Much ado about flu: A mixed methods study of parental perceptions, trust and information seeking in a pandemic

Background Effective public health messaging is essential in both the planning phase and duration of a pandemic. Objectives This study aimed to gain an understanding of parental information seeking, trusted sources and needs in relation to pandemic influenza A 2009 (pH1N1) to inform future policy planning and resource development. Patients/Methods We conducted a mixed methods study; parents from 16 childcare centres in Sydney, Australia, were surveyed between 16 November and 9 December 2009, and interviews were conducted with participants from six childcare centres between June 2009 and May 2011. Results From 972 surveys distributed, 431 were completed; a response rate of 44%. Most parents (90%) reported that doctors were “trusted a lot” as a source of influenza information, followed by nurses (59%), government (56%) and childcare centres (52%). Less trusted sources included media (7% selected “trusted a lot”), antivaccination groups (6%) and celebrities (1%). Parents identified a range of key search terms for influenza infection and vaccine. From 42 in‐depth interviews, key themes were as follows: “Action trigger,” “In an emergency, think Emergency,” “Fright to hype” and “Dr Google and beyond.” Parents relied heavily on media messages, but cynicism emerged when the pandemic was milder than expected. Parents viewed a range of information sources as trustworthy, including doctors, authoritative hospital or government websites, and childcare centres and schools. Conclusions A user‐centred orientation is vital for pandemic communications including tailored information provision, via trusted sources based on what parents want to know and how they can find it.

in Australia in September 2009 and freely available for those aged 10 years and above 8 ; this was later extended in December 2009 to include children aged 6 months to 9 years. 9 There were approximately 37 000 laboratory-confirmed cases of pH1N1, and 5000 hospitalisations and 191 deaths due to pH1N1 in Australia in 2009. 7 The median age of those who died was much lower than in preceding influenza seasons (53 years, rather than 83 years). 10 An international systematic review revealed a higher pH1N1 attack rate in children compared to older adults who had some immunity from previous exposure. 11 More than 100 children were hospitalised in Australia during the pandemic period and 11 children died. 12 Effective communication and public health messaging is a key component in both the planning phase for a health emergency such as a pandemic and during the emergency itself. 13,14 Public cooperation during a pandemic is essential to minimise disease spread, ensure compliance and support for hygiene and social distancing measures and vaccination efforts, and avoid unnecessary overload on the health system. 15,16 As a pandemic progresses, messages may need to be modified according to the changing context. 15  Public health messages are received and interpreted contextually and according to individual experiences. 15,17 Parents are the key enablers of whether or not their children comply with public health measures, including vaccination, and trust plays a key role in decisionmaking in both pandemic 16,18,19 and non-pandemic periods. 20 It is therefore important to know who and what parents trust to provide information about influenza and influenza vaccine in a pandemic context. Thus, this study aimed to explore what information sources parents trusted and used to obtain information about pH1N1, during both the acute and post-pandemic phase. Further, it examined how parents searched for information on influenza infection and influenza vaccine. An understanding of parental information needs and searching preferences could provide valuable insights to inform future pandemic planning and information campaigns.

| ME THODS
This study, conducted in Sydney, Australia, was part of a broader study examining the health, social and economic impacts of vaccinating children attending childcare against influenza.
We used mixed methods-a quantitative survey and qualitative semi-structured interviews with parents of children aged 6 months to 5 years. Use of mixed methods combines the strengths of both quantitative and qualitative approaches and allows for a more robust exploration of an issue. 21 The timing of the survey distribution and interviews in relation to external pH1N1 events is contained in Table 1.

| Quantitative survey
To inform the questions for the quantitative survey, the first author (CK) conducted five pilot qualitative interviews between 18 and 25 June 2009 at a Sydney metropolitan childcare centre. Parents were asked whether they had sufficient information about pandemic influenza, where they had obtained the majority of their information, and who or what they trusted to give them reliable information about pandemic influenza. The term "swine flu" was specifically used in interview questions as this closely mirrored the terminology commonly used in the media at the time.
Responses from these interviews formed the basis for quantitative questions on information sources and trust. These questions were designed by three authors (CK, MC and JL). Parents were asked to rate their level of trust in nine information sources using a Likert scale. Parents also had the option to use free text responses to nominate any other trusted sources, and to indicate how they would undertake an Internet search for information on influenza infection and influenza vaccine. Wording was kept general so parents could answer for either seasonal or pandemic influenza (swine flu) as, by November 2009, pH1N1 was the predominant circulating strain in Australia. 22 Full methods for the questionnaire sampling, distribution and collation have been previously described. 23 The quantitative survey was conducted in 16 Sydney metropoli-

