Inappropriate antibiotic prescription in acute respiratory infections (ARI) is a major problem in developed countries(1-4). Evidence on the extent of the problem in developing countries is still limited. Also, it is not known whether the level of development influences antibiotic prescription as it seems to influence ARI prevalence(5). ARI is a major cause of childhood mortality, hospitalization and consultation in Brazil(6) As children under 5 years of age have frequent episodes of ARI, on average 5-8 episodes a year(7), ARI is also a frequent cause of drug and antibiotic prescripton. If antibiotics are inappropriately prescribed, ARI may contribute substantially to the increase of drug resistance ultimately incrementing costs for families and for the health sector. To control ARI mortality and decrease severe morbidity, the Brazilian Ministry of Health started the ARI National Program(8), which was evaluated in a health facility survey in 1996. In this article, we present partial results from this survey documenting the rate of inappropriate antibiotic prescription to children under 5 years of age with ARI, the types of antibiotic prescribed by health workers in ARI and whether they counsel mothers in antibiotic use at home, and to assess the availability of antibiotics in health facilities.

Six Brazilian state capitals located in the northern (Belém), northeastern (Aracaju, Fortaleza and Recife), central-western (Brasilia) and southeastern (Rio de Janeiro) regions of the country were surveyed. In all capitals, the ARI Program had started in the 1980’s. Health facilities from the public sector, providing outpatient care to children were eligible. Hospitals with no outpatient clinics were excluded. In each state, a representative sample of health facilities was selected using a cluster sampling method based on the mean number of consultations of children under 5 years of age in each facility(9). Children under 5 year of age with cough in the last 15 days were eligible and selected consecutively if mothers agreed to participate. Health workers’ management of children with ARI was compared to the management proposed in the ARI National guidelines. Whenever an antibiotic was prescribed to a child with ARI and the recommendation in the guidelines was not to prescribe a antibiotic, it was considered inappropriate. Assessment by a pediatrician trained in the ARI Brazilian guidelines was taken as the gold standard. The choice of antibiotic and the dose used were not included as criteria for determining appropriatness of therepy. The ARI National gudelines were based on WHO Guidelines to manage ARI in children less than 5 years of age. According to these guidelines, antibiotics are not indicated for mild ARI episodes that are mostly caused by a viral infection and children with fast breathing according to their age or chest indrawing should receive antibiotics as they might have a bacterial pneumonia.

Data were collected in each capital for a two-week period during May to September 1996, after a pilot study was conducted in two health facilaities not selected for the main study. Each site had a team with a local coordinator and 6 to 8 local data collectors (3-4 physicians and 3-4 nurses) who were trained and agreement of 90% among them was attained beforehand. In each health facility the team worked in pairs: the nurse observed health workers attending children and, after the consultation, the physician examined children and classified them according to the ARI National Guidelines. In each facility, the local pharmacy was visited and the administrator interviewed.

Data were processed in standard forms, compiled and summarized in tables. For the inappropriate antibiotic prescription rates 95% confidence intervals were calculated and the rates were presented for all types of ARI as a whole. All parents and health workers gave consent to participate in the study, which was approved by the Brazilian Ministry of Health and local Secretaries of Health.

156 health facilities from the public sector were studied: 35 hospitals, 112 health centers and 9 health posts. These facilities represented 76.7% of the hospitals; 47.9% of the health centers and 20.0% of the health posts existing in the six capitals. Out of the 1565 children included in the study, 94 were below 2 months of age, 489 between 2-11 months and 982 between 1-4 years. The most common ARI classification was cough or cold - common cold - (77.5%), followed by pneumonia (16.4%) and very severe disease or severe pneumonia (6.1%). Children were seen by 393 health workers, 97.7% of them physicians and 2.3% nurses.

Among all children, 499 (31.8%) received antibiotics, 139 (11%) with cough and cold. Antibiotics were inapropriately prescribed to 9.2% of the children overall. 28% of the total antibiotic prescriptions by the health workers were inappropriate. The rates of inappropriate antibiotic prescriptions in each capital are presented in Table I. In some capitals, such as Rio de Janeiro, the rate among all children was much lower (7.8%; 95% CI: 5.0-8.9), compared to the rate among those receiving antibiotics (61.7%; 95% CI: 43.6-77.3). The most frequently prescribed antibiotics are shown in Table II. Most health workers explained to parents how to use antibiotics at home. However, only a few demonstrated the first dose (Table III).Antibiotics were available, on average, in 84% of all health facilities: 61.9% in Aracaju, 63.6% in Belem, 95.8% in Brassilia, 80.0% in Fortaleza, 93.3% in Recife and 100% in Rio de Janeiro.

