Key issues for consideration in the development of a data strategy: A review of the literature

This report, supported by funding from The Atlantic Philanthropies and completed in 2008, highlights key issues for consideration in the development of a data strategy. The Irish Government has stipulated that each Department should develop a formal data or statistics strategy as part of its information strategy. Chapter 1 highlights the purpose of a comprehensive data strategy. It is important that the value of data is understood within an organisation, both for the end users and for the production of data itself. In addition, there is a need for all stakeholders to understand the role of research and data within an organisation, and to this end three models of research utilisation are discussed – the problem-solving model, the interactive model and the dialogical model. Chapter 2 explores the management challenges of data systems, with an examination of data quality issues. The RADAR model, developed by the United Nations Statistics Commission, is proposed to aid in effectively developing and improving the performance of data or statistics strategies. It is necessary to consider, agree and implement the well-managed documentation of data flow within an organisation as part of a data strategy. The definition of ‘data quality’ and ways in which it can be measured are also discussed . Two existing data quality frameworks are presented, one from Canada and the other from New Zealand. The 2005 Data Quality Framework by the Canadian Institute for Health Information (CIHI) uses 5 dimensions of data quality: accuracy, timeliness, comparability, usability and relevance. The New Zealand Ministry of Health’s data quality framework is based on the CIHI model and aims to achieve consistent and accurate assessment of data quality in all national health data collections so as to enable improved decision-making and policy development in the health sector. Chapter 3 explores issues of data integration, which refers to a set of processes that allows organisations to access and locate fragmented data, as well as to aid in the creation of accurate and consistent views of the issue that needs to be addressed. There are three main approaches or models of data integration – data warehousing, metadata and data brokering. Each of these have specific advantages and challenges, highlighted within the discussion. Following on from this, a discussion of the legal issues concerned with data integration focuses on the issues of consent, anonymisation, data ownership, research, electronic health governance, liability issues and other systems issues. In terms of data protection, international examples are presented to illustrate the issues with specific regard to children. Finally, ethical issues of data integration are explored, concluding with a discussion on the challenges to the development and use of electronic health records. Chapter 4 illustrates three examples of data strategies from international sources. Data strategies from the Netherlands, Australia and the USA are explored to provide insight into the development of a data or statistics strategy. This report highlights that the development and production of a data strategy is both complex and multifaceted. The cooperation of multiple key players, as well as interdepartmental collaboration, is essential. There are statistical programs and companies available to be third-party members of a data strategy. Nevertheless, the development of a data strategy for children’s statistics raises many legal and ethical challenges that require detailed consideration.