REVIEW OF THE YELLOW CARD SYSTEM
The following is a response to questions raised in the consultation letter
- submitted to this review in November 2003,
by Andrew Herxheimer FRCP and Charles Medawar
Should data submitted in confidence to MHRA/CSM be made more widely available for uses outside monitoring and control of medicines safety?
The question rather implies that monitoring and control of medicines safety is the exclusive preserve of the regulators. This responsibility should, in our view, be widely shared. The regulators alone have neither the expertise nor capacity to make the fullest use of the data provided. It follows that these data should be made as widely available as possible.
The fact that the data are submitted in confidence does not alter our view on this. The nature of the confidentiality relates strictly to the need to guarantee the complete anonymity of doctor, patient and any other individual. Subject to this most important consideration, the data itself should be available for analysis. By analogy, a doctor should not be prohibited from discussing medical matters arising from contact with individual patients, provided the patients personal privacy is protected.
To underline the above point about the limited capacity of the regulators to investigate the significance of the data they receive, we refer to the attached table, Suicidal behaviour number of reports in 5 ADR categories, (Herxheimer, 2003). Our point is that such data would never normally be generated within the scheme, and that such analyses should be encouraged, rather than effectively prohibited. In this particular case, the significance of the results for suicidal ideation may not be clear but they clearly merit further investigation. There is far too much raw data for the regulators to be able to generate all relevant outputs on their own.
What are the implications for patient safety, public health and the future of the Scheme of widening access to data or changing the reporting procedures?
If sensibly and sensitively handled, we would think that the wider availability and scrutiny of data would ultimately improve patient safety and public health, both directly, and by increasing public confidence in this mainstay of post-marketing drug safety monitoring, not least by introducing an element of public accountability.
What are the ethical issues raised in offering wider access to use of Yellow Card data?
In our view the major ethical challenge is to use the available data in the best possible ways. Offering wider access will facilitate this, whereas restrictions will waste good opportunities and obstruct drug safety research.
If the data were to be used for research, how should the ethics of individual research proposals be evaluated?
This question implies [a] that these data should be released only if required in connection with some formal research proposal; and [b] that the present guardians of these data should be the ones to make decisions on ethical issues. We would suggest that [a] such data be released as a matter of course, provided there were no reason to think they were to be used for some improper purpose; and [b] that any decision to withhold data should be subject to appeal to an independent body. It is a cardinal principle of openness that those who keep the secrets should not, in the last analysis, be those who decide what should be kept secret.
Should research be permitted with or without the agreement of the patient whose suspected adverse event has been reported?
The question conveys to us a mindset determined to try to use every available method to prevent disclosure. Provided patient anonymity is assured, there seems to be no reason to get patients consent any more than there would be if a doctor wrote up a case report in a medical journal.
Should research that includes genetic information or analysis be permitted?
As a general rule, we see no reason why not. Exceptions to that rule should be made only if the information or analysis compromised rights of personal privacy.
Controls on use of Yellow Card Data
What safeguards are required to ensure use of the data is likely to promote public health?
The main safeguards would seem to us to relate to putting available data (properly anonymised) into the public domain. Requesters who might benefit from help in understanding and interpreting those data should be offered such help.
Should aggregated data be made available to NHS trusts/primary care trusts for their use? If so, in what form and for what purposes?
We see no reason why such aggregated data should not be made available to these trusts, among others. The form in which those data were supplied should be determined by utility. The purposes should be to improve understanding and to promote safer and more effective and rational drug use.
Conditions on access
If these data are to be more widely available to researchers external to the MHRA, what conditions of access [ethical/financial/potential impact on public health] should be applied?
The objective should be to reduce conditions of access to an absolute minimum, subject to the considerations mentioned above.
Should data from the Yellow Card Scheme be made available for direct or indirect financial gain?
