One method of screening for signals for drugs in the United Kingdom that are under intensive monitoring (''Black Triangle'' drugs) is to use both the PRR and the chi-square statistic. A cut-off for each can be used; for example, a PRR > 2 and chi-squared > 4 and the number of reports > 2. When this method is first used with these criteria on an existing database, many of the signals generated will already be known problems. In the first usage at the UK MCA slightly more than 60% were known; e.g. uveitis with rifabutin. About 15% were not believed to be caused by the drug but were events—effects of disease or a function of the patient population being treated, e.g. haemoptysis with dornase alpha. About 25% were new signals that required more detailed evaluation, e.g. renal failure with losartan.
In general, the method is used for continuing monitoring so that known problems will not
Was this article helpful?