Skip Navigation

A Multifaceted Intervention to Improve Antimicrobial Prescribing for Upper Respiratory Tract Infections in a Small Rural Community

  1. Michael A. Rubin1,
  2. Kim Bateman3,
  3. Stephen Alder2,
  4. Sharon Donnelly3,
  5. Gregory J. Stoddard1, and
  6. Matthew H. Samore1
  1. 1Departments of Internal Medicine and University of Utah, Salt Lake City, Utah
  2. 2Departments of Family and Preventive Medicine, University of Utah, Salt Lake City, Utah
  3. 3Departments of HealthInsight, Salt Lake City, Utah
  1. Reprints or correspondence: Dr. Michael A. Rubin, Div. of Clinical Epidemiology, Dept. of Internal Medicine, University of Utah School of Medicine, 300 N. 1900 East, Rm. AC-230A, Salt Lake City, UT 84132 (Michael.Rubin{at}hsc.utah.edu).
 
Next Section

Abstract

Background. Antibiotic prescribing for upper respiratory tract infections (URTIs) is widespread, is often inappropriate, and may contribute to antibiotic resistance among community-acquired pathogens, such as Streptococcus pneumoniae.

Methods. A multifaceted intervention involving health care professionals and patients was introduced to a small rural Utah community and included the repetitive use of printed diagnostic and treatment algorithms by professionals. Data on the quantity and class of antibiotic prescribing, which were collected from multiple sources, were measured for the intervention period (from January through June) in 2001 and compared with data for the baseline period during the same months in 2000.

Results. Medicaid claims data revealed that the percentage of patients in the community who received antibiotics for URTIs during the intervention period was 15.6% less than that for the baseline period, whereas the percentage in the rest of rural Utah was relatively stable, with a 1.5% decrease (P = .006). The greatest impact of the intervention was on prescribing for acute bronchitis (decreases of 56.1% and 1.7% in the community and rural Utah, respectively; P = .024) and on prescribing of macrolides (decreases of 13.4% and 0.2% in the community and rural Utah, respectively; P < .001). Community pharmacy data likewise revealed a 17.5% decrease in the rate of antibiotic prescribing during the intervention period (P < .001), with the largest decrease observed for macrolide prescribing (50.9%; P < .001). Chart review data, in contrast, revealed no significant decrease in the percentage of patients with URTI who were prescribed an antibiotic (3.8%; P = .49), although there was a significant decrease of 11.2% in macrolide use (P = .045).

Conclusions. A multifaceted intervention involving the repetitive use of printed algorithms resulted in modest improvements in antibiotic prescribing for outpatient URTIs, although one data source did not corroborate this. However, macrolide prescribing decreased sharply, irrespective of the source of data.

Antimicrobial agents are the second most commonly prescribed group of medications in the United States [1]. The overwhelming majority of such prescriptions in community practice are for the treatment of acute upper respiratory tract infections (URTIs), including rhinosinusitis, pharyngitis, bronchitis, otitis media, and nonspecific URTIs [2], despite the fact that viruses cause most of these illnesses. This widespread use of antibiotics has contributed to the spread of antibiotic resistance among community-acquired pathogens, such as Streptococcus pneumoniae.

Numerous interventions to alter antimicrobial prescribing practices have been reported, with varying results [3–6]. No single intervention appears to have a superior efficacy, although the use of small-group outreach visits (i.e., academic detailing) may be the most effective strategy for achieving change in physician behavior, as demonstrated in numerous studies [6–8]. Many experts now agree that combinations of interventions are more effective [9–12], and strategies that target health care professionals and patients (or parents of young patients) appear to have achieved success at reducing antimicrobial prescriptions for these predominantly viral conditions [13–17]. Most of these interventions were centered on educational sessions and materials for health care professionals and patients in urban and suburban settings.

In the present article, we describe the impact of a multidimensional intervention on antimicrobial prescribing for treatment of URTIs in a small rural town where the frequency of exposure to cephalosporins (particularly cefixime) among children was known to be high [18], although the total antimicrobial consumption was comparable to that in other rural communities. In addition to educational components, our intervention included algorithms for the diagnosis and management of acute URTIs. By including this component to facilitate changes in the decision-making process used by the health care professionals, we hypothesized that the effect of our education program would be enhanced, leading to significant reductions in the number of antimicrobial prescriptions and more appropriate selection of antimicrobials for cases that do require treatment (as determined by published evidence, guidelines, and expert opinion). Furthermore, by asking that health care professionals use the algorithms a minimum number of times in the treatment of successive patients with URTIs, we hypothesized that the repetition would help health care professionals break clinical habits that were inconsistent with judicious antibiotic prescribing.