The rate of inappropriate antibiotic prescription to children with ARI in Brazil was not high, although there was substantial variation among capitals. In the literature, these rates vary considerably which seems to be related to factors such as the geographical area, the type of attending physician and the age of patients. For instance, a rate of 10.5% was observed among pediatricians in Toronto (Canada) after examination of 3,585 patient’s visits(10). However, another large study from Toronto(1) reported much higher rates, 49% for children with upper respiratory infections. Another study(11) in Hamilton (Canada) also found differences on inappropriate antibiotic prescription for children with ARI among pediatricians and family doctors. In the US, antibiotics were prescribed to 44% of patients under 18 years of age with a common cold(2). It was also noted that antibiotics were prescribed more often for children aged 5 to 11 years and rates were lower for pediatricians compared to non-pediatricians. In 10 Spanish hospitals, the rate was 67% for ARI and non-specified infections among all ages, with significant inter-center variability(12). In Brazil, the profile of inappropriate request for restricted therapeutic antimicrobials was investigated in a General Hospital(13). Among 3,389 requests, 17.8% were considered inappropriate. We observed that the most frequently prescribed antibiotics by health workers were those recommended by the ARI National Program. This finding may reflect the availability of these antibiotics in the health facilities but also may be a consequence of the ARI Program training activities. It is relevant to mention that approximately 60% of all antibiotics prescribed in children with ARI were those recommended by WHO to treat pneumonia in children under 5 years of age(14).

Drug resistance may be related to the inappropriate use of antibiotics by parents and this may depend on the correct understanding of prescriptions. ARI Programs might influence the ability of doctors to provide advice to the caretakers regarding home care of sick children(15). We found that most health workers explained to parents about use of antibiotics at home but did not demonstrate the first dose. Demonstration of the first dose was not a part of the ARI National guidelines but it is part of the Integrated Management of Childhood Illness (IMCI) strategy(16) which has been adopted by the Brazilian government since 1997 and is being implemented in some states.

The use of antibiotics may be determined by several (cultural, economic, and microbiological patterns) that should be considered when planning actions to decrease inappropriate prescribing practices. For instance, there is evidence that parent’s

Table I__Citywise percentage of inappropriate antibiotic prescription, Brazil 1996.

Aracaju

2.1 (0.9-4.6)

8.9 (3.9-18.1)

Belém

16.7 (12.6-21.9)

37.2 (28.7 - 46.4)

Brasília

11.2 (7.2-16.8)

24.1 (15.8-34.7)

Fortaleza

9.9 (6.7-14.2)

27.0 (18.8-36.9)

Recife

9.1 (6.0-13.6)

29.1 (19.7-40.5)

Rio de Janeiro

7.8 (5.0-11.9)

61.7 (43.6-77.3)

Total

9.2 (7.8-10.7)

28.8 (27.8-36M.6)

Table II__Type of antibiotics prescribed to children with ARI
			

N              (%)

N        (%)

N      (%)

N  	    (%)

N           (%)

Aracaju

31

(39.7)

7

(9.0)

13

(16.7)

0

(0)

27

(34.6)

Belém

81

(67.0)

6

(5.0)

21

(17.0)

0

(0)

13

(11.0)

Brasìlia

13

(14.9)

19

(21.8)

0

(0)

21

(24.1)

34

(39.2)

Fortaleza

33

(33.0)

30

(30.0)

6

(6.0)

13

(13.0)

18

(18.0)

Recife

26

(32.9)

22

(27.8)

21

(26.6)

1

(1.3)

9

(11.4)

Rio de Janeiro

12

(35.3)

9

(26.5)

5

(14.7)

0

(0)

8

(23.5)

Total

196

(39.3)

93

(18.6)

66

(13.2)

35

(7.0)

109

(21.8)

Table III__ARI Cases Receiveing Instructions about Antibiotic Use at Home

Aracaju

28 (35.8)

0 (0)

8 (10.2)

Belém

80 (66.1)

6 (5.0)

9 (7.4)

Brasília

68 (78.1)

0 (0)

21 (29.2)

Fortaleza

78 (78.0)

0(0)

6 (6.0)

Recife

71 (89.9)

8 (10.1)

31 (39.2)

Rio de Janeiro

34 (100.0)

0 (0)

11 (32.3)

Total

359 (71.9)

14 (3.9)

86 (17.2)