Overall, this study will determine (1) the effect of vaccine type (FluMist vs. TIV) on immunization rates, (2) assess the feasibility of school immunization clinics, and (3) inform a larger study to understand the effects of these two vaccines on reducing influenza.

Our study partnership (Public Health Agency of Canada [PHAC] / Canadian Institutes of Health Research [CIHR] Influenza Research Network [PCIRN], Public Health Ontario [PHO], and PCCHU), with the support of Ontario's Ministry of Health and Long-Term Care [MOHLTC], will conduct an open-label randomized trial in elementary schools within the geographic boundaries of PCCHU during the 2013-14 influenza vaccination campaign.

In order to address the four study objectives in our comparison of LAIV vs. TIV via SBII, we will use multiple methodologies: questionnaires, interviews, and household surveillance.

Overall study design PCCHU implemented SBII in elementary schools from 2000 to 2010, and we have approval from the Kawartha Pine Ridge District School Board (KPRDSB) to renew this program for the upcoming influenza season (2013-14). Ten schools (each school having an average of 300 students) will be recruited to participate in this study. The schools will be randomized on a 1:1 basis (i.e., five schools in each arm) to having its students offered (at no cost) 0.2 mL of LAIV or 0.5 mL of TIV at SBII clinics organized by PCCHU.

Intervention (LAIV): Healthy children and adolescents between Junior Kindergarten (JK) and Grade 8 in the intervention schools will be immunized with LAIV (FluMist influenza vaccine) recommended for the 2013-14 the influenza season. A 0.2 ml dose of the vaccine will be administered intranasally. TIV will also be available for children whose parents decline LAIV but would still like their child immunized against influenza, and for children who, according to 2012-13 NACI recommendations, have contraindications for LAIV: severe asthma (defined as currently on oral or high-dose inhaled glucocorticosteroids or active wheezing) or medically attended wheezing in the 7 days prior to vaccination; receiving aspirin or aspirin-containing therapy because of the association of Reye's syndrome with aspirin and wild-type influenza infection; and immune-compromising conditions due to underlying disease and/or therapy, as the vaccine contains live attenuated virus.

Control (TIV): Healthy children and adolescents between JK and Grade 8 in schools assigned to TIV will be immunized with inactivated influenza vaccine (Vaxigrip vaccine) recommended for the 2013-14 influenza season. A 0.5 ml dose of the vaccine will be administered intramuscularly. Children with one or more of the following contraindications, according to NACI recommendations, will not be eligible for vaccination with TIV: an anaphylactic reaction to a previous dose; an anaphylactic reaction to any of the vaccine components, with the exception of egg; or Guillain-Barré Syndrome (GBS) within six weeks of a previous influenza vaccination.

Children in either arm of the study who are younger than 9 years of age and have not previously received influenza vaccines require two doses; following the dose administered at school, they will be given a reminder letter to inform their parent/guardian that a second dose can be received at healthcare settings including their family physician's office or a community-based public health clinic, four weeks after the first dose as per influenza immunization recommendations.

For the purpose of this study, the vaccines will only be offered to children enrolled at participating schools; other groups such as school staff, and students' parents and siblings may obtain TIV at no cost at the settings listed above or LAIV from a physician or pharmacy (paid for privately).

Objective 1: To determine whether offering LAIV vs. TIV via SBII leads to higher vaccine coverage

Study design: A web-based questionnaire will be sent to all parents/guardians of students in participating schools for whom we have contact information, to ask, among other questions, about vaccination status of their child, and if the child was vaccinated, where this occurred.

Study procedures: All parents/guardians who provided an email address will be contacted in January 2014 via email with an invitation to complete an online (or paper-based, if preferable) Parents' Vaccination Questionnaire on behalf of their child. Prior to beginning the questionnaire, the parent/guardian will access a screen which describes the purpose of the questionnaire, which data are being collected, and why. They will also be informed that all questions (except for whether or not their child was vaccinated, and with which influenza vaccine) are optional, and that completion of the questionnaire implies consent. The questionnaire will include questions about student demographics, presence of chronic conditions and other important risk factors for influenza, receipt of influenza vaccines during previous influenza seasons, and receipt of influenza vaccines outside of the school setting (e.g., doctor's office, pharmacy, PCCHU community-based clinic) during the current influenza season. We will also request health card numbers and permission for future linkage to provincial health administrative data sets housed at the Institute for Clinical Evaluative Sciences (ICES) to assess healthcare utilization, in order to determine which of TIV and LAIV is associated with fewer physician and emergency room visits. In cases where the health card number is not provided or is incorrect, the first and last name, sex, and birthdate of the student will be used for linkage. In April 2014, the questionnaire will be sent out again to parents/guardians of children who, based on the questionnaire sent in January, had not received influenza vaccine, to capture anyone whose vaccination status changed between January and April, and to also provide another opportunity for parents who did not yet complete the questionnaire. After the email with the questionnaire web link is sent out in both January and April 2014, two email reminders will be sent over the next three weeks. For parents/guardians who provide their phone number to the study team, an RA will also remind them by telephone to complete the questionnaire.