Previous SectionNext Section

Participants and Methods

Design. The study was a prospective, nonrandomized trial to determine the effect of a multidimensional behavior-change strategy for prescribing antibiotics in cases of acute URTI (e.g., pharyngitis, rhinosinusitis, otitis media, bronchitis, and nonspecific URTI). The intervention occurred from January through June of 2001 (i.e., the “intervention period”). Data were then collected and analyzed retrospectively for the intervention period and for January through June of 2000 (i.e., the “baseline period”).

Setting and subjects. The study was conducted in a rural Utah community (Community A) of <10,000 residents that was chosen because previous data showed a high frequency of cefixime exposure among children [18]. Although health care professionals had previously received feedback concerning this finding, our baseline data indicate that cefixime continued to be heavily prescribed. In year 2000, the median age for residents of Community A was 27.7 years; 50.8% were male, and 61.3% were adults aged >18 years [19]. All patients presenting to their primary care professional with URTI symptoms were included in the study regardless of age, race, ethnicity, sex, or insurance coverage.

Health care for Community A is provided entirely by 2 family practice groups; 1 health care professional declined participation. No other groups practice (or prescribe antibiotics) in this area. Two community pharmacies dispense all antibiotics prescribed by these practitioners. To our knowledge, no other groups were involved in any antibiotic awareness campaigns in this community, aside from the ongoing nationwide campaign coordinated by the Centers for Disease Control and Prevention (Atlanta, GA) over approximately the past decade [20].

Intervention. The intervention consisted of 4 components: patient education materials, a media campaign to increase public awareness, 1 small group session involving physicians, and physician use of URTI algorithms. Patient education materials consisted of informational brochures displayed in the clinic office, and the media campaign consisted of references in the local media to the project and to antimicrobial resistance among community-acquired pathogens.

The small group session involving physicians included two 30-min discussions. The first provided an overview of antibiotic resistance in community-acquired pathogens such as S. pneumoniae, including trends in local resistance rates gathered from another study [18], as well as a review of appropriate antibiotic use for URTIs. The second introduced the algorithms and provided a tutorial on their use.

Two algorithms were created using published evidence, guidelines, and expert opinion outlining the diagnosis and treatment of URTIs in children and adults. Each fit on a single piece of paper for efficient, point-of-care use via check boxes for symptoms, signs, testing, and treatment options. Key components of these algorithms are provided in table 1. Health care professionals were asked to use the algorithms with ⩾200 consecutive patients presenting with URTI symptoms beginning in January 2001. This number of iterations was chosen because it represented an achievable target that would provide adequate exposure to patients with a variety of URTIs and sufficient repetition to help physicians break their clinical habits; it is feasible, however, that adequate behavior change could be achieved with fewer iterations. De-identified, completed algorithms were stored apart from patient charts and were collected at the end of the study period as a measure of their use. One physician completed 121 algorithms, and all others completed >200 (maximum, 268 algorithms).

Figure 1
View larger version:
Figure 1

Change in antibiotic prescribing in Community A in rural Utah from 2000 (baseline period, January through June) to 2001 (intervention period, January through June), assessed using local community pharmacy data. P values were determined by comparing prescription rates in 2000 with those in 2001 using the binomial probability mid-P exact test for rate data, with adjustment for multiple comparisons by means of the Holm procedure.

Figure 2
View larger version:
Figure 2

Percentage of patients with urinary tract infection in Community A in rural Utah who were treated with antibiotics in 2000 (baseline period, January through June), compared with those who were treated in 2001 (intervention period, January through June), assessed using manual chart review. P values were determined by means of Pearson's χ2 test (TOTAL) and Pearson's χ2 test with adjustment for multiple comparisons using the Holm procedure (MACROLIDES).

Table 1
View larger version:
Table 1

Key components of algorithms provided to physicians at a community practice in Utah for diagnosis and treatment of upper respiratory tract infections (URTIs)—2000–2001.

