Knowledge, attitudes, and practices of Hong Kong population towards human A/H7N9 influenza pandemic preparedness, China, 2014.

Background Since SARS epidemic in 2003, Hong Kong has experienced several major epidemic risks, but how general community might react to the repeated infectious diseases health risks have not been studied. In 2013, imported human H7N9 influenza infected cases from China were reported. Our study aims to assess the knowledge, attitude and practice (KAP) concerning A/H7N9 among Hong Kong general population regarding pandemic preparedness in early 2014. Methods A cross-sectional, population-based telephone survey study was conducted among the Cantonese-speaking population aged over 15 years in Hong Kong in February 2014. The study survey was composed of 78 KAP questions. Factors associated with individual and household pandemic preparedness were analyzed. Results Final study sample was 1,020 with a response rate of 45.9 %. Among the respondents, most of them believed personal hygiene and avoidance of avian contacts were effective in preventing H7N9 infections. The majority of respondents had satisfactory hand hygiene practices and avoided touching avian species but did not employ other preventive measures. Female, 25 years old or older, white collar workers, people with chronic diseases and people living in the city center tended to report better hygiene practices. The average State-Trait Anxiety Inventory score was 1.85, similar to that of the period during the first wave and at the start of the second wave of the H7N9 epidemic. Self-reported face masks wearing when having influenza-like illness in general population dropped from 92.4 % during H5N1 period in 2007 to 39.0 % in this study. Conclusion Hong Kong citizens show a low level of anxiety, misconceptions regarding the novel strains as well as gaps between perceived usefulness and practice of preventive measures towards influenza outbreaks. Educational campaigns and framing the issue to increase public and media awareness are crucial in preventing the current public fatigue towards outbreaks.


Background
The experience of SARS coronavirus epidemic in 2003 has radically changed the concept and catalyzed actions globally and locally for combating infectious diseases outbreaks at the community level [1,2]. Different health agencies globally are now striving to increase community resilience in combating this emerging infectious disease challenge. In particular, in the Hong Kong Special Administration Region of China, where a number of infectious diseases epidemic originated, resources were invested for enhancing infrastructure, policies and health education for pandemic preparedness and infection control. Whilst population resilience level in Hong Kong in handling infectious disease challenges are thought to be higher than other global cities, the mild clinical outcome of A/H1N1 influenza pandemic in 2009 and the limited transmissibility of the A/H5N1 avian influenza outbreaks in humans in 2007 may have modified citizens' attitude and perceived risk towards similar outbreak incidences [3][4][5], and may have led them to underestimate the emerging A/H7N9 outbreak at the community level. A study conducted between 2006 and 2010 suggested that prolonged warning of a future pandemic was likely to cause pandemic fatigue in the public [6,7]. For example, studies showed a decrease in self-reported mask use in case of influenza-like illness symptoms since the SARS in 2003 and post-SARS period [8]. There is currently limited literature to examine how prolonged warning may affect community perception and response to these large scale infectious disease risks.
In March 2013, a human infected H7N9 influenza A virus (A/H7N9) case was first identified in Eastern China and caught global attention [9,10]. Serious response level of the Preparedness Plan for Influenza Pandemic was activated in Hong Kong in Dec 2013 [11]. As of August 2014, a total of 450 confirmed cases were reported in various provinces in China [12]. A recent study in Hong Kong during the first wave of H7N9 epidemic in April 2013 showed low population anxiety levels [13]. Although current evidence of sustained human-to-human transmission for A/H7N9 is rare, a small cluster of infected cases within the family showed hospitalization might be required [14][15][16]. Confirmed case fatality rate was around 20 %, although the estimated symptomatic case fatality risk was lower [17]. At that time, there were a substantial number of confirmed cases of human infected cases of H7N9 avian influenza during the second wave of the epidemic (Fig. 1). Meanwhile, the infection rate of seasonal influenza was high in Hong Kong, as reflected by the Government Center for Health protection sentinel surveillance system [4]. Some personal behaviors had been suggested to be effective to control pandemics such as SARS and H5N1 in Hong Kong. Previous studies suggested that close contact with live poultry, visiting public places and the places having confirmed cases were risk factors, while frequent mask use, hand washing, disinfection, and avoidance of eating poultry were protective factors [13,18,19]. In this study, we conducted a representative, population-based telephone survey in Hong Kong 1) to investigate the current situation regarding the self-reported knowledge, attitude and practice of pandemic responses at individual level on the second wave human infected A/H7N9 avian influenza outbreak; and 2) to investigate factors associated with the self-reported personal hygiene practices.

Study design and study population
A cross-sectional, population-based telephone survey was conducted from 15-28 February 2014 in Hong Kong.
The study population was the population aged 15 years or above residing in Hong Kong, including residents holding valid work or study visas. Exclusion criteria included i) non-Cantonese-speaking respondents; ii) overseas visitors holding tourist visas to Hong Kong; iii) 2-way permit holders from mainland China; and iv) those who were unable to be interviewed due to medical reasons.

