Potential of routinely collected farm data to monitor dairy cattle welfare.

In the EU Welfare Quality® (WQ) project, validated measures have been developed to assess the welfare status of dairy farms. The on-farm welfare assessment, however, is a time consuming and therefore expensive activity. Routinely collected data of various farm variables relating to health, fertility, productivity and product quality, may have potential to signal impaired animal welfare on dairy farms. To determine this potential, an analysis of the associations between such farm variables and the thirty-seven variables that are part of the WQ dairy cattle welfare assessment is required. Three WQ variables, i.e. somatic cell count, mortality and culling rate, are already routinely collected as part of regular farm records in the Netherlands. The other WQ variables are not collected as such, but it was expected that they could be associated with farm variables. The objective of this study was to explore associations between farm variables and WQ variables in a literature review. Literature showed that nearly all WQ variables are associated with farm variables. Often, single WQ variables are related to multiple farm variables. For example, the WQ variable prevalence of lameness can be associated with the farm variables milk yield, fat and protein contents of the milk, calving to first service interval, non-return rate, number of services, average age of the herd, and culling rate. Inversely, single farm variables are associated with multiple welfare variables. Similar association patterns were found for other variables and will be presented. It is concluded that routinely collected farm data show potential to signal impaired dairy cattle welfare. Due to the multifactorial nature of dairy cattle welfare, intelligent combinations of farm variables are most promising to detect farms at risk. In a next step, associations will be verified in a field study.