The impact of influenza infection on young children, their family and the health care system

Background Influenza is a major cause of respiratory illness in young children. Assessing the impact of infection on children and the community is required to guide immunisation policies. Objectives To describe the impact of laboratory‐proven influenza in young children and to compare its impact with that of other respiratory viruses on the child, their family and the health care system. Methods Preschool children presenting for care or admission to a tertiary paediatric hospital during the 2008‐2014 influenza seasons were tested for respiratory virus by polymerase chain reaction and culture. Parental surveys were used to determine the impact of infection on illness duration, medication use, absenteeism and health service utilisation. Multivariate regression analyses were used to assess the impact of influenza and to evaluate the association between influenza status and outcomes. Results Among 1191 children assessed, 238 had influenza. Among children with influenza, 87.8% were administered antipyretics and 40.9% antibiotics. 28.6% had secondary complications. 65.4% of children missed school/day care, and 53.4% of parents missed work. When influenza and other viruses were compared, significant differences were noted including duration of illness (influenza: 9.54 days, other viruses: 8.50 days; P = 0.005) and duration of absenteeism for both the child (23.1 vs 17.3 hours; P = 0.015) and their parents (28.5 vs 22.7 hours; P = 0.012). Conclusions Influenza infection in young children has a significant impact on medication use, absenteeism and the use of health care service. Significant differences are identified when compared with other ILI. These data demonstrate that influenza prevention strategies including immunisation are likely to have wide and significant impacts.


| INTRODUC TI ON
Influenza is a major cause of respiratory illness in young children, 1 with those <5 years at greatest risk of disease. The rate of influenzaassociated hospitalisations in this age group exceeds that seen in the elderly. 2 Assessing the impact of influenza infection in young children is challenging as it co-circulates with other respiratory viruses, from which it is clinically indistinguishable without laboratory testing. 3,4 It is well recognised that estimates of influenza morbidity and mortality are likely to be underestimated. 5 The majority of influenza-related disease in children is seen in outpatient settings. 1,6 Despite lower severity, it is speculated that non-hospitalised influenza represents a greater burden overall. 6,7 Influenza in children is associated with the frequent use of antipyretics, antibiotics, 6,8,9 and parental and child absenteeism. 6,10-14 Health care visits due to influenza illness have been previously estimated to cost the Australian health care system AU$115 million annually (2008). 15 Indirect costs from parental absenteeism are thought to be the largest contributor to the economic impact of influenza in young children. 10,14,16 The impact of influenza in children has been well described in the Northern Hemisphere, 6,10,14,17 yet impact in Australia remains unknown. To date, data on the impact of influenza have been limited to hospitalised children, 18,19 single influenza seasons 16,[18][19][20] or studies with small numbers of influenza-positive children. 16,20 In Australia, influenza vaccine is not included on the National Immunisation Program (NIP), except for Aboriginal and Torres Strait Islander children for whom it was funded in 2015. 21 Western Australia (WA) introduced a free influenza programme for children Commencing in 2008, the WA Influenza Vaccine Effectiveness (WAIVE) study was established to assess vaccine effectiveness in young children 22,[24][25][26] and to assess the impact of influenza on young children and their families. We describe the impact of laboratoryproven influenza on the child, their family and health care services, and compare the impact of influenza with other respiratory viruses presenting as influenza-like illness (ILI).

| Study population
As previously described, 24

| Study procedures
Following written consent, bilateral flocked nasopharyngeal swabs were obtained (Copan Diagnostics Inc. Murrieta, CA). Children from whom a nasopharyngeal aspirate (NPA) was already taken at presentation did not require additional testing. Using previously published methods, samples were tested for respiratory viruses including influenza, respiratory syncytial virus (RSV), human metapneumoviruses, differences were noted including duration of illness (influenza: 9.54 days, other viruses: 8.50 days; P = 0.005) and duration of absenteeism for both the child (23.1 vs 17.3 hours; P = 0.015) and their parents (28.5 vs 22.7 hours; P = 0.012).
Conclusions: Influenza infection in young children has a significant impact on medication use, absenteeism and the use of health care service. Significant differences are identified when compared with other ILI. These data demonstrate that influenza prevention strategies including immunisation are likely to have wide and significant impacts.

| Exclusions
Enrolled children for whom a respiratory sample was not successfully collected (n = 16), a second questionnaire was not completed (n = 1946) or a second questionnaire was completed either too early or too late to accurately capture impact data (<6 days or >40 days after enrolment; n = 360) were excluded from the analysis. In addition, to minimise parents reporting the impact from consecutive and/or concurrent viruses not detected by the respiratory sample, children who reportedly had a continuous fever commencing >7 days prior to enrolment were also excluded (n = 102).

