The Christmas bundle reveals how the MCA/CSM responded to the risks discussed in ADWEB, to late March 1998. Contrary to the impression given in correspondence, it is clear that the 'drug safety' machine did go into action. The bundle mainly comprises the MCA's Assessment Report (around 250 pages, including appendices) and the expurgated minutes (6-pages) of the CSM meeting at which the Assessors' conclusions and recommendations were discussed. Only the CSM minutes are published here in full.
To obtain a copy of the Assessors' report, please apply to the MCA. The report does include some interesting detail and, refreshingly, makes minimal reference to the effectiveness of antidepressants in treating dread and widespread disease. The one-page summary of the Assessment has been posted on ADWEB, but not the whole report - because the MCA/CSM's main conclusions have repeatedly been reflected in correspondence with Social Audit, from July 1997. Briefly, "the Committee agreed that the review had revealed no evidence of dependence associated with SSRI and related antidepressants", also "no evidence to suggest ... that withdrawal reactions with these drugs had public health implications".
The following notes consider some wider implications of this disclosure, in relation to:
- the abuse of secrecy in government;
- the nature of drug safety investigation;
- the quality of risk assessment;
- influence of pharmaceutical companies; and
- risk of dependence on SSRIs and other antidepressants (ADs).
1. Re: secrecy in government
[a] After numerous refusals and long delays, the MCA finally disclosed this bundle only after Social Audit had complained to the Parliamentary Commissioners for Administration (Ombudsman).
[b] It is quite clear from the documentation provided that only a fraction of the information requested was in fact 'confidential' within any credible meaning of the Code. As ever, the regulators seem addicted to their secrets to the point of abuse.
[c] It is equally clear that deletion of sensitive material was (or could have been) easily accomplished. The main deletions in the CSM minutes were of members' and manufacturers' names - a few minutes work with a bottle of correcting fluid, no more. One large and one small chunk of text had been deleted from the Assessors' report: submissions from manufacturers, and from other regulatory agencies within the EU. Again, deletions were easily made.
[d] For all the excuses about the importance of not damaging international relations etc., the main reason for non-disclosure of data from other EU countries seems almost certainly to be the need to avoid embarrassment about the paucity of data supplied. This exemption should be challenged in due course.
[e] The wholesale deletion of a 13-page summary of comments from manufacturers seems very disturbing - the triumph of property rights over truth and integrity in medical science, openness and accountability in public life, and over patients' rights to safety and information. The only paragraph to remain intact in the 13-page summary suggested undue reliance by commercial interests on the work of Dr S Montgomery (See correspondence with CSM).
[f] This disclosure appears to establish a general right of access at least to expurgated CSM minutes and working papers.
2. Re: nature of drug safety investigation
[a] The MCA Assessors clearly went to considerable trouble to gather available data - but drug 'safety' assessments still seem dangerously dependent on process and otherwise misconceived. No amount of data can substitute for focused, critical thinking.
[b] Over-reliance on spontaneous adverse drug reaction reporting systems (eg Yellow Cards) maintains a gross bias in perceptions of drug benefits to risks - especially when unfamiliar or unexpected reactions are involved.
[c] Regulators persist in estimating the extent of risk by contrasting relatively small numbers of adverse reaction reports with vast numbers of prescriptions. Thus, the doubling of the number of prescriptions for antidepressants in the 1990s seems to cause offical relief, rather than concern - because when the denominator doubles, the risk appears halved.
[d] The investigative process appears misconceived for its reliance on "current evidence"; it does not systematically define what evidence is needed and missing, and tends not to recognise risk in the absence of clear evidence of harm.
[e] The process seems geared more to examining the adequacy of product labelling, than to critical scrutiny of clinical practice. Assessment involves retrospective analysis of received medical wisdoms. It fails ultimately because is intended to reflect current medical opinion, rather than lead or challenge it.
[g] On this evidence, the MCA pretty much runs the show. The 6-page expurgated minutes suggest the CSM had a negligible input. "The Committee complimented the assessors on a very good report", and pretty much endorsed the conclusions. Slight data sheet amendments were proposed: they missed the point.
3. Re: quality of risk assessment
[a] This Assessment appears primarily addressed to the question: 'Do we need to do anything?' More fundamental questions have not been considered - notably the risk that the long-term effectiveness of antidepressants has been wrongly inferred from trials which have invariably classified withdrawal-induced depression as 'relapse'.
[b] The Assessment included a review of The Antidepressant Web; this amplifies the earlier press release, proceeding more ponderously in the same vein. To this most biased observer, the analysis seemed superficial, sometimes obtuse; the style ranged from lofty to disparaging. The authors appear deeply committed to close scrutiny of essentially niggling and trivial points, but largely fail to focus on basic issues. It's as if the MCA/CSM had carefully picked their way through minefields, then declared them safe enough. It was particularly irritating to see that they felt they had little to learn from all the patient reports on the Internet of antidepressant withdrawal/dependence problems: "These reports are lacking in detail and difficult to interpret".