Data sources. We sampled data from multiple sources: Medicaid claims, local community pharmacies, and chart review. Medicaid claims were obtained for 2000–2001, from which we isolated URTI episodes during the specified periods using codes from the ninth revision of the International Classification of Disease (ICD-9), including (with appropriate subgroups) 034, 381.00–381.03, 381.4, 382, 388.7, 461, 473, 462, 463, 472, 474, 466, 487, 490, 491, 786.2, 460, 464, 465, 476, 519.9, and 786.1. We then identified all cases with an associated prescription claim. From local community pharmacies, we gathered anonymous, gross counts of antibiotic prescriptions filled during the specified periods. Manual chart review was performed at each clinic by trained data abstraction personnel, who randomly selected charts on the basis of ICD-9 codes provided by the clinics. All target URTI diagnoses that occurred during the specified periods were identified in chart visit notes; abstractors then collected anonymous data, including signs and symptoms, tests performed, test results, and any antibiotics prescribed. For both Medicaid and chart review data, a second visit by a patient ⩽30 days after the initial visit was interpreted as the same episode of illness and was excluded from analysis. Data from all sources were analyzed using Access 2000 (Microsoft).

Outcome measures. The primary outcome measure was the proportion of URTI episodes treated with an antibiotic (i.e., the number of episodes associated with a prescription divided by the total number of episodes during the specified period). Community pharmacy data included only gross numbers of antibiotic prescriptions and were not associated with any diagnosis. For these data, the primary outcome was the prescription rate, calculated as the total number of prescriptions per 1000 population during the specified period. Secondary outcomes included antibiotic prescribing for patients with particular URTI diagnoses and overall prescribing of the major antibiotic classes (penicillins, cephalosporins, macrolides, and quinolones).

Statistical methods. For proportional data, we compared periods using the χ2 test or Fisher's exact test, as appropriate. For data with >2 unordered categories, we used the Fisher-Freeman-Halton test (which is the Fisher's exact test extended to tables larger than 2 × 2 cells).

To compare community-level change in proportions across periods, we used a community-by-time interaction term in an ordinary (unconditional) logistic regression model. To control for specific diagnoses while making this same comparison, we used a conditional logistic regression model with specific diagnoses as the stratification variable. Comparisons within individual diagnosis strata (and individual antimicrobial class strata) were similarly made using the same interaction term of an unconditional logistic regression model, with P values adjusted for multiplicity using Holm's stepdown procedure for P value adjustment [21].

For rate-based data with person-time denominators, we compared periods using the binomial probability mid-P exact test [22]. Comparisons within individual antimicrobial class strata were similarly made, with reported P values adjusted for multiplicity using Holm's stepdown procedure for P value adjustment [21]. All statistical analyses were performed using the Stata statistical software package, version 8 (StataCorp).

Previous SectionNext Section

Results

Medicaid claims. The proportion of all URTIs treated with an antibiotic during the baseline and intervention periods, as measured using Medicaid claims data, is displayed in table 2. At baseline, health care professionals in Community A prescribed antibiotics for all URTIs at a proportion higher than that observed for the rest of rural Utah (72.6% vs. 63.0%). We accounted for this baseline difference in subsequent analyses by assessing the difference in the number of prescriptions between the baseline period (2000) and the intervention period (2001). During the intervention period, there were statistically significant decreases in the proportion of URTIs treated with antibiotics both in Community A (15.6%; P = .002) and in rural Utah (1.5%; P = .047), compared with the baseline period. The much larger decrease in prescribing observed in Community A was statistically significant, compared with the decrease seen in rural Utah (P = .006; P = .004 when controlled for specific diagnoses).

Table 2
View larger version:
Table 2

Change from 2000 (baseline period, January through June) to 2001 (intervention period, January through June) in the proportion of episodes of upper respiratory tract infection (URTI) treated with antibiotics, using data from Medicaid claims.

When data were stratified according to diagnosis, we observed a decrease in the frequency of antibiotic prescribing for most URTIs in Community A, along with little change in prescribing in rural Utah (table 2). The most significant decrease in prescribing in Community A was for acute bronchitis (56%, compared with 1.7% in rural Utah; P = .024). Although prescribing in Community A also decreased for acute pharyngitis, acute sinusitis, and nonspecific URTI, these changes were not statistically significant, compared with the changes observed in rural Utah.

