Live‐attenuated influenza vaccine effectiveness against hospitalization in children aged 2–6 years, the first three seasons of the childhood influenza vaccination program in England, 2013/14–2015/16

Abstract Introduction In 2013, the United Kingdom began to roll‐out a universal annual influenza vaccination program for children. An important component of any new vaccination program is measuring its effectiveness. Live‐attenuated influenza vaccines (LAIVs) have since shown mixed results with vaccine effectiveness (VE) varying across seasons and countries elsewhere. This study aims to assess the effectiveness of influenza vaccination in children against severe disease during the first three seasons of the LAIV program in England. Methods Using the screening method, LAIV vaccination coverage in children hospitalized with laboratory‐confirmed influenza infection was compared with vaccination coverage in 2–6‐year‐olds in the general population to estimate VE in 2013/14–2015/16. Results The overall LAIV VE, adjusted for age group, week/month and geographical area, for all influenza types pooled over the three influenza seasons was 50.1% (95% confidence interval [CI] 31.2, 63.8). By age, there was evidence of protection against hospitalization from influenza vaccination in both the pre‐school (2–4‐year‐olds) (48.1%, 95% CI 27.2, 63.1) and school‐aged children (5–6‐year‐olds) (62.6%, 95% CI 2.6, 85.6) over the three seasons. Conclusion LAIV vaccination in children provided moderate annual protection against laboratory‐confirmed influenza‐related hospitalization in England over the three influenza seasons. This study contributes further to the limited literature to date on influenza VE against severe disease in children.


| INTRODUCTION
The United Kingdom previously had a long-standing selective influenza vaccination program that targeted populations at higher risk of severe disease due to influenza. Historically, this approach had been targeted at those 65 years of age and over, and those less than 65 years of age in clinical risk groups, including pregnant women, and healthcare workers, and aimed to directly protect these groups. 1 Despite the program, there was recognized to still remain a considerable burden of disease, both in the targeted groups, largely due to limited effectiveness of vaccination, and, prior to the COVID-19 pandemic, no substantial increases in uptake for many years, as well as on-going burden and transmission in the non-targeted groups, such as children. 1 It is estimated that between 10% and 30% of children are infected with influenza annually [2][3][4] and, although most often influenza infection is self-limiting, complications leading to hospitalization can occur, particularly in those under 5 years of age and in children with chronic medical conditions. In the United Kingdom, the youngest children have the highest influenza-related admission rates of all ages. [5][6][7] In addition, children are recognized to play a key role in the transmission of the influenza virus in the wider community. [8][9][10] In 2012, the Joint Committee on Vaccination and Immunisation (JCVI) considered the evidence for extending influenza immunization to healthy children using the newly licensed live-attenuated influenza vaccine (LAIV) and recommended universal annual vaccination of all children aged 2-16 years for influenza with LAIV in England. 11 The program was introduced incrementally from the start of the 2013/2014 influenza season, during which initially children aged 2-3 years were targeted nationally through general practice, as well as primary school aged children in seven geographical pilot areas in England. Since then, the program has been rolled out incrementally by adding additional age cohorts each season.
An important component of any new vaccination program is measuring how effective it is. The direct effect of a vaccine can be assessed after introduction in the targeted population using observational vaccine effectiveness (VE) studies. 12 LAIVs have been previously shown to provide good protection against influenza illness, although more recent observational studies have provided mixed results, with VE varying widely across seasons and other countries. [13][14][15][16]  This study aims to extend the findings of these studies by    (Table 1).