| Exposure ascertainment and outcomes
Primary exposure was defined as laboratory-proven influenza A or B by either PCR or culture. Two comparison groups were used: (a) influenza-negative children and (b) influenza-negative children testing positive for other respiratory viruses. The second comparison group was used to reduce the probability of including false-negative influenza infection due to inadequate specimen collection, storage or transport and to exclude children whose symptoms were due to non-infectious causes. 24 The outcomes examined were parental reported duration of symptoms (resolution being defined as the majority of symptoms resolving and the child being back to usual activities); prevalence and duration of medication use (over the counter [OTC] and prescription); diagnosis of secondary infections or complications, child absenteeism from playgroup, day care or school; parental absenteeism from work; and health care service use including hospitalisation, visits to general practitioners (GP) or other health care professionals and the performance of diagnostic tests or procedures.  Influenza-infected children were older (mean age 2.85 vs 1.99 years, P < 0.001), more likely to be unvaccinated against influenza (86.1% vs 67.6%, P < 0.001) and more likely to attend day care or school ≥4 hours a week (72.2% vs 60.3%, P = 0.001) compared with children with other respiratory viruses.

| Impact of influenza and other respiratory viruses
Impact study outcomes are shown in Tables 2 and 3. As similar results were found, only the comparison between the influenza-positive and other respiratory virus-positive groups is shown. The mean duration of influenza illness was 9.54 days, compared to 8.50 days in other respiratory viruses (P = 0.005 [ Table 3]).
The majority of children with influenza were given antipyretics (paracetamol and/or ibuprofen; 87.8%) with an average duration of 5.25 days and 4.91 days, respectively. Antibiotics were administered in 41% and decongestant/cough medicine in 22% (Tables 2,   3). Antipyretic and decongestant/cough medicines were used more frequently in children with influenza than with any other viruses (an-    Table 3).
Two-thirds of children with influenza were reported to have visited a GP or out of hours GP (

| D ISCUSS I ON
This is the first Australian study to quantify the impact of laboratoryconfirmed influenza infection in preschool children over several consecutive influenza seasons. Our findings illustrate that influenza causes a substantial impact on young children, their families and medical services. Although minor, differences exist between the impact of influenza and other common respiratory viruses presenting as ILI.
Mean duration of influenza illness was found to be 9.5 days, which was approximately 1 day longer than other respiratory viruses.
The increased duration of the use of antipyretics and decongestant/ cough medicine, and of child and parental absenteeism observed, is likely to be associated with the longer duration of illness.   34 and the increasing prevalence of antibiotic resistance. 35 In contrast, the infrequent use of antivirals highlights that empiric antiviral therapy, as opposed to antibacterial therapy, is uncommon in WA, particularly those seen in the emergency department. Of the nine children prescribed oseltamivir, six were shown to have influenza and three did not. Only 13% of children hospitalised with influenza were treated with antivirals. The use of rapid diagnostic tests, which are infrequently used in Australia, may increase the use of antivirals, decrease antibiotic prescriptions and reduce the influenza burden in those children at risk of severe disease and mortality. 36,37 A substantial social impact of influenza on the family was demon-  Multiple visits to medical practitioners were required in many cases: around two-thirds of the children with influenza who presented to ED were taken to the GP for the same illness and a third required ≥2 GP visits. In addition, around one-fifth required two or more ED visits. Those re-presenting to ED were significantly younger both among those with influenza and overall (data not shown). This represents a high burden on health services with associated economic costs. The rate of hospitalisation for children with influenza and other viruses must be interpreted with caution, as children presenting to ED and hospitalised were both actively recruited. The higher rate of hospitalisation seen with other respiratory viruses was not significant when only children enrolled in ED prior to hospitalisation were included. Despite the limitations, these data are still useful in comparing the impact between influenza and non-influenza ILI, and also for ascertaining average duration of hospitalisation among children admitted with influenza, which was found to be 2.23 days.
As the study population was limited to children presenting to ED or hospitalised at a single metropolitan tertiary hospital, there are limitations on the generalisability of the findings. Enrolled children had a higher rate of comorbidities and prematurity than the general population. For example, while only 8.5% of babies born in Australia were preterm in 2012, 38 14.5% of our sample were reported as preterm.
Also, parents of children who were severely unwell may not have been approached by research assistants or less willing to participate, thereby affecting the overall disease severity of the sample. Other limitations included data collection being via parental report rather than through medical record verification, decreasing accuracy of the reported clinical outcomes. Additionally, due to clinical duty of care, parents were informed whether their child was influenza-positive or influenza-negative prior to completion of the second questionnaire, which may have affected the reported outcomes.
These data indicate a high impact of influenza illness in WA children and can be used to re-assess the current economic impact of influenza in young children in Australia. 15 The WAIVE study is funded by the Department of Health of Western Australia.

CLI N I C A L TR I A L R EG I S TR ATI O N
None.