We also examined data stratified according to the 4 major antimicrobial classes (cephalosporins, macrolides, penicillins, and quinolones; table 2). Antimicrobial prescribing remained stable throughout rural Utah from the baseline period to the intervention period. In Community A, however, we observed a decrease of 5.7% in the number of URTIs for which oral cephalosporin was prescribed and increases of 3.8% and 1.7% in the number of URTIs for which oral penicillin and quinolone were prescribed. None of these changes were statistically significant, compared with the corresponding changes seen in rural Utah. There was, however, a significant decrease in oral macrolide prescribing in Community A, compared with the decrease in rural Utah (13.4% vs. 0.2%; P < .001). This pattern of change was concordant with the recommendations in the treatment algorithms, which emphasized the use of narrower-spectrum agents, such as penicillins, over broader-spectrum agents, such as macrolides and some cephalosporins.

To determine whether the change in the frequency of antibiotic prescribing was associated with a change in the frequency of URTI diagnoses encountered (a so-called diagnosis shift), we examined the incidences of various URTI diagnoses for the baseline and intervention periods using Medicaid claims, as shown in table 3. Although the incidence of URTIs in rural Utah decreased by >11% from 2000 to 2001, the distribution of URTI diagnoses remained stable during each period. In contrast, in Community A, there was a decrease in the frequency of acute bronchitis (from 14% to 6%) and nonstreptococcal pharyngitis (from 23% to 14%) from 2000 to 2001, alongside an increase in the frequency of otitis media (from 33% to 43%) and nonspecific URTI (from 19% to 25%). The change in the distribution of these disease frequencies in Community A from 2000 to 2001 was statistically significant (P = .016). The overall incidence of URTIs in Community A, however, remained stable from 2000 to 2001 (175 vs. 179 episodes).

Table 3
View larger version:
Table 3

Distribution of upper respiratory tract infection (URTI) diagnoses recorded from 2000 (baseline period, January through June) to 2001 (intervention period, January through June), using data from Medicaid claims.

Community pharmacy data. Antibiotic prescription data collected from the community pharmacies were also analyzed for changes between the baseline and intervention periods. The change in the total number of antibiotic prescriptions filled in Community A from 2000 to 2001, as well as the change in the number of prescriptions of drugs from each of the 4 main antibiotic classes, is shown in figure 1. In Community A, there was a 17.5% decrease in all antibiotic prescriptions (P < .0001). Macrolides accounted for the majority of this decrease: 50.9% fewer prescriptions of macrolides occurred in 2001 than in 2000 (P < .0001). There was also a concomitant increase of 41.2% in the number of cephalosporin prescriptions (P < .0001) and a decrease of 10.8% in the number of penicillin prescriptions (P = .037).

The change in the prescription rates for selected antibiotics, as measured by community pharmacy data, is shown in table 4. From 2000 to 2001, there were decreases in the prescription rates for each of the 3 most commonly prescribed macrolides, with the largest decrease involving azithromycin (35.4 fewer prescriptions per 1000 population; P < .001). Additionally, there was a significant decrease in the prescription rate for the broad-spectrum agent amoxicillin/clavulanate (8.5 fewer prescriptions per 1000 population; P = .002). We also noticed that prescription rates for cefixime and loracarbef decreased, whereas the prescription rates for cephalexin and cefdinir increased (table 4); all of these findings were consistent with the recommendations specified in the treatment algorithms. The prescribing rates for the quinolones remained stable from 2000 to 2001, although rates for individual agents fluctuated to some extent (table 4).

Table 4
View larger version:
Table 4

Change in prescribing of selected antibiotics for upper respiratory tract infections (URTIs) in Community A from 2000 (baseline period, January through June) to 2001 (intervention period, January through June), using data from local community pharmacies.

Manual chart review. We manually reviewed 155 and 149 charts from the baseline and intervention periods, respectively, for URTI diagnosis and prescribing information. The results for overall antibiotic prescribing, as well as prescribing data for each of the 4 main antibiotic classes, are provided in figure 2. Unlike the previous findings, data compiled from chart review showed no significant change from 2000 to 2001 in the proportion of URTIs treated with an antibiotic (3.8% fewer URTIs were treated with antibiotics in 2001; P = .49). Likewise, there was no significant change from 2000 to 2001 in the proportion of URTIs treated with a cephalosporin (0.4% increase; P = 1.000), penicillin (4.9% increase; P = 1.000), or quinolone (0.2% increase; P = 1.000). There was, however, a significant decrease of 11.2% in the prescription of macrolides for treatment of URTIs (P = .045), which is consistent with the findings from Medicaid claims and community pharmacy data sources.