Outcome: The number of students at each school for whom the Parents' Vaccination Questionnaire is completed will serve as the denominator for the vaccine coverage measurement, and the number of students who receive LAIV or TIV and have the questionnaire completed by their parent/guardian will serve as the numerator. Separate estimates of vaccine coverage will be calculated for healthy children and those with chronic conditions that place them at elevated risk of serious influenza infection. As a secondary measure of immunization rates, we will use vaccination data obtained by PCCHU at the beginning of January 2014, of the number of doses (by vaccine type) given during the school-based clinics (numerator). We will ask each principal for the number of students at their school as of November 1, 2013, to use as the denominator.

Sample size: To calculate the sample size required to detect a 10% increase in coverage with LAIV compared to TIV (i.e., primary outcome), we used the sample size calculation for two proportions based on 80% power and an alpha level of 0.05 (two-sided). The unit of randomization will be the school level while the unit of analysis will be the individual students in the schools. To account for the clustering effect of schools, we inflated the sample size by 20% (i.e., variance inflation factor), resulting in a sample size of 451 students/group. We anticipate that we should have an adequate sample size since we only need a response rate of 30% (902/3000), and we hope that with the incentives that we will have a response rate of over 70%.

Data analysis: Bivariate analysis of statistical significance of continuous variables will be performed using two-tailed Student's t-tests. Proportions will be compared using Chi-Square tests or Fisher's exact tests. We will identify predictors of the outcome (vaccinated or not) using stepwise regression methods. To account for the clustering effect in schools, we will use generalized estimating equations or mixed models to determine statistical significance of continuous or dichotomous variables adjusted for covariates and interaction terms. The exposure of interest will be the vaccine type (LAIV or TIV). The null hypothesis will be rejected fixing the probability of type I error at 0.05 and a type II error at 0.20. Two-tailed probability distributions will be used for all statistical analyses unless specified. Variables will be excluded from the model if they lead to complete separation of data, have >10% missing values, exhibit collinearity, or have no observations in one group. We will also examine the residuals to identify if model assumptions are invalid or are affected by influential observations. We will use STATA version 10.0 (2007, StataCorp, LP, College Station, TX) to analyze the data.

Objective 2: To assess the acceptance of administering LAIV vs. TIV via SBII based on school administrator, public health, and parent/guardian perceptions

Study design: In order to assess perceptions of the LAIV and TIV SBII programs, we will interview school management staff, and public health managers, and administer questionnaires to parents/guardians of students in the participating schools.

Inclusion/exclusion criteria: Interviewees must be members of the management staff (principal, vice principal) of the participating schools, or PCCHU manager or immunization nurse, and have been involved in some aspect of the coordination of the SBII clinic. The Parents' Vaccination Questionnaire will be used to gauge parental perceptions of the different vaccine types, and will be administered to all parents/guardians for whom we have contact information, regardless of whether or not their child was vaccinated.

Study procedures:

• Interviews with school management staff immediately following the SBII clinic days to understand their experiences with coordinating the program, any facilitators/barriers to its continued existence, and also experiences/interactions with parents regarding this program. The Information Letter and preliminary interview guide are outlined in Appendices H and I. These interviews will take place over the telephone in November or December, 2013.
• Interviews with public health immunization staff (manager(s) and nurses) to gauge perceptions of the program, and of administration of LAIV vs. TIV, based on logistical factors including time to coordinate the program, communication with parents, and staff time. These interviews are scheduled to take place in November or December, 2013; however, if the public health staff are unavailable at this time due to conducting vaccination clinics, we will conduct these telephone interviews in January 2014 instead.
• Questions related to parent/guardian experiences with SBII clinics, as well as their perceptions of LAIV vs. TIV, specifically regarding safety, effectiveness, and acceptability (invasiveness, concerns about administration, adverse events, etc.), will be included in the Parents' Vaccination Questionnaire.

