Antibiotic use among children in British Columbia, Canada

Transcription

1 Journal of Antimicrobial Chemotherapy (2006) 58, doi: /jac/dkl275 Advance Access publication 18 August 2006 Antibiotic use among children in British Columbia, Canada Fawziah Marra 1,2 *, David M. Patrick 1,2, Mei Chong 2 and William R. Bowie 1,3 1 University of British Columbia, Vancouver, BC, Canada; 2 British Columbia Centre for Disease Control, Vancouver, BC, Canada; 3 Division of Infectious Diseases, Department of Medicine, Vancouver Hospital and Health Sciences Centre, Vancouver, BC, Canada Introduction Received 27 April 2006; returned 19 May 2006; revised 2 June 2006; accepted 5 June 2006 Background: In North America use of second-generation macrolides such as clarithromycin and azithromycin is popular due to favourable dosing and adverse event profiles. However, studies have also shown that use of second-generation macrolides promotes carriage of macrolide-resistant nasopharyngeal and oral streptococci. The present study was undertaken to characterize overall antibiotic use including macrolide antibiotics among children in British Columbia. Methods: Population-based data from British Columbia were analysed to determine antibiotic prescribing patterns for outpatient prescriptions from 1996 to Antibiotic prescription rates per 1000 children per year were evaluated by age (0 4, 5 9, 10 14, <15 years old), sex and physician diagnosis. Results: From 1996 to 2003, the overall BC prescription rate in children <15 years old decreased by onethird from 720 to 488 per 1000 children. The decrease in the rate of antibiotic consumption over time was seen across all age strata; however, the largest decrease (33%) was seen in children between the ages of 0 4 years. From 1996 to 2003, use of penicillins and cephalosporins decreased by 40% and 30%, respectively. This trend of decreasing antibiotic use with b-lactams was seen in all age groups but the greatest decline was in the age group of 0 4 years (P value <0.05). During this time, macrolide use increased significantly (24%) from 102 to 126 per 1000 children (P value <0.001). This increase was seen in all age groups but again the greatest increase was seen in children of age between 0 and 4 years. Within the macrolides, use of erythromycin decreased by 72% (from 83 to 23 per 1000 children) while clarithromycin use increased by almost 3-fold (18 67 per 1000 children) and azithromycin use increased 81-fold ( per 1000 children). In 2003, antibiotics were primarily being used for the treatment of upper respiratory tract infections, acute otitis media and bronchitis. Conclusions: Overall antibiotic use has declined in children; however, there is increased use of macrolides which may have ramifications on macrolide-resistant streptococci, including Streptococcus pneumoniae and group A streptococci. A large proportion of antibiotic use in children is for upper respiratory tract infections and bronchitis, indications where there is a high likelihood that the aetiology is viral rather than bacterial. Keywords: paediatrics, antibiotic prescriptions, macrolides In Canada, antibiotics are the fourth mostly commonly prescribed class of agents, after cardiovascular medications, psychotherapeutic drugs and hormones. 1 In the 1990s, the number of antibiotics dispensed per year increased steadily and reached a peak of 27.3 million prescriptions in However, campaigns targeting healthcare professionals on decreasing antibiotic use and a greater public awareness on antibiotic misuse has resulted in a 14% decrease in the number of prescriptions dispensed such that in million prescriptions of antibiotics were dispensed in Canada. 1 Despite the decreasing use of antibiotics within Canada, the prevalence of antimicrobial-resistant organisms, particularly Streptococcus pneumoniae, is increasing and has become a significant threat to public health. 3 S. pneumoniae is responsible for a high proportion of community-acquired infections in the paediatric population, including meningitis, otitis media, sinusitis... *Corresponding author. Tel: ; Fax: ; fawziah.marra@bccdc.ca Ó The Author Published by Oxford University Press on behalf of the British Society for Antimicrobial Chemotherapy. All rights reserved. For Permissions, please journals.permissions@oxfordjournals.org