| Vaccination uptake of cases
Vaccination history of cases was obtained by sending a standard data collection proforma by post to the cases' General Practitioner in England. The date of administration and whether the vaccine was administered by injection or intranasally were also collected. This information was used to determine the proportion of cases vaccinated (PCV). This data collection was limited to the seasons included in this study.
A case was considered vaccinated if they received at least one dose of influenza vaccine (LAIV) in the relevant influenza season more than 14 days before disease onset, as this was considered the minimum time period to achieve maximum protection. When onset date was missing, the sample date minus 2 days was used as a proxy (based on the median time among those cases in whom the information was available) and if sample date was unknown then the test date minus 3 days was used (again based on the median time among those cases in whom the information was available).
Cases vaccinated by injection (i.e., by IIV) were excluded since population vaccine uptake is not available by vaccine type. All of the children in the school aged cohort will have been offered LAIV; however, a small number of 2-4-year-olds may have received IIV if they were contraindicated to recieve LAIV (due to severe immunodeficiency, those receiving salicylate therapy, those who have active wheezing at the time of vaccination or severe asthma and some with egg allergy 1 ).
Cases where the vaccination history was unknown, or they were vaccinated less than 14 days before onset of symptoms, were also excluded from the analysis. When the date of vaccination was missing, cases who were hospitalized after mid-January in the respective seasons were assumed to be vaccinated at more than 14 days before onset, because the vast majority of vaccinations are completed by mid-January. In addition, cases where the interval between onset of illness and swab date exceeded 7 days were also excluded due to well documented reduced test sensitivity for longer time periods between these two-time points. 25-28

| Reference population vaccine uptake
Data on population vaccination coverage was obtained from ImmForm: the routine national vaccine uptake monitoring system in England. 29 For the pre-school ages, weekly data were extracted from ImmForm on the number of children registered in primary care and the number of children who received seasonal influenza vaccination during the study seasons. Data were available by age group, week, geographical area (CCG), and presence/absence of a clinical risk factor. Data were not available by vaccine type.
T A B L E 1 Target childhood vaccination groups in England, 2013-15, setting of vaccine delivery and method of vaccine uptake data collection This information was used to determine the proportion of the population vaccinated (PPV).
For the crude analysis, VE can be calculated as 1À the odds of vaccination in cases/odds of vaccination in the population, or as below: where PPV is the proportion of the reference group vaccinated, and PCV is the proportion of the influenza cases vaccinated. The presence/absence of risk group was collected on all hospitalized cases although population vaccine uptake by presence/absence of risk group was only available for the pre-school age groups (2-4-year-olds). In a subgroup analysis for 2-4-year-olds, where risk factor information was available, PPV was assigned to cases that matched to each case by presence/absence of risk factor, as well as age group, geographical area, and week used above. Separately, because data completion on risk factor information was poor, a sensitivity analysis was also undertaken for the above group. In the first instance, all cases aged 2-4 years were assumed to have a risk factor, then second, they were all assumed not to have a risk factor. Stratified VE estimates were also estimated for those with/without a risk factor.

| Description of reference population
The population vaccine coverage for England for the population groups eligible for vaccination in the study seasons are shown in Table 3. Uptake was generally higher in all population groups compared with the PCV (Table 3).

| VE estimates
The crude and adjusted VE estimates for preventing influenzahospitalized cases in children by season and age group, for all influenza types, are shown in Tables 3 and 4. In a further subgroup analysis restricted to 2-4-year-olds on whom risk factor status was known, VE estimates were also adjusted for presence/absence of a risk factor ( Abbreviations: CI, confidence interval; VE, vaccine effectiveness. *Adjusted for risk factor in addition to CCG, week, and age. vaccine uptake for presence of a risk group was used. The adjusted VE in this instance was 69.9% (95% CI 57.6, 78.6) for all 2-4-yearolds (Table 5). Alternatively, the adjusted VE when cases with missing risk group status were assumed to have no risk factor was 57.6% (95% CI 34.1, 72.7) in all 2-4-year-olds (Table 5).

| CONCLUSIONS
In this study, the screening method was used to evaluate the effec-  should be restricted to 7 days or less, which is likely to also be applicable to the screening method. 35 In summary, we have found that vaccination with LAIV provided

PEER REVIEW
The peer review history for this article is available at https://publons. com/publon/10.1111/irv.12990.

DATA AVAILABILITY STATEMENT
The data that support the findings of this study are available from the corresponding author upon reasonable request.