Signs and symptoms are graded in much the same way. Three criteria are assessed;
• Pharmacological plausibility: are the signs and symptoms suggestive of a pharmacological effect of the drug (i.e. a type A reaction), that could be reproduced experimentally?
• Other causes: have other reasonable causes for the event been looked for and eliminated? By reasonable, one means most (90%?) of the usual causes for the disease. There has been much discussion on what reasonable means, and this is probably where the consensus conference criteria are most useful.
• Is there a laboratory test that is specific to the drug-reaction pair, and is it positive or negative? The criteria for specificity may vary. For example, if there were signs of toxicity, elevated or null plasma concentrations of a drug would qualify (within pharmacokinetic timeframes, of course). This would not apply for an allergic reaction, though null plasma concentration with sufficient sensitivity could perhaps qualify as a negative laboratory test if it effectively eliminates drug exposure within the appropriate timeframe.
Again, the results are fed into a three-way table (Table 16.3), resulting in a semiology grading from S1 (doubtful) to S3 (very suggestive). Most cases are S2 (non-specific reaction, no other reasonable cause, no specific laboratory test), or S1 (same but
Table 16.3. Semiological imputability.
Signs and symptoms
Very suggestive of drug involvement or interaction Compatible
Alternate non-drug explanation:
• Possible or present S3 S2 S1 S3 S1 S1
other causes not looked for usually because reaction to the drug is known, and all signs abated when the drug was stopped, before further investigations were made).
This method is not very precise, and is probably much less specific than other methods, and especially the Bayesian approaches. It has a number of merits, however:
* It is more of a triage method, and can be applied extremely rapidly in the vast majority of cases if there is the appropriate information.
* It is, in fact, extremely useful to ensure that the proper information on a case report is retrieved on an ongoing basis. Using the causality method on a routine basis helps tremendously in making sure all relevant information is retrieved when discussing a case with a reporter. In this it improves the quality of the data, and the later application of any causality method, be it the same with refined criteria, as would be used in an official investigation, or any other, since all methods rely on mostly the same information.
* Its use by all persons involved in the system facilitates communication, by the use of a common language. This was and remains indispensable in a network-based system, where harmonisation of practice is essential.
In conclusion, the French System is based on a number of specific items, which made its success, and should be included or at least understood in other systems that want to use a Regional Centre approach:
* The existence of a real network, where alert investigation is done in the Regional Centres, which are not only a glorified mailbox sending data into a distant administrative black box, never to see them again.
* The use of common procedures, to ensure quality of data, including the use of the causality method.
* The integration of the Centres in clinical pharmacology and in the local life of the regional hospitals, with emphasis on the information centre function, as a continuing resource for healthcare professionals: retrieving good cases is just a by-product of good information.
Was this article helpful?