It is obviously not desirable that the primary reason for requesting data should be financial gain yet one can envisage situations in which such motives might be constructive (e.g. where an intermediary agency sold data that it had processed or analysed on to a third party whose interests related primarily to health). We do not think that the desire to make financial gains should be an absolute disqualification to access bearing in mind that the reasons data from the Yellow Card scheme are at present withheld have, at least indirectly, a great to do with economic factors and associated financial gain.
Does the existing legal framework sufficiently protect data subjects and reporters or in any way inhibit the reporting of adverse drug reactions?
We are not aware of any gap in the current legal framework. We believe that the full protection of subjects can be achieved through administrative measures within the existing legal framework.
If so what changes are required for improvement?
As the Government has a statutory responsibility to have such a scheme in place, how should financial consequences of wider access to data be addressed?
Primarily with a view to encouraging wider access with a view to extending and improving the interpretation of data and dissemination of useful information.
What are the operational implications of wider access e.g. for the maintenance of the database and its continued utility for drug safety monitoring?
We are not well placed to answer this question, other than to observe that the six-month wait we have so far experienced, following a request for ASPPs for certain antidepressants, suggests the need for substantial improvements. We cannot say how far operational or political questions have led to this impasse; we believe that they would have been promptly supplied if requested by the pharmaceutical company concerned.
What implications might changes to the use of the data have on reporting to the scheme?
Our analysis of paroxetine ASPPs suggests that wider access would greatly improve the quality of analysis, interpretation and utility of data, both directly and through the greater accountability of the present organisers of the Yellow Card scheme.
Strengthening of the Yellow Card Scheme
Our analysis of paroxetine ASPPs (Medawar & Herxheimer, 2003) suggests a range of benefits to a scheme whose present operation appears to us to be in several respects "chaotic and misconceived".
How might the reporting Scheme better promote and protect public health if changes are made to access to the data?
The wider availability of ADR data could be expected to encourage useful, if not essential developments, in both reckoning with, and effectively communicating, the uncertainty of risk. The Yellow Card scheme as presently operated has largely failed to do this. We believe that the system, and the analysis of data produced, should be focused much more on models for assessing the probability of risk; at present it seems slavishly geared to establishing the certainty of harm.
The present operation of the Yellow Card scheme is unsatisfactory also, in focusing on the detection of rare, esoteric, new adverse drug reactions while purporting to monitor the impact of ADRs overall. (See MHRA/CSM guidance notes on the scheme). We believe the scheme is important and essential, but cannot accept its sufficiency as a mainstay of post-marketing drug safety surveillance.
How might the reporting Scheme be strengthened?
Any other issues
t is notable and regrettable that the Review makes no reference to the role of the Internet and its potential future importance in pharmacovigilance, nor to the now rapidly changing relationship between providers and users of medicines. We believe the Review should be setting its sights on building for the longer-term future at least adequate to the needs of the next decade (circa 2015). In our view, any attempt to restrict access to these data would contribute substantially to loss of confidence by users in the will and ability of the professions, regulators and manufacturers to see medicine develop in the public interest responsive to scientific and democratic imperatives alike. We draw to the attention of the Review, the concluding remarks in the annex to this submission, a short note on Information management, conflicts of interest and the safety of SSRIs. This Review provides an opportunity to recommend changes that would make the Yellow Card system truly the best in the world, and by far. We expect nothing less.
Are there any other issues we need to consider in the context of altered access to the Yellow Card data?
Andrew Herxheimer FRCP
Medawar C, Herxheimer A,Risk of Suicidality and Dependence with Paroxetine: Comparing Yellow Card Reports of Suspected Adverse Drug Reactions with Reports from Users, Int. J. Risk & Safety in Medicine, 2003, 16, 3-17.
Medawar C., Information management, conflicts of interest and the safety of SSRIS, note posted to www.socialaudit.org.uk in August 2003
Herxheimer A, Unpublished, 2003: Table: Suicidal behaviour number of reports in 5 ADR categories (attached)