Incidence of medically attended influenza and influenza virus infections confirmed by serology in Ningbo City from 2017–2018 to 2019–2020

Abstract Objectives In mainland China, the disease burden of influenza is not yet fully understood. Based on population‐based data, we aimed to estimate incidence rates of medically attended influenza and influenza virus infections in Ningbo City. Methods We used data for outpatient acute respiratory illness (OARI) from a platform covering all health and medical institutes in Yingzhou District, Ningbo City. We applied generalized additive regression models to estimate influenza‐associated excess incidence rate of OARI by age. We recruited local residents aged ≥60 years in the autumn of 2019 and conducted follow‐up nearly 9 months later. Every survey, the sera were collected for testing hemagglutination inhibition antibody. Results From 2017–2018 to 2019–2020, the annual average of influenza‐associated incidence rate of OARI in all ages was 10.9%. The influenza‐associated incidence rate of OARI was the highest in 2017–2018 (16.9%) and the lowest in 2019–2020 (4.8%). Regularly, influenza‐associated incidence rates of OARI were the highest in children aged 5–14 years (range: 44.1–77.6%) and 0–4 years (range: 8.3–46.6%). The annual average of excess OARI incidence rate in all ages was the highest for influenza B/Yamagata (3.9%). The overall incidence rate of influenza infections indicated by serology in elderly people was 21% during the winter season of 2019–2020. Conclusions We identified substantial outpatient influenza burden in all ages in Ningbo. Our cohort study limited in elderly people found that this age group had a high risk of seasonal influenza infections. Our study informs the importance of increasing influenza vaccine coverage in high‐risk population including elderly people.


| INTRODUCTION
Influenza viruses cause a substantial burden on morbidity and mortality worldwide. 1,2 In the developed countries, the disease burden associated with influenza has been increasingly understood. 3,4 Mortality, rate of hospitalization, incidence of cases seeking medical care, and incidence of symptomatic and asymptomatic infection with influenza usually are used to understand the profile of influenza-associated disease burden. 5 China is a large country with a diversity of climatic zones. The previous studies have characterized seasonality patterns of influenza epidemics by type, subtype, or lineage that varied across the country. 6,7 The disease burden of influenza in China is not yet fully understood.
Although hospital-based influenza-like illness (ILI) sentinel surveillance network has been built in mainland China to monitor influenza activity, 6,7 there is a challenge in estimating influenza-associated outpatient disease burden for the surveillance network. The challenge may arise from difficulty in determining well-defined catchment populations of surveillance sentinel hospitals and underrepresentation of sentinel hospitals as most of sentinel hospitals are ones at tertiary level and located in prefectural cities. 8 Additionally, few data are available about disease burden of infections with seasonal influenza in elderly people at community level in mainland China.
Ningbo is an economically developed coastal city located in the southeast of China and has two peaks of influenza activity in November to February and July to September. 6,7 Since 2015, an Integrated Platform of Health and Medical Big Data (IPHMBD) covering all public and private health and medical institutes in Ningbo City has been established. 9 The platform aims to improve the management of health and medical big data and exploit fully these data to provide evidence for decision making of public health and medical care. The platform in Yingzhou District of the city is the first one with similar function that was completed and put into use in China. The objective of this study was to estimate influenza-associated incidence rate of outpatient acute respiratory illness (OARI) based on the populationbased data platform. Additionally, we established a community-based cohort for people aged ≥60 years to qualify the incidence rate of sea-  The platform collects data from hospital information system of the all public and private medical and health institutes located in Yingzhou District for the purpose of disease surveillance and control. 9 The outpatient and inpatient data in these medical and health institutes can be transferred in real time to the platform. The   records with the diagnosis of OARI related to influenza under the   International Classification of Diseases, Tenth Revision (ICD-10) code were extracted. The ICD-10 codes related to acute respiratory illness include J11.101, J06.900, J06.901, J06.903, J02.801, J02.900, J03.900, J00.x02, and J00.x03. The number of OARI was compiled by the designated data manager of IPHMBD and then was offered to us. The data we received did not include any identification information of patients.
To remove the repeated visits of patients due to the same episode of illness, we define an OARI with the time interval of 7 days or less between two consecutive medical visits as the same episode of influenza-related respiratory illness. An OARI with the time interval of more than 7 days since the past medical visit is defined as a new episode of influenza-related respiratory illness. For those repeated visits due to the same episode of illness within the time interval of 7 days, we only counted them as one episode of influenza-related illness. Incidence rates of OARI in 0-4, 5-14, 15-14, 25-59, and ≥60 years and all ages combined were calculated by the number of OARI divided by the size of resident population in these age groups of Yingzhou District.

