Location of live poultry wet markets (open food stall markets) in relation to population density, Hong Kong, China, June–November 2009.
In 2009, a cross-sectional survey of 360 poultry workers in Hong Kong, China, showed that workers had inadequate levels of avian influenza (H5N1) risk knowledge, preventive behavior, and outbreak preparedness. The main barriers to preventive practices were low perceived benefits and interference with work. Poultry workers require occupation-specific health promotion.
In 1997, a zoonosis in humans caused by a highly lethal strain of avian influenza virus (H5N1) was reported in Hong Kong. Live-poultry markets were the source of this outbreak (
Because members of the average household in Hong Kong shop in wet markets on a habitual basis, these markets are located in the most densely populated areas (
Location of live poultry wet markets (open food stall markets) in relation to population density, Hong Kong, China, June–November 2009.
Despite occupational risk for exposure to avian influenza (
An anonymous, cross-sectional survey was conducted during June–November 2009. Interviewers approached 132 licensed live-poultry retail businesses in wet markets and 23 wholesale establishments. The final sample was 360 poultry workers (194 retailers and 166 wholesalers; response rate 68.1%).
Respondents were asked about their demographics, past month’s work and preventive behavior, and avian influenza–related knowledge on the basis of a World Health Organization factsheet (
Summative scores were computed for avian influenza–related knowledge, current preventive behavior patterns, outbreak preparedness, and various perception domains. Higher scores reflected more beneficial levels of each domain. Unconditional multilevel regression indicated no evidence of clustering effect by poultry market. Standard multivariable linear regression was conducted by using SAS version 9.1.3 (SAS Institute, Cary, NC, USA) with knowledge, practice, and preparedness scores as outcomes and potential predictors showing p<0.25 in unadjusted analyses as input variables. Distribution of standardized residuals and their association with predicted values were examined to assess model assumptions.
Most (208, 60.1%) respondents were men 35–54 years of age, of whom 192 (55.3%) had worked a mean of 16.1 years in the poultry industry. Respondents showed low mean summative scores for knowledge of avian influenza (
A total of 242 (69.1%) workers reported that consuming undercooked poultry could transmit the virus, and 210 (59.7%) knew that infection could result from touching bird feces. For other transmission routes, awareness was lower, ranging from 14.0% (48) for eating undercooked eggs to 29.1% (102) for slaughtering poultry.
Ninety-six (27.4%) workers were unsure whether avian influenza virus (H5N1) infection had occurred in humans in Hong Kong, 198 (58%) incorrectly believed that nearly everyone survives this infection, and 110 (32.8%) incorrectly believed that a human vaccine for avian influenza was available. Most (208, 89.9%) respondents were familiar with influenza-like symptoms of avian influenza virus (H5N1) infection such as fever, but fewer workers were aware of respiratory and gastrointestinal symptoms of virus infection.
The Internet and other sources (e.g., health talks) of information about avian influenza were strong independent predictors of greater knowledge. However, training did not result in higher knowledge levels.
Poultry workers reported low-to-moderate levels of compliance with hand hygiene and other preventive measures (ranging from 7.3% [36] using eye protection to 65.2% [245] using handwashing with soap after slaughtering poultry). Working in the poultry industry ≥10 years, lower perceived barriers to preventive behavior, and retail poultry work were independent predictors of higher preventive behavior scores.
With regard to avian influenza–related perceptions, lack of training (277, 83.4%) and the view that compliance with all infection regulations is difficult during peak hours (218, 64.9%) were the most frequently cited barriers to adoption of preventive behavior. A total of 154 (46.4%) workers believed that face masks reduced business, and 153 (46.1%) believed that vaccination was expensive.
