals, SmithKlineBeecham, and Roche to help fund a patient educational initiative in 2000.

References
1.
Gonzales  R, Steiner  JF, Sande  MA.  Antibiotic prescribing for adults with colds, upper respiratory tract infections, and bronchitis by ambulatory care physicians.  JAMA. 1997;278:901-904. PubMedGoogle ScholarCrossref
2.
Linder  JA, Stafford  RS.  Antibiotic treatment of adults with sore throat by community primary care physicians: a national survey, 1989-1999.  JAMA. 2001;286:1181-1186. PubMedGoogle ScholarCrossref
3.
Piccirillo  JF, Mager  DE, Frisse  ME, Brophy  RH, Goggin  A.  Impact of first-line vs second-line antibiotics for the treatment of acute uncomplicated sinusitis.  JAMA. 2001;286:1849-1856. PubMedGoogle ScholarCrossref
4.
McCaig  LF, Besser  RE, Hughes  JM.  Trends in antimicrobial prescribing rates for children and adolescents.  JAMA. 2002;287:3096-3102. PubMedGoogle ScholarCrossref
5.
Snow  V, Mottur-Pilson  C, Gonzales  R.  Principles of appropriate antibiotic use for treatment of nonspecific upper respiratory tract infections in adults.  Ann Intern Med. 2001;134:487-489. PubMedGoogle ScholarCrossref
6.
Snow  V, Mottur-Pilson  C, Gonzales  R.  Principles of appropriate antibiotic use for treatment of acute bronchitis in adults.  Ann Intern Med. 2001;134:518-520. PubMedGoogle ScholarCrossref
7.
Takata  GS, Chan  LS, Shekelle  P, Morton  SC, Mason  W, Marcy  SM.  Evidence assessment of management of acute otitis media, I: the role of antibiotics in treatment of uncomplicated acute otitis media.  Pediatrics. 2001;108:239-247. PubMedGoogle ScholarCrossref
8.
Austin  DJ, Kristinsson  KG, Anderson  RM.  The relationship between the volume of antimicrobial consumption in human communities and the frequency of resistance.  Proc Natl Acad Sci U S A. 1999;96:1152-1156. PubMedGoogle ScholarCrossref
9.
Not Available.  Resistance of Streptococcus pneumoniae to fluoroquinolones—United States, 1995-1999  MMWR Morb Mortal Wkly Rep. 2001;50:800-804. PubMedGoogle Scholar
10.
Hyde  TB, Gay  K, Stephens  DS,  et al.  Macrolide resistance among invasive Streptococcus pneumoniae isolates.  JAMA. 2001;286:1857-1862. PubMedGoogle ScholarCrossref
11.
Gonzales  R, Barrett Jr  PH, Crane  LA, Steiner  JF.  Factors associated with antibiotic use for acute bronchitis.  J Gen Intern Med. 1998;13:541-548. PubMedGoogle ScholarCrossref
12.
Gonzales  R, Barrett Jr  PH, Steiner  JF.  The relation between purulent manifestations and antibiotic treatment of upper respiratory tract infections.  J Gen Intern Med. 1999;14:151-156. PubMedGoogle ScholarCrossref
13.
Hemminki  E.  Review of literature on the factors affecting drug prescribing.  Soc Sci Med. 1975;9:111-116. PubMedGoogle ScholarCrossref
14.
Soumerai  SB, Ross-Degnan  D, Avorn  J, McLaughlin  T, Choodnovskiy  I.  Effects of Medicaid drug-payment limits on admission to hospitals and nursing homes.  N Engl J Med. 1991;325:1072-1077. PubMedGoogle ScholarCrossref
15.
Wazana  A.  Physicians and the pharmaceutical industry: is a gift ever just a gift?  JAMA. 2000;283:373-380. PubMedGoogle ScholarCrossref
16.
Avorn  J, Soumerai  SB.  Improving drug-therapy decisions through educational outreach: a randomized controlled trial of academically based "detailing."  N Engl J Med. 1983;308:1457-1463. PubMedGoogle ScholarCrossref
17.
Avorn  J, Solomon  DH.  Cultural and economic factors that (mis)shape antibiotic use: the nonpharmacologic basis of therapeutics.  Ann Intern Med. 2000;133:128-135. PubMedGoogle ScholarCrossref
18.
National Center for Health Statistics.  Public Use Microdata File Documentation, National Ambulatory Medical Care Survey, 1997 Hyattsville, Md: National Technical Information Service; 1999.
19.
National Center for Health Statistics.  Public Use Microdata File Documentation, National Ambulatory Medical Care Survey, 1998 Hyattsville, Md: National Technical Information Service; 2000.
20.