| Qualitative interviews
The first author (CK) conducted an additional 37 semi-structured interviews between 24 November 2009 and 24 May 2011, across an additional five childcare centres. Interviews continued until theoretical saturation of the topics was reached. Four centres were utilised for both survey distribution and interviews.
Each interview was recorded with participant consent and then transcribed word-for-word. Qualitative research software, NVIVO 10, was used by the first author (CK) to assign codes to both the pilot and subsequent interviews. Using a thematic analysis approach informed by elements of grounded theory, 24 interviews were coded initially by the first author using a line-by-line methodology.
Initial analysis was completed soon after each set of interviews.
Subsequent coding phases compared initially coded items with new interview data and examined the relationships between emerging themes. All co-authors analysed a subset of the interviews to compare, refine and finalise themes.

| Quantitative survey
There were 431 completed surveys from 972 distributed, a response rate of 44%. Demographic details of participants have been previously reported in full and found to be more highly educated than the general population (in which the rate for a university qualification is 24%) 23 ; in contrast, participants in our study were predominantly highly educated (postgraduate qualification 48%; undergraduate 27%) mothers (90%) aged between 31 and 40 years (70%). A chisquared analysis revealed no significant associations between demographic variables and information sources, with the one exception of parental education level and trust in natural therapists (defined as Complementary and Alternative Medicine [CAM] practitioners, which include naturopaths, homeopaths and herbal medicine practitioners; χ 2 = 5.58, df = 1, P = .02). Parents with a university education were less likely to trust their natural therapist, compared with parents without a university education (61% vs 76%, OR 0.5, 95% CI 0.3-0.9).
Parents reported that people they "trusted a lot" with regard to influenza information included their doctor (90%), nurses (59%), government (56%) and childcare centres (52%). The media was only "trusted a lot" by 7% of participants. Celebrities and antivaccination groups were not well trusted. These results are more fully explored in Figure 1.
In relation to parental search terms, 384/431 (89%) participants provided responses on how they would use a search engine to find information on "influenza infection" and 366/431 (85%) participants provided responses on how they would search for information on "influenza vaccine." The 10 most commonly used individual parental search terms for "influenza infection" and "influenza vaccine" are available in Tables 2 and 3, respectively.
Of note is the strong preferential use of "flu" rather than "influenza" by parents, despite the potential priming effect of using "influenza" in the questions. The main synonyms mentioned by parents for vaccine included "shot," "jab," "needle" and "inoculation." Also interesting was the use by some parents of geographic limiters, for example, Australia or Sydney, suggesting a preference by some for local information. When listing terms for "influenza infection," a few parents used the colloquial term "bug" and one parent noted "I would not Google this, who has the time?". When listing terms for "influenza vaccine," there were a few sophisticated responses including a search for "official trials" and "put the name of the vaccine [and] search in PubMed."

| Qualitative interviews
Including the pilot interviews, 42 interviews were conducted with parents (41 mothers and one father). The majority of the parents were between 31 and 40 years of age (n = 33), Australian born (n = 26), university educated (n = 28) and working outside the home (n = 28). Demographic details have been published previously. 25 To preserve confidentiality, pseudonyms were assigned to all participants for reporting purposes. Key emergent themes in relation to parental experiences of the pandemic are discussed below.

| Action trigger
The majority of parents thought they had not received enough in-

| In an emergency, think Emergency
Parents trusted general practitioners (GPs), but in the pandemic con- This perception that hospitals were the best place to take a child with pH1N1 caused one parent to deliberately avoid taking her young child to hospital during a prolonged vomiting illness as she was concerned the child would be exposed to pH1N1.

