Estimating the Effect of Influenza Vaccination on Nursing Home Residents' Morbidity and Mortality

Abstract

Objectives: To estimate the effect of influenza vaccination on hospitalization and mortality in nursing home (NH) residents.

Design: Retrospective cohort study.

Setting: Medicare claims data linked to NH Minimum Data Set assessments and Centers for Disease Control and Prevention (CDC) surveillance data from 122 U.S. cities.

Participants: More than 1 million Medicare fee-for-service, long-stay NH residents between 2000 and 2009.

Measurements: Weekly facility outcome aggregates of NH resident pneumonia and influenza (P&I) hospitalizations and all-cause mortality and city-level P&I mortality as reported by the CDC were created. The seasonal vaccine match rate for influenza A/H1N1, A/H3N2, and B strains was calculated, and each outcome was compared in seasons of high and low match rates using facility fixed-effects regression models separately for full-year and nonsummer months.

Results: Average weekly all-cause mortality varied across seasons from 3.74 to 4.13 per 1,000 NH residents per week, and hospitalization for P&I varied from 2.05 to 2.43. Vaccine match rates were invariably high for H1N1 but variable across seasons for the other two types. The association between vaccine match and reduction in overall mortality and P&I hospitalizations was strongest for A/H3N2, the influenza strain typically responsible for the most-severe influenza cases. Given the approximately 130,000 deaths and 77,000 P&I hospitalizations of long-stay NH residents during the 32 nonsummer weeks, the model estimated that a 50-percentage-point increase in the A/H3N2 match rate (from <25% to >75%) reduced long-stay NH resident deaths by 2.0% and P&I hospitalizations by 4.2%.

Conclusion: Well-matched influenza vaccine prevents P&I hospitalizations and mortality in NH residents.

Keywords: hospitalization; influenza vaccination; nursing home; vaccine match rate.

Conflict of interest statement

Conflict of Interest: The editor in chief has reviewed the conflict of interest checklist provided by the authors and has determined that the authors have no financial or any other kind of personal conflicts with this paper.

© 2015, Copyright the Authors Journal compilation © 2015, The American Geriatrics Society.

Figures

Figure 1

Relationship between influenza A/H3N2 weekly prevalence and weekly pneumonia and influenza (P&I) hospitalization rate for bad vaccine match versus good vaccine match years. Increasing weekly prevalence of influenza strain A/H3N2 affects the rate of P&I hospitalization during nonsummer weeks for well-matched versus poorly matched seasons. In years with a poor match (solid line), weeks with high influenza A/H3N2 prevalence increased P&I hospitalization, but this relationship was barely observable in good match years (dotted line). Match rate varies from 1% to 100%; low (bad) match rate (≥25%) influenza seasons: 2003–04, 2004–05, 2006–07, and 2007–08. LS = long-stay; NH = nursing home.