Rapid Responses to:
|
|
Rapid Responses: Rapid Responses published:
-
![[Read Rapid Response]](/icons/misc/sectionDOWN.gif)
Sponsored drug trials show more favourable outcome
- Hugh Boardman (16 February 1999)
-
Sponsored drug trials show more-favorable outcomes
- Michael J Tidd (27 February 1999)
-
Sponsored trials not necessarily more favourable
- J N S Matthews (5 March 1999)
-
Conflict of Interest in Abuse Referrals?
- Brian Morgan (8 April 1999)
-
Editorial independence
- Robert F Garry (27 May 1999)
|
|
|||
|
Hugh Boardman
Send response to journal:
|
Editor Wahlbeck and Adams report an interesting finding about drug company sponsored clinical trials as a result of their work on systematic reviews for the Cochrane Group [1]. It is unfortunate that they abandon their objectivity and fail to apply scientific method when proposing an explanation for their observation that drug company sponsored prospective, randomised, double blind, placebo controlled, clinical trials are more likely to find positive results than those run by the academic community. The pharmaceutical industry has immense scientific resources at its disposal and is obliged to use them scrupulously in designing, running, monitoring and analysing clinical trials on its products. The findings of its research have to be above reproach because they are scrutinised, not only prior to publication in peer review journals, but also prior to licensing by regulatory authorities around the world. Some authorities, such as the American FDA, insist on being supplied with the raw data so that the results can be checked. Some authorities such as the European EMEA and the American FDA issue carefully researched guidelines on how studies should be conducted and even specify which rating scales should be used to assess specific diseases. Sadly in academia things are not so tightly regulated or so well resourced. When involved in a drug company sponsored research project I have ample access to statistical advice and sponsoring companies pay for data to be collected systematically and audited independently to ensure their veracity. Which academic units are able to lavish such luxuries on their in house projects? Perhaps the reason that drug companies are more able to demonstrate that their products are effective is that they are more careful about how they design their research projects. They also have the resources to run them more efficiently and thus get cleaner, more reliable data to analyse. In short, they get better results because they have exercised more care. Is it surprising that the most recent scandals about fraudulent research in Britain and America have centred on academic research and not on drug company sponsored clinical trials? If the Cochrane Groups want to produce reliable systematic reviews they should first determine who collects the most reliable data. It is not good enough to let preconceived ideas bias an objective evaluation of the best available evidence. Dr Hugh Boardman MA MB BChir FFPM Honorary Research Fellow University of Surrey HPRU Guildford. [1] Wahlbeck K, Adams C. Sponsored drug trials show more-favourable outcomes. BMJ 1999 ; 318 : 465. |
|||
|
|
|||
|
Michael J Tidd
Send response to journal:
|
Editor: Wahlbeck and Adams (letters: 13 February 1999) conclude that there may be something sinister about industry sponsorship of clinical trials being associated with more frequent positive outcomes. They do not consider the possibility, remote though it may seem to some, that industry-sponsored trials may be better designed, supported, executed and analyzed. They say nothing about the comparative quality of studies from these two arenas - for example, statistical power considerations. The resources available to industry, and the tight regulation and compulsive attention to detail required of industrial sponsors are very expensive, but greatly reduce the risks of poor design, undetected erroneous data and misconduct. Few independent centers can come close to matching industry in these respects. Certainly, authors and editors of any affiliation like to see positive outcomes, but in industry a clearly negative result from a high quality study is generally regretfully, but quickly, recognized for what it is ? a signal to cease investment. Health care professionals and regulatory authorities are not fools. Neither are their practicing colleagues. Marginally effective or ineffective new drugs are soon seen for what they are in the market place, and usually well before. Industry has no sane or rational motive for investing in them. As Wahlbeck and Adams recognize, conflicts of interest are experienced in academia as well ? the interests may be different but the conflicts are still there, and just as strong. ‘Independence’ does not automatically confer greater objectivity or competence any more than industrial affiliation immediately confers loss of ethical standards. To imply that ‘independent’ investigations are somehow invested with greater integrity or reliability by virtue of their lack of industrial affiliation does not accord with my experience. Michael J. Tidd, B.Sc., M.B., B.S. President, Norwich Clinical Research Associates Ltd P.O. Box 91 Plymouth NY 13832 USA Competing Interest: My company designs and executes clinical research studies on drugs and medical devices. We have no financial interest in the outcome of such studies. |
|||
|
|
|||
|
J N S Matthews, Professor of Medical Statistics University of Newcastle upon Tyne
Send response to journal:
|
