School absenteeism among school‐aged children with medically attended acute viral respiratory illness during three influenza seasons, 2012‐2013 through 2014‐2015

Background Acute respiratory illnesses (ARIs) are common in school‐aged children, but few studies have assessed school absenteeism due to specific respiratory viruses. Objective To evaluate school absenteeism among children with medically attended ARI due to common viruses. Methods We analyzed follow‐up surveys from children seeking care for acute respiratory illness who were enrolled in the influenza vaccine effectiveness study at Marshfield Clinic during the 2012‐2013 through 2014‐2015 influenza seasons. Archived influenza‐negative respiratory swabs were retested using multiplex RT‐PCR to detect 16 respiratory virus targets. Negative binomial and logistic regression models were used to examine the association between school absence and type of respiratory viruses; endpoints included mean days absent from school and prolonged (>2 days) absence. We examined the association between influenza vaccination and school absence among children with RT‐PCR‐confirmed influenza. Results Among 1027 children, 2295 days of school were missed due to medically attended ARIs; influenza accounted for 39% of illness episodes and 47% of days missed. Mean days absent were highest for influenza (0.96‐1.19) and lowest for coronavirus (0.62). Children with B/Yamagata infection were more likely to report prolonged absence than children with A/H1N1 or A/H3N2 infection [OR (95% CI): 2.1 (1.0, 4.5) and 1.7 (1.0, 2.9), respectively]. Among children with influenza, vaccination status was not associated with prolonged absence. Conclusions School absenteeism due to medically attended ARIs varies by viral infection. Influenza B infections accounted for the greatest burden of absenteeism.

more school than those with respiratory illnesses of other etiologies. 4,5 Few studies have assessed the burden of school absenteeism due to laboratory-confirmed influenza, [5][6][7] and only one was conducted after the recommendation for annual influenza vaccination in children in the United States. 5 Absenteeism data on the relative contribution of ARIs caused by viruses other than influenza are lacking. Non-influenza viral illnesses may be less common, but also likely to disrupt usual activities and cause increased school absenteeism. Quantifying school absenteeism due to specific viruses can help target prevention or treatment strategies to reduce burden in school-aged children.
We utilized data from an observational influenza vaccine effectiveness study to evaluate parental-reported school absenteeism across three seasons among children with medically attended ARI due to various viruses. Specifically, we aim to estimate the average days absent for specific respiratory viruses in children, identify risk factors for prolonged (>2 days) absence from school due to viral ARIs, and evaluate the association between influenza vaccination and prolonged absence among children with influenza.

| METHODS
We conducted an analysis using follow-up surveys from children with ARIs. Individuals in the defined cohort were eligible for recruitment to the influenza VE study if they were ≥6 months old and presented with symptoms of cough lasting no more than seven days at the time of their visit. Consenting patients completed an enrollment interview and provided a nose and throat swab for influenza testing. From the enrollment interviews, we obtained information on age, race/ethnicity, self-reported health status prior to the onset of illness, illness onset date, and symptoms. Vaccination status was obtained from the validated vaccine registry that serves the population. 11 Approximately 1 week following enrollment, all influenza-positive patients and approximately 50 influenza-negative patients per week were contacted for a follow-up interview. For children, interviews were conducted with a parent or guardian and included questions about when the child returned to normal activities, how many days the child missed school due to the illness (numerically open-ended), and medications prescribed and taken. The answer to the question, "how many days of school did your child miss, due to this illness" was used to assess school absenteeism. Responses larger than the maximum number of weekdays between illness onset and follow-up interviews were truncated to the maximum weekdays within the interval (nine observations were truncated).

| Laboratory testing
Combined nose and throat swabs collected at the time of enrollment in the influenza VE study were tested for influenza virus (type and subtype) using real-time reverse-transcription polymerase chain reaction (RT-PCR). 12

