PropertyValue
is nif:broaderContext of
nif:broaderContext
is schema:hasPart of
schema:isPartOf
nif:isString
  • A detailed description of the data sources and analytic approach is given in Information S1 and summarized below. Mortality data and excess mortality approach: We compiled monthly pneumonia and influenza (P&I) deaths and population estimates for individuals ≥65 yrs during 1978–2006 from the Japanese Ministry of Health, Labor and Welfare and the US National Center for Health Statistics (Table S2). As in previous studies, we concentrated on trends in P&I mortality rates, a specific indicator of influenza-related mortality [25], [17]. We focused on the 1978–2006 time period to obtain reliable information on vaccine coverage ([17] and Table S1) and to avoid the period of adaptation following the 1968 influenza A/H3N2 pandemic, during which mortality in Japan and the US declined steeply even in the absence of vaccination [16]. Mortality data were stratified by five senior age groups (65–69, 70–74, 75–79, 80–84, 85–89 years) to allow careful adjustment for changing post-war demographics. Data on dominant influenza subtypes circulating in Japan and the US during the study period were obtained from [5], [17]. To estimate seasonal influenza-related mortality rates, we applied Serfling cyclical regression models to monthly P&I death rates for each country and age-group, as in [5], [26]. These models provide seasonal baseline levels of expected mortality in the absence of influenza virus circulation. Mortality observed in excess of the baseline during influenza-epidemic months is attributed to influenza and termed “seasonal excess mortality”. Estimating excess mortality reduction among seniors: To account for long-term changes in mortality unrelated to influenza and for baseline differences between Japan and the US, we standardized winter-seasonal excess P&I mortality rates for changes in population structure and summer mortality rates (Information S1). This approach has been used in the past to adjust for differences in socioeconomic status and access to healthcare between high and middle-income countries [27]. We chose the US summer P&I mortality rate and population structure in 2000 as a reference. To quantify the indirect benefits of the Japanese vaccination program, we assessed differences in crude and adjusted excess P&I mortality rates during the schoolchildren vaccination period (1978–1994) and the period immediately after the program was discontinued (1995–2006). We modeled age-stratified seasonal excess death rates using multivariate negative binomial regression, using a log-link with a dummy indicator for time period, adjusting for age and dominant virus subtype. A formal description of the model used and model diagnostics are provided in the Fig. S3, Table S8. We checked that interaction terms between the time period and age variables were not significant in the model. We also compared influenza-related mortality trends in Japan with those in the US using the same methodology, utilizing the US experience as a concurrent control that did not implement the vaccination of schoolchildren.
rdf:type