TL;DR
This study compares two methods for estimating influenza-associated deaths: using all-cause mortality versus specific causes of death. They found that all-cause mortality provides a more comprehensive picture of excess deaths during influenza seasons and is more timely, making it a valuable tool for real-time public health monitoring during epidemics and pandemics.
What Problem Does This Paper Address?
Accurately and quickly estimating the mortality burden of seasonal influenza is crucial for public health preparedness and response. Traditional methods often rely on cause-specific death data (like pneumonia and influenza), but these can be delayed due to the complex process of coding and validating death certificates. This delay hinders timely decision-making during an outbreak. All-cause mortality data, on the other hand, is more readily available but can be multicausal, potentially overestimating influenza's direct impact. This paper addresses the question of whether all-cause mortality can serve as a reliable and timely proxy for influenza-associated mortality, and how it compares to cause-specific estimates.
Methods
The researchers employed the EuroMOMO model, a time-series regression model that uses a Poisson distribution to estimate weekly excess mortality. This model was applied to mortality data from Denmark, Spain, and the United States for the 2015/2016 and 2016/2017 influenza seasons. They analyzed several categories of death: all-cause, respiratory and circulatory, circulatory, respiratory, and pneumonia and influenza (P&I). The model was adjusted for time trends and seasonal variations. They compared the weekly and seasonal cumulative excess mortality estimates across these categories and countries, and also compared their findings to official estimates from national health authorities.
Datasets
The study utilized mortality data from:
- Denmark: Individual notifications with underlying cause of death, dates of death, and dates of birth from the Danish Civil Registry.
- United States: Weekly aggregated age group-specific data from the National Center for Health Statistics (NCHS).
- Spain: Weekly aggregated age group-specific data from computerized civil registers covering 92% of the total Spanish population, obtained through the Centro Nacional de Epidemiología, Instituto de Salud Carlos III (CNE-ISCIII).
Population data for Denmark and Spain were sourced from Eurostat, and for the United States from the U.S. Census Bureau.
Key Insights and Ideas
- All-cause mortality captures broader trends: While cause-specific data might seem more direct, all-cause mortality appears to better reflect the overall excess mortality during influenza seasons, likely encompassing indirect effects and deaths exacerbated by influenza.
- Timeliness is paramount: The availability of all-cause mortality data with minimal delay makes it a superior tool for real-time monitoring and rapid response compared to cause-specific data, which can have reporting lags of up to two years.
- Proportionality varies by country and season: The relationship between all-cause excess deaths and cause-specific excess deaths (particularly P&I) is not constant. This suggests that while all-cause data can be a useful indicator, direct translation to specific causes requires country-specific and season-specific calibration.
- Age stratification reveals nuances: The ratios of excess mortality between different causes of death varied significantly across age groups, highlighting that a one-size-fits-all approach might not be appropriate when inferring specific causes from all-cause data.
Key Results
- Similar timing of peaks: For all three countries, the periods of weekly excess mortality for all-cause and cause-specific deaths (especially respiratory and circulatory) showed similar chronological patterns, including the start of excess periods, peak timing, and duration.
- Higher all-cause excess mortality: Seasonal all-cause excess mortality rates were consistently higher than those for respiratory and circulatory causes, often by a factor of two to three. For example, in the US during the 2016/2017 season, all-cause excess mortality was 14.3 per 100,000, while respiratory and circulatory was 6.9 per 100,000.
- Spain showed higher overall burden: Spain consistently reported higher excess mortality rates across all categories compared to Denmark and the United States, particularly in the 2016/2017 season.
- Ratios of P&I to all-cause varied: The ratio of P&I excess deaths to all-cause excess deaths was generally higher in Spain and Denmark than in the US, indicating different patterns in how influenza mortality manifests across these countries. For instance, in Spain, all-cause excess deaths were 10-18 times higher than P&I excess deaths annually.
Limitations and Future Work
- Limited countries: The study analyzed data from only three high-income, temperate countries, limiting the generalizability of findings to other regions with different healthcare systems, demographics, and climate.
- Model limitations: The EuroMOMO model, while robust, does not inherently include influenza activity indicators, which could lead to residual confounding from other respiratory viruses or temperature effects.
- Proportionality challenges: The study found that the proportionality between all-cause and cause-specific excess mortality is not constant across seasons or age groups, making direct inference challenging.
- Future work: The authors suggest exploring the proportionality further to potentially allow countries to approximate cause-specific excess mortality from all-cause data in near real-time. They also highlight the need for continued development of timely cause-specific data and the integration of influenza transmission indicators into mortality monitoring systems.
Summary updated