Good Practice of Secondary Data Analysis (GPS): guidelines and recommendations

Abstract

In 2005, the Working Group for the Survey and Utilisation of Secondary Data (AGENS) of the German Society for Social Medicine and Prevention (DGSMP) and the German Society for Epidemiology (DGEpi) first published "Good Practice in Secondary Data Analysis (GPS)" formulating a standard for conducting secondary data analyses. GPS is intended as a guide for planning and conducting analyses and can provide a basis for contracts between data owners. The domain of these guidelines does not only include data routinely gathered by statutory health insurance funds and further statutory social insurance funds, but all forms of secondary data. The 11 guidelines range from ethical principles and study planning through quality assurance measures and data preparation to data privacy, contractual conditions and responsible communication of analytical results. They are complemented by explanations and practical assistance in the form of recommendations. GPS targets all persons directing their attention to secondary data, their analysis and interpretation from a scientific point of view and by employing scientific methods. This includes data owners. Furthermore, GPS is suitable to assess scientific publications regarding their quality by authors, referees and readers. In 2008, the first version of GPS was evaluated and revised by members of AGENS and the Epidemiological Methods Working Group of DGEpi, DGSMP and GMDS including other epidemiological experts and had then been accredited as implementation regulations of Good Epidemiological Practice (GEP). Since 2012, this third version of GPS is on hand and available for downloading from the DGEpi website at no charge. Especially linguistic specifications have been integrated into the current revision; its internal consistency was increased. With regards to contents, further recommendations concerning the guideline on data privacy have been added. On the basis of future developments in science and data privacy, further revisions will follow.