Estimation of marginal benefits of antibiotic treatment for otitis media in Iowa Medicaid children using instrumental variable techniques

Background. In the past several years, antibiotics have been considered overused for otitis media (OM). The Iowa Department of Public Health (IDPH) conducted an educational intervention on August 30, 1999 to reduce antibiotic use. A recent study reported the IDPH intervention contributed to reduction of antibiotic use by 20% in Medicaid population. However, it is not known whether this intervention lowered antibiotic use for OM patients and how many OM patients were not cured as a result of the intervention.; Objectives. This study identifies whether the IDPH intervention and physician supply per capita had a significant influence on antibiotic choice for OM in the Iowa Medicaid children from 1997 through 2002, assesses the validity of using the variable of the IDPH intervention and physician supply per capita as an instrument, and estimates average treatment benefits of antibiotics relative to watchful-waiting for marginal OM children using instrumental variable techniques with a valid instrument.; Methods. The Iowa Medicaid Claims database was used to analyze across the patient demographics, comorbidities, the years and seasons at diagnosis, physician specialty, physician supply, and the IDPH intervention. The study population consisted of 36,585 children less than 13 years of age with the index OM between August, 1996 and August, 2002.; Results. Physician supply per capita were significantly associated with the probability of receiving antibiotic treatment for OM (p < 0.0001) but not the IDPH intervention (p = 0.94). Antibiotics were more likely prescribed for the OM patients living in areas with higher physician supply per capita (OR = 1.154, 95% CI = 1.094--1.217). Physician supply per capita was found to be able to use as an instrument but not the IDPH intervention. The antibiotic treatment relative to watchful-waiting for marginal OM patients whose treatment choices were affected by difference in physician supply per capita was not significantly effective (antibiotic effectiveness = 4.94%, 95% CI = -0.25%--0.35%).; Conclusions. Physician supply had an effect on antibiotic choice for OM. If the marginal patients represented patients who did not receive antibiotics as a result of the IDPH intervention, lowering antibiotic use would not have decreased antibiotic benefits among OM children in Iowa Medicaid.