Instrument
A structured Chinese questionnaire was constructed and used for data collection. The questionnaire consisted of 78 closed ended questions that aimed to collect the following information from the respondents: seeking, and their current anxiety level concerning the outbreak (total 30 questions). The 6 questions short form of the State-Trait Anxiety Inventory (STAI) was used for measuring anxiety [20]. A five point Likert-type scale were used to ascertain level of agreement or disagreement for the questions (from 1 to 5, 1 = strongly disagree, 2 = disagree, 3 = uncertain, 4 = agree, 5 = strongly agree). 4. Practices of preventive measures against H7N9 influenza A infections, including personal hygiene practices, avoidance of contact with avian species, household preparedness and vaccination practice (total 18 questions). A four point Likert-type scale were used to ascertain level of the practices (from 1 to 4, 1 = always, 2 = often, 3 = sometimes, 4 = never). 5. Active personal hygiene practice was defined by practicing 6 or more of the following 9 personal hygiene measures, including wash more hands, wash hands with soap, do not share utensils, wear mask when sick, bring own utensils during meals, avoid going to public places and avoid using public transport, avoid contacting with live poultry, avoid eating poultry and avoid going to places that had H7N9 confirmed cases. Active personal hygiene attitude was defined by regarding 6 or more of the above 9 personal hygiene measures above as useful for prevention H7N9.
Each interview took approximate 15-25 min to complete. The survey questionnaire was pilot tested in January 2014 (n = 50) to ensure practicability, validity, and interpretability of answers. The questionnaire was slightly refined for wording and format before distribution to the study sample based on the results of the pilot study. Results about vaccination practice and channels for disease surveillance have been reported elsewhere and not included in the analysis.

Data collection
Randomly generated telephone numbers from a list of all land-based telephone numbers in Hong Kong was used as the sample frame [3,8]. The telephone interviews were conducted by trained interviewers. The telephone calls were made in the evening on weekdays (6 pm-10 pm) and during daytime and evening on weekends (10 am-10 pm) to prevent over representation of the unemployed population. The subjects undergoing the interview were chosen based on the "last birthday method" [21,22], in which the household member who was present in the household during the survey and whose last birthday was closest to the interview date was invited to participate. The subjects were invited on the basis of the proportion of age, gender and living district from the 2011 Hong Kong Population Census data. The sampling would continue until the quotas for each stratum were met. If the selected participant was busy or not there, up to 4 follow-up calls would be made. The numbers were called for a maximum of 5 times before being classified as unanswered.

Statistical analysis
Differences in proportions between demographic characteristics in this survey and the Hong Kong Population Census data in 2011 were examined. Descriptive statistics for knowledge, attitude and practice of H7N9 influenza A prevention in the study population were calculated. Differences in proportions between groups were tested using Pearson's Chi-square test or Fisher's exact test. Likert-type scale results were collapsed to binary outcomes for analysis. Cut off point for questions with 5 point scales were defined as >3 and for questions with 4 point scale as >2. Univariate analysis was first conducted to identify apparent associations between individual pandemic preparedness practices with sociodemographic factors, perceived health risk, knowledge and attitudes towards A/H7N9 influenza virus. Backward multivariable logistic regression analyses were then conducted to identify factors associated with actual pandemic preparedness. All variables were retained in the final model if they had P-value < 0.05. All statistical analyses were conducted using R (version 3.0.2, R Development Core Team 2011). Statistical significance was set at α = 0.05.

Research ethics
This study was approved by the Survey and Behavioural Research Ethics Committee of The Chinese University of Hong Kong. Oral consent was obtained from each of the participant at the beginning of the study. All collected data were anonymous.

Results
The final number of respondents who completed the survey was 1020, and the response rate was 45.9 % (1,020/2,221). Figure 2 detailed the data collection algorithm and response rate. Table 1 shows the sociodemographic characteristics of the study population compared with the general population in Hong Kong in 2011 [23].

Knowledge of A/H7N9 avian influenza
Of the respondents, only 25.3 % thought that they had sufficient knowledge to manage the risks that H7N9 avian influenza brings to their health and security. Majority of the respondents (62.5 %) mistook H7N9 influenza A virus as the common seasonal influenza of the current year and many respondents did not demonstrate accurate knowledge of known transmission routes of H7N9. Only 29.0 % of respondents correctly stated H7N9 can be transmitted via animals but not by insects and human feces. More than one fourth (27.5 %) thought insects can be vector for transmission and more than half (53.9 %) thought human feces was a possible transmission route. Meanwhile more than half thought those seasonal influenza transmission routes (human saliva 50.5 %, air 64.2 %, direct hand contact 59.9 %, indirect hand contact 55.0 %) can also transmit H7N9 avian influenza (Table 2).