| Statistical analysis
For this study, we included children 5-17 years old at the time of enrollment in the influenza VE study, whose parent or guardian had participated in the follow-up interview and who attended school outside the home.
Demographic and clinical characteristics were compared using χ 2 tests. Absenteeism rate was estimated by dividing the total number of days missed by the number of children with the infection. Negative binomial regression models were used to estimate the average days absent and 95% confidence interval (CI) for each viral infection category because absenteeism was skewed to the right. All variables potentially associated with days absent were added to the initial model. These variables included sex, race/ethnicity, reported health status prior to current illness, influenza vaccination status at the time of illness onset, and receipt of antivirals (after enrollment in the influenza VE study). Backward elimination with a cutoff of P value=.05 was conducted to determine inclusion in the final model. Age was included a priori. Children with coinfections with viruses other than influenza were excluded.
Similar multivariate methods (with the same variables listed above) were performed to identify risk factors for prolonged absence from school using logistic regression. Age was included a priori. Prolonged absence was defined as >2 days absent from school.
We performed a separate analysis to estimate the mean days absent for vaccinated and unvaccinated children and examine the association of influenza vaccination and prolonged absence due to influenza. This analysis was restricted to children infected with influenza, and each influenza subtype and lineage (A/H1N1, A/H3N2, B/ Yamagata, and B/Victoria) were evaluated separately. Children who were vaccinated within 14 days of illness onset or were not considered adequately vaccinated according to the Advisory Committee on Immunization Practices (ACIP) were excluded. 14- 16  season. 17,18 All analyses were performed using sAs statistical software (version 9.3; SAS Institute, Cary, NC, USA).

| RESULTS
During Respondents and non-respondents were similar with regard to age, sex, race/ethnicity, reported symptoms, influenza vaccination status, and time from illness onset to influenza VE study enrollment (data not shown). Respondents were more likely to be enrolled in the influenza VE study during the period after influenza circulation peaked (63% vs 46%, P<.0001). Among respondents, 55 (5%) were excluded because they did not attend school outside the home (n=37), had influenza B with no lineage data (n=12), had influenza A and B or B/Yamagata and B/Victoria coinfection and could not be classified into a single influenza virus group (n=3), or had missing data on days absent from school (n=3).
Among the 1027 children included in this study, viral infections were detected in 747 (73%) ( Table 1)

| Mean days absent
In total, 2295 days of school were missed by our study population over three influenza seasons; 175 (17%) children did not miss any days. The 2012-2013 season, reported fair/poor health status prior to illness, enrollment during the peak influenza season, and receipt of antivirals were associated with greater mean days absent from school ( Table 2).
Longer duration of illness, being fatigued, having a fever, being short of breath, and having a sore throat were also associated with greater mean days absent. There were no differences in mean days absent by age group, sex, race/ethnicity, or influenza vaccination status.

| Risk factors for prolonged absence
Most children (62%) missed two or fewer days of school due to their medically attended ARI. The 2012-2013 season and peak influenza season were associated with prolonged absence of >2 days ( Table 2).

| DISCUSSION
In this study, we evaluated the association between specific viral infections and school absence. Influenza, RSV, coronaviruses, and rhinoviruses were the most commonly detected viruses in children 5-17 years old with medically attended acute respiratory illness, and viral illness accounted for about 75% of missed school days. Influenza contributed to a substantial burden on school absenteeism, averaging over one missed school day per illness and over 40% missing >2 days.
School absenteeism due to ARI caused by other viruses contributed to fewer days missed, but was common.
Influenza accounted for almost half of days absent by children with medically attended acute respiratory illness. Seasons predominated by A/H3N2 tend to be associated with increased morbidity and mortality, [19][20][21] but there was no difference in the average days missed due to influenza between children with different subtypes or lineage in our study population. However, children with influenza B/Yamagata infection were more likely to have prolonged absence than children with influenza A. The reason for this is unclear; influenza B infection rates are highest in school-aged children, and those with influenza B were more likely to seek medical attention, 1,22 but more cases of A/H3N2 were identified in our study. In contrast, previous studies have found lower, but similar absenteeism rates by influenza type 6,7  in milder illness, leading to fewer outpatient visits. Our study population was restricted to respiratory illnesses that were severe enough to seek medical attention, and was not able to assess mild illness due to vaccination that may have resulted in fewer school days missed.
However, a prior randomized placebo-controlled trial in children with influenza B found no difference in duration of illness between those receiving the vaccine and those receiving a placebo. 31 For children under five years old, RSV causes significant morbidity, with hospitalization rates higher than those reported for influenza. [32][33][34] For older children and adults, RSV typically causes more mild illness than influenza. 13 In our study, the prevalence of RSV infection and school absenteeism due to RSV was lower than that among children with influenza, consistent with milder illness compared to influenza.
However, this contrasts a previous study among children attending the emergency room in Italy, where there was no difference in median days missed from school between children infected with RSV and influenza. 35 The median days missed were much higher (10-12 days) than those in our outpatient study, suggesting emergency room visits may have been more severe in general and may explain discrepancies between the two studies.
RSV, coronavirus, and rhinovirus were prevalent during all seasons examined, but these viruses contributed to fewer missed school days. This is not surprising as coronaviruses and rhinoviruses are the Data on school absenteeism due to RSV should be considered when estimating the direct and indirect impact of potential RSV vaccine policies.