Evaluation of effects of metritis management in a complex dairy herd health management program.

Evaluating the effects of all interventions in a dairy herd, including the effects of various herd health management programs (HHMP), is highly relevant. A traditional randomized controlled trial is the gold standard but is likely practically impossible or prohibitively expensive to use for a general evaluation of a HHMP. Generalizability may also be poor because of the dynamics of the production contexts. In this study, we demonstrate an approach for evaluating the effects of an HHMP in the field, specifying an intervention theory for an ongoing HHMP in the context of the Danish dairy industry. As an example, we suggest one coherent analytical approach for studying the possible effects on milk production of systematic postpartum examinations of vaginal discharge, which is supposed to improve detection and treatment of metritis or endometritis. This routine is one component of the HHMP. The data consisted of 121 herds and 76,953 lactations over a 15-yr period. For parity group 1, the negative effects of metritis (despite treatment) on 305-d milk production after a normal calving were reduced by 116 kg of energy-corrected milk after enrollment in the HHMP. For parity group 2 and parity group >2, enrollment in the HHMP resulted in a 129-kg and an 80-kg energy-corrected milk yield increase in milk production, respectively. The results indicate that effects of the HHMP existed, which were mediated through improved metritis detection. This study demonstrates the importance of a clear-cut intervention theory, although even with a theory, the research question can be too herd and context specific. In such a case, a within-herd randomized controlled trial study design seems to be the only way to achieve a valid result for a given herd, and acquiring valid results from an observational multi-herd study will be very difficult.