Causality assessment of adverse drug reactions using decision support and informatics tools

Abstract

Abstract Informatics information organizing techniques have been applied to the causality assessment of adverse drug reactions since the 1970s. There has not appeared an easily applied and accurate method. The problem has to do with low quality and poorly organized input information or the attempted application of methods that are too rigid for some problems and inapplicable for others. Potentially the best current method, based on Bayesian probability, is difficult to use because of low quality or quantity of information. This handicap requires extrapolations which tend to result in a form of numerical global introspection. We suggest that the problem might be approached by using a generalized decision support algorithm which can be used with many different decision methods and proper informatics support for the immediate environment of the assessor in drug safety. The decision support algorithm we apply is based on an analytical hierarchical process algorithm implemented in the software CRITERIUM©. We give as an example the application of the algorithm to a suspected adverse drug reaction, cholestatic jaundice, as a result of treatment by chloropromazine. However, to support any decision‐making method, the information and informatics environment around the decision‐maker must be planned and developed. A set of requirements that best reflect adverse drug reaction causality assessment needs is discussed.