Königsweg 63
14163 Berlin
+49 30 838 62676
gefluegelkrankheiten@vetmed.fu-berlin.de
We analyzed 2644 laying hens who survived whole observation (56 weeks) using RFID sensors and were subject to necropsy at the end of their laying period for the presence of some diagnosed infectious diseases. We calculated time series of: 1) the time that individuals spent at the lower feeder, upper feeder, in the nest boxes, and on the range area in minutes; 2) each potential contact of every 2 individual hens (animals as nodes) which are in the range of a given antenna with duration in minutes; 3) each movement of a single hen between two consecutively visited antennae being nodes. Using logistic regression models, we have tested the ability of mobility and contact patterns for prediction of Spotty Liver Disease (bacterial infection), Ascaridia galli and Cestodes infections (both parasites) as assessed by postmortem investigation. We also proposed a SEIR model for a hypothetical infectious disease with characteristics similar to common bacterial and viral infections. We applied social network analysis, representing the hens and antennas with nodes in the network, to map/predict/detect infectious diseases spread and within the aviary system. For a hypothetical disease model we have tested whether the point of introduction of the disease (by interaction with the wildlife vs. no adequate biosecurity adherence of workers) is manifested in variables measured in PLF context (i.e. laying productivity and hen mortality). From actual registered diseases, we have identified a possible exposure due to going outside of the shed with a 3.15 odds ratio per hour of ranging daily for Ascaridia galli infection chance (p<0.001). Up to our knowledge it is the first approach of modeling within-farm transmission network dynamics using empirical contact data (from sensors) in poultry (with AUC>0.65). We will discuss the ongoing experiments with computer vision for monitoring contact.