Previous SectionNext Section

Discussion

In recent years, great emphasis has been placed on curbing excessive antibiotic use, particularly for outpatient treatment of URTIs. Given that many of these prescriptions are unnecessary, considerable effort is being spent on interventions that promote judicious outpatient prescribing. Little focus has been placed on prescribing in rural communities, however. Here we demonstrate that a multifaceted intervention involving physicians and patients in a rural Utah community had a significant impact on antibiotic prescribing in 2 of 3 data sources. The greatest impact was on macrolide prescribing, which decreased regardless of the measurement source.

The level of change in antibiotic prescribing seen in our study is comparable to that seen in other studies that attempted to alter prescribing for URTIs via interventions targeting patients and health care professionals [13–16]. Gonzales et al. [13] used a multidimensional approach to demonstrate a 26% decrease in the rate of prescribing for acute bronchitis in 4 urban practices, without an increase in return visits for treatment of bronchitis or pneumonia. Similarly, an intervention at an urban urgent care clinic, which included a patient-focused computer education module, led to a 34% decrease in the percentage of patients for whom an antibiotic was prescribed for acute bronchitis and a 14% decrease in the rate of prescribing for all acute URTIs [15]. Other studies targeting antibiotic use for URTIs in children used multiple educational interventions for health care professionals and parents in urban and suburban practice settings, resulting in decreases in prescribing rates of up to 16% [16] and 11% [14].

Two other recent intervention trials were performed in rural communities in Wisconsin [17] and Alaska [23], and they achieved similar results. Belongia et al. [17] employed multiple clinician and community education tactics to reduce overall numbers of prescriptions by 11% and 23% for solid and liquid antibiotics, respectively, using sampled retail pharmacy data, although a separate analysis using chart review data revealed no decrease in prescribing for acute respiratory infections in children attending child care facilities. Hennessy et al. [23] also measured antibiotic prescribing for respiratory infections among rural Alaskan villages after an intervention targeting health care professionals and residents and found a 27% decrease in the rate of antibiotic prescribing per respiratory infection visit [23].

One of our concerns during this study was that a significant shift in the distribution of URTI diagnoses would occur in Community A in favor of those diagnoses for which antibiotics are better justified (such as acute otitis media, acute sinusitis, and streptococcal pharyngitis). In fact, there was a significant shift in disease distribution in Community A in 2001 (table 3), which was not seen in the rest of rural Utah. Whereas acute bronchitis and nonstreptococcal pharyngitis (predominantly viral diagnoses) decreased in frequency from 2000 to 2001, there was a concomitant increase in the frequency of otitis media and, of interest, nonspecific URTI. Although this suggests a diagnosis shift away from some illnesses that do not require antibiotics, our data do not allow us to distinguish between shifting on the part of the physicians and random or nonrandom fluctuations in the incidence of different URTIs. Furthermore, our approach to this analysis is, in effect, unbiased by diagnostic shift, because total antibiotic use was used as the primary outcome measure.

The use of multiple data sources to measure prescribing is one unique aspect of our study. However, the intervention's effect on antibiotic prescribing varied across the 3 data sources. Given the small sample sizes (particularly with respect to Medicaid and chart review data), some degree of random variability is not unexpected. Each data source also measures different aspects of antibiotic use. Both the Medicaid and chart review data measure diagnosis-specific prescribing but target different patient populations. Medicaid recipients have demographic characteristics unlike those of the general population; in our Medicaid data, 56.5% and 64.8% (in 2000 and 2001, respectively) of URTIs in Community A were in children ⩽5 years of age (data not shown), whereas in the chart review data, these proportions were 22.8% and 26.7% (in 2000 and 2001, respectively). Furthermore, community pharmacy data reflect the overall change in antibiotic prescribing, irrespective of diagnosis. Although such data are robust to the effects of diagnosis shifting, they are influenced by fluctuations in care-seeking behavior and disease incidence. Nonetheless, a consistent finding from each data source was a sharp decrease in macrolide use.