Low anxiety about illness was reported by 242 (76.6%) respondents. In the event of a local outbreak, workers expressed various levels of acceptance for precautionary actions, ranging from 15.8% (56) for reducing work hours to 82.4% (290) for seeking medical care for influenza-like symptoms. Ninety-six (27.4%) respondents anticipated taking oseltamivir. Greater perceived benefit of preventive behavior was the strongest independent predictor of higher preparedness scores (
Similar to other regions (
Higher levels of knowledge demonstrated by workers who accessed health information sources (e.g., Internet) that provide detailed information suggest that comprehensive, occupation-relevant information should be more widely accessible. However, occupational practices of animal workers might not be amenable to change solely on the basis of improvements in knowledge. Only 129 (42.1%) respondents reported that poultry workers could realistically adhere to all government guidelines (
Despite the ongoing government regulations regarding avian influenza in Hong Kong (
Implications of these findings extend to other poultry-borne pathogens, such as
In the spirit of the One Health Commission, which calls for an integrated, interdisciplinary approach to human–veterinary–environmental health challenges (
We thank the poultry workers for participating in the study and Terry Wong for assisting with preliminary data collection.
This study was supported by the Hong Kong Research Fund for the Control of Infectious Diseases.
Medline reports the first author should be "Olsen SJ" not "Olsen S" in reference 10 "Olsen, Laosiritaworn, Pattanasin, Prapasiri, Dowel, 2005".
| Item† | Value |
|---|---|
| Virus can infect | |
| Wild birds? [Y] / Live poultry [Y] / Domestic birds? [Y] | 65.4/60.1/44.8 |
| Humans? [Y] / Other animals? [Y] | 50.0/17.0 |
| Virus can be transmitted from 1 place to another on | |
| Bird feces? [Y] / Poultry cages? [Y] / Bird feed? [Y] / Clothing and shoes? [Y] | 82.7/26.8/16.9/11.9 |
| Humans can be infected by | |
| Eating behavior: eating poultry cooked well done? [N] / Eating runny eggs? [Y] | 69.1/14.0 |
| Infected bird contact: bird feces? [Y] / Breathing near birds? [Y] / Swimming with birds? [Y] | 59.7/22.1/19.4 |
| Occupational risks: slaughtering poultry? [Y] / Defeathering poultry? [Y] | 29.1/14.9 |
| Symptoms of infection in humans | |
| Dermatologic: Hair loss? [N] / Rash? [N] | 99.4/98.2 |
| Influenza-like: Fever? [Y] / Cough? [Y] / Muscle pain? [Y] / Runny nose? [Y] | 89.8/67.1/67.1/54.5 |
| Respiratory: Difficulty in breathing? [Y] / Crackling breathing sounds? [Y] | 59.2/18.8 |
| Other: Vomiting? [Y] / Diarrhea? [Y] / Nose bleeding? [Y] | 25.0/14.3/2.9 |
| There have been reported human cases | |
| Somewhere? [Y] / Asia? [Y] / People’s Republic of China? [Y] / Hong Kong? [Y] | 96.8/91.1/78.2/72.6 |
| 100% effective, commercially available vaccines exists | |
| For birds? [N] / For humans? [N] | 80.2/67.2 |
| Duration virus can survive outside body | |
| <6 h / 6h– 24h / several days / [ | 29.0/22.7/18.0/30.3 |
| Mortality rate for humans | |
| Almost everyone survives / <50% / [50%–89% die] / | 57.9/26.3/13.5/1.2/1.2 |
| Overall knowledge score, range 0–36, mean, SD | 6.7, 6.43 |
| Knowledge score multivariable linear regression model, β (95% CI), p value‡§ | |
| Educational level, primary or less = referent, F1–3, | 0.97 (0.07–1.87), 0.035 |
| Household monthly income >20,000 Hong Kong dollars | 1.61 (0.09–3.13), 0.038 |
| Received prevention information from the Internet | 4.35 (2.58–6.13), <0.0001 |
| Received prevention information from other sources¶ | 3.86 (1.10–6.62), 0.006 |
| In the past month, how frequently did you? Almost always / Sometimes? / Never? | |
| Handle live chickens with bare hands | 37.5/27.5/35.0 |
| Handle dead chickens with bare hands | 10.3/27.4/62.3 |
| Wear eye protection when handling chickens | 7.3/22.3/70.4 |
| Wear face mask when handling chickens | 25.3/35.2/39.5 |
| Wear PPE (e.g., apron, mask) when handling chickens | 51.2/22.4/26.4 |
| Sterilize your clothes | 52.9/31.8/15.3 |
| Wash hands with soap after killing chickens | 65.2/24.6/10.2 |
| Overall preventive practice score, mean ± SD (range)# | 8.16 ± 3.26 (0–14) |
| Practice score multivariable linear regression model, β (95% CI), p value‡** | |
|
| 1.45 (0.39–2.51), 0.010 |
| Retail shop worker; wholesale is referent | 1.11(0.08–2.14), 0.034 |
| Below median perceived barriers score; above median is referent | 1.44 (0.43–2.46), 0.006 |
*Values are % responding correctly unless otherwise indicated. Y, yes; N, no; PPE, personal protection equipment; CI, confidence interval. †Correct answers are indicated in [brackets]. ‡Variance inflation factors (VIF) diagnostics indicated no evidence of colinearity (all VIF <1.2) among variables in final models. Model fit analysis showed that standardized residuals of models were normally distributed and not associated with standardized predicted values. §Final model constant for knowledge score, α (95% CI) 13.70 (11.8–15.6). The following candidate covariates had the following β coefficients and p values before removal from knowledge score model: age, β = 0.25, p = 0.57; <10 years in poultry industry, β = 1.08, p = 0.20; newspaper information source, β = 0.79, p = 0.284; health workers information source, β = −1.06, p = 0.441; poster information source, β = −0.04, p = 0.650. ¶Other health information sources included health talks, seminars, school, radio, flyers, and other poultry workers. #Scored 2 = always, 1 = sometimes, 0 = never for computing summative score. Items 1 and 2 about chickens are reverse coded. **Final model constant for practice score, α (95% CI) 6.18 (5.08–7.29). The following candidate covariates had the following β coefficients and p values before removal from practice score model: monthly income >20,000 Hong Kong dollars, β = −0.036, p = 0.956; above median avian influenza (H5N1) susceptibility score, β = 0.171, p = 0.797; above median avian influenza (H5N1) perceived severity score, β = −0.965, p = 0.143.
| Item | Value |
|---|---|
| Perceived benefits of preventive measures | |
| Influenza vaccination for poultry | 69.8 |
| Handwashing with soap | 68.4 |
| Used gloves | 59.4 |
| Killed all live poultry in market by end of every day | 52.4 |
| Used N95 face masks | 38.4 |
| Two wet market rest days a month for cleaning | 38.0 |
| Made sure poultry are healthy before buying | 31.1 |
| Sterilized cutting boards and surfaces | 27.2 |
| Stayed | 19.2 |
| Took antiviral drugs | 14.4 |
| Used goggles | 10.1 |
| Perceived benefit summative score, mean ± SD (range) | 4.05 ± 2.33 (0–11) |
| Perceived severity | |
| Anxiety toward severity of symptoms: low/medium/high | 76.6/15.5/7.9 |
| Anxiety toward severity of infection: less than SARS/similar to SARS/more than SARS | 46.0/45.4/8.6 |
| Perceived severity summative score, mean ± SD (range) | 2.37 ± 1.42 (0–4) |
| Perceived susceptibility | |
| Government has sufficient measures to prevent infection in humans | 65.8 |
| I have immunity to avian influenza | 48.4 |
| Virus is transmitted from birds to humans | 32.7 |
| General public is susceptible to avian influenza | 15.8 |
| An epidemic will occur in Hong Kong | 14.7 |
| Poultry workers are highly susceptible to avian influenza | 13.9 |
| Perceived susceptibility summative score, mean ± SD (range) | 1.91 ± 1.19 (0–6) |
| Perceived self-efficacy | |
| I know how to protect myself from avian influenza | 82.4 |
| I can reduce the risk for transmission in the community | 76.6 |
| I am confident that I know how to handle infected poultry | 48.3 |
| Perceived self-efficacy summative score, mean ± SD (range) | 2.05 ± 0.93 (0–3) |
| Perceived cues to action | |
| Received prevention information from mass media | 93.3 |
| Public announcements are effective reminders of risk behavior | 61.2 |
| Exposed to worksite cues of action (health workers, posters, employer) | 41.7 |
| Cues to action summative score, mean ± SD (range) | 2.04 ± 0.75 (0–3) |
| Perceived barriers toward preventive measures | |
| Never received any infection control training | 83.4 |
| Following hygiene guidelines is difficult during peak hours | 64.9 |
| It is difficult to attend training on prevention | 57.6 |
| Wearing face masks when working will reduce business | 46.4 |
| Influenza vaccination is too costly | 46.1 |
| Wet market does not provide sufficient cleaning facilities | 35.3 |
| Influenza vaccination is inconvenient | 33.3 |
| Perceived barrier summative score, mean ± SD (range) | 3.69 ± 1.66 (0–7) |
| Preparedness | |
| Know who to contact for a suspected outbreak at work? | 71.1 |
| In the past year, have you been vaccinated for influenza? | 28.8 |
| In the event of a local outbreak in birds, are you likely to | |
| Increase sanitation measures at work | 79.7 |
| Wash hands more often | 72.6 |
| Accept influenza vaccination | 67.5 |
| Prevent customers from direct contact with birds | 62.4 |
| Get influenza vaccination | 62.2 |
| Wear a face mask during work | 57.3 |
| Wear more PPE during work | 30.8 |
| Stay away from chickens | 24.3 |
| Reduce work until condition improves | 15.8 |
| In the event of a small local human outbreak, will you | |
| See a doctor right away if you have symptoms | 82.4 |
| Wash hands more often | 68.5 |
| Get influenza vaccination | 62.2 |
| Wear a face mask during work | 59.4 |
| Wear a face mask in public | 38.9 |
| Take oseltamivir | 27.4 |
| Stay away from chickens | 24.1 |
| Quarantine yourself if you feel sick | 17.9 |
| Preparedness summative score, mean ± SD (range) | 9.22 ± 3.77 (0–18) |
| Preparedness score multivariable linear regression model, β (95% CI), p value† | |
| Above median perceived barriers score; above or equal to median is referent | 1.56 (0.64–2.47), 0.001 |
| Above or equal to median perceived susceptibility score; below median is referent | 0.98 (0.21–1.75), 0.013 |
| Above or equal to median perceived benefit score; below median is referent | 3.42 (2.61–4.22), <0.001 |
| Above or equal to median knowledge score; below median is referent | 1.26 (0.46–2.07), 0.002 |
*Values are % agree/yes unless otherwise indicated. Wet market, open food stall market; SARS, severe acute respiratory syndrome; PPE, personal protection equipment; CI, confidence interval.
†Variance inflation factors (VIF) diagnostics indicated no evidence of colinearity (VIF<1.2) among variables in final models. Model fit analysis showed that standardized residuals of models were normally distributed and not associated with standardized predicted values. Final model constant for preparedness score α (95% CI) 5.64 (4.49–6.80); not significant at p<0.05. The following candidate covariates had the following β coefficients and p values before removal from the final backward elimination model: cues to action above median, β = 0.101, p = 0.840; avian influenza (H5N1) training, β = 0.432, p = 0.502;
Dr Kim is a research scientist in the School of Public Health and Primary Care at The Chinese University of Hong Kong. Her research interests are biomedical and social epidemiology.