National Center for Health Statistics.  Public Use Microdata File Documentation, National Ambulatory Medical Care Survey, 1999 Hyattsville, Md: National Technical Information Service; 2001.
21.
Not Available.  International Classification of Diseases, Ninth Revision, Clinical Modification. Washington, DC: Public Health Service, US Dept of Health and Human Services; 1988.
22.
Gonzales  R, Malone  DC, Maselli  JH, Sande  MA.  Excessive antibiotic use for acute respiratory infections in the United States.  Clin Infect Dis. 2001;33:757-762. PubMedGoogle ScholarCrossref
23.
Ioannidis  JP, Contopoulos-Ioannidis  DG, Chew  P, Lau  J.  Meta-analysis of randomized controlled trials on the comparative efficacy and safety of azithromycin against other antibiotics for upper respiratory tract infections.  J Antimicrob Chemother. 2001;48:677-689. PubMedGoogle ScholarCrossref
24.
Cabana  MD, Rand  CS, Powe  NR,  et al.  Why don't physicians follow clinical practice guidelines? a framework for improvement.  JAMA. 1999;282:1458-1465. PubMedGoogle ScholarCrossref
25.
Schwartz  RK, Soumerai  SB, Avorn  J.  Physician motivations for nonscientific drug prescribing.  Soc Sci Med. 1989;28:577-582. PubMedGoogle ScholarCrossref
26.
Metlay  JP, Shea  JA, Crossette  LB, Asch  DA.  Tensions in antibiotic prescribing: pitting social concerns against the interests of individual patients.  J Gen Intern Med. 2002;17:87-94. PubMedGoogle ScholarCrossref
27.
Epstein  AM, Read  JL, Winickoff  R.  Physician beliefs, attitudes, and prescribing behavior for anti-inflammatory drugs.  Am J Med. 1984;77:313-318. PubMedGoogle ScholarCrossref
28.
Carrin  G.  Drug prescribing: a discussion of its variability and (ir)rationality.  Health Policy. 1987;7:73-94. PubMedGoogle ScholarCrossref
29.
Cockburn  J, Pit  S.  Prescribing behaviour in clinical practice: patients' expectations and doctors' perceptions of patients' expectations—a questionnaire study.  BMJ. 1997;315:520-523. PubMedGoogle ScholarCrossref
30.
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;315:1211-1214. PubMedGoogle ScholarCrossref
31.
Wilkes  MS, Bell  RA, Kravitz  RL.  Direct-to-consumer prescription drug advertising: trends, impact, and implications.  Health Aff (Millwood). 2000;19:110-128. PubMedGoogle ScholarCrossref
32.
Hamm  RM, Hicks  RJ, Bemben  DA.  Antibiotics and respiratory infections: are patients more satisfied when expectations are met?  J Fam Pract. 1996;43:56-62. PubMedGoogle Scholar
33.
Butler  CC, Rollnick  S, Pill  R, Maggs-Rapport  F, Stott  N.  Understanding the culture of prescribing: qualitative study of general practitioners' and patients' perceptions of antibiotics for sore throats.  BMJ. 1998;317:637-642. PubMedGoogle ScholarCrossref
34.
Chew  LD, O'Young  TS, Hazlet  TK, Bradley  KA, Maynard  C, Lessler  DS.  A physician survey of the effect of drug sample availability on physicians' behavior.  J Gen Intern Med. 2000;15:478-483. PubMedGoogle ScholarCrossref
35.
Bell  RA, Wilkes  MS, Kravitz  RL.  Advertisement-induced prescription drug requests: patients' anticipated reactions to a physician who refuses.  J Fam Pract. 1999;48:446-452. PubMedGoogle Scholar
36.
Rosenthal  MB, Berndt  ER, Donohue  JM, Frank  RG, Epstein  AM.  Promotion of prescription drugs to consumers.  N Engl J Med. 2002;346:498-505. PubMedGoogle ScholarCrossref
37.
Chren  MM, Landefeld  CS, Murray  TH.  Doctors, drug companies, and gifts.  JAMA. 1989;262:3448-3451. PubMedGoogle ScholarCrossref
38.
Chren  MM, Landefeld  CS.  Physicians' behavior and their interactions with drug companies: a controlled study of physicians who requested additions to a hospital drug formulary.  JAMA. 1994;271:684-689. PubMedGoogle ScholarCrossref
39.
Lurie  N, Rich  EC, Simpson  DE,  et al.  Pharmaceutical representatives in academic medical centers: interaction with faculty and housestaff.  J Gen Intern Med. 1990;5:240-243. PubMedGoogle ScholarCrossref
40.