In summary, we employed a multifaceted, algorithm-based intervention to improve antibiotic prescribing for URTIs among office-based physicians in a rural Utah community. Our intervention stressed repetition as a method of ingraining behavior change and used multiple data sources to measure antibiotic prescribing. Although antibiotic prescribing decreased significantly according to Medicaid claims and community pharmacy data, chart review data could not confirm this. Macrolide use, however, decreased sharply regardless of the data source analyzed. Our data also indicate that outpatient antibiotic prescribing still has tremendous room for improvement, despite recent evidence that nationwide calls for judicious antibiotic use are beginning to be heard by many primary care clinicians [24–26]. Outpatient antibiotic prescribing continues to have an important influence on antimicrobial resistance in the community and remains a rich target for continued efforts to change prescribing behavior at the level of clinical practice.

Previous SectionNext Section

Acknowledgments

We'd like to thank Duane Parke, Jan C. Jacobsen, and Oscar Fuller (Division of Health Care Finance, Utah Department of Health; Salt Lake City) for providing Medicaid data for our study; and Terri Rose and Joleen Rischer (HealthInsight; Salt Lake City) for their assistance with chart review.

Financial support. Centers for Disease Control and Prevention (grant RS1 CCR820631).

Potential conflicts of interest. All authors: no conflicts.

