Marika K. Iwane, PhD MPH¹, Nidhi Jain, MD, MPH¹, Susan Dolan, RN, MS, CIC², Carolyn Bridges, MD¹, Paul Gargiullo, PhD¹, John R. Copeland, MS¹, Felicita David, MS¹, Kristin Kenyan, RN, MPH¹, Zack Moore, MD, MPH¹, Ann-Christine Nyquist, MD MPH², James Todd, MD², Jim Alexander, MD MPH¹, and Jane Seward, MBBS MPH¹. (1) National Immunization Program, Centers for Disease Control and Prevention, 1600 Clifton Rd, MS E61, Atlanta, GA 30333, 404-639-8769, miwane@cdc.gov, (2) Department of Epidemiology, The Children's Hospital, 1056 East 19th Avenue, B276, Denver, CO 80218

Objectives: The 2003-04 influenza season was characterized by widespread activity in early November in parts of the country. Additionally, the predominant influenza A/H3N2 virus circulating was antigenically distinct from the H3N2 vaccine strain. To estimate the vaccine effectiveness (VE) of the 2003-04 TIV, we conducted a rapid assessment among hospital employees using a retrospective cohort study design.

Methods: Influenza viruses circulated at high levels among TCH’s patients November 1- December 27, 2003. During December 11-17, we asked TCH employees to complete a questionnaire on influenza vaccination and influenza-like illness (ILI). ILI was defined as subjective fever plus cough or sore throat after November 1st. The exposure cohort time was November 1 until the survey completion date or an ILI. We determined vaccinated and unvaccinated person-time and conducted two analyses considering person-time 0-13 days after vaccination as a)unvaccinated or b)excluded. VE was calculated as [1- Incidence Rate Ratio (IRR)]; IRR equals ILI incidence rate among vaccinated/ ILI incidence rate among unvaccinated.

Results: Of 3100 employees, 1886 (61%) responded. Of these, 1424 (78%) received TIV, with 1009 (71%) vaccinated before November 1st. ILI was self-reported by 217 (16%) respondents. VE estimates were a) -0.10 [95% Confidence interval (CI)= -0.41 to +0.14] and b) -0.15 (95%CI= -0.52 to +0.13) adjusted for age group, high-risk conditions, and patient contact.

Conclusions: In this rapid assessment, we demonstrated no VE of TIV against ILI. However, VE against severe illness or laboratory-confirmed influenza is likely to be higher. Prospective VE studies with laboratory-confirmed outcomes are optimal and should be conducted annually.

Learning Objectives:

• Develop a retrospective cohort study design,
• describe a vaccine effectiveness study,
• perform a person-time analysis, and
• recognize the limitations involved in using this study design to determine vaccine effectiveness

Presenting author's disclosure statement:
I do not have any significant financial interest/arrangement or affiliation with any organization/institution whose products or services are being discussed in this session.