| Influenza surveillance data
We used weekly influenza laboratory surveillance data collected by a laboratory and two sentinel hospitals based in Ningbo City as part of the Chinese Influenza Sentinel Surveillance Network from October 2017 through September 2020. The two sentinel hospitals for influenza surveillance are at tertiary level: One is a comprehensive hospital, and the other is a children hospital.
According to the National Influenza Surveillance Guideline, each sentinel hospital reports weekly count of outpatients with ILI by age group (ILI is defined as body temperature ≥38 C with either cough or sore throat) to a web-based national information system. Each sentinel hospital is also required to collect 20 or more pharyngeal swab specimens from ILI outpatients per week.
The respiratory specimens are transported to the network laboratory for influenza detection. These sentinel hospitals and the net- Service Center to participate in our study via face-to-face invitation from mid-October to mid-November 2019 (pre-season survey). When these elderly people agreed to participate, they were asked to complete a questionnaire including demographics and underlying medical conditions. Serum samples were collected from the participants by trained nurses. The elderly participants were followed up from mid-June to early-August 2020

| Laboratory test
The paired sera were tested in parallel by HI assays using five repre-   The influenza-associated excess incidence rates of OARI were estimated by subtracting the predicted incidence rate assuming that influenza activity for a specific type or subtype to zero from the predicted incidence rate from the model based on the reported weekly influenza activity. 14 The influenza-associated excess incidence rates of OARI was estimated as the number of influenza-associated excess respiratory illness seeking outpatient consultants per 100 population.
The 95% confidence intervals (CIs) for excess incidence rates were estimated with a bootstrap approach. We estimate the cumulative inci-    Figure 2B).

| Incidence rates of influenza virus infections indicated by serology
Our cohort study recruited 500 people aged ≥60 years from mid-October to mid-November 2019; 476 elderly people (95%) were followed up from mid-June to early-August 2020 and provided preseason and post-season sera. The characteristics of subjects participating in pre-season and post-season surveys were similar (Table 2).
Among participants followed up, 46% were male, and 51% were 70 years or older; 100% of participants followed up had tertiary education or above; and 45% of them were retired or did not have a temporary or stable job. Thirteen of the participants received seasonal influenza vaccine during the 2019-2020 influenza season (Table 2).

| DISCUSSIONS
The IPHMBD in Ningbo is the first information management and sur-  2017-2018 to 2019-2020. In 2019-2020, the overall outpatient burden attributable to influenza was the lowest (4.8%, 95% CI: ILI case definition may include the presence of fever (or another systemic symptom) in addition to one or more respiratory symptoms.
OARI case definitions used in our study usually do not require an obligatory presence of fever or feverishness. Only a portion of all symptomatic influenza cases are captured by ILI case definitions. 19 The surveillance systems using OARI case definitions are more suited to describe and capture the burden of disease of influenza. 15 Similar to other studies, the age-specific outpatient burden attributable to influenza among children of age 0-4 and 5-14 years was constantly the highest in each year. 8,[16][17][18] Although the overall incidence rate of OARI in children aged 0-4 years was higher than that in children aged 5-14 years, the influenza-associated outpatient burden for the two age groups presented the opposite situation. This could be associated with the a higher baseline level of OARI among children aged 0-4 years during the non-epidemic period and a higher propor- In conclusion, we found considerable outpatient burden attribut- resources; supervision.

PEER REVIEW
The peer review history for this article is available at https://publons. com/publon/10.1111/irv.12935.

DATA AVAILABILITY STATEMENT
The datasets in our study are available from the first author and correspondence author.