Editor's Choice | This Week in BMJ | Press releases
BMJ No 7107 Volume 315 Editorial Saturday 30 August 1997
Adverse drug reactions: finding the needle in the haystackPharmacovigilance is improving: now we need to ensure that patients benefitPatients need to be sure that the medicines they take are as safe and effective as possible. Concerns over a product's safe use must be discovered, evaluated, and acted on, and the results made available for patient care as expeditiously as possible. In this process spontaneous reports of adverse reactions play an important part, but the problem they present is that of finding needles - true adverse reactions - in large haystacks of suspicions. There is no substitute for spontaneous adverse drug reaction reports for providing early signals of problems with drugs. The article in this week's issue by Lee et al shows that reporting by pharmacists can make a difference to the number of meaningful reports from hospitals (p 519).(1) Britain thus joins 36 other countries in the World Health Organisation's programme on international drug monitoring that accept reports from pharmacists. Britain is already in the top eight countries for reporting, with a rate of >200 reports/million inhabitants,(2) so what is the advantage of more reports? There is relative under-reporting from British hospitals compared with general practice,(3) so there is scope for new important information since adverse reactions to new drugs and those used in special disease categories are most likely to be seen in hospitals. Also serious adverse reactions cause hospital admissions. Experience internationally with pharmacist reporting has been variable, both quantitatively and qualitatively.(2) The need for medical evaluation of a suspected adverse reaction can be a drawback because of the extra work, particularly if this necessitates correspondence after the first report is submitted (a problem that Lee et al avoided). At worst, the submission of more reports which are clinically unsubstantiated may simply add to the size of the haystack without providing any more needles. The dilemma of spontaneous reporting is that, to make as sure as possible that nothing is missed, we ask for all (serious) suspicions to be reported: this inevitably means a large haystack. Actual reporting requirements vary between countries and include direct reporting by patients in some countries.(2) The result is that national regulatory and pharmaceutical industry databases are crowded with associations between drugs and reactions that have little value in raising new general concerns. Information technology has now made it possible easily to share information in the different databases throughout the world, but this has also exacerbated the problems with duplicate reports. The problem of getting early and useful information out of this huge mass of data is one that has taxed members of the WHO monitoring programme since its inception. The Uppsala centre, which collects adverse reactions reported from all over the world, now has nearly 2 million records stored according to a format established by the WHO programme. The centre has developed a data mining tool based on Bayesian, mutual information logic within a neural network(4) that allows the strength of all drug-reaction associations to be quantified. Effectively the whole database is being used as the control, so that any new positive drug-reaction association highlighted implies a significant difference from the global reporting experience: in this case the size of the haystack becomes an advantage. Nevertheless, even a significant report is only a concern about a drug and a reaction: a direct causal relation remains unproved. Many other methods, such as toxicological testing and pharmacoepidemiology, may need to be used to determine the nature of the reaction, and reactions must be evaluated individually for their clinical importance before action is taken.(5) Finally, although we have made progress both in finding adverse drug reactions and analysing them, has this benefited patients? There remains a gap between the development of pharmacovigilance knowledge and its use in practice. Some drugs are too widely used when there are safer alternatives(6-7); others have been taken off the market when they still have a useful place(8); and we are some way from satisfactory, open, communication, particularly with patients,(9) on benefit and risk in therapeutic choices.(10) The gaze of pharmacovigilance professionals has been preoccupied by gathering complete information and developing a more certain science. We must also be sure that individual patients benefit as much as possible from the information they, the patients, give us. I Ralph Edwards Director
Uppsala Monitoring Centre, References 1 Lee A, Bateman D N, Edwards C, Smith J M, Rawlins M D. Reporting of adverse drug reactions by hospital pharmacists: pilot scheme. BMJ 1997;315:519. 2 Olsson S, ed. National pharmacovigilance systems. Uppsala: Uppsala Monitoring Centre, 1997. 3 Smith C C, Bennet P M, Pearce H M, Harrison P I, Reynold D J M, Aronson J K, et al. Adverse drug reactions in a hospital general medical unit meriting notification to the Committee on Safety of Medicines. Br J Clin Pharmacol 1996;42:425-9. 4 Lansner A, Holst A. A higher order Bayesian neural network with spiking units. International Journal of Neural Systems 1996;7:115-28. 5 Meyboom R H B, Egberts A C G, Edwards I R, Hekster Y A, de Koning F H P, Gribnau F W J. Principles of signal detection in pharmacovigilance. Drug Safety 1997; 6 Ju1:355-65. 6 Bateman D N. Tiaprofenic acid and cystitis. BMJ 1994;309:552-3. 7 Savage R, Lindquist M, Pettersson M, Edwards I R, Sanderson J H, Taylor N F A, et al. How does cystitis affect a comparative risk profile of tiaprofenic acid with other non-steroidal anti-inflammatory drugs? An international study based on spontaneous reports and drug usage data. Pharmacology and Toxicology 1997;80:211-7. 8 Brandt C, Beermann B, Hambn C, eds. Novalgin. Information från Läkemedelsverket 1995;6:437-39. 9 Herxheimer A. Side effects: freedom of information and the communication of doubt. In: Aronson J K, van Boxtel C J, eds. Side effects of drugs annual 19. Amsterdam: Elsevier, 1995. 10 Edwards I R, Hugman B. The challenge of effective benefit-risk information communication. Drug Safety (in press).
Home | Current issue | Past issues | Classified ads | Career Focus | Feedback Collections | About this site | About the BMJ | BMA | Medline
|