[c] The conclusion that pretty much all is well appears to be mainly supported by these beliefs: that doctors would generally be well aware of any problems that might arise; that SSRIs are not to be compared with BDZs, since dependence on the latter is characterised by abuse; that SSRI withdrawal symptoms are typified by physical rather than psychological symptoms; that a drug is safe enough if most people usually find it so; and that objections to such assumptions can be adequately dealt with by parrying with a counter citation or two.
[d] In short, the MCA/CSM and Social Audit would seem to agree that the other's arguments might have some merit, but for their foundation on thin evidence, flawed understandings, inconclusive analyses, and inconsistency of argument.
[e] Time will tell which party came closer to the mark. Significantly, the MCA/CSM concluded that "an article in 'Current Problems in Pharmacovigilance" should be published", and the bundle includes a March 1998 draft. It was never published, so maybe the penny began to drop.
4. Re: influence of manufacturers on drug regulation
[a] The MCA Assessors' report appears to have been first discussed at a meeting of the CSM's Sub-Committee on Pharmacovigilance (SCOP). It is not clear what influence SCOP had on the decision made by the CSM as a whole. This matters, partly because only two members of the CSM sit on this 8-person sub-committee, both with possible conflicts of interest to declare.
(The SCOP Chairman was Prof Alasdair Breckenridge, who had a personal 'non-specific' interest with SmithKline Beecham, the manufacturers of paroxetine (Seroxat/Paxil). Because of this, Prof Breckenridge had to leave the room for at least some part of the meeting, when the CSM discussed the Assessors' report. It is not clear whether any possible conflict arose at the SCOP meeting, nor what would have happened if it did. A further request for disclosure will be made, to obtain minutes of the relevant SCOP meeting and clarification of Prof Breckenridge's role).
[b] Seven of the 18 CSM members who discussed the MCA Assessor's report declared some possible conflict of interest, and subtle differences are evident in the Chairman's response. See the response to declarations of personal 'non-specific' interests by three CSM members, below. (Names have been deleted from the minutes, but the 1997 CSM annual report indicates they were Professors Breckenridge, Jacobs and Smith. Breckenridge was then Deputy Chairman of the CSM, and became Chairman of the CSM in 1999. One is left to guess which response applies to whom).
" . at the request of the Chairman contributed freely to the discussion and left the room for the decision", whereas " answered questions from the Chairman and left the room for the decision", and " left the room".
[c] The MCA Assessors relied overwhelmingly on definitions of withdrawal derived from "a consensus panel of experts" (without mentioning that the Eli Lilly Corporation had convened them). Thus they effectively discounted the risk of depression as a symptom of withdrawal.
[d] Significant that " had been approached recently by a representative of one of the manufacturers regarding the issue, but that had declined to enter into any discussions".
[e] Interesting that one of the manufacturers had "expressed concern at the use of the term 'withdrawal reaction' and had suggested the use of the term 'discontinuation reactions'." Even more interesting that the CSM considered this request as a specific agenda item - but to their great credit: "It was agreed that it would be inappropriate to change medical terminology in this way".
[f] The MCA/CSM appear to make modest demands of the manufacturers: "The possibility for specific targeted investigation of the frequency and severity of withdrawal reactions, and of dependence should be discussed with (them)". Come to think of it, why did MCA/CSM never require manufacturers of SSRIs to provide evidence relating to drug withdrawal from the outset - if indeed, "it has been known for many years that symptoms can occur on withdrawal of antidepressants"?
[g] The manufacturers' ability to advertise and orchestrate drug benefits cannot be reconciled with the absolute secrecy to which they are entitled when discussing their views of risk. This diminishes the scientific basis of risk assessment and casts long shadows over the whole process of regulation.
5. Re: risk of dependence on SSRIs and other antidepressants
[a] These words will come back to haunt the CSM/MCA: "The Committee agreed that the review had revealed no evidence of dependence associated with SSRI and related antidepressants", also "no evidence to suggest ... that withdrawal reactions with these drugs had public health implications", and "the Committee complimented the assessors on a very good report"
[b] Evidence published in 1998 (from the four Lilly-sponsored trials) demands urgent re-examination of these conclusions. Granted, most people find antidepressants unpalatable. It is nevertheless clear that, after relatively short-term use, about 30% of patients get withdrawal symptoms and that "depression" is a major one - and doctors' lack of awareness of this threatens only to compound the problem of dependence. It follows that the long-term effectiveness of antidepressants has been mistakenly inferred, and that the drug regulatory system has failed. There are public health implications indeed.
CM, 28 December 1998