As of July 1 2009, 436 confirmed human H5N1 avian influenza (AI) cases, resulting in 262 deaths (case fatality rate, 60.1%), have been reported since November 2003, most of them in southeast Asia [1]. There is general agreement on the potential spread to humans with devastating consequences [2]. Case-control studies during the 1997 outbreak in the Hong Kong Special Administrative Region (SAR) of China [3], Thailand [4], Vietnam [5] and China [6] during 2004-2005 revealed the visiting of poultry markets and direct contact with sick or dead poultry to be significant H5N1 risk factors.

In mainland China, 37 confirmed H5N1 cases have been reported since October 2005 [7]. The Chinese H5N1 cases comprised two different groups regarding patterns of exposure to poultry: 71.4% (10/14) of urban patients had visited wet poultry markets (live or freshly slaughtered poultry) before onset of the illness, whilst 81.8% (18/22) of rural patients had previously been exposed to backyard poultry and to sick or dead poultry [6,8].

To date, there are limited data on the knowledge, attitudes and practices (KAPs) associated with AI in the general population [9,10]. We conducted a two-stage, household-based cluster survey of KAPs regarding AI in an urban area (city of Shenzhen), where poultry markets are common, and in a rural area (Xiuning County), where backyard poultry is common. Focal points of the survey were: types of information commonly held, awareness of AI, emotional reactions to human AI, and practices associated with its prevention.

4,414 (98%) of the 4,504 eligible participants agreed to participate in the study and completed the questionnaire. Fifteen participants in Shenzhen and one in Xiuning were excluded from the analysis due to incomplete responses. Residents in the Shenzhen group were significantly younger than those in Xiuning (p < 0.001), with a median age of 32 years and 46 years, respectively. The level of education (high school or above) was much higher among the urban residents compared to the rural villagers: 760 (42%) vs. 199 (8%), respectively (p < 0.001). A similar difference was observed in occupation levels: 244 (13%) of participants were homemakers in Shenzhen vs. 39 (2%) in Xiuning (p < 0.001).

The most common source of information about human AI was television, with 94% of urban residents and 92% of rural residents (p = 0.04) identifying television as the main source of information on AI. Urban Shenzhen residents were more likely than rural Xining residents to be informed through newspapers (45% vs. 7%, p < 0.001) or the internet (21% vs. 2%, p < 0.001). More residents in the Xiuning group obtained information through health professionals (26% vs. 7%, p < 0.001) and family/friends (55% vs. 29%, p < 0.001).

A significant percentage of participants in both groups requested more information (62% for the Shenzhen group and 76% for the Xiuning group, p < 0.001). However, approximately a quarter of the population in both groups did not care about this problem (Shenzhen 28% vs. Xiuning 22%, p < 0.001). Official announcements of a poultry outbreak were less frequently demanded, especially by the Xiuning group (p < 0.001). Television was the preferred source of further information in 92% of the Shenzhen group and 95% of the Xiuning group (p < 0.001) (Table 1).

Almost three-of all participants were aware AI is an infectious disease that can be prevented; 40% of participants in the Shenzhen group vs. 31% in the Xiuning group (p < 0.001) were aware the infection could be transmitted from poultry; however, only 4% of urban and 6% of rural participants were aware that humans could fail to recover fully with treatment. Overall, a greater proportion of urban participants were considered to be knowledgeable about human infection with AI (Shenzhen group 69% vs. Xiuning group 56%; p < 0.001) (Table 2). A comparison of knowledge scores associated with AI for all interviewed respondents in Shenzhen city and Xiuning county found the same result; that urban residents were more knowledgeable about AI than the rural villagers (p < 0.001) (Table 3).

Significant single factors for AI knowledge scores including gender, age, education level, occupation, and requests for AI information were incorporated as independent variables into a multivariate analysis. The results showed, for urban residents, age, education level, occupation, and requests for AI information were possible influencing factors of AI knowledge. For rural villagers, gender, age, education level, occupation, and requests for AI information were possible influencing factors of AI knowledge (Table 4).

Overall, the rural group was significantly more concerned (p < 0.001) than the urban group about human infection with AI virus, as measured by concern about AI infection of family and friends, fear of visiting public places and catching the virus, and general concern about human infection with the virus. A comparison of attitude scores associated with AI for all interviewed respondents in Shenzhen city and Xiuning county also found that rural villagers were more concerned than urban residents (p < 0.001) (Table 5).

Significant single factors for AI attitude scores including gender, age, education level, occupation, requests for AI information, and knowledge scores were incorporated as independent variables into a multivariate analysis. The result showed that, for urban residents, gender, age, education level, occupation, requests for AI information, and knowledge scores were possible influencing factors of AI concern. For rural villagers, gender, education level, requests for AI information, and knowledge scores were possible influencing factors of AI concern (Table 6).

Among the practices used to prevent human AI transmission, 1,152 (64%) of urban participants used soap for hand washing vs. 965 (38%) among the rural group (p < 0.001). Other practices (use of detergent for dish washing, anti-bacterial solutions, washing powder and shampoo) were similarly, but rarely, adopted by the two groups (7% to <1%). Regarding eating habits, 45% (714/1,601) of Shenzhen participants reported eating less poultry since becoming aware of this infectious disease, whereas 33% (732/2,213) of participants in the Xiuning group reported eating less poultry (p < 0.001).

