Neighborhood Contribution to Childhood Influenza Vaccination and Mortality.

Abstract

The neighborhood in which a child resides has increasingly been identified as a risk factor for infectious diseases. The contribution of area-level measures of the neighborhood on pediatric influenza hospitalizations has been studied, but little is known about how these measures may impact pediatric influenza vaccination and death. Such information would help target public health interventions including education and vaccination campaigns for prevention and hospitalization and early antiviral medication use for treatment of influenza. To investigate neighborhood factors, census-derived measures were linked to pediatric influenza vaccination records from the Michigan Care Improvement Registry and to influenza-associated pediatric deaths reported to public health from an 18 state sample. We first found that county-level poverty (odds ratio, (OR)=0.63 for a 10% increase, 95% confidence interval (CI) [0.49, 0.80]) and Black raced residents (OR=0.87, 95% CI [0.84, 0.91]) helped explain pediatric influenza vaccination. These findings highlight the importance of area-level poverty and race as indicators of pediatric influenza vaccination. Second, census tract-level urbanicity (incidence rate ratio, (IRR)=1.04 for a 10% increase, 95% CI [1.00, 1.06]), proportion of college graduates (IRR=1.44 comparing census tracts with <15% vs. 41-100%, 95% CI [1.01, 2.04]), and county-level proportion of children without insurance (IRR=1.06 for a 10% increase, 95% CI [1.03, 1.09]) were significantly associated with pediatric influenza mortality. These factors also contributed to the increased rates of observed pediatric influenza mortality for Black and Hispanic children compared to White children. These findings suggest that area-level factors have an independent association with pediatric influenza mortality and that they are relevant determinants of a race/ethnic disparity. Finally, census tract-level proportion of Black residents (IRR=0.63 comparing census tracts with >15% vs. <5%, 95% CI [0.45, 0.89]) was predictive of the interval from influenza illness onset to death for children. Neither area-level factors nor individual factors were associated with the timing of influenza acquisition during epidemic periods. These findings highlight a previously unidentified disparity in the interval from onset to death and provide an opportunity for enhanced medical interventions such as antiviral medication and early hospitalization. Taken together, the results suggest area-level measures can impact influenza vaccination and mortality in children.