  • Received June 14, 2004.
  • Accepted October 12, 2004.
Previous Section
 

References

  1. ↵
    1. Nelson CR
    . Drug utilization in office practice: National Ambulatory Medical Care Survey, 1990. Adv Data 1993;232:1-12.
    Medline
  2. ↵
    1. McCaig LF,
    2. Hughes JM
    . Trends in antimicrobial drug prescribing among office-based physicians in the United States. JAMA 1995;273:214-9.
    Abstract/FREE Full Text
  3. ↵
    1. Christakis DA,
    2. Zimmerman FJ,
    3. Wright JA,
    4. Garrison MM,
    5. Rivara FP,
    6. Davis RL
    . A randomized controlled trial of point-of-care evidence to improve the antibiotic prescribing practices for otitis media in children. Pediatrics 2001;107:15.
    CrossRef
    1. Beilby J,
    2. Marley J,
    3. Walker D,
    4. Chamberlain N,
    5. Burke M
    . Effect of changes in antibiotic prescribing on patient outcomes in a community setting: a natural experiment in Australia. Clin Infect Dis 2002;34:55-64.
    Abstract/FREE Full Text
    1. O'Connell DL,
    2. Henry D,
    3. Tomlins R
    . Randomised controlled trial of effect of feedback on general practitioners' prescribing in Australia. BMJ 1999;318:507-11.
    Abstract/FREE Full Text
  6. ↵
    1. Solomon DH,
    2. Van Houten L,
    3. Glynn RJ,
    4. et al
    . Academic detailing to improve use of broad-spectrum antibiotics at an academic medical center. Arch Intern Med 2001;161:1897-902.
    Abstract/FREE Full Text
    1. Avorn J,
    2. Soumerai SB
    . Improving drug-therapy decisions through educational outreach: a randomized controlled trial of academically based “detailing”. N Engl J Med 1983;308:1457-63.
    MedlineWeb of Science
  8. ↵
    1. Ilett KF,
    2. Johnson S,
    3. Greenhill G,
    4. et al
    . Modification of general practitioner prescribing of antibiotics by use of a therapeutics adviser (academic detailer). Br J Clin Pharmacol 2000;49:168-73.
    CrossRefMedlineWeb of Science
  9. ↵
    1. Greco PJ,
    2. Eisenberg JM
    . Changing physicians' practices. N Engl J Med 1993;329:1271-3.
    CrossRefMedlineWeb of Science
    1. Grol R
    . Personal paper: beliefs and evidence in changing clinical practice. BMJ 1997;315:418-21.
    FREE Full Text
    1. Davis DA,
    2. Taylor-Vaisey A
    . Translating guidelines into practice: a systematic review of theoretic concepts, practical experience and research evidence in the adoption of clinical practice guidelines. CMAJ 1997;157:408-16.
    Abstract/FREE Full Text
  12. ↵
    1. Smith WR
    . Evidence for the effectiveness of techniques to change physician behavior. Chest 2000;118:8-17.
    FREE Full Text
  13. ↵
    1. Gonzales R,
    2. Steiner JF,
    3. Lum A,
    4. Barrett PH Jr
    . Decreasing antibiotic use in ambulatory practice: impact of a multidimensional intervention on the treatment of uncomplicated acute bronchitis in adults. JAMA 1999;281:1512-9.
    Abstract/FREE Full Text
  14. ↵
    1. Perz JF,
    2. Craig AS,
    3. Coffey CS,
    4. et al
    . Changes in antibiotic prescribing for children after a community-wide campaign. JAMA 2002;287:3103-9.
    Abstract/FREE Full Text
  15. ↵
    1. Harris RH,
    2. MacKenzie TD,
    3. Leeman-Castillo B,
    4. et al
    . Optimizing antibiotic prescribing for acute respiratory tract infections in an urban urgent care clinic. J Gen Intern Med 2003;18:326-34.
    CrossRefMedlineWeb of Science
  16. ↵
    1. Finkelstein JA,
    2. Davis RL,
    3. Dowell SF,
    4. et al
    . Reducing antibiotic use in children: a randomized trial in 12 practices. Pediatrics 2001;108:1-7.
    Abstract/FREE Full Text
  17. ↵
    1. Belongia EA,
    2. Sullivan BJ,
    3. Chyou PH,
    4. Madagame E,
    5. Reed KD,
    6. Schwartz B
    . A community intervention trial to promote judicious antibiotic use and reduce penicillin-resistant Streptococcus pneumoniae carriage in children. Pediatrics 2001;108:575-83.
    Abstract/FREE Full Text
  18. ↵
    1. Samore MH,
    2. Magill MK,
    3. Alder SC,
    4. et al
    . High rates of multiple antibiotic resistance in Streptococcus pneumoniae from healthy children living in isolated rural communities: association with cephalosporin use and intrafamilial transmission. Pediatrics 2001;108:856-65.
    Abstract/FREE Full Text
  19. ↵
    1. US Census Bureau
    . Census 2000 data for the State of Utah. Available at: http://www.census.gov/census2000/states/ut.html. Accessed 16 August 2004.
  20. ↵
    1. Centers for Disease Control and Prevention
    . National campaign for appropriate antibiotic use. Available at: http://www.cdc.gov/drugresistance/community/. Accessed 1 May 2004.
  21. ↵
    1. Sankoh AJ,
    2. Huque MF,
    3. Dubey SD
    . Some comments on frequently used multiple endpoint adjustment methods in clinical trials. Stat Med 1997;16:2529-42.
    CrossRefMedlineWeb of Science
  22. ↵
    1. Rothman K,
    2. Greenland S
    1. Greenland S,
    2. Rothman K
    . Introduction to categorical statistics. In: Rothman K, Greenland S, editors. Modern epidemiology. 2nd ed. Philadelphia: Lippincott-Raven Publishers; 1998. p. 249-50.
  23. ↵
    1. Hennessy TW,
    2. Petersen KM,
    3. Bruden D,
    4. et al
    . Changes in antibiotic-prescribing practices and carriage of penicillin-resistant Streptococcus pneumoniae: a controlled intervention trial in rural Alaska. Clin Infect Dis 2002;34:1543-50.
    Abstract/FREE Full Text
  24. ↵
    1. McCaig LF,
    2. Besser RE,
    3. Hughes JM
    . Trends in antimicrobial prescribing rates for children and adolescents. JAMA 2002;287:3096-102.
    Abstract/FREE Full Text
    1. Halasa NB,
    2. Griffin MR,
    3. Zhu Y,
    4. Edwards KM
    . Decreased number of antibiotic prescriptions in office-based settings from 1993 to 1999 in children less than five years of age. Pediatr Infect Dis J 2002;21:1023-8.
    CrossRefMedlineWeb of Science
  26. ↵
    1. Finkelstein JA,
    2. Stille C,
    3. Nordin J,
    4. et al
    . Reduction in antibiotic use among US children, 1996–2000. Pediatrics 2003;112:620-7.
    Abstract/FREE Full Text
    1. Centor RM,
    2. Witherspoon JM,
    3. Dalton HP,
    4. Brody CE,
    5. Link K
    . The diagnosis of strep throat in adults in the emergency room. Med Decis Making 1981;1:239-46.
    FREE Full Text
| Table of Contents

This Article

  1. Clin Infect Dis. (2005) 40 (4): 546-553. doi: 10.1086/427500
  1. AbstractFree
  2. » Full Text (HTML)Free
  3. Full Text (PDF)Free

Classifications

Share

  1. Email this article

Navigate This Article

Current Issue

  1. March 1, 2012 54 (5)
  1. Clinical Infectious Diseases
  1. Alert me to new issues

Most