Publication Abstract

Risk based surveillance (RBS) is undertaken to improve the efficiency of resource allocation. By focusing on high risk animals, farms and regions the likelihood of detecting disease or abnormal mortality is increased. Risk ranking farms or farming areas is fundamental to a RBS system. Cefas has developed a model to risk rank freshwater fish farms, based on the likelihood of disease introduction and spread, in part to fulfil the requirements of EC directive 2006/88 for RBS. To this end farms or farming areas have to be categorised as high, medium or low risk. A project identifying routes of introduction for marine non-native species collated data on depuration plants, the location of ports, marinas, bird colonies and movements of animals for aquaculture. It is important that resources used to develop RBS do not outweigh any efficiency savings. Thus existing datasets were used to develop an RBS for shellfish farms. It was judged most appropriate to risk rank at the farming area, as there are no physical barriers between farms, they are not epidemiologically isolated. Routes for pathogen introduction and spread were classified under 4 themes: i) live animal movements; ii) water, ii) birds and iv) vessels. More than one dataset were used for some themes, for example, under the water theme the number of depuration plants, shellfish holding facilities and proximity to other shellfish farming areas were used. Some data were transformed to minimise excess leverage by outlying values. Each route was scored, on a scale of 0 to 1. Themes, and risk factors within themes, were weighted using expert opinion. Separate scores for introduction and spread were calculated by adding the weighted scores for each theme. Results were plotted on a scattergram (Fig 1) and observed clustering was used to define thresholds between low, medium and high categories. The next step is to obtain weight estimates through a formal expert elicitation process (e.g. Delphi panel). The model can be further developed by ranking farms within farming areas and parameterising the model for specific pathogens. Outputs from hydrodynamic models would provide better estimates of spread by water.