An Assessment of the Use of Death Certificate Record Linkage to Supplement Death Ascertainment in the Year Following Influenza-Associated Hospitalization Among Utah Adults, 2009-2013

BACKGROUND:  Utah law requires that providers report influenza-associated deaths among children, but not adults. Current guidelines suggest death assessment should occur during the first month following influenza-associated hospitalization (IAH).  EDEN, Utah’s Electronic Death Entry Network, conducts surveillance of death certificate records (DCR) for underlying and contributing causes of death which include pneumonia and influenza.  The efficacy of this system and timeframe to conduct surveillance for deaths following IAH is not well known.  The objective of this analysis was to use DCR linkage to evaluate the completeness of current mechanisms for adult influenza death assessment and determine trends in causes of death one year following IAH in Utah.

METHODS:  Electronic record linkage was utilized to match IAH cases reported in 2009-2013 for those aged 18 years or greater in TriSano, Utah’s disease surveillance system, with vital statistics DCRs and pneumonia- and influenza-related death data captured by EDEN. The variables used for record linkage were first and last name, gender, and birthdate.  Deaths occurring one year or less from IAH were identified and compared to deaths reported by EDEN.  Causes of death were categorized and tabulated for frequency by the amount of time from hospitalization and death rates were determined.  

RESULTS:  A total of 2,098 adults were hospitalized for influenza between 2009 and 2013. EDEN identified 90 pneumonia- and influenza-related deaths among these individuals. Electronic record linkage to DCR identified an additional 119 individuals with mortality one year or less following IAH.  Of those with mortality within one year, 29% were attributed to underlying and 40% were attributed to contributing pneumonia and influenza (death rate=2,859 per 100,000 IAH).  The majority (86%) of deaths attributed to pneumonia and influenza were identified during the month following hospitalization.  The rates of malignant neoplasms, heart disease, and chronic lower respiratory diseases as underlying causes of death among this population were 1,906, 1,715, and 1,048 per 100,000 IAH, respectively.  

CONCLUSIONS:  Relying solely on pneumonia and influenza death data in EDEN may underestimate the true prevalence of mortality following IAH.  DCR linkage may permit more effective but less timely ascertainment of all-cause mortality following IAH.  Death rates among this population for pneumonia and influenza, malignant neoplasms, heart disease, and respiratory disease following IAH far exceeded that seen among the general population.  The results of this analysis support the assessment of influenza-associated death during the first month following hospitalization; however, greater sensitivity may be achieved by increasing this timeframe to six months.