The roles that hospital and community pharmacists currently play and ways in which their role might be extended were two of the pharmacovigilance issues discussed at a symposium to commemorate the work of the late Dr Sue Wood, the former director of the post-licensing division at the Medicines Control Agency
Pharmacovigilance was a mission for Sue Wood, Dr Keith Jones (executive director, Medicines Control Agency) said. She had developed it in the interests of patients and the wider public health.
Dr Jones paid tribute to Dr Wood's special vision, not only across Europe, but globally. This was reflected by the introduction of an information management system (known as ADROIT) which was later networked to the industry. This, in turn, had provided a quality control process and a control procedure for licensing and pharmacovigilance, and widespread access to pharmacovigilance data. Adverse drug reactions (ADRs) could now be reported electronically directly to regulatory authorities.
This vision also foresaw the need for a common language by which the component parts could exchange information. The medical dictionary of defined terms (known as MedDRA) was developed from this vision, and had become an international terminology for regulatory affairs.
The environment in which pharmacovigilance was being practised was changing, Dr Jones said. Five years ago, about 1,500 completed case reports were required for each marketing authorisation application (MAA). Since then, drug development had changed to include biologically-derived materials and products from genomics, products which had narrow indications and which were applicable to much smaller populations of patients. As a result, there would now be many fewer than 1,500 completed case report forms for each MAA. Pharmacovigilance was a means of monitoring these new entities.
Professor Sir Michael Rawlins, a former chairman of the Committee on Safety of Medicines, described the importance, success, and cost-effectiveness of the yellow card scheme, introduced in 1964, was stressed. The pharmacovigilance scenario in Europe, however, was often different. For example, two member states of the European Union did not even acknowledge receipt of ADR reports and made no effort to stimulate reporting. This was a major problem in trying to achieve a single European pharmaceutical market, and Professor Rawlins emphasised the importance of achieving harmonisation of reporting methods across Europe.
The value of randomised controlled clinical trials in developing ADRs was stressed by Sir Michael. However, one spectacular failure of this method had been the failure to identify cough as an ADR in angiotensin-converting-enzyme inhibitors.
Probably the biggest recent advance had been the availability of the electronic data linkage collection, and especially the General Practice Research Database (GPRD, see below). Such databases allowed for case- control studies, and could provide estimates of relative and absolute risk. One of their limitations was the failure to capture socio-economic data.
Professor Rawlins also reflected on the poor quality of patient information leaflets. He stated that the Plain English Society would be appalled by the quality of many of these.
Pharmacovigilance activities in Europe were reviewed by Dr Mary Teeling (medical director, Irish Medicines Board). The vital role played by Dr Sue Wood as chairman of the Committee for Proprietary Medicinal Product's pharmacovigilance working party was noted by Dr Teeling.
The activities of this working party included dealing with informal referrals by member states about safety issues. This process was especially useful for specific safety issues, such as therapeutic class labelling and updating safety data. Other activities of the working party included organisational matters, production of guidance documents, development of procedures, training, and, perhaps most importantly, the ongoing exchange of information and expertise . All of these were intended to create consensus within Europe.
Dr Murray Lumpkin (deputy director, Centre for Drug Evaluation and Research, Food and Drug Administration) described some of the changes in the environment of the global market.
One of the biggest changes recently had been instantaneous global communication, with many people with many different agendas having access to various databases. This was particularly important in drug safety issues. The media was one of the parties accessing this information. Those involved in drug safety issues had to accept that the "soundbite" was important, and that instantaneous communications did not always allow a thoughtful, iterative process.
The safety profiles of many drugs had been publicly debated over the past two years. There was an impetus for international co-operation from this. The fundamental challenges were risk detection, risk management decision-making, and risk communication and reassessment.
Various issues had been harmonised. For example, for spontaneous reports under the Council of International Medical Sciences and International Conference on Harmonisation, definitions of "serious", "expected", and "listed" had been agreed. There had also been harmonisation of data elements considered important in carrying out medical assessments of individual cases. The development of MedDRA and its translation into other EU languages was vital, as was harmonisation of the international electronic transfer of case safety reports, Dr Lumpkin said.
Dr Patrick Waller (specialist in pharmacovigilance, post-licensing division, MCA) stated that spontaneous ADR reporting had been conceived as a direct response to the failure to recognise thalidomide phocomelia until around 10,000 foetuses had been affected. A notification system should have recognised this association of a specific drug with a very rare disorder at least three orders of magnitude more effectively (ie, with perhaps 10 or fewer cases).
A question was asked as to whether the value of spontaneous reporting could be increased and, if so, how.
Dr Waller said that spontaneous ADR reporting systems had three distinct phases: obtaining information; data processing; and analysis and interpretation. The first was the most problematic and was becoming more difficult, perhaps because of competing demands in modern health care systems. In the second and third areas, great strides had been made in the past decade, but there was still more that could be done, Dr Waller said.
In the UK, there had been a steady increase in the number of spontaneous ADR reports over the first 25 to 30 years of the yellow card scheme, reaching more than 20,000 in 1992. Numbers had declined slightly since then, mainly due to competitive pressure on reporters' time. Consequently, the MCA had broadened the base of reporters by including pharmacists and specialists in therapeutic areas, such as HIV medicine, for which reporting was low. It had also developed an electronic yellow card and improved the feedback to reporters.
Should patients report ADRs, Dr Waller asked. There had been relatively little research into this and its value should be regarded as unproven for the present, he thought. Experience from the United States showed that such reporting could markedly increase the number of reports received. The key test was whether important hazards could be recognised more quickly. This remained an open question.