2 Paediatric antibiotic use in Canada and pneumonia. 4 6 Current evidence suggests that the use of certain antibiotics such as extended-spectrum penicillins, cephalosporins and macrolides 7,8 has contributed to the increasing development of penicillin-resistant and macrolide-resistant pneumococci. 9 Although the data shows a decline in overall prescribing of antibiotics within Canada, 1 there are limited data on antibiotic prescribing in children within primary care. Our initial study on antibiotic prescribing rates for the entire population of British Columbia suggested that antibiotic prescribing was decreasing over time; however, the use of second-generation macrolides and fluoroquinolones was increasing. 10 We conducted a similar study in children to determine trends across various classes of antibiotics in children less than 15 years of age within British Columbia. Methods Antibiotic prescription data Data were obtained from the PharmaNet (College of Pharmacists) and Medical Services Plan (MSP) databases 11 in British Columbia; the reliability and validity of both datasets are high with underreporting and misclassification being minimal. The PharmaNet database contains a record for all ambulatory care prescriptions dispensed to the population of British Columbia. Data within this dataset include the date of dispensing, drug name and dose, duration of treatment, age, sex and geographic location of the pharmacy. Prescriptions included in the database were from the Anatomical Therapeutic Chemical (ATC) categories J01 Antibacterials for Systemic Use. 12 The annual rates of antibiotic prescriptions per 1000 children during the years were reported for all antibiotics for systemic use (J01) and for the six major classes of antibiotics including tetracyclines (J01A), penicillins (J01C), cephalosporins (J01DA), sulphonamides and trimethoprim (J01E), macrolides, lincosamides and streptogramins (J01F), and quinolones (J01M). Within the category of macrolides, we categorized the second-generation macrolides as spiramycin, clarithromycin, azithromycin and telithromycin. The PharmaNet database was used to determine antibiotic consumption in children over time. Data were obtained for all children less than 15 years of age in primary care and further stratified by age (0 4 years, 5 9 years, years) and sex. Data were expressed as the rate of prescriptions per 1000 children rather than daily defined dose to better reflect prescribing activity since the daily defined doses are not accurate for children whose antibiotic dosing is dependent on weight. The population estimates were obtained from Health Data Warehouse and defined as children within that age-sex group. The Medical Services Plan database has patient age, sex, date of visit, physician reimbursement claims (according to International Classification of Diseases, Ninth Revision), 13 physician specialty and geographic location of the physician office. For the present study, we used the following diagnostic categories: acute otitis media, bronchitis, upper respiratory tract infection, pneumonia, laryngitis or pharyngitis, sinusitis, lower urinary tract infection and skin/soft tissue infection. Records from the PharmaNet database were linked with the MSP data by a third party using the patient s personal health number; however, personal identifiers were removed in the final dataset used for analysis. Record linkages were for the years and each antibiotic prescription was linked to the most recent physician visit during the 5 days preceding dispensing of the prescription. Dispensing of prescriptions which could not be linked to a visit (e.g. refills or phone prescriptions) was excluded from the analysis. The linked database was used to determine for what indications the antibiotics were being prescribed. Statistical analysis Trends in prescribing rates and differences between age groups were analysed for statistical significance using generalized linear models. In order to stabilize the variance of the consumption rates that follow a Poisson distribution, square root transformation on the rates were used as the dependent variable in the modelling. All analyses were conducted using SAS statistical software, version 9.1 (SAS Institute, Cary, NC, USA). A P value of <0.05 was considered statistically significant. Results Overall antibiotic consumption The PharmaNet database showed that prescriptions were dispensed from 1996 to 2003 for children less than 15 years of age. After linking with the MSP database, a total of records remained; thus, 83% of the prescriptions dispensed were linked to a prior visit to a physician office. The rest of the records were not analysed and would have been prescriptions from physicians by phone, refills or out of town residents. In 1996, the annual rate of antibiotic consumption in children less than 15 years of age was 720 per 1000 children. This decreased by 32% to 488 prescriptions per 1000 children in 2003 (P value <0.001) (Figure 1). The lowest rate of consumption, however, was