Attitudinal determinants towards the impact of A/H7N9
Around half (51.8 %) of the respondents regarded Hong Kong as susceptible to infectious diseases outbreaks. The majority (86.0 %) believed that H7N9 would not acquire the ability of human-to-human transmission within the current year. Most (72.0 %) of them thought the chances of getting infected by H7N9 within the current year were low or very low. Only a small portion believed the H7N9 outbreak will cause high or very high impact for their health (28.9 %) and economic status (17.0 %). Only 37.8 % believed H7N9 will have high or very high impact affecting society. Respondents believed that H7N9 spread was controllable in general. 26.6 % of respondents did not believe the spread of H7N9 could be prevented at government policy level, and around half (22.6 %) believed that the spread could not be prevented at household and individual level.
The average anxiety level among the respondents towards the current infectious disease outbreak was 1.85 (IQR: 1.33, 2.33) on a scale of 1 ("not anxious at all") to 4 ("very anxious"), which was similar to the first wave of H7N9 in April 2013 and during the first imported case in Hong Kong in December 2013 [13].

Attitude and practice towards A/H7N9 personal hygiene prevention
Overall, most of the respondents thought that active personal hygiene measures reduced the risk of influenza transmission: Washing hands (87.2 %), using soap for washing hands (88.5 %), not sharing utensils (83.9 %), wearing mask when having respiratory infections (94.4 %), bringing their own utensils during meals (62.0 %), and to a lesser extent, avoiding public places and using public transport (58.4 %). They also had active attitude towards avoiding contact with live poultry (93.1 %), avoiding travel to previous human H7N9 infected areas (80.9 %) and to a lesser extent, avoiding eating avian species (56.0 %) ( Table 3).
While for the actual hygiene practice, most respondents did not practice well the above control measures as shown in Table 3. Not all measures in practice were dependent with the perceived usefulness, such as washing hands, wearing mask when sick and not sharing utensils during meals. Only hand hygiene (97.1 % and 72.6 %) and avoiding contact with live poultry (75.2 %) were widely practiced. More than half (61.0 %) did not wear or just occasionally wore masks when they had respiratory infections. More than half shared utensils (54.1 %), and very few (1.6 %) brought their own utensils during meals when outside home. Over 90 % would not avoid going to public places or using public transport. Around half (55.8 %) would still visit places that had confirmed H7N9 cases.

Factors affecting personal hygiene practice
Univariate analysis of different socio-demographic, attitude factors and active attitude towards preventive measures against H7N9 that were associated with active personal hygiene practice (defined as practicing 6 or more out of 9 measures above) was included in the multivariable analysis. In terms of attitudinal factors, better personal hygiene practices were associated with those who had active attitude towards personal hygiene practices (OR = 2.83, 95 % CI: 1.60-5.01, P < 0.001), those with higher anxiety level towards infectious disease outbreak

Discussion
We assessed the knowledge, attitude and actual practice of the Hong Kong population for pandemic preparedness during the second wave of human infected H7N9 epidemic in mainland China and when the seasonal influenza activity was high in the community. The result showed respondents still harbored misconceptions of H7N9 influenza, were generally not anxious about the H7N9 outbreak, and were not sufficiently practicing protective measures against infections.

Demographic patterns and knowledge gaps
Our study found several demographic determinants coincided with other studies in previous literature of respiratory epidemics and pandemics [24]. Female, 25 years old or older, having chronic disease, living in city center district and white collar workers were associated with better preventive behaviors. Similar to other previous influenza studies [25,26], respondents did not have detailed understanding of H7N9 avian influenza. A lot of them mixed up A/H7N9 with other types of influenza or other respiratory virus. Mixing up the characteristics may lead to inappropriate health behaviors, making the protection suboptimal or even risky. Although some precautionary measures are similar for preventing H7N9 influenza and other respiratory infections, specific precautions against avian types of influenza, in particular those avoiding contacts with live poultries and their feces should be emphasized to increase individual awareness for better informed protective measures.

Population anxiety and risk perception
Our study was conducted during the second wave of H7N9 epidemic, at the time, Hong Kong had 5 imported cases with high case fatality ratio, during the winter peak of seasonal influenza. However we found the population still did not take H7N9 seriously. This was reflected in their attitudinal determinants towards the impact of A/H7N9 and their low anxiety level (STAI of 1.85). The population anxiety level was similar to a recent study in Hong Kong during the first wave of H7N9 epidemic in April 2013, and did not increase comparing to the beginning of the second wave when the first case imported in Hong Kong in December 2013 [13]. The anxiety score was also similar to that during the first wave of H1N1 in 2009. In contrast, the anxiety score was much higher (STAI 2.50) during SARS [5]. 74.8 % of Hong Kong population would avoid going to crowded places, and 74.7 % and 71.8 % would avoid going to mainland China and hospitals respectively during SARS period in 2003 [27]. In current study, only 7 % would avoid going to public places and using public transport, and 55.8 % avoid going to places having H7N9 confirmed cases. The findings may indicate public fatigue in response to prolonged warning of influenza pandemic, with people reporting less adoption of infection control behaviours. Though low anxiety level is expected, the misconceptions, inactive attitudes and lack of preventive practices are not optimal for preparedness of future epidemics.

