Estimating the burden of influenza‐associated hospitalizations and deaths in Chile during 2012‐2014

Background Influenza is a vaccine preventable disease that causes important morbidity and mortality worldwide. Estimating the burden of influenza disease is difficult. However, there are some methods based in surveillance data and laboratory testing that can be used for this purpose. Objectives Estimating the burden of serious illness from influenza by means of hospitalization and death records during the period between 2012 and 2014, and using information from Severe Acute Respiratory Illness (SARI) surveillance. Methods To estimate the Chilean rate of influenza‐associated hospitalizations and deaths, we applied the influenza positivity of respiratory samples tested in six SARI surveillance sentinel hospitals to the hospitalizations and deaths from the records with ICD‐10 codes from influenza and pneumonia. Results Annually, 5320 people are hospitalized for influenza and 447 die for this cause. The annual influenza‐associated hospitalization rate for the period was 71.5/100 000 person‐year for <5 years old, 11.8/100 000 person‐year for people between 5 and 64 years old; and 156.0/100 000 person‐year for ≥65 years. The annual mortality rate for the period was 0.08/100 000 person‐year for <5 years; 0.3/100 000 person‐year for people between 5 and 64 years; and 22.8/100 000 person‐year for ≥65 years. Conclusions This is the first study of influenza burden in Chile. Every year an important quantity of hospitalizations and deaths result from influenza infection. In countries in temperate zones, it is important to know the burden of influenza in order to prepare the health care network and to assess preventive intervention currently in practice and the new ones to implementing.

annual vaccination campaigns against influenza, following the World Health Organization (WHO) recommendation for the composition of the Southern Hemisphere influenza vaccine. The target groups for vaccination have been children between 6 to 24 months of age and 2 to 5 years of age (since 2015), as well as adults 65 years of age and older, pregnant women, chronically ill persons, and the healthcare personnel. [2][3][4] In Chile, target groups for vaccination were incorporated early into the country's national policies. 5 As awareness of the risks associated with influenza infection and the importance of influenza monitoring has increased, the surveillance system for influenza and other respiratory viruses has been strengthened in Chile. The system is currently based on sentinel surveillance for influenza-like illness (ILI), severe acute respiratory infection (SARI), and monitoring of visits to emergency wards; additionally, a decentralized laboratory network plus the Instituto de Salud Pública of Chile (ISP), which is a National Influenza Center recognized by WHO, 6 conduct virologic surveillance of influenza. The sentinel influenza-proxy syndromic surveillance for SARI has been progressively implemented in the country since 2011 with the aim of characterizing severe acute respiratory infections associated with influenza and of monitoring the severity of influenza virus subtypes. 7 However, the burden of influenza disease is difficult to estimate, because influenza virus produces a wide range of symptoms and syndromes not very specific and it is unlikely that laboratory confirmation of cases is carried out routinely. 8,9 Additionally, the seasonal increase in influenza virus coincides with that of other respiratory viruses that can cause a similar clinical picture and often bacterial superinfection occurs. This situation, in addition to making it difficult to measure the burden of disease caused by influenza, hinders the estimation of the impact of existing prevention and treatment measures, as well as, the analysis of health interventions. 8,9 Population-based cohort studies likely provide the most accurate direct measurement of ambulatory influenza disease burden; however, cohorts are not practical to measure hospitalized influenza burden and are extremely costly. There are some strategies that allow for the estimation of influenza burden of disease using data from epidemiological surveillance when it is associated with the use of laboratory techniques for confirmation. In the Americas, several estimation analyses have been carried out using these methods, mainly in the United States, but also in Central America and the Caribbean. [10][11][12][13][14][15][16][17] In Chile, despite having a stable surveillance system with quality data, to date, no burden estimates have been made using these strategies.
Thus, the objective of this analysis was to estimate the burden of serious illness from influenza through analysis of hospitalization and death records for the period during 2012-2014 and laboratory surveillance among SARI cases in Chile. These estimates will allow national authorities to have relevant information for the impact of prevention and control strategies.

| METHODS
To estimate the burden of influenza-associated hospitalizations and deaths in Chile, a WHO published methodology was used. 12,18 Virologic surveillance data and the number of respiratory hospitalizations and deaths, at national level, were paired to estimate the amount of influenza circulating. Representativeness of virologic and SARI sentinel surveillance was assumed.

| Information sources
The following data were used to perform this analysis:

4.
The SARI sentinel surveillance database 22 that contains the records from the six SARI sentinel sites was obtained. These sites are distributed across three of the four geographic-administrative macro zones of the country (north, center, and center south), where more than 80% of the population lives. The personnel in charge of surveillance is trained annually in SARI case identification definition using the following case definition-a patient who is hospitalized with fever or history of fever of 38°C or higher, cough, and respiratory distress (ie, polypnea and decreased oxygen saturation). 7 For patients meeting the case definition, a standard form is completed that includes demographic information, date of onset of symptoms, date of clinical sample collection and sample testing, influenza vaccination history, presence of risk factors for adverse outcome, and history of use of the healthcare system. A nasopharyngeal aspirate or a nasopharyngeal swab is also collected, as per the ISP laboratory protocol. 23

| Estimation of the rate of hospitalizations and deaths associated with influenza by age-group
For the analysis, the counts of the number of persons hospitalized or deceased (obtained from records of national hospitalizations and deaths) due to influenza or pneumonia (International Classification of Disease [ICD] 10 codes J09 to J18) as a primary or basic cause of death, for the period between January 1, 2012 to December 31, 2014, were included. ICD 10 codes were chosen to compare our result with previous studies from the USA and Central America and Caribbean.
The census population was stratified into three age-groups: <5 years old, 5-64 years of age, and ≥65 years of age and older.
The rate of influenza-associated hospitalizations was calculated applying the percentage of monthly samples positive for influenza obtained from SARI surveillance to the monthly number of hospital discharges due to influenza and pneumonia (J09-J18) at the national level. For the estimation of influenza-associated deaths, the same monthly influenza positivity obtained from SARI surveillance was used and applied to the monthly number of mortality records for pneumonia and influenza. These monthly counts were summed to estimate a total number of influenza-associated hospitalizations or deaths for the specified year. To determine the denominator of the calculation of the annual hospital and death rate, it was assumed that each inhabitant had the same risk of developing a serious illness by influenza and of being hospitalized and as such the census population for the country was used as the denominator.
The hospitalization rate calculation is summarized by the following formula: where: I i : hospital incidence associated with influenza for each age- The analysis was carried out using anonymized databases under the epidemiological surveillance model, respecting the national legislation on the management of clinical data and confidentiality, and therefore did not require approval of an Ethics Committee.

| RESULTS
A total of 185 379 hospitalizations with diagnosis of pneumonia or influenza were registered in Chile during 2012-2014; of these, 34.9% were children younger than 5 years of age, 23.6% were people between 5 and 64 years of age, and 41.5% were people older than 65 years of age (Table 1).
During the same period, 11 828 SARI cases were reported and tested for influenza from the six surveillance sentinel hospitals; of these, 51.9% were children younger than 5 years of age, 25.5% were persons aged between 5 and 64 years of age, and 22.6% were 65 years of age and older. The overall positivity for influenza was 3.8% among children <5 years of age, 15.0% among those 5-64 years of age, and 13.2% among those 65 years of age and older. Table 2   Overall, the risk by age was similar between type and subtypes of influenza virus; however, there were some small differences observed.
The risk of hospitalization for the group of 5-64 years of age was higher with influenza A(H1N1)pdm09 than influenza A(H3N2) when comparing years and the relation with the two another age-groups.
Otherwise, the difference in risk between age-groups was smaller. Table 4 shows the rates and distribution of the influenza-associated hospitalizations by type and subtype year and age-group.

| DISCUSSION
This is the first analysis to estimate the burden of severe influenza in Chile. An important number of people throughout all ages are hospitalized for influenza annually; however, the risk of serious influenza illness was greatest in the extreme ages, being about six times as high as and 13 times as high as among children younger than 5 years of age and the elderly of 65 and more years old, as compared to people between 5 and 64 years of age.
As influenza is known to cause more severe disease among the age extremes as compared to persons aged 5-64 years of age, our finding is consistent with previously published findings (CDC). Furthermore, a higher percent positivity for influenza was reported among the older age-group as compared to the younger age-groups, which provides further evidence of the greater burden in this age-group. This pattern is maintained for all viral subtypes; however, the differences tend to decrease with influenza A(H1N1) perhaps due to the published finding that influenza(H1N1)pdm09 affects younger adults more than older ones due to absence of immune-protection against this virus (USA, Mexico, Brazil, four European countries). [24][25][26] Our estimates for persons aged 65 years and older are greater than those reported in an analysis from Central America using the same data source (ie, pneumonia and influenza codes) and rate  persons-years) 27 and lower than the estimates from Central American countries (113 influenza-associated hospitalization per 100 000 persons-year). 12 Again this is likely attributable to various country-level behaviors and surveillance system characteristics as explained above. perform SARI surveillance apply the same protocols for capturing cases, there may be variations that are related to the intensity to maintain and prioritize sampling throughout the year, especially in adults.

| CONCLUSION
Annually, between 4000 and 6500 hospitalizations associated with influenza are expected at the national level in Chile, with the highest risk group being those over 65 years of age followed by those under 5 years of age. In relation to mortality, between 450 and 500 deaths are expected to be associated with influenza in Chile, being mostly among those 65 years of age and older.
This analysis highlights the importance of maintaining risk-based prevention and control strategies including the use of the influenza vaccine; further analyses should assess vaccine impact as well as estimate economic burden and medical burden during more seasons.
Additionally, resources should continue to be devoted to maintain SARI surveillance, which allows the monitoring of patterns of severe influenza disease.