seen in 2002 with 450 prescriptions per 1000 children. The decrease in the rate of antibiotic consumption over time was seen across all age stratifications; however, the largest decrease was seen in children aged between 0 and 4 years (0 4 years: 33%; 5 9 years: 31%, years: 25%; all P values <0.05) (Figure 1). In 1996, the prescription rate of antibiotics in children between 0 and 4 years was two times higher than the year group. A similar trend was also seen in 2003, with 688 prescriptions per 1000 children being used in the 0 4 year age group compared with 342 prescriptions per 1000 children being used in the year age group (i.e. twice as much). When evaluating differences in antibiotic use according to sex, there was no statistically significant difference between males and females, and in both cases antibiotic prescribing rates decreased by 32% from 1996 to Class-specific consumption Table 1 shows the class-specific consumption rates for all children less than 15 years of age but stratified according to the six major ATC classes of antibiotics. For all children less than 15 years of age, the largest decrease in consumption rate was seen with trimethoprim/sulfamethoxazole. In 1996, the use of trimethoprim/ sulfamethoxazole for treatment of bacterial infections was 93 prescriptions per 1000 children; this rate decreased by 63% to 34 prescriptions per 1000 children in The consumption rate of b-lactams (i.e. penicillins and cephalosporins) also decreased significantly over time. In 1996, the rate of penicillin use was 414 prescriptions per 1000 children but fell by 40% to 249 prescriptions per 1000 children in Similarly for the cephalosporins, the rate decreased by 30% from 106 per 1000 children in 1996 to 74 per 1000 children in This trend of decreasing antibiotic use with trimethoprim/sulfamethoxazole 831

3 Marra et al Prescription rate per 1000 population < Figure 1. Prescription rates of all antibiotics (J01) for children less than 15 years of age. Significant difference between 1996 and 2003 (P value <0.05) for children less than 15 years of age, and between the ages of 0 4 years, 5 9 years and years. Table 1. Trends in antibiotic prescribing rates for British Columbian children less than 15 years of age, Year Prescription rate per 1000 children in each year Tetracyclines (J01A) tetracyclines (J01AA) Penicillins (J01C) a penicillins with extended spectrum (J01CA) b-lactamase-sensitive penicillins (J01CE) b-lactamase-resistant penicillins (J01CF) combinations of penicillins (J01CR) Other b-lactams (J01D) a cephalosporins (J01DA) Sulphonamides and trimethoprim (J01E) a trimethoprim and derivatives (J01EA) short-acting sulphonamides (J01EB) sulphonamides, intermediate acting (J01EC) combination of trimethoprim and sulphonamides (J01EE) Macrolides, lincosamides, streptogramins (J01F) b macrolides (J01FA) lincosamide (J01FF) Quinolones (J01M) fluoroquinolone (J01MA) other quinolones (J01MB) a Significant decreasing trend from 1996 to 2003 (P value <0.05). b Significant increasing trend from 1996 to 2003 (P value <0.05). 832

4 Paediatric antibiotic use in Canada and b-lactams was seen in all age groups but the greatest decline for both antibiotic classes was in the age group of 0 4 years. For trimethoprim/sulfamethoxazole, there was a 65% decrease in use while the b-lactams decreased by 38% in the 0 4 year age category. In the 5 9 year age group, trimethoprim/ sulfamethoxazole consumption decreased by 60% and b-lactams by 36%. Antibiotic consumption declined by the smallest percentage in the year olds with trimethoprim/ sulfamethoxazole and b-lactams decreasing by 56% and 32%, respectively. Ninety-nine per cent of tetracycline use was contributed largely by the year old children. The consumption rate of tetracyclines has remained relatively stable over the 7 year study period (3.6 per 1000 children in 1996 to 3.7 per 1000 children in 2003). The rate of antibiotic consumption increased for both the macrolide/lincosamide/streptogramin and quinolone antibiotic classes. From 1996 to 2003, macrolide consumption increased by 24% from 102 to 126 prescriptions per 1000 children. This increasing trend seen in the consumption of macrolides was reflected in all three age stratifications but once again the greatest increase was seen in the 0 4 year group (44% increase in the 0 4 years; 19% increase in the 5 9 years; and 12% increase in the years). Quinolone consumption also increased over the 7 year study period from 0.59 to 0.84 prescriptions per 1000 children (i.e. 42%); however, the increasing trend for quinolones was only significant for age groups 5 9 (15%) and (36%). Table 1 and Figure 2 show