EDITOR - Wahlbeck and Adams [1] report a review of randomized studies comparing clozapine with other antipsychotic drugs for the treatment of patients with schizophrenia. They argue that trials which reported some connection with the manufacturer of clozapine (sponsored trials) were associated with more-favourable outcomes for the drug than trials in which no connection with the manufacturer could be discerned (non-sponsored trials). Unfortunately the information the authors present does not support their claims. It is stated that the odds of relapsing on clozapine are clearly less than on its comparators in sponsored trials (odds ratio 0.5, 95% confidence interval 0.3 to 0.7) but there are equivocal findings in the non-sponsored studies (odds ratio 0.4, 95% confidence interval 0.1 to 1.4). However, an unequivocal difference in the sponsored trials and an equivocal difference in the unsponsored trials does not establish a difference between the two types of study. To claim that it does amounts to comparing subgroups on the basis of their P-values and this is known to be flawed [2]. The confidence intervals show that the odds of relapse on clozapine relative to its comparators lie within wide intervals for both types of trial and the data are certainly compatible with the effect of clozapine being the same in sponsored and unsponsored studies. A formal assessment would entail forming the ratio of the odds ratios, namely 0.5/0.4=1.25 and constructing a 95% confidence interval for this quantity, which is approximately from 0.3 to 5. This shows that there could be marked differences in either direction in the effects found in the two types of trial, but certainly does not establish that there is a difference between them. Similar comments apply when the outcome is clinical improvement or early withdrawal. JNS Matthews, Professor Department of Statistics University of Newcastle Newcastle upon Tyne NE1 7RU 1 Wahlbeck K, Adams C. Sponsored drug trials show more-favourable outcomes. BMJ 1999;318:465. (13 February) 2 Matthews JNS, Altman DG. Interaction 2: compare effect sizes not P values. BMJ, 1996; 313:808. Note: I have no conflict of interest to declare. In particular I have no connection with and receive no support from any pharmaceutical company. |
|||
|
|
|||
|
Brian Morgan, Freelance Journalist Cardiff
Send response to journal:
|
Questions have arisen concerning payment of fees to paediatricians for writing reports and appearing in court in connection with care proceedings arising from referrals they have themselves made to social services after clinical assessments made originally during NHS consultations. There is a anecdotal evidence that such payments are being made and that they are kept by the paediatrician as part of his or her private remuneration. The questions are: 1. Is this true? 2. If true, is there a financial conflict of interest? 3. How should the paediatrician involved in making a referral to social services as part of his or her NHS practice be remunerated for the additional testimony that may arise? 4. Is there any official guidance on this subject? 5. Do parents who have been accused or will be accused of child abuse need to be reassured that no doctor making such allegations is influenced by the expectation of earning significant additional remuneration as a result of the diagnoses? 6. If the basis for concerns that have been brought to my attention prove to be well founded what might the profession's response be? Brian Morgan Conflict of Interest: the questions arise from research currently under way for journalistic purposes. |
|||
|
|
|||
|
Robert F Garry, Professor of Microbiology and Immunology Tulane Medical School
Send response to journal:
|
To an editor other than Kassirer or Angell: Kassirer was correct in his assessment that “editorial independence sorely needs further defining.” (1) Perhaps unwittingly, his editorial also touched on several circumstances where dismissal of an editor may be justified. I urge medical editors not to forsake clarification of this facet of the controversy as well. I very much agree with Kassirer’s statement that “medical editors must be held accountable in terms of their competence and their overall judgment.”(1) Two recent NEJM editorials call in to question the competence and judgment of its editors. In the first instance, executive editor Marcia Angell compared trials to develop strategies to prevent perinatal transmission of AIDS in developing nations to the infamous Tuskegee syphilis experiments.(2) Such hyperbole in an area where the editor has little if any documented research competence is troubling and as recently discussed “the analogy to Tuskegee entailed a gross distortion.” (3) Yet, the only recourse to two editorial board members who disagreed with Angell’s editorial and the implicit implication that they agreed with her needlessly inflammatory public statements was to resign. In the second instance, Kassirer and Angell penned an editorial largely dismissing the risks of obesity.(4). In a letter submitted to the journal, but not published, former Surgeon General C. Everett Koop and two distinguished co-authors called the latter editorial an “irresponsible use of editorial privilege.” To whom, if anyone, are NEJM editors accountable? Kassirer also wrote that “To ensure objectivity, editors must have no financial conflicts of interest that might bias them in choosing reviewers, making editorial decisions, or issuing public statements.”(1) Who, if anyone, decides if an editor at the journal has made a biased choice of reviewer or other editorial decision? Furthermore, Angell has presented public lectures for the Independent Women’s Forum and the CATO Institute, both for a fee. Both organizations represent the legal and political interests of the medical device industry. Furthermore, she has written a book about an ongoing medicolegal controversy surrounding a medical device, breast implants, the sale of which was promoted at these media events. Who, if anyone, decides what constitutes a conflict of interest for an NEJM editor? Recently, I called on the editors of the British Medial Journal to lead an initiative to have Journal Editors periodically disclose there own conflicts of interest.(5) Can NEJM editors support this as well? Robert F. Garry, Ph.D. Professsor of Microbiology and Immunology Tulane Medical School New Orleans, LA 70112 504-587-2027 504-584-1994 rgarry@tmcpop.tmc.tulane.edu 1. Kassirer, J.P. 1999. Editorial independence. NEJM 340: 1671. 2. Angell, M. 1998. The ethics of clinical research in the Third World. NEJM 337: 847. 3. Fairchild, A.L. and Bayer R. 1999. Uses and Abuses of Tuskegee. Science 284: 919-921. 4. Kassirer J. P. and Angell M. 1997. Losing weight -- an ill-fated New Year's resolution. NEJM 338: 52. 5. Garry, R.F. 1999. BMJ's editors should publish their own conflicts of interest regularly. British Medical Journal 318: 464. |
|||