The usefulness of the computerized medical records of one practice for research into pregnancy loss in dairy cows

Abstract

The use of routinely collected data for epidemiological research has been advocated to reduce the cost of field research. The disadvantage of such an approach is the lack of control over the collection and quality of the data. This study was undertaken to describe the quality of practice-generated computerized medical records and to assess their usefulness as a source of data for research. In particular, this study evaluated the quality of the medical records generated by one software program for research into pregnancy loss in dairy cows. The records of approximately 4000 cows in 70 herds in one private dairy practice were analyzed. Various methods were used to assess data quality. The task of preparing the medical records for analysis was complex. However, the programs written for this study could be adapted for use elsewhere allowing the generation of similar data-sets from other practices. The quality of the data available from these medical records was acceptable. More information is needed to set standards for quality in data-sets such as this. The observed tendency of cows with complete data to have worse reproductive performance has divergent implications for practising veterinarians. More work is needed to refine this association.