the use of the specific type of penicillins from 1996 to In British Columbia, extendedspectrum penicillins account for the majority of antibiotic use within the class of penicillins. For example, in 2003, extendedspectrum penicillins (e.g. amoxicillin and ampicillin) accounted for 85% of penicillin prescriptions while the b-lactamasesensitive penicillins (e.g. penicillin V) were used 6% of Prescription rate per 1000 children the time and combinations of penicillins (e.g. amoxicillin/ clavulanate) account for 5% of all penicillin prescriptions. Figure 3 shows the use of the specific types of macrolides within British Columbia over time. In 1996, erythromycin use accounted for 83 prescriptions per 1000 children. But this has declined significantly, and in 2003 the erythromycin prescription rate was 23 per 1000 children ( 72%). In contrast, use of clarithromycin increased by 269% from 18.2 (1996) to 67.0 (2003) prescriptions per 1000 children (P < 0.001). Use of azithromycin has also increased from 0.4 prescriptions per 1000 children to 34.9 prescriptions per 1000 children (an increase of 8056%). Antibiotic prescription rates according to specific diagnosis Children less than 15 years of age. Table 2 shows the change in population-based rates of antibiotic prescribing for each condition over the 7 year study period. From 1996 to 2003, antibiotic prescribing rates decreased significantly for all indications except treatment of pneumonia, which saw an increase in prescribing rates. The largest decrease was seen in the use of antibiotics for the treatment of acute otitis media, which decreased by approximately one-half by In 1996, use of antibiotics for upper respiratory tract infections was the second most common indication but by 2003 it was the most common reason for antibiotic prescriptions in children less than 15 years of age, particularly acute pharyngitis (41 prescriptions per 1000 children) and tonsillitis (24 prescriptions per 1000 children). Antibiotic prescriptions for treatment of bronchitis also declined by approximately one-third in the 7 year time frame; the majority of this was accounted for by a decrease in prescriptions for acute bronchitis and bronchiolitis ( 36%). Table 3 shows the trend in antibiotic prescribing rates from 1996 to 2003, according to the diagnostic indications J01CR02 - Amoxicillin and enzyme inhibitor J01CF05 - Flucloxacillin J01CF02 - Cloxacillin J01CE02 - Phenoxymethylpenicillin J01CA06 - Bacampicillin J01CA04 - Amoxicillin J01CA02 - Pivampicillin J01CA01 - Ampicillin Year Figure 2. Prescription rates of antibiotics within the class of penicillins (J01C) in children less than 15 years of age,

5 Marra et al Prescription rate per 1000 children Year (ICD-9 codes) and age stratifications of 0 4 years, 5 9 years and years. All three cohorts of children had a substantial drop in the rate of antibiotic prescribing for acute otitis media, bronchitis and upper respiratory tract infection. The decrease in antibiotic prescribing rates was similar across the three age groups with acute otitis media prescriptions decreasing by 43% in the 0 4 year old children, 48% in the 5 9 year old children and 42% in the year old children. Prescriptions for treatment of bronchitis decreased by 34% in the 0 4 and 5 9 year old children and 29% in the year old children while prescriptions for treatment of upper respiratory tract infections decreased by 28% in all age groups (30% in the 0 4 year olds; 25% in the 5 9 year olds; 26% in the year olds). In contrast, antibiotic prescribing increased in all three age groups for the treatment of pneumonia with the largest increase seen in the 0 4 year olds (39%). Children 0 4 years of age. In 2003, prescriptions for the treatment of acute otitis media (185 prescriptions per 1000 children), particularly suppurative and unspecified otitis media (163 prescriptions per 1000 children), accounted for the majority of antibiotic use in the 0 4 year olds (Table 3). In this age group, the class of penicillins followed by the class of macrolides were the most commonly used antibiotics (Table 4). Amoxicillin was used most commonly (107 prescriptions per 1000 children) followed by clarithromycin (18 prescriptions per 1000 children) and azithromycin (16 prescriptions per 1000 children). However the trend from 1996 to 2003 for antibiotic use in the treatment of acute otitis media for children between the ages of 0 and 4 years was a reduction in the use of penicillins by 42% but an increase in the use of macrolides by 146%. In 2003, macrolides were the preferred agents for treatment of bronchitis and pneumonia in the 0 4 year olds. For bronchitis, clarithromycin