Whether surveillance is based on serological tests, the exact number of detected cases may remain unknown because identifying seroconversion can be difficult. Moreover, incomplete cases / outbreaks detection is a recurrent issue in surveillance. Our study addresses these two issues, regarding equine viral arteritis (EVA) surveillance. Our goals were to establish suitable rules for identifying seroconversion to estimate the number of EVA outbreaks detected by the French breeding stock surveillance system between 2006 and 2013, and to assess surveillance sensitivity by estimating the total number of outbreaks that occurred during this period using a capture-recapture model. Data from mares with at least one positive result using viral neutralization test between 2006 and 2013 were used for analysis (n=1,645). Data consisted of annual antibody titers and mares’ location (towns). Seroconversion was defined as a change in antibody titer from negative to at least 32 or a three-fold increase. The number of seroconversions was counted for each town and modeled using a zero-truncated binomial capture-recapture model with R software. From 2006 to 2013, 239 cases of seroconversion located in 177 towns (outbreaks) were identified. Total number of outbreaks in breeding stock was estimated at 215 (CrI95% 195249) and the surveillance sensitivity at the town level at 82% (CrI95% 71%-91%). The estimated surveillance sensitivity is relatively high. The proposed rules may be used to analyze other serological surveillance data, including testing before sales or international trade. This study shows how capture-recapture methods may help to estimate surveillance sensitivity and to valorize surveillance data. Introduction Equine viral arteritis (EVA) is an equine respiratory and reproductive disease which can lead to abortions and neonatal deaths. EVA is caused by a virus of the Arteriviridae family and is mainly transmitted horizontally by aerosols or venereal contact, including frozen semen (1, 2). EVA is monitored in many countries in breeding stock to avoid its spread during breeding activities (3). In France, the breeding stock surveillance (BSS) is mainly based on serological tests, but difficulties in interpreting certain series of results may impair the estimation of the number of outbreaks. Moreover, only a part of breeding horses are tested, depending on the studbooks’ regulations. The first objective of this study was to establish suitable rules for identifying seroconversion in order to estimate the number of EVA cases and outbreaks detected by the BSS between 2006 and 2013. The second goal was to estimate the sensitivity of the BSS, after having estimated the total number of out-breaks that occurred in breeding stock during this period (including undetected outbreaks) using a capture-recapture method.

Monitoring of vector-borne diseases : methods, strengths / weak point. Recent outbreaks of vector-borne diseases in France over the past decade have illustrated the need for efficient surveillance methods to monitor the onset and / or spread of these diseases. Epidemiological surveillance can detect the emergence of new or exotic diseases or monitor the epidemiological situation of an already established disease in an area to adapt control measures. Epidemiological surveillance can take many forms that can be combined to monitor a disease : Eventdriven surveillance (also called passive surveillance), programmed surveillance (or active surveillance) and syndromic surveillance. The specificities of the vector-borne diseases surveillance are based on the frequent complexity of their epidemiological cycles and the surveillance of the vectors. Entomological surveillance uses specific methods that are sometimes heavy and expensive to implement. It can track the geographical distribution of vectors, periods of vector activity, but due to a frequently low vector competence, usually cannot monitor diseases transmitted by these vectors.