The country of Sweden has an unusually long tradition of population registration, dating back to 1749 when a specific governmental institution, “Tabellverket” [1] was founded for this specific purpose. Since then, several nationwide population-based disease or effect registers have been developed [2]. The drop-out rates are typically very low, usually under 5%. The registers are frequently utilized for research, evaluation, planning and other purposes by a variety of users [3]. Furthermore, there is a unique system of individual so-called personal registration numbers (PRN) for the identification of every Swedish resident [3]. This, together with availability of modern computer-based techniques, provides facilities for easy linkage of exposure and outcome data down to the level of individual patients [2]. Population-based administrative registers are practical and cost-effective when utilized for epidemiological purposes [4]. However, there are caveats when using such databases for scientific research [5]. Since the primary purpose of many such registers is administrative rather than scientific, this may reflect negatively on data quality [2].