| Fright to hype
Parents reported high media usage and reliance throughout the pan- it just seemed to fizzle out, and … as the months went by, it didn't seem to be turning out as they expected it to, so yes, it didn't worry me that much after that.
Indeed, the mildness of the pandemic compared to initial predictions, the intensity of the media coverage and the lack of exposure to pH1N1 by them or their family or friends led to a sense of disillusionment and loss of trust in many parents. As recalled by Diane in

there was all this hype and panic about it, but then it was like 'oh we don't actually know how bad it is, but let's freak out anyway'.
A small number of parents expressed being either unconcerned or disinterested in the H1N1 pandemic. This was often proximity related; for example, a lack of concern was linked to a lack of personal or immediate contact exposure to pH1N1.

| Dr Google and beyond
While the survey indicated that the mass media ranked low on  were not sufficiently getting through to the public. 27 The Australian government did provide information at key points throughout the pandemic period, 7 yet the overwhelming perception of parents in this study was that, apart from initial messages about hygiene measures, little other information was forthcoming. Another

This mixed methods study provides insights into
Australian study conducted early in the pandemic in May 2009 revealed that 44% of those interviewed felt they did not have enough information about the pandemic. 28 An analysis of television coverage of the pandemic in Australia between 25 April and 9 October 2009 found that information provided included discussion of the potential seriousness of pH1N1, the changing alert level and infection rates. This supports the findings from our qualitative study where parents recalled receipt of infection control messages but then little else.
Our study found that parents increasingly reported feeling that pandemic risks had been exaggerated and/or sensationalised in the media. This finding is supported by studies in both Australian and international contexts. [30][31][32] This increased cynicism could be due to the eventual mildness of the pandemic and therefore the perception of conflicting messages, in combination with the lack of personal experience with or exposure to pH1N1. 18

| Limitations
Limitations of this study include the questionnaire response rate of 44%. As previously described, this is within the response rate range of other parental vaccine surveys. 23 A further potential limitation is that the questionnaire did not specify seasonal or pH1N1 influenza.
While this was a deliberate choice given that pH1N1 was the dominant strain by the time of the questionnaires, it may have resulted in some parental confusion, and it is difficult to know whether parents were definitively answering for seasonal or pH1N1 influenza.
Another limitation was limited generalisability due to selection bias. Many of the participants in both the questionnaire and interviews were highly educated, which may have influenced responses, including the specificity of the search methods they used. Social desirability bias in which idealised answers are provided could also have impacted on the results. This is less likely in a self-administered questionnaire compared to a researcheradministered questionnaire. 37

| Recommendations for public health
The results of this study have implications for pandemic preparedness; in particular, parental preference for presentation to hospital rather than GPs has the potential to overwhelm hospital resources in These suggest the need for clear, carefully crafted and tailored messages with a key role for health professionals. 38,39 Trusted sources such as doctors, government health department representatives and researchers could be utilised in both traditional media spaces and in non-conventional settings such as on popular programs. Providing and promoting a hotline staffed by trusted sources such as doctors and nurses could assist in disseminating advice to guide appropriate presentation at Emergency Departments.
In addition, factsheets developed by trusted sources (including hospitals) could be available physically in GP surgeries, hospitals, schools and childcare centres, and on websites.
Resources for use during a pandemic should take into account the preferred search terminology expressed by parents, for example, the use of the more informal "flu" rather than "influenza." To optimise search engine retrieval, metadata underpinning resources could use this as a variant term so that resources can be effectively located by parental Internet searches. 40

| CON CLUS ION
Understanding and considering the range of views, information needs, and preferences for searching and sources expressed by parents during the pandemic period provides useful context for developing tailored information materials and messages. Using and further promoting trusted sources via the media, as well as using existing trusted sources such as childcare centres and schools, could assist in disseminating public health messages in the event of future pandemics.

CO M PE TI N G I NTER E S TS
The authors have no competing interests.

E TH I C A L A PPROVA L
The study was granted ethics approval from the Human Research Ethics Committee of The Children's Hospital at Westmead, Australia.
Informed consent was obtained for participation in the study.

ACK N OWLED G EM ENTS
We gratefully acknowledge the Paediatric Influenza Vaccine Outcome Trial team, the participating families and KU Children's Services staff for their assistance with this project. We also wish to thank Donna Armstrong for editing support.