Gaps between perceived usefulness and practice of preventive measures
There are significant gaps between perceived usefulness and practice of mask wearing and do not share utensils during meals. The population showed high compliance with hand washing. 94 % of the Hong Kong population believed that hand washing was efficacious in preventing human-to-human avian influenza during H5N1 period [19] while 97.8 % perceived washing hands frequently as an effective public health measure during H1N1 period [8]. In the current study, this figure was 87.2 %. This might be due to public health announcements on mass media, emphasizing the importance of washing hands frequently. In contrast, for mask wearing and utensils sharing, the proportion of actual practices was far lower than perceived usefulness. 92.4 % of the Hong Kong residents wore face masks when having influenza-like illness during H5N1 period in 2007 [19] and 88.7 % in 2009 during H1N1 period [8]. In the current study, it dropped to 39.0 %. Barriers for taking action may be partly due to discomfort when wearing mask, traditional Chinese norm of sharing utensils when dining, and the inconvenience. Nevertheless, our study results showed the main reason hindering individual control behaviors were that people did not consider H7N9 outbreak a serious threat. Their attitude could be reflected by their low perceived severity and infectivity of the virus. Targeted interventions may focus on increasing the awareness of the population towards the outbreak according to current situations. And further analysis will be needed to Active personal hygiene attitude was defined by practicing 6 or more of the following 9 personal hygiene measures, including wash more hands, wash hands with soap, do not share utensils, wear mask when sick, bring own utensils during meals, avoid going to public places and avoid using public transport, avoid contacting with live poultry, avoid eating poultry and avoid going to places that had H7N9 confirmed cases b COR, crude odd ratio in the univariate analysis c AOR, Adjusted odd ratio in the multivariable analysis identify other specific barriers and to increase the selfefficacy in overcoming them, thus changing the health behavior for better protection.

Limitations
Limitations in the study included the methodological limitations of telephone survey. Firstly, households with no possession of a land-based telephone service may be missed. Nonetheless, the penetration rate of residential fixed line service in Hong Kong was 102.6 % in November 2013. The residential fixed line penetration is calculated by dividing the number of residential fixed lines by the number of households in Hong Kong [28]. Almost all households have at least one home based telephone service in Hong Kong. Secondly, our sample population had higher household income and education level than the Hong Kong general population. This may overestimate the overall results as these groups are more knowledgeable and have better practices for pandemic preparedness in general. Thirdly, there may be some reporting bias. Data were self-reported and data from non-respondents could not be obtained.
Our results may not be generalizable to other countries or cities that have not experienced previous severe epidemics such as SARS. This study presents as a unique reference for pandemic preparedness behavioral response during a second wave epidemic in an urban setting. External factors may influence the stability of the responses during the survey period. Nevertheless, we managed to finish the field data collection within a short period (two weeks) to produce a stable response.

Conclusion
Similar to previous studies, Hong Kong citizens show a low level of anxiety, misconceptions regarding the novel strains as well as gaps between perceived usefulness and practice of preventive measures towards influenza outbreaks. Educational campaigns and framing the issue to increase public and media awareness is crucial in preventing the current public fatigue towards outbreaks.

Availability of data and materials
Not applicable.

Competing interests
The authors declare that they have no competing interests.
Authors' contributions EYYC and CKYC participated in planning the study, data collection, and interpreting the results. EYYC, PYL, GT and ZH participated in data collection, interpreting the results and final paper writing. EYYC obtained research funding. EYYC and CKYC wrote the first draft of the article. All authors were involved in planning the article, critical review and editing of the first draft, and subsequent revisions to the paper. All authors read and approved the final manuscript. Active personal hygiene practice was defined by practicing 6 or more of the following 9 personal hygiene measures, including wash more hands, wash hands with soap, do not share utensils, wear mask when sick, bring own utensils during meals, avoid going to public places and avoid using public transport, avoid contacting with live poultry, avoid eating poultry and avoid going to places that had H7N9 confirmed cases b Active personal hygiene attitude was defined by regarding 6 or more of the above 9 personal hygiene measures above as useful for prevention H7N9 c COR, crude odd ratio in the univariate analysis d AOR, Adjusted odd ratio in the multivariable analysis