was the most common agent to be used (17 prescriptions per 1000 children) while azithromycin use was 11 prescriptions per 1000 children. J01FF01 - Clindamycin J01FA15 - Telithromycin J01FA10 - Azithromycin J01FA09 - Clarithromycin J01FA02 - Spiramycin J01FA01 - Erythromycin Figure 3. Prescription rates of antibiotics within the class of macrolides, lincosamides, streptogramins (J01F) in children less than 15 years of age, Similarly, for pneumonia, clarithromycin use accounted for six prescriptions per 1000 children and azithromycin use was three prescriptions per 1000 children. Between 1996 and 2003, the use of macrolides for the treatment of bronchitis and pneumonia increased 33% and 233%, respectively, for the 0 4 year olds. Children 5 9 years of age and years of age. In 2003, the majority of antibiotic use for the 5 9 and year olds was for treatment of upper respiratory tract infections (Table 3), particularly for the treatment of acute pharyngitis (5 9 year olds, 44 prescriptions per 1000 children; year olds, 40 prescriptions per 1000 children). For this indication, penicillins were the most commonly used antibiotic in both age groups, particularly amoxicillin (5 9 year olds, 67 prescriptions per 1000 children; year olds, 50 prescriptions per 1000 children). Macrolide use accounted for 21% of all antibiotic use for this indication in both age groups. Specifically, in the 5 9 year olds, clarithromycin and azithromycin use was 13 and 5 prescriptions per 1000 children, respectively, and in the year olds clarithromycin and azithromycin accounted for 10 and 4 prescriptions per 1000 children, respectively. Although the use of penicillins accounts for the majority of antibiotic use in 2003 for the treatment of upper respiratory tract infections, this has declined significantly since 1996 (5 9 year olds, 27%; year olds, 30%) while the use of macrolides has remained stable for both age cohorts between 1996 and For the 5 9 year olds and year olds, the use of macrolides was more common than penicillins in the treatment of bronchitis or pneumonia; here macrolide use accounted for 57% of overall antibiotic use for bronchitis and 75% of overall antibiotic use for pneumonia. Compared with 1996, the use of macrolides for the treatment of pneumonia increased by 133% for the 5 9 year olds and 81% for the year olds but for 834

6 Paediatric antibiotic use in Canada Table 2. Trends in antibiotic prescribing rates for selected diagnoses in children less than 15 years of age, Prescription rate per 1000 children in each year Age group Diagnosis (ICD-9 code) <15 years Acute otitis media a non-suppurative otitis media and Eustachian tube disorders (381) suppurative and unspecified otitis media (382) Bronchitis a acute bronchitis and bronchiolitis (466) bronchitis, not specified as acute or chronic (490) chronic bronchitis (491) Upper respiratory tract infection a streptococcal sore throat and scarlatina (034) acute pharyngitis (462) acute tonsillitis (463) acute laryngitis and tracheitis (464) acute upper respiratory infections of multiple or unspecified site (465) Pneumonia b pneumococcal pneumonia (481) other bacterial pneumonia (482) pneumonia due to other specified organism (483) pneumonia in infectious diseases classified elsewhere (484) bronchopneumonia, organism unspecified (485) pneumonia, organism unspecified (486) Sinusitis acute sinusitis (461) chronic sinusitis (473) Lower urinary tract infection a cystitis (595) urethritis, not sexually transmitted, and urethral syndrome (597) other disorders of urethra and urinary tract (599) Skin/soft tissue infection a erysipelas (035) carbuncle and furuncle (680) cellulitis and abscess of finger and toe (681) other cellulitis and abscess (682) impetigo (684) other local infections of skin and subcutaneous tissue (686) a Significant decreasing trend (P value <0.05). b Significant increasing trend (P value <0.05). bronchitis has remained stable over the 7 year time frame of the present study. Discussion Our population-based study included more than 3 million prescriptions of antibiotics for all children in the province of British Columbia from 1996 to Our results show that antibiotic use within the community declined by one-third from This decrease was seen across all age groups of children but the largest drop was seen in children between the ages of 0 and 4 years. Despite this drop seen in the youngest age cohort of our study, antibiotic prescriptions were twice as high in the 0 4 year olds as the year olds. Our study also found that the largest decrease in the use of antibiotics occurred with the class of sulfamethoxazole and trimethoprim, which declined by 63% 835

