BACKGROUND:
The 2003-2004 influenza season in Colorado was early (11/1 through 12/31) and severe. As part of an RCT of registry-based reminder recall, we attempted universal immunization in healthy 6-21 month olds in 5 pediatric practices.

OBJECTIVE:
1) Describe influenza immunization rates and timing of immunization in healthy 6-21 month olds and 2) assess the effectiveness of one versus two doses of influenza vaccine in preventing influenza-like illnesses (ILI).

METHOD:
We analyzed billing and immunization registry data from a cohort of 5193 healthy 6-21 month olds from 5 Denver pediatric practices. ILI was defined from ICD-9 codes for office visits from 11/1 to 12/31. Definitions: fully vaccinated (FV) = two shots >14 days prior to first ILI; partially vaccinated (PV) = one shot during current season and >14 days prior to first ILI. A Cox hazard model accounting for age, gender, and immunization status as a time-varying variable was used to calculate hazard ratios.

RESULT:
Over the season, 36.3% were unvaccinated (UV), 23.5% were PV, and 40.2% were FV, with 5.8% FV by 11/1 and 36.3% FV by 1/1. 28% had an ILI and 5% had pneumonia or influenza (P&I), a subset of ILI, during the season. Hazard ratios (HRs) for PV vs. UV were 1.4 (CI 1.2-1.6) and 1.1 (CI 0.80-1.4) for ILI and P&I respectively. HRs for FV vs. UV were 0.45 (CI 0.38-0.52) and 0.15 (CI 0.10-0.25) corresponding to vaccine efficacy (1 – HR) of 55% and 85% for ILI and P&I.

CONCLUSION:
Due to the early onset of the epidemic, a low proportion of children were fully vaccinated prior to its onset. While two doses of vaccine were 55% effective against ILI and 85% effective against P&I, one dose of vaccine didn't appear to prevent illness-associated visits.

LEARNING OBJECTIVES:
1) Describe influenza Iz rates in healthy 6 to 21 month olds and, 2) evaluate influenza vaccine effectiveness.