Estimating the Disease Burden of Influenza from Global Surveillance Data: A Meta-Analysis and Case Study from Indonesia

Influenza viruses are important respiratory pathogens that cause both annual seasonal epidemics and occasional pandemics in human populations. Pandemics result from the emergence of novel influenza viruses that cross over to humans from avian, swine, and other zoonotic hosts. While many diagnostic tests for influenza infection are available for clinical and public-health decision-making, their use is limited in lower-income settings. This lack of testing results in evidence gaps at the global level to effectively guide influenza prevention and control efforts. Limited global testing of influenza and other respiratory viruses also poses a risk for under-detection of newly emerging and potentially pandemic influenza strains and other novel respiratory viruses.

Global efforts are underway to address these gaps in evidence, with a focus on measuring medically attended influenza-associated illnesses and hospitalizations. This work draws data from a global network of influenza surveillance platforms to monitor circulating influenza viruses, detect novel viruses, and estimate the total disease burden of influenza, including severe disease, hospitalizations, and deaths. However, these platforms rely on individuals to seek care for their illness, and therefore community-based studies are also critical for understanding the true burden of influenza disease in a population.

This project addressed three objectives to better understand influenza disease burden globally, and as a case study in one set of communities in Southeast Asia. First, as part of a broader initiative to estimate medically attended influenza illnesses globally, we estimate global influenza-associated hospitalizations by age group, as a percent of all acute respiratory hospitalizations and as a total number. Second, to explore trends in the source data for the global models, we explore inpatient and outpatient influenza surveillance trends over a three-year period, in an urban community in Jakarta, Indonesia. Finally, to provide a community-based perspective to better contextualize the facility-based findings, we estimate the proportion of individuals with symptomatic respiratory disease that seek care at public and private facilities in two communities in Java, Indonesia.

To estimate the percent of acute respiratory hospitalizations associated with influenza among pediatric and adult populations, we conducted a systematic review and meta-analysis. Source data for the meta-analysis included both published and unpublished sources from 1996-2016 that fit a set of strict inclusion criteria. The percentages from these meta-analyses were then applied to global estimates of total respiratory hospitalizations, to estimate the total number of influenza hospitalizations worldwide, by age group and geographic region.

In Indonesia, we conducted enhanced surveillance over a three-year period (2011-2014) to systematically identify and test both inpatients and outpatients for novel and seasonal influenza viruses, in an urban district where influenza A(H5N1) was frequently identified among poultry populations. In addition to this surveillance activity, we conducted a cross-sectional community-based survey among households in two districts to identify and describe care-seeking patterns for respiratory illnesses.

Globally, among children less than 18 years, 10% (95% Confidence Interval (CI): 8%‒11%) of acute respiratory hospitalizations were influenza-associated. Among adults 18 years and older, 14% (95% CI: 12%‒17%) of acute respiratory hospitalizations were influenza-associated. Applying these estimates to the total number of global acute respiratory hospitalizations results in over eight million influenza-associated hospitalizations each year, with highest rates among children less than one year (284 per 100,000 person-years (p-y) (95% CI: 200–409) and adults 65 years and older (437 per 100,000 p-y (95% CI: 265–612).

In Indonesia, 15% (95% CI: 12%‒17%) of respiratory hospitalizations were influenza-associated across all age groups over a three-year surveillance period, with only one human case of avian influenza A(H5N1). Community-based data indicated that among those with respiratory disease that resulted in a visit to a healthcare provider, public clinics were the most common facilities chosen (33% of outpatient and 53% of inpatient visits), but private clinics and other providers also captured a substantial number of illnesses.

Through these surveillance platforms, we were able to quantify global influenza hospitalizations by age group, demonstrating the important role this virus plays in severe disease in both temperate and tropical settings. These estimates are crucial in describing the magnitude of severe disease in a population, as well as estimating the economic costs of influenza to society. We also describe epidemiological trends of seasonal influenza in Indonesia, both in inpatient and outpatient settings, and the important role of private sector healthcare in this community. Methods to estimate influenza disease burden should consider both public and private sector inpatient facilities, as well as severe illnesses that are not medically attended.