7 Marra et al. Table 3. Trends in antibiotic prescribing rates for selected diagnoses Prescription rate per 1000 children in each year Age group Diagnosis years Acute otitis media a Bronchitis a Upper respiratory tract infection a Pneumonia b Sinusitis a Lower urinary tract infection Skin/soft tissue infection years Acute otitis media a Bronchitis a Upper respiratory tract infection a Pneumonia Sinusitis Lower urinary tract infection Skin/soft tissue infection years Acute otitis media a Bronchitis a Upper respiratory tract infection a Pneumonia Sinusitis Lower urinary tract infection Skin/soft tissue infection <15 years Acute otitis media a Bronchitis a Upper respiratory tract infection a Pneumonia b Sinusitis Lower urinary tract infection a Skin/soft tissue infection a a Significant decreasing trend (P value <0.05). b Significant increasing trend (P value <0.05). from 1996 to Similarly, the use of b-lactams also decreased in the 7 year study period, with all age cohorts seeing an 38% decline in use of penicillins and cephalosporins. Although our results show a dramatic decline in antibiotic consumption over time, we also see an increase in the use of second-generation macrolide antibiotics in British Columbian children. In British Columbia, the provincial formulary only carries two types of second-generation macrolides clarithromycin and azithromycin. Overall macrolide consumption increased by 24% between 1996 and 2003 and the majority of this increase was accounted for by use of the second-generation macrolides such as clarithromycin and azithromycin. Within the class of macrolides, clarithromycin use increased 3-fold while azithromycin use increased by 81-fold during the 7 year study period. Most of the increase in the use of macrolides was seen in the 0 4 year olds children in this age group saw a 44% rise in the use of macrolides from 1996 to 2003 while children between the ages of 5 and 9 years and 10 and 14 years saw a 19% and 12% increase, respectively. The overall decrease in antibiotic use in children has now been reported by a number of jurisdictions in Europe and North America Like our study, these investigators report a drop of about one-third in overall antibiotic prescriptions, with the greatest drop occurring in the youngest cohort (i.e. 0 4 year olds). Of note, one of these studies was a Canadian study that used the Manitoba population-based database and researchers showed an overall decline in antibiotic use by 28% from 1995 (1201 prescriptions per 1000 children) to 2001 (864 per 1000 children). 19 In addition, some of these studies also report a shift from use of the old narrowspectrum antibiotics to use of more broad-spectrum agents such as the extended-spectrum penicillins (e.g. amoxicillin) and secondgeneration macrolides such as clarithromycin and azithromycin in the paediatric population. 15,18 20 Our study findings corroborate what others have seen around the world For example, Stille et al. found a 241% increase in the dispensing of second-generation macrolides for children less than 6 years of age between 1996 and 2000 while the National Ambulatory Medical Care Survey (NAMCS) found a 320% increase in macrolide use in children less than 5 years of age from 1993 to ,21 The increase in use of the second-generation macrolides seen in our study is slightly higher given that our overall increase over a period of 7 years was 450%. The increase seen with the use of macrolides is concerning since there is new 836

8 Paediatric antibiotic use in Canada Table 4. Specific antibiotics used for selected diagnoses in children less than 15 years of age, Prescription rate per 1000 children in each year Age group Diagnosis years Acute otitis media a penicillins cephalosporins trimethoprim/sulfamethoxazole macrolides Bronchitis a penicillins cephalosporins trimethoprim/sulfamethoxazole macrolides Upper respiratory tract infection a penicillins cephalosporins trimethoprim/sulfamethoxazole macrolides Pneumonia b penicillins cephalosporins trimethoprim/sulfamethoxazole macrolides years Acute otitis media a penicillins cephalosporins trimethoprim/sulfamethoxazole macrolides Bronchitis a penicillins cephalosporins trimethoprim/sulfamethoxazole macrolides Upper respiratory tract infection a penicillins cephalosporins trimethoprim/sulfamethoxazole macrolides Pneumonia penicillins cephalosporins trimethoprim/sulfamethoxazole macrolides years Acute otitis media a penicillins cephalosporins trimethoprim/sulfamethoxazole macrolides Bronchitis a penicillins cephalosporins trimethoprim/sulfamethoxazole macrolides Upper respiratory tract infection a penicillins cephalosporins

