Rapid Responses to:
|
|
Rapid Responses: Rapid Responses published:
-
![[Read Rapid Response]](/icons/misc/sectionDOWN.gif)
Observational data on harm should complement systematic reviews of benefit
- Jan P Vandenbroucke (15 July 2003)
-
Screening and immunizations
- Scott L. Johnston (15 July 2003)
-
Observational Data on Harm is Already Included in Systematic Reviews
- Tom Jefferson, Vittorio Demicheli (18 July 2003)
|
|
|||
|
Jan P Vandenbroucke, Professor of Clinical Epidemiology Leiden University Medical Centre, PO Box 9600 , 2300 RC Leiden, The Netherlands
Send response to journal:
|
The meeting on “rare events”, mentioned by Cuervo and Clarke (1) was a wonderful opportunity to bridge two worlds: the world of systematic reviews of randomised trials which investigate benefits of treatment, and the world of pharmacoepidemiology which mainly investigates the harms of the very same treatments. At the end of the meeting there was quasi unanimity that there was no escape: observational evaluation of harms should eventually have to enter systematic reviews to present a balanced picture of benefit and risk. Otherwise, systematic reviews will only tell half the story. At the meeting, participants discovered each other’s field. Pharmacoepidemiologists learned that their skills were needed, and that they should perhaps become involved in systematic reviews. They should do a better job at explaining how observational research works in detecting and verifying rare or delayed adverse effects of treatment, and why there is little alternative to detection by case reports and verification by case-control and cohort studies. They should take responsibility to ensure that “the best quantitative evidence” on harm enters systematic reviews. Participants working at systematic reviews of randomised trials discovered that there already exists an area of research on adverse effects for 40 years, with its methods, textbooks, journals, departments and publications upon which clinicians and government agencies rely for important decisions. Still, bridging will not be easy. On one side there are entrenched views about the superiority of randomised evidence, leading to prejudices that there has been no reliable progress except for randomised trials, that case reports are biased (despite evidence to the contrary about adverse effects (2)), or that observational data-bases are intrinsically problematic. This attitude ignores established differences between discovery and verification, and between investigating intended and non- intended effects of treatments (3, 4). On the other hand, discussions about epidemiologic methods as applied to adverse drug reactions can become so arcane that only a few people still follow (5). For sure, judging observational research on harm will involve more subject matter knowledge, and will be less easily codified than judgements about randomised trials. However, participants from both fields felt the need to think about incorporating the others in future activities, perhaps by giving young people dual training in both types of research. This should lead to a future where several types of evidence can be entered in a single systematic review, to present a true balance of risks and benefits. Jan P Vandenbroucke, co-organiser of the meeting “rare events”. (1) Cuervo LG, Clarke M. Balancing benefits and harms in health care. BMJ 2003;327:65-66. (2) Chalmers I. Geoffrey Venning’s classic BMJ paper. http://bmj.com/cgi/eletters/326/7403/1346#33806 (3) Jick H. The discovery of drug-induced illness. N Engl J Med 1977;296:481-5. (4) Miettinen OS. The need for randomization in the study of intended effects. Stat Med 1983;2:267-71. (5) Skegg DCG. Third generation oral contraceptives. BMJ 2000; 321: 190- 191. Competing interests: None declared |
|||
|
|
|||
|
Scott L. Johnston, Wright Health Care Clinic Family Physician PO Box 689 Wright, Wyoming 82732
Send response to journal:
|
A timely and important discussion is certainly needed. The best example of under reported harm is in the area of laboratory testing, especially screening tests. Most clinicians focus on the potential benefit and only rarely consider the harm of screening. Lab tests are reported with sensitivity and specificity numbers that are useless to most clinicians. Requiring tests to report predictive values coupled with data on differing populations would alleviate much of that problem. Immunizations bring about another area of contention. Currently many immunizations have only rare side effects but are seldom beneficial to any one patient. Accurately reporting the benefits and risks will encourage more parents to immunize their children. Improving patient care in these two areas should be a goal set to the forefront of most research. Competing interests: None declared |
|||
|
|
|||
|
Tom Jefferson, Coordinator Cochrane Vaccines Field, Vittorio Demicheli
Send response to journal:
|
EDITOR - We read the correspondence on the issue of including observational data in systematic reviews with some surprise. The tone of the prose was that of an event which might happen in the future. It has been happening for some time, precisely for the reasons that Johnston (screening and immunisation) illustrates so cogently for the field of vaccinations. Allegations about possible causal associations between exposure to one or more vaccines and unexpected events have been made since the time of Jenner. Recently, an increasing number of potentially damaging allegations were seen to threaten national and international immunisation programmes. Evidence was scattered, seldom assessed by its methodological quality and sometimes included in descriptive reviews. In response to this requirement we used allegations of the association between MMR vaccines and different adverse events and Hepatitis B (HB) vaccines and demyelinating disease to develop methods to identify, assess and synthetise evidence from studies of different designs, ranging from randomised controlled trials to case-only designs (1,2, 3). Assessment of possible rare and unforeseen adverse events following vaccines is methodologically particularly difficult, as in most cases there are no independent controls, as the majority of the population is already vaccinated and those who are not, are likely to be unrepresentative of the reference population. We believe that such difficulties can be overcome by inclusion of studies with no independent controls such a before-and-after and case cross-over designs. Development of quality assessment criteria for such studies was a worthy challenge. Broadening the focus of systematic reviews is not an optional feature which may come about in the future, it's here already. Tom Jefferson
References 1. Demicheli V, Rivetti A, Di Pietrantonj C, Clements CJ, Jefferson T. Hepatitis B vaccination and multiple sclerosis - Evidence from a systematic review. Journal of Viral Hepatitis 2003 (in press). 2. Jefferson TO, Price D, Demicheli V, Bianco E. Unintended Events following immunisation with MMR: a systematic review. Vaccine 2003 (in press). 3. http://www.who.int/vaccines- surveillance/ISPP/IssuesofInterest.shtml. Competing interests: None declared |
|||