Outcome: The outcome of this study component is stakeholder (school staff, public health, and parents) perceptions of the LAIV and TIV SBII programs, specific to coordination, facilitators/barriers, and overall experiences.

Sample size: We anticipate that we will interview approximately one principal or vice-principal from each of the 10 schools, as well as five PCCHU managers or staff members. However, the number of interviews conducted with each of these stakeholder groups will ultimately be based on the number of individuals who were involved in planning of the SBII clinics; therefore, fewer or more interviews may be conducted, as appropriate. The Parents' Vaccination Questionnaire will be used to gauge parental perceptions of the different vaccine types, and will be administered to all parents/guardians for whom we have contact information.

Data analysis: The interviews will be imported into qualitative analysis software (N-Vivo Version 10.0) to aid in data organization, review, coding, and analysis, and to facilitate an exploration of themes that emerge from the data. Themes will be analyzed by type of vaccine given, and other factors including the school's population size. At least two team members will read each transcript once to get an overall sense of the data, and then again to identify major topics or issues in the data. Consensus decision-making will be used when needed to arrive at mutually agreed-upon coding.

We will use the two-proportion z-test to compare the proportion of parents/guardians with children in the LAIV group and those with children in the TIV group, based on perceptions of vaccine safety and effectiveness, and overall preference for vaccine type. We will also assess these outcomes for parents/guardians who did not vaccinate their children this season. To account for multiple comparisons, we will adjust the p values using Benjamini-Hochberg's method for controlling false discovery rates (FDR), which controls for the number of comparisons falsely declared significant.

Objective 3: To compare the costs associated with delivering LAIV vs. TIV via SBII

Study design: To obtain costing data associated with coordination of the LAIV and TIV SBII programs, principals at participating schools and immunization program managers at PCCHU will be interviewed.

Inclusion/exclusion criteria: We will use the same criteria, as for Objective 2, since the cost data will be obtained during interviews.

Study process: During the interviews with school board and school management staff, and public health managers, we will ask about the costs associated with administering LAIV and TIV, including but not limited to training costs, equipment costs, and staff resource use (nursing time for coordinating the campaign, communication with parents/guardians, delivery of vaccine and post-vaccination monitoring, as well as data entry clerk time, where appropriate).

Outcome: The outcome of this component of the study is the costs that PCCHU, participating schools, and the MOHLTC incurred in order to deliver LAIV and TIV through SBII programs.

Sample size: We anticipate that we will interview approximately one principal or vice-principal from each of the 10 schools, as well as the immunization program manager at PCCHU. However, we may need to conduct additional or fewer interviews, depending on how many individuals are able to provide us with costing data.

Objective 4: To inform a future RCT to evaluate the direct and indirect benefits (protection against influenza) of LAIV vs. TIV via SBII

Study design: From each of the 10 study schools, 25 vaccinated students and their households will participate in the Household Surveillance Study, involving surveillance for laboratory-confirmed influenza during the 2013-2014 influenza season, to assess the effect of LAIV and TIV on influenza incidence. The purpose of this component is to generate parameter estimates (e.g., incidence of laboratory-confirmed influenza, intraclass correlations) to inform a future RCT.

Inclusion/exclusion criteria: Households will be eligible for the Household Surveillance Study if: i) they include a child who was vaccinated against influenza at one of the 10 school-based clinics; and (ii) all household members consent/assent to participating. However, as stated above, we will only maintain inclusion criterion (ii) if feasible. Additionally, the students selected from each school must have received the vaccine to which their school was assigned (e.g., a student at an LAIV-assigned school must have received LAIV).

Study procedures: Once consent/assent has been obtained for the household, the HRA will also collect and record baseline data from each household member, including demographics, risk factor status, and current influenza vaccination status. We are collecting this information to be able to describe the baseline characteristics of our population, as is reported on a standard basis in scientific publications. In addition, we will need to control for demographic and risk factor variables when we are analyzing trends in the data, to better compare the effects of each type of vaccine. Each individual will be asked the baseline questions in turn, and in private, to help encourage honesty. Each household will be provided with either paper or electronic diaries for recording daily symptoms as well as a digital thermometer (which they will be taught to use). They will also be supplied with Copan flocked nasal swabs, and will be taught by the HRA how to self-swab nasally. Participating households will be asked to:

i. Complete the daily diary to record whether any household member has cold, influenza, or acute respiratory illness symptoms, illness history (e.g., hospitalizations related to lower respiratory tract infections and pneumonia, physician visits for respiratory illness), medical history (e.g., medication use), and missed days of school or work due to acute respiratory infection. The diary will be available in both paper and electronic form.

ii. Have any household member exhibiting any one of the three symptoms for influenza (cough, sore throat, fever) or at least two other related symptoms perform an intranasal self-swab as soon as symptoms appear (within 48 hours), and call a Research Assistant (RA).