9 Marra et al. Table 4. (continued) Prescription rate per 1000 children in each year Age group Diagnosis trimethoprim/sulfamethoxazole macrolides Pneumonia penicillins cephalosporins trimethoprim/sulfamethoxazole macrolides a Significant decreasing trend (P value <0.05). b Significant increasing trend (P value <0.05). evidence now to suggest that overuse of macrolides not only leads to the development of macrolide-resistant streptococci but also S. pneumoniae and Streptococcus pyogenes that are resistant to both penicillins and macrolides. 22,23 This may be particularly true for azithromycin, which has a long elimination half-life and can therefore reside in tissues for a prolonged period of time at subinhibitory concentrations, resulting in an environment that selects for resistant organisms; the same may not be the case for clarithromycin. 24 Our study is unique in that it evaluated reasons for prescribing antibiotics at a population level. In 2003, use of antibiotics for the treatment of upper respiratory tract infections accounted for 24% of overall antibiotic use. This is a change from 1996 when the most common indication for prescribing antibiotics was for acute otitis media (26% of overall antibiotic use). This decrease in the use of antibiotics for otitis media is likely due to clinicians using antibiotics judiciously or using the watchful waiting approach in the treatment of otitis media. The use of antibiotics for upper respiratory tract infections and bronchitis in children would indicate that more work is required in British Columbia to educate physicians on judicious use of antibiotics and in particular promote the use of penicillins rather than second-generation macrolides. For example, for the treatment of acute otitis media high dose amoxicillin is just as effective as the extended-spectrum macrolides. Most upper respiratory tract infections, including acute pharyngitis and laryngitis, are viral in origin and therefore the use of antibiotics should be minimal for this category If the infection is bacterial, then once again the use of macrolides is unnecessary since these infections are mostly due to S. pyogenes, which is presently susceptible to narrow-spectrum penicillins such as penicillin V. 25 Our study did not evaluate the reasons for antimicrobial prescribing and overuse; however, other studies have found several reasons for over prescribing including perceived or real pressure from patients and parents of patients to prescribe drugs, inadequate knowledge of the proper indications for some drugs, lack of awareness of prescription guidelines, lack of time to explain to the patient that antimicrobials are not needed, lack of education about resistance patterns in the community, fee-forservice remuneration of physicians and availability of product on the national formulary. 28 In order to change prescribing behaviour of clinicians and patient expectations of receiving an antibiotic prescription, multifaceted interventions are required. 28 In other words, passive education through conferences or use of written material alone will not change physician or patient behaviour. 29 Public health will need to educate all clinicians and the public through multiple means, such as use of written material, pre-printed guidelines, pre-printed prescription pads, written guidelines, reminder methods on treatment guidelines, media campaigns for the public and teaching of healthcare professionals, daycare staff and school children on the need to reduce antibiotic prescribing and when it is necessary to use narrow-spectrum agents. Conclusions For the past 5 years, the public health community and infectious diseases physicians have been delivering a unified message to family physicians and the public on decreasing antibiotic use and this has had an impact in British Columbia. However, our study suggests that a new message is needed to decrease the use of extended-spectrum macrolides and increase the use of b-lactams when antibiotics are necessary. In addition, we need to continue decreasing the use of antibiotics for the treatment of upper respiratory tract infections and bronchitis. Acknowledgements Accepted as a slide presentation at the Interscience Conference on Antimicrobial Agents and Chemotherapy, Washington, DC, We thank the College of Pharmacists of British Columbia and the BC Ministry of Health for providing the data to the BC Centre for Disease Control. Financial support: the study was funded by a BC Medical Services Foundation (Vancouver Foundation) operating grant. Transparency declarations None of the authors on this manuscript has received funds for speaking at symposia organized on behalf of industry nor have they received funds for research from industry. None of the authors holds any stocks in industry. References 1. IMS Health Canada. Top 10 Prescribed Therapeutic Classes, (December 2005, date last accessed). 838