The majority of urban residents (1,693, 93%) never had direct contact with poultry or their cages when purchasing them. A significant majority of those who had contact with either poultry or cages never subsequently touched their mouth, nose, or eyes (92/112, 82%). Similarly, most rural residents stated they infrequently had direct contact with sick or dead poultry (2,409, 95%). The majority of those who had direct contact with sick or dead poultry washed their hands after contact (119/133, 89%).

Despite certain limitations in the study methodology, mainly due to the need to produce simple questions understandable to all groups, our study revealed a number of important differences between KAPs of urban and rural populations for AI. Multivariate analysis demonstrated that age and level of education were likely the main factors giving rise to these differences. The study showed a high degree of awareness of human AI in both urban and rural regions. Awareness evaluations in Thailand [12] and Cambodia [13] after their 2004-2005 AI outbreaks gave similarly high percentages.

Television was found to be the most effective way to disseminate information on AI in both groups, followed by newspapers in the urban group and family/friends in the rural community. A significant percentage of both groups requested additional information; television, again, being the preferred source. These results confirm findings in Thailand where television proved to be the most efficient source of information [12].

The percentage of participants with 'knowledge associated with human AI' was low, especially in the rural areas where education levels were lower. Olsen et al [12] also found very low percentages of basic knowledge of human AI in rural Thailand before public education campaigns. However, in Italy, Abbate et al [11] found 64% of 284 poultry workers correctly defined AI as a contagious infection caused by a virus that can affect all species of birds. Nearly all workers identified poultry and wild birds as common vectors. These percentages are much higher than our results here. A recently published Italian study showed that knowledge of AI was greater among more highly educated populations, with a similar percentage aware that AI was preventable [14].

The urban group in this survey was less concerned about human infection with AI than their rural counterparts, probably due to their greater awareness of AI, in part, resulting from their higher level of education. In the Hong Kong SAR, with its high level of education, a telephone survey found 36% of respondents agreed that buying live chickens was a health risk and only 9% estimated a >50% likelihood of resultant sickness [9]. In addition, a telephone survey conducted in Hong Kong SAR in 2005 [15], which evaluated perceptions related to human AI and its association with anticipated psychological and behavioral responses in an outbreak, revealed that only 19.9% of participants avoiding visiting hospitals due to fear of AI, and a general education level (42.4% was matriculated or above) higher than in our study population.

Multivariate analysis results suggested, for both populations in our study, a possible relationship between AI knowledge, some of the demographic characteristics, and requests for AI information existed, as well as a possible relationship between AI attitude, some of the demographic characteristics, knowledge scores, and requests for AI information. However, a quantitative relationship cannot be evaluated from the data in our present study, and such an issue could be explored by further investigations in the future.

Our study showed a higher level of proper hygienic practice among urban residents. A study by Olsen et al. in Thailand before the 2004 outbreak demonstrated similar results among rural residents [12]. Another study conducted in Italy among the general population showed low compliance with precautionary behavior [14]. All implied the practice of precautionary activities in avoiding infection by AI virus needs to be strengthened.

Both groups in the present study reported a higher proportion of eating habit changes after awareness of the spread of AI, especially the rural villagers, who had more concern about AI, comparison with a survey conducted by Lau et al [15]. Facing AI, a new, emerging infectious disease with a high fatality rate, it is expected that people will have varying degrees of concern, ranging from indifference to panic. Timely and comprehensive public-risk communications from the government or other professional agencies are necessary to appease the possible negative social psychological influences such an outbreak would bring, in addition to the importance of persuading the public to take appropriate attitudes towards their practices of disease control and prevention.

Despite the novelty and significance of these findings, one methodological consideration that ought to be highlighted when interpreting these results is that the scores of AI knowledge and attitudes were not validated in China.

Studies conducted in China, Thailand and Vietnam revealed that risk factors for H5N1 infection included recent exposure to live poultry, direct contact with dead poultry that had died of unknown causes [8], and the presence in the household of dead or sick poultry [3-6,9,10,13]. This study investigated the levels of knowledge, attitude and practices regarding these risk factors and could provide scientific support to assist the Chinese government in developing strategies and health education campaigns to prevent transmission of the AI virus among the general population. These campaigns should include such advice as avoidance of direct contact with sick or dead poultry, and use of protective equipment such as gloves and masks when contact is unavoidable. Such campaigns should utilize television as the primary medium of dissemination in all localities, with newspapers and the internet being the secondary source in urban areas, and local opinion leaders, such as family/friends and local doctors, taking a significant role in rural areas.

The authors declare that they have no competing interests.

HY contributed to study design; NX, YS, JW, and SZ participated in the field investigations, collected data from participants for the study, and helped to analyze the data; HY drafted the manuscript; JM helped to revise the manuscript. All other co-authors participated in collection and management of data. All authors read and approved the final manuscript.

The pre-publication history for this paper can be accessed here:

http://www.biomedcentral.com/1471-2334/10/34/prepub