Epsilon Open Archive

Abstract

Nordic examples of the use of secondary small animal and equine databases include studies based on data from tumour registries, breeding registries, young horse quality contest results, competition data, insurance databases, clinic data, prescription data and hunting ability tests. In spite of this extensive use of secondary databases, integration between databases is less common. The aim of this presentation is to briefly review key papers that exemplify different ways of utilizing data from multiple sources, to highlight the benefits and limitations of the approaches, to discuss key issues/challenges that must be addressed when integrating data and to suggest future directions. Data from pedigree databases have been individually merged with competition data and young horse quality contest data, and true integration has also been done with canine insurance data and with equine clinical data. Data have also been merged on postal code level; i.e. insurance data were merged to a digitized map of Sweden and additional meteorological information added. In addition to all the data quality and validity issues inherent in the use of a single database, additional obstacles arise when combining information from several databases. Loss of individuals due to incorrect or mismatched identifying information can be considerable. If there are any possible biases affecting whether or not individuals can be properly linked, misinformation may result in a further reduction in power. Issues of confidentiality may be more difficult to address across multiple databases. For example, human identity information must be protected, but may be required to ensure valid merging of data. There is a great potential to better address complex issues of health and disease in companion animals and horses by integrating information across existing databases. The challenges outlined in this article should not preclude the ongoing pursuit of this approach.