Unity from diversity: the evidential use of anecdotal reports of adverse drug reactions and interactions.

Anecdotal case reports contribute about one-third of the published literature on adverse drug reactions and interactions, but are regarded as providing poor-quality evidence. However, they can occasionally provide proof of cause and effect, and there are many other reasons for publishing them. Because an anecdote is a narrative, narratological paradigms from literature, art, and music can show how we can make evidential use of anecdotes. Useful paradigms are the dramatic unities (of time, place, and action), comprehensive catalogues, and pattern formations. Here I give examples of each of these types of paradigm and show how they can be used to interpret anecdotes about adverse drug reactions and interactions. The dramatic unities show how a proper classification of adverse drug reactions can be achieved, according to dose-relation, time-course, and susceptibility factors; use of this classification should improve the evidential use of anecdotal reports. A high background incidence of the effect (the medical equivalent of subplots, which violate the unity of action) makes it more difficult to detect adverse drug effects using anecdotal reports. To make best evidential use of the corpus of anecdotal reports of adverse drug reactions, comprehensiveness is important: each suspected adverse reaction should be reported in detail and reactions should be reported in sufficient numbers for proper classification and for patterns to be recognized. One form of pattern recognition, teleoanalysis of data, should, when possible, include not only randomized controlled trials and observational studies, but also case series and anecdotal reports.