UC Irvine

Objectives

To determine temporal associations of local measures of influenza morbidity and mortality by the Centers for Disease Control and Prevention (CDC) with influenza hospitalizations in nursing home residents.

Design

Retrospective, longitudinal panel study.

Setting and participants

Long-stay nursing home residents, aged 65 years or older in 823 nursing homes from 2011 to 2015.

Measures

CDC-reported rates of influenza and pneumonia mortality and laboratory-confirmed influenza hospitalizations. We compared the CDC measures to nursing home resident hospitalizations due to (1) all-cause, (2) a primary diagnosis of respiratory or circulatory illness, and (3) a primary diagnosis of pneumonia or influenza based on Medicare Part A Claims data.

Results

Our final sample included 273,743 unique residents in 819 nursing homes in 108 cities. National laboratory-confirmed influenza-associated hospitalizations for the group aged 65 and older occurred 0 to 1 week prior to nursing home resident influenza-related hospitalizations (Spearman ρ = 0.54). CDC-reported influenza hospitalizations occurred 3 weeks prior to CDC-reported influenza deaths (ρ = 0.59). Nursing home resident influenza hospitalizations occurred 2 weeks before local CDC-reported pneumonia and influenza deaths occurred (ρ = 0.44).

Conclusions/implications

Publicly reported CDC measures correlate well with nursing home hospitalizations for pneumonia and influenza. Rates of laboratory-confirmed influenza hospitalizations (as reported by the CDC) may be a useful surrogate for nursing home influenza outbreaks but should be considered along with local indicators of disease outbreaks. Early community signals could be clinically leveraged as a trigger for increased infection control measures in nursing homes.