ABSTRACT
The use of historical data in clinical trials has attracted interest in recent years, although its origin is much older. Incorporation of historical controls was originally proposed in the 1970s [6, 47], continued throughout the 1980s [49, 67], and has become increasingly important ever since [9, 22, 36, 43, 46, 55, 58, 68]. Concurrently, substantial methodological work was developed for incorporation of historical data in non-clinical toxicology studies [23, 28, 31, 51, 59, 60, 72]. However, the topic would likely have remained a niche area if there were not several fundamental changes happening in technology, statistics, and in the medical community during the 1990s.
