[Herd-specific estimation of milk yield reduction due to recurrent clinical mastitis].

To decide about investments in preventive measures improving udder health it is important that the farmer knows the monetary loss due to mastitis on his specific farm. The Saxon dairy herd improvement association (Sächsischer LandeskontrolIverband e.V., LKV) is developing a computer program called "Daten- und Informationsplattform Tier (DIT)"which determines - among other analyses - the milk loss caused by mastitis for a specific herd based on individual cow's mastitis and daily milk yield data. The present article introduces the linear mixed model which is taken as a basis for calculating the reduction in performance through clinical mastitis. The short and long term decrease in daily milk yield is described by expanding the lactation curve model of Ali and Schaeffer (1987). For falculating the short term drop, the model includes the laps of time in days since the mastitis incident as secon-ddegree polynomial. The coefficients are estimaged specifically for the first respectively every following case of mastitis (class of episode). Classes of episode are also considered calculating the long term decrease by estimating lactation curves without mastitis as well as corresponding to the classes of episode. By integrating the statistic software R (R Development Core Team, 2012) into the processes of the DIT the estimation of the farm specific model parameters is largely automated on the servers of the LKV.Thereby, milk yield can be estimated for every day in milk according to episode number and laps of time since the incident or with no mastitis incident respectively for a particular period of time in a specific dairy herd. The loss resulting from reduced performance due to clinical mastitis is specified by adding up the differences and can serve as a valuable basis for management decisions.