Estimated Influenza Illnesses, Medical visits, Hospitalizations, and Deaths in the United States — 2017–2018 influenza season

Introduction

This web page provides estimates on the burden of influenza in the United States for the 2017–2018 influenza season. For the past several years, CDC has used a mathematical model to estimate the numbers of influenza illnesses, medical visits, and hospitalizations ^1-4. The methods used to calculate the estimates have been described previously ¹. CDC uses the estimates of the burden of influenza in the population to inform policy and communications related to influenza.

2017–2018 Estimates

Influenza activity in the United States during the 2017–2018 season began to increase in November and remained at high levels for several weeks during January–February⁵. While influenza A(H3N2) viruses predominated through February, and were predominant overall for the season, influenza B viruses were more commonly reported starting in March, 2018. The season had high severity with unusually high levels of outpatient influenza-like illness, hospitalizations rates, and proportions of pneumonia and influenza-associated deaths.

CDC estimates that the burden of illness during the 2017–2018 season was also high with an estimated 48.8 million people getting sick with influenza, 22.7 million people going to a health care provider, 959,000 hospitalizations, and 79,400 deaths from influenza (Table 1). The number of cases of influenza-associated illness that occurred last season was the highest since the 2009 H1N1 pandemic, when an estimated 60 million people were sick with influenza⁶.

The 2017–2018 influenza season was additionally atypical in that it was severe for all ages of the population⁵. The burden of influenza and the rates of influenza-associated hospitalization are generally higher for the very young and the very old, and while this was also true during the 2017–2018 season, rates of hospitalization in all age groups were the highest seasonal rates seen since hospital-based surveillance was expanded in 2005 to include all ages (Table 2). This translated into an estimated 11.5 million cases of influenza in children, 30 million cases of influenza in working age adults (aged 18-64 years), and more than 7.3 million cases in adults aged 65 years and older.

Our estimates of hospitalizations and mortality associated with the 2017–2018 influenza season continue to demonstrate how severe influenza virus infection can be. We estimate overall, there were 959,000 hospitalizations and 79,400 deaths during the 2017–2018 season. More than 48,000 hospitalizations occurred in children (aged < 18 years); however, 70% of hospitalizations occurred in older adults aged ≥65 years. Older adults also accounted for 90% of deaths, highlighting that older adults are particularly vulnerable to severe disease with influenza virus infection. An estimated 10,300 deaths occurred among working age adults (aged 18–64 years), an age group that often has low influenza vaccination.

Deaths in children with laboratory-confirmed influenza virus infection are reportable in the United States and 183 deaths were reported for the 2017–2018 season. However, influenza-associated pediatric deaths are likely under-reported as not all children whose death was related to an influenza virus infection may have been tested for influenza ^{7, 8.} Therefore, we used a mathematical model to estimate the total number of pediatric deaths based on hospitalization rates and the frequency of death in and out of the hospital from death certificates, estimating that there were more than 600 deaths associated with influenza in children.

Conclusion

CDC estimates that influenza was associated with more than 48.8 million illnesses, more than 22.7 million medical visits, 959,000 hospitalizations, and 79,400 deaths during the 2017–2018 influenza season. This burden was higher than any season since the 2009 pandemic and serves as a reminder of how severe seasonal influenza can be.

Limitations

These estimates are subject to several limitations.

First, rates of influenza-associated hospitalizations and in-hospital death were adjusted for the frequency of influenza testing and the sensitivity of influenza diagnostic assays. However, data on testing practices during the 2017–2018 season were not available at the time of estimation. We used data on testing practices from the 2014–2015 influenza season as a proxy, as the seasons were similar with respect to overall severity and media attention that may impact influenza testing practices. Burden estimates from the 2017–2018 season will be updated at a later date when data on contemporary testing practices become available.

Second, estimates of influenza-associated illness and medical visits are based on data from prior seasons, which may not be accurate if patterns of care-seeking have changed.

Third, we used a mathematical model to estimate influenza-associated deaths, which relies on information about location of death from death certificates. However, death certificate data during the 2017–2018 season were not available at the time of estimation. We have used death certification data from the 2014–2015 influenza season as a proxy, as the 2014–2015 season had similar circulating viruses and death certificate data were available from the National Center for Health Statistics. Furthermore, our model uses the frequency of influenza-associated deaths that have cause of death related to pneumonia or influenza (P&I), other respiratory or cardiovascular (other R&C), or other non-respiratory, non-cardiovascular (non-R&C). These frequencies were not available from the 2017–2018 season at the time of estimation, so we used the 2014–2015 frequencies as a proxy. Analysis comparing the frequency of P&I, other R&C, and non-R&C causes of death across the 2012–2013 through 2015–2016 influenza seasons suggests that the distribution of cause of death within an age group does not vary substantially between seasons.

Fourth, estimates of burden were derived from a mathematical model based on rates of influenza-associated hospitalization, which is a different approach than the statistical model used in previously published reports. This makes it difficult to compare to previous reports, though the estimates from the mathematical model are largely consistent for similar recent years (9-11). Furthermore, some of the previous published models have estimated influenza-associated hospitalizations and deaths back as far as the 1970s, and that level of historic data is not available for this current model.  However, it is useful to keep in mind that direct comparisons to influenza disease burden decades ago are complicated by large differences in the age of the US population and the increasing number of adults over age 65.