A method for integrating animal welfare measures into an overall welfare index that reflects expert opinion

Aggregating multiple welfare indicators into an overall welfare index remains a challenge. We aimed to develop an integration system in which the main welfare issues are aggregated into a balanced, discriminative overall index that corresponds with expert opinion. This index is calculated as the average of a severity score (S) multiplied by the prevalence (P) for each welfare issue. This average can be corrected to reduce compensation of (extremely) bad scores on certain indicators (e.g. SxP exceeds a threshold). The development of the method is illustrated using data of Welfare Quality® assessments of 491 dairy cattle herds. Five major welfare problems were selected based on their importance for overall welfare as perceived by 15 trained users of the Welfare Quality® protocol. In addition to these experts, 8 ILVO dairy cattle researchers and 173 veterinary students (after having been explained how these welfare problems are scored) indicated on a continuous scale from 0-100 how severely each problem compromises cow welfare. The mean of these scores divided by 100 was used as the severity score (S) for each problem. Students and ILVO-researchers allocated an overall welfare score on a 100mm tagged visual analogue scale to 6 example farms for which the prevalence of each welfare problem relative to all other herds was marked on a graph. Correspondence between these scores and the overall index (calculated as the mean SxP) was checked. Using a newly developed app, the students simulated the effects of different compensation-reduction methods on the calculated overall index for all herds and on the correspondence with their own subjective overall welfare scores for the 6 example herds. Although currently limited to five welfare problems that can be expressed as a percentage, this integration method seems to hold great promise for calculating a discriminative welfare index that corresponds with expert opinion.