Prioritizing live bird markets at risk of avian influenza H5N1 virus contamination for intervention: a simple tool for low resource settings.

Live bird markets (LBMs) are at risk of contamination with the avian influenza H5N1 virus. There are a number of methods for prioritizing LBMs for intervention to curb the risk of contamination. Selecting a method depends on diagnostic objective and disease prevalence. In a low resource setting, options for prioritization are constricted by the cost of and resources available for tool development and administration, as well as the resources available for intervention. In this setting, tools can be developed using previously collected data on risk factors for contamination, and translated into prediction equations, including decision trees (DTs). DTs are a graphical type of classifier that combine simple questions about the data in an intuitive way. DTs can be used to develop tools tailored to different diagnostic objectives. To demonstrate the utility of this method, risk factor data arising from a previous cross-sectional study in 83 LBMs in Indonesia were used to construct DTs. A DT with high specificity was selected for the initial stage of an LBM intervention campaign in which authorities aim to focus intervention resources on a small set of LBMs that are at near-certain risk of contamination. Another DT with high sensitivity was selected for later stages in an intervention campaign in which authorities aim to detect and prioritize all LBMs with the risk factors for virus contamination. The best specific DT achieved specificity of 77% and the best sensitive DT achieved sensitivity of 90%. The specific DT had two variables: the size of the duck population in the LBM and the human population density in the LBM's district. The sensitive DT had three variables: LBM location, whether solid waste was removed from the LBM daily and whether the LBM was zoned to separate the bird holding, slaughtering and sale areas. High specificity or sensitivity will be preferred by authorities depending on the stage of the intervention campaign. The study demonstrates that simple tools utilizing DTs can be developed to prioritize LBMs for intervention to control H5N1-virus. DT tools are simple to apply, suitable for low-resource settings and can be tailored to the particular needs and stage of the disease control program.

Environmental sampling for avian influenza virus A (H5N1) in live-bird markets, Indonesia.

To identify environmental sites commonly contaminated by avian influenza virus A (H5N1) in live-bird markets in Indonesia, we investigated 83 markets in 3 provinces in Indonesia. At each market, samples were collected from up to 27 poultry-related sites to assess the extent of contamination. Samples were tested by using real-time reverse transcription-PCR and virus isolation. A questionnaire was used to ascertain types of birds in the market, general infrastructure, and work practices. Thirty-nine (47%) markets showed contamination with avian influenza virus in ≥ 1 of the sites sampled. Risk factors were slaughtering birds in the market and being located in West Java province. Protective factors included daily removal of waste and zoning that segregated poultry-related work flow areas. These results can aid in the design of evidence-based programs concerning environmental sanitation, food safety, and surveillance to reduce the risk for avian influenza virus A (H5N1) transmission in live-bird markets.