Ziegler  MG, Lew  P, Singer  BC.  The accuracy of drug information from pharmaceutical sales representatives.  JAMA. 1995;273:1296-1298. PubMedGoogle ScholarCrossref
41.
Lambert  BL, Salmon  JW, Stubbings  J, Gilomen-Study  G, Valuck  RJ, Kezlarian  K.  Factors associated with antibiotic prescribing in a managed care setting: an exploratory investigation.  Soc Sci Med. 1997;45:1767-1779. PubMedGoogle ScholarCrossref
42.
Avery  AJ, Walker  B, Heron  T, Teasdale  SJ.  Do prescribing formularies help GPs prescribe from a narrower range of drugs? a controlled trial of the introduction of prescribing formularies for NSAIDs.  Br J Gen Pract. 1997;47:810-814. PubMedGoogle Scholar
43.
Mather  JL, Bayliff  CD, Reider  MJ, Hussain  Z, Colby  WD.  The impact of formulary reservations on drug utilization: a controlled trial.  Can J Hosp Pharm. 1994;47:111-116. PubMedGoogle Scholar
44.
Harrold  LR, Field  TS, Gurwitz  JH.  Knowledge, patterns of care, and outcomes of care for generalists and specialists.  J Gen Intern Med. 1999;14:499-511. PubMedGoogle ScholarCrossref
45.
Walker  S, McGeer  A, Simor  AE, Armstrong-Evans  M, Loeb  M.  Why are antibiotics prescribed for asymptomatic bacteriuria in institutionalized elderly people? a qualitative study of physicians' and nurses' perceptions.  CMAJ. 2000;163:273-277. PubMedGoogle Scholar
46.
Safavi  KT, Hayward  RA.  Choosing between apples and apples: physicians' choices of prescription drugs that have similar side effects and efficacies.  J Gen Intern Med. 1992;7:32-37. PubMedGoogle ScholarCrossref
47.
Hasty  M, Schrager  J, Wrenn  K.  Physicians' perceptions about managed care restrictions on antibiotic prescribing.  J Gen Intern Med. 1999;14:756-758. PubMedGoogle ScholarCrossref
48.
Hadley  J, Mitchell  JM, Sulmasy  DP, Bloche  MG.  Perceived financial incentives, HMO market penetration, and physicians' practice styles and satisfaction.  Health Serv Res. 1999;34:307-321. PubMedGoogle Scholar
49.
Reschovsky  J, Reed  M, Blumenthal  D, Landon  B.  Physicians' assessments of their ability to provide high-quality care in a changing health care system.  Med Care. 2001;39:254-269. PubMedGoogle ScholarCrossref
50.
Mainous 3rd  AG, Hueston  WJ, Eberlein  C.  Colour of respiratory discharge and antibiotic use.  Lancet. 1997;350:1077. PubMedGoogle ScholarCrossref
51.
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;158:1813-1818. PubMedGoogle ScholarCrossref
52.
Mainous III  AG, Hueston  WJ, Love  MM.  Antibiotics for colds in children: who are the high prescribers?  Arch Pediatr Adolesc Med. 1998;152:349-352. PubMedGoogle ScholarCrossref
53.
Hutchinson  JM, Foley  RN.  Method of physician remuneration and rates of antibiotic prescription.  CMAJ. 1999;160:1013-1017. PubMedGoogle Scholar
54.
Oxman  AD, Thomson  MA, Davis  DA, Haynes  RB.  No magic bullets: a systematic review of 102 trials of interventions to improve professional practice.  CMAJ. 1995;153:1423-1431. PubMedGoogle Scholar
55.
Bero  LA, Grilli  R, Grimshaw  JM, Harvey  E, Oxman  AD, Thomson  MA, for the Cochrane Effective Practice and Organization of Care Review Group.  Closing the gap between research and practice: an overview of systematic reviews of interventions to promote the implementation of research findings.  BMJ. 1998;317:465-468. PubMedGoogle ScholarCrossref
56.
Soumerai  SB, Majdumar  S, Lipton  HL.  Evaluating and improving physician prescribing.  In: Strom  BL. ed.  Pharmacoepidemiology. New York, NY: John Wiley & Sons Inc; 2000:483-503.Google Scholar
57.
Soumerai  SB, Avorn  J.  Principles of educational outreach ("academic detailing") to improve clinical decision making.  JAMA. 1990;263:549-556. PubMedGoogle ScholarCrossref
58.
Gonzales  R, Steiner  JF, Lum  A, Barrett Jr  PH.  Decreasing antibiotic use in ambulatory practice: impact of a multidimensional intervention on the treatment of uncomplicated acute bronchitis in adults.  JAMA. 1999;281:1512-1519. PubMedGoogle ScholarCrossref
×