10 Paediatric antibiotic use in Canada 2. IMS Health Canada. Antibiotic Use on the Decline as 2000 Flu Season Starts. (December 2005, date last accessed). 3. Powis J, McGeer A, Green K et al. In vitro antimicrobial susceptibilities of Streptococcus pneumoniae clinical isolates obtained in Canada in Antimicrob Agents Chemother 2004; 48: Jacobs MR. Streptococcus pneumoniae: epidemiology and patterns of resistance. Am J Med 2004; 117 Suppl 3A: 3 15S. 5. Chavez-Bueno S, McCracken GH Jr. Bacterial meningitis in children. Pediatr Clin North Am 2005; 52: Sinaniotis CA, Sinaniotis AC. Community-acquired pneumonia in children. Curr Opin Pulm Med 2005; 11: Granizo JJ, Aguilar L, Casal J et al. Streptococcus pyogenes resistance to erythromycin in relation to macrolide consumption in Spain ( ). J Antimicrob Chemother 2000; 46: Granizo JJ, Aguilar L, Casal J et al. Streptococcus pneumoniae resistance to erythromycin and penicillin in relation to macrolide and b-lactam consumption in Spain ( ). J Antimicrob Chemother 2000; 46: Mera RM, Miller LA, Daniels JJ et al. Increasing prevalence of multidrug-resistant Streptococcus pneumoniae in the United States over a 10-year period: Alexander Project. Diagn Microbiol Infect Dis 2005; 51: Patrick DM, Marra F, Hutchinson J et al. Per capita antibiotic consumption: how does a North American jurisdiction compare with Europe? Clin Infect Dis 2004; 39: Medical Services Plan. Information Resource Management Branch, Knowledge Management & Technology Division. Ministry of Health Services, (8 August 2006, date last accessed). 12. WHO Collaborating Centre for Drug Statistics Methodology, Oslo. ATC/DDD Index (28 September 2005, date last accessed). 13. International Classification of Diseases, Ninth Revision, Clinical Modification. Doves, DE. American Medical Association, Hogberg L, Oke T, Geli P et al. Reduction in outpatient antibiotic sales for pre-school children: interrupted time series analysis of weekly antibiotic sales data in Sweden J Antimicrob Chemother 2005; 56: Otters HB, van der Wouden JC, Schellevis FG et al. Trends in prescribing antibiotics for children in Dutch general practice. J Antimicrob Chemother 2004; 53: Schindler C, Krappweis J, Morgenstern I et al. Prescriptions of systemic antibiotics for children in Germany aged between 0 and 6 years. Pharmacoepidemiol Drug Saf 2003; 12: Finkelstein JA, Stille C, Nordin J et al. Reduction in antibiotic use among US children, Pediatrics 2003; 112: Linder JA, Bates DW, Lee GM et al. Antibiotic treatment of children with sore throat. JAMA 2005; 294: Kozyrskyj AL, Carrie AG, Mazowita GB et al. Decrease in antibiotic use among children in the 1990s: not all antibiotics, not all children. CMAJ 2004; 171: Stille CJ, Andrade SE, Huang SS et al. Increased use of second-generation macrolide antibiotics for children in nine health plans in the United States. Pediatrics 2004; 114: Hyde TB, Gay K, Stephens DS et al. Macrolide resistance among invasive Streptococcus pneumoniae isolates. JAMA 2001; 286: Garcia-Rey C, Aguilar L, Baquero F et al. Pharmacoepidemiological analysis of provincial differences between consumption of macrolides and rates of erythromycin resistance among Streptococcus pyogenes isolates in Spain. J Clin Microbiol 2002; 40: Bergman M, Huikko S, Pihlajamaki M et al. Finnish Study Group for Antimicrobial Resistance (FiRe Network). Effect of macrolide consumption on erythromycin resistance in Streptococcus pyogenes in Finland in Clin Infect Dis 2004; 38: Blondeau JM, Borsos S. Compararative minimum inhibitory and resistance prevention concentrations (RPC) of azithromycin, clarithromycin and erythromycin against 191 susceptible strains of Streptococcus pneumoniae. In: Abstracts of the Forty-fifth Interscience Conference on Antimicrobial Agents and Chemotherapy, Washington, DC, Abstract C American Society for Microbiology, Washington, DC, USA. 25. Bisno AL. Acute pharyngitis. N Engl J Med 2001; 344: Gerber MA. Diagnosis and treatment of pharyngitis in children. Pediatr Clin North Am 2005; 52: Reveiz L, Cardona AF, Ospina EG. Antibiotics for acute laryngitis in adults. Cochrane Database Syst Rev 2005; 1: CD Arnold S, Straus S. Interventions to improve antibiotic prescribing practices in ambulatory care. Cochrane Database Syst Rev 2005; 4: CD Nash DR, Harman J, Wald ER et al. Antibiotic prescribing by primary care physicians for children with upper respiratory tract infections. Arch Pediatr Adolesc Med 2002; 156: