This study suggests that backyard flocks are no exception to avian influenza exposure and that Maryland flocks may have been exposed to AI from wild birds or pests. Pests are defined as both mammals and invertebrates. AI vaccination was ruled out based on survey data, as all owners denied vaccinating flocks once on the premises. AI vaccination practices are also rare in the U.S. and require USDA licensure and approval from both state and federal governments prior to field deployment 
. To date, only a handful of studies based in industrialized countries have evaluated the seroprevalence of avian influenza in unvaccinated backyard flocks. While one study in New Zealand found a flock seroprevalence of 20.8% (5/24), comparable to 23.1% (9/39) in this study, a Minnesota team only detected one flock out of 150 (0.66%) for AI antibodies 
. In Switzerland, researchers reported a higher seroprevalence of AI at 37.5% (15/40) in fancy breeding flocks 
. However, many variables contribute to sample prevalence rates such as testing method, time of year, climate differences, migratory trends, species and age of waterfowl, and backyard flock exposure and management practices.
Earlier studies focusing on the Delaware Bay and Maryland's Eastern shore have shown the prevalence of AI reservoir species ranging from May to November. The Delaware Bay has been identified as a “hotspot” for AIV prevalence, from May to June, in shore birds, particularly the ruddy turnstone, however, the surveying time period excludes this population. Migratory waterfowl also travel up the Atlantic Flyway and arrive late July through October with peak AIV prevalence detected in August 
. A study on the Eastern Shore of Maryland sampled cloacal swabs from resident ducks for 3 weeks between May 28 and Sept 2, 1998. Results suggested that influenza A viruses were introduced or increased in prevalence in resident waterfowl between July 15 and Aug 27 as AIV positives were detected from August 27 to September 2 at a prevalence of 13.9% 
While no AI RNA was detected in backyard poultry flocks, serological analysis indicated that almost a quarter of flocks had been previously exposed. Detection of antibodies against AI also allowed for screening of poultry that were infected prior to the sampling period. Detectable levels of antibodies against AI appear one to two weeks after infection and can last for several months 
. Sera positive for antibodies were also screened for hemagglutinin (HA) subtypes H5, H7, and H9 which are thought to have the greatest pandemic potential by the World Health Organization as they, although rare, are transmissible from birds to humans 
. However, these HA subtype specific antibodies were not found in this study which is consistent with other publication findings. Previous influenza surveillance studies conducted in Maryland waterfowl have reported the presence of HA subtypes H2, H3, H6, H9, H11, and H12, whereas the majority of North American subtypes consist of H3, H4, and H6 
It is believed that all of the AI seropositive chickens identified in this study were exposed to LPAI viruses as the birds survived the infection and owners did not report any significant mortalities in their flocks as a result of disease. The majority of circulating strains are low pathogenic viruses which may produce subtle or no signs of clinical infection to mild respiratory distress. Other signs may include diarrhea, decrease in egg production, and inactivity. However, these signs are not specific to AI infection and are often present in other poultry diseases 
. Almost half of owners (46%) with an AI positive test observed diarrhea in their flock within the past six months. One third of AI seropositive flock owners reported a decrease in egg production or soft/misshapen eggs in the previous six months and only one AI seropositive flock exhibited coughing, sneezing, nasal secretions, or swollen sinuses. Another indication that flocks may have been exposed to LPAI viruses was the negative HI assay result for H5 and H7 influenza subtypes, which are the exclusive subtypes associated with naturally occurring virulent isolates 
. The lack of a secure housing environment and location near water sources, which serve as a congregation point for wild birds, waterfowl, and pests, increases the likelihood of disease transmission. These potential risks associated with disease reservoirs and vectors are similar with findings from other studies. For example, wild birds most frequently reported visiting poultry houses were sparrows and European starlings, both of which are susceptible to experimental highly pathogenic H5N1 infection and excrete high viral titers 
. Another study conducted in an artificial barnyard setting found that mallards recently infected with H5N2 and H7N3 could transmit influenza A virus to chickens, blackbirds, rats, and pigeons demonstrating the potential for disease to spread by wild birds and pests 
. All owners of AI seroconverted flocks, as well as most AI seronegative flocks, also allowed visitors onto their poultry premises. A higher volume of traffic on the premises potentially increases the risk of introducing disease via fomites as visitors' vehicles, boots, and clothing may carry pathogens. Several outbreak investigations have linked fomites in connection with disease spread, such as the 1983 HPAI H5N2 outbreak in Pennsylvania and Virginia commercial poultry which was associated with human and equipment traffic from New York live bird markets 
To the authors' knowledge, this is the first study to report associations between biosecurity management practices and disease prevalence/seroprevalence of AI among backyard flocks located within close proximity to the Delmarva commercial poultry region. However, this study was subject to some limitations. The overall response rate of this study (4.1%) was relatively poor, but believed to stem from the concern over the mandatory reporting of flock positives to the State Veterinarian and potential repercussions, such as “Hold Orders” that restrict the movement of birds onto or off the premises, as well as the stigma attached to having an infectious disease. A larger sample size may have also increased the ability of this study to detect significant associations between biosecurity risk factors and disease prevalence. While association could be hypothesized based on proportional analysis, wide confidence intervals indicate that these estimates have low precision from an inadequate sample size and therefore associated risk results should be interpreted cautiously in this preliminary study. Although methods of convenience sampling are often assumed to be representative of a population, sampling biases (most notably selection bias) do occur, making it difficult to develop statistically valid estimates of disease prevalence, regardless of how many birds are sampled. Another constraint was the lack of detail collected in the wild bird-domestic poultry interface such as type of wild bird/waterfowl species identified on the property as well as the means of exposure (i.e. nose to nose, adjacent habitat, droppings only) which may have provided greater insight to the exposure risk and should be included in future studies. Widening the sample collection time frame from May to October could have improved the chances of obtaining a more representative data set in relation to the transmission of AI from wild birds to poultry. This study was also limited to a population of backyard flock owners that had registered with the MDA. It is believed that AI prevalence estimates reported in this study are lower than the true population as most owners with clinically ill birds would be reluctant to participate. Due to the low response rate and potential biases, this study cannot be generalized to other backyard flock populations.
Surveillance is a dynamic process that requires continuous observation, collection, and analysis of data in order to identify the presence of a disease and contain its spread. While migratory waterfowl have been the main target of disease investigations, domesticated poultry warrant consideration as well. This surveillance study aimed to capture the prevalence and seroprevalence of AI during an outbreak-free period and to illustrate baseline levels of exposure in this growing population. As a result, data from this project has provided a better understanding of AI ecology and transmission relationships within backyard flocks. As demonstrated in this study, education is essential for backyard flock owners especially with non-commercial poultry ownership's recent increase in popularity. Several flock owners did not practice biosecurity methods, many of which are simple, practical, and affordable. Therefore, it is recommended that proactive biosecurity education highlight prevention measures such as protecting poultry from wild birds and waterfowl particularly during the spring and summer months when migration season is at its peak and implementing a pest control plan. Targeted education and surveillance strategies will help protect the health of U.S. poultry flocks, minimize economic effects of the disease, and greatly reduce the health risks to the U.S. public.