Abstract
Background: Antibiotic prescribing rates for young children have declined dramatically in recent years. We aimed to determine what drove this decline in 8 communities throughout Massachusetts.
Methods: We obtained automated claims data from four large health insurers on antibiotic dispensings and visit diagnoses in children 3 to <24 months of age and 24 to <48 months of age from Sept 1998-August 2003, and created diagnosis categories for common infections. We determined diagnosis rates (# visits for each diagnosis/total # of person-years [p-y]), antibiotic treatment rates (# visits for each diagnosis associated with an antibiotic/ total # of p-y) and the percent treated (# visits for each diagnosis associated with an antibiotic/total # of visits for each diagnosis). We additionally assessed antibiotic dispensings not associated with a specific visit.
Results: We analyzed a total of 78,253 p-y. Among those aged 3 to <24 months, overall treatment rates decreased from 2.8 to 2.2 antibiotics/p-y (22%). Acute otitis media (AOM) was the most common diagnosis (2.1/p-y) and the most frequent reason for antibiotics (1.4/p-y) in 1998–1999. These decreased by 19% and 20%, respectively, by 2002–2003 (P<0.0001), accounting for 45% of the overall decrease. Antibiotics that were not linked to a visit accounted for another 28%. Among those aged 24 to <48 months, overall treatment rates decreased from 1.7 to 1.5 antibiotics/p-y (11%). AOM was again the most common diagnosis (1.0/p-y) and the most frequent reason for antibiotics (0.7/p-y) in 1998–1999. These decreased by 12% and 11%, respectively, by 2002–2003 (P<0.0001). AOM treatment accounted for 43% of the overall decrease in antibiotics and unlinked antibiotics accounted for another 52%. The percent of AOM associated with antibiotics remained stable in both age groups.
Conclusions: In spite of new recommendations for ‘watchful waiting’ for some cases, the percent of AOM treated, once diagnosed, remained stable. The substantial overall decline in antibiotic prescribing was driven by declines in rates of diagnosis of AOM and antibiotics not linked to a visit. Unlinked antibiotics included refills and prescriptions after telephone encounters. Decreases in antibiotic use appear to be the result of changes in diagnostic thresholds rather than treatment of diagnosed conditions.
- Received September 11, 2008.