Data collection for the Household Surveillance study will begin when the weekly percentage of tests positive for influenza from the Peterborough area (tested via PHO) exceeds 5%, while the end date will be when the weekly percentage of tests positive for influenza from the Peterborough area falls below 5%, indicating the end of the 2013-14 influenza season. Once the study begins, an RA will contact each household to let them know to start recording their data daily in the symptom diaries. For households who opt for paper diaries, the RA will provide them with paper copies of the symptom diary to complete on a daily basis. The RA will contact each of these households twice weekly by telephone to obtain the data for each household member, and will record this directly into FluidSurveys. Weekly emailed reminders will also be sent to these participating households. Those households who opt for electronic diaries will receive a daily link to the patient diary in FluidSurveys. This email will be sent to the main contact of each household, who will fill out the form for each household member. The SRC will monitor diary entries, and if no entries are submitted for a household for three consecutive days, an RA will call that household to remind them to complete the diary. If any individual is experiencing at least two cold or influenza symptoms or any one of the three specific symptoms for influenza [cough, sore throat, fever]), the SRC will let the RA know to contact the individual to: i) remind them to self-swab; and ii) arrange a household visit to take a nasal swab of that individual within 48 hours of the symptom onset (as reported by the individual). Thus there should be two swabs per episode. A repeat swab will be collected 7 days after the collection of the initial swab (indicating a new episode) if the individual has at least one additional symptom, as well as the initial 2 symptoms; or if the individual has at least two new symptoms (different from the initial two symptoms and the first two have resolved).

The RAs will collect all nasal swabs during household visits, and transport them to the regional laboratory in Peterborough, where they will be stored in a freezer until transported to the central laboratory in Toronto where testing will take place.

At the end of the study, the RA will request that the parent/guardian for each participating household complete the Parents' Vaccination Questionnaire for their child, if they have not yet done so.

Sample size: The Household Surveillance study will include 25 vaccinated students and their households, from each of the 10 participating schools. The students must have been vaccinated during the SBII clinic at their school.

Laboratory testing: The specimens will be tested by means of a reverse-transcriptase polymerase chain reaction (RT-PCR) assay to detect the presence of influenza A or B virus. We will estimate primary and secondary attack rates, based on the proportion of household contacts in whom influenza developed (as determined by RT-PCR assay), as well as rates of acute respiratory illness.

PHO has a regional laboratory in Peterborough so all the specimens will be dropped off there and sent along with the routine clinical specimens to the central laboratory in Toronto for PCR testing. We will test the specimens in batches since real-time results are not required for the purposes of this study. We will culture the specimens that test positive for influenza on PCR and explore strain identification by the National Microbiological Laboratory at a future date.

Data analysis: To calculate influenza incidence among household members, we will count only the first infection by a specific influenza type/subtype (i.e., A/H3N2, A/H1N1, B) as the numerator (age-specific), while the denominator is age-specific uninfected person-days. If a participant is swabbed more than once within 14 days, and if the same virus is detected on more than one occasion in this time frame, only the first episode detected will be used in the analysis. This will be calculated for both influenza and non-influenza viruses. We will calculate incidence using events per 1000 person-days and estimate differences in rates assuming a Poisson distribution.

Children in schools randomized to LAIV cannot receive this vaccine if they have any of the following contraindications: severe asthma (defined as currently on oral or high-dose inhaled glucocorticosteriods or active wheezing) or medically attended wheezing in the 7 days prior to vaccination; receiving aspirin or aspirin-containing therapy because of the association of Reye's syndrome with aspirin and wild-type influenza infection; or immune-compromising conditions due to underlying disease and/or therapy, as the vaccine contains live attenuated virus. These children will receive TIV instead, unless they have a contraindication to receiving TIV. Children in schools assigned to TIV cannot receive this vaccine if they have any of the following contraindications: an anaphylactic reaction to a previous dose; an anaphylactic reaction to any of the vaccine components, with the exception of egg; or Guillain-Barré Syndrome (GBS) within six weeks of a previous influenza vaccination.

Children from junior kindergarten to Grade 8, from 10 elementary schools will participate in this study.