Professor Stephen Evans, (principal consultant statistician, Quintiles) listed some of the many sources of electronic pharmacovigilance data. These included central databases of GP-derived data (eg, GPRD, see below), prescription-event monitoring data, and prescribing data held by hospital pharmacies. Prescribing data was available from the Prescription Pricing Authority and commercial sources. The data had been linked into a system in Dundee which combined prescribing and medical event data.
Hospital data on prescribing was starting to be brought together commercially, but not nationally in the UK. The absence of this data made it extremely difficult to interpret spontaneous reports for some classes of drugs, especially as UK spontaneous reporting in hospitals was less efficient than in the community. However, the increasing sophistication of hospital computer systems increased the potential for more widespread recording of linked prescription and medical event data.
Dr Elizabeth Andrews (director of world-wide epidemiology, Glaxo Wellcome) suggested that evaluation of early development data by the industry might detect actual signals or anticipate potential safety issues. She thought that a "toolkit" would be invaluable to all partners (industry, government, and academia) in evaluating safety data, one component of which would be pharmacoepidemiology.
Prior and proposed models of collaboration were illustrated by reference to lamotrigine (Lamictal), used in the treatment of epilepsy. Sudden unexplained death was experienced during clinical development in a small number of cases through named-patient studies that were long-term extensions of clinical trial development. However, it was not possible to carry out risk evaluation as there were no similar extension studies in comparator groups.
Three studies had therefore been initiated by Glaxo Wellcome to evaluate and quantify the risk of sudden unexplained death in epilepsy. Collaboration with the FDA led to an assessment of sudden unexplained death in users and non-users of lamotrigine. Evidence from a variety of sources enabled an accurate assessment of the risks of this event in lamotrigine use, and contributing risk factors. Serious cutaneous ADRs were also seen with lamotrigine and compared with the risk occurring in other anti-epileptic drugs. Pharmacoepidemiology data contributed much valuable information to this assessment.
Pharmacoepidemiology had also played an important role in other examples of collaborative assessment of safety data, including those associated with HIV therapy, Dr Andrews said. Seven companies had pooled resources and data on 13 HIV products. The FDA encouraged companies with new products to join this registry for antiretroviral products. This had ensured that all products received comparable monitoring. An advisory committee had been formed with expertise from government, companies and academia.
One of the main concerns with HIV therapy had been the transmission of the virus to the foetus. Consequently, one aim of the monitoring was to measure pregnancy outcome data in HIV treatment. This had generated data on 761 cases. Similar collaborative efforts were being used in the assessment of long-term highly aggressive antiretroviral therapy (HAART).
Disease-specific databases which could address multiple ADR signals were also described by Dr Andrews. One example was the EuroSIDA study, a cohort study of HIV patients, monitoring disease progression from 50 sites in 17 countries and with 7,000 patients.
The role of the General Practice Research Database was explained by Dr Louise Wood (GPRD project manager, MCA). The GPRD was the world's largest computerised database of anonymised longitudinal patient records, containing more than 30 million patient-years of data. The GPRD was a critical resource for pharmacovigilance, she said, and had great potential for improving public health.
However, there were a number of problems: declining numbers of general medical practitioners were making reports, the database had deficiencies, there was a lag time between collection and research of data, and access to the data was expensive and resource intensive. To overcome these problems, a new, user-friendly, database with timely access was to be developed. This would support a range of services, possess state-of-the-art analytical capability, and be a tool to support innovative research, Dr Wood said.
The impetus for changes in the provision of health care information came from a wide range of sources. These included public expectations, the growth in the use of the internet, greater freedom of information, an increasingly global regulatory environment, the focus on primary health care, and evidence-based medicine.
According to Professor Alexander Walker (professor of epidemiology, Harvard school of health) pharmacovigilance was based upon the clinically detectable toxic effects of drugs, a prior hypothesis of what adverse effects there might be, a known mechanism, and no alternative causal explanation.
Informal inference of the ADR meant that there was a single observer who recognised a pattern in the treatment producing the ADR. The pattern would be sensitive, non-specific in "noisy" settings, and delayed in the absence of a prior hypothesis.
Factors which created difficulties in detecting spontaneous ADRs include reactions which were delayed, unanticipated, infrequent, readily expected in the absence of a drug, and multiple drug treatments.
Epidemiology could be used as a language for evaluating the risk of ADRs, Professor Walker said. It provided explicit definitions for the concept of risk, measured for comparison, and gave a vocabulary for describing multiple determinants of disease. The ADR event needed to be described in epidemiology, and in the context of drug use, patient characteristics, observational alertness, and the perceived importance for reporting.
Dr David McNamee (executive editor, Lancet) explained the role of the media in trying to communicate the efficacy of a drug in relation to its safety profile. A doctor, journalist, or consumer group had to put complex information about relative and absolute risks, for instance, into terms understandable to the patient or public.
In September, 1996, several parties interested in communicating drug safety issues met in Verona, under the auspices of the World Health Organisation's collaborating centre for international drug monitoring. The meeting recognised the importance and interdependence of all those players involved in the dissemination and understanding of drug safety data.
Academic medical journals were traditionally placed near the end of the line from research to publication as most research results had little immediate impact on medical practice. However, for drug safety data, someone, somewhere, would want to know and act on the findings, and be able to act quickly.
Most editors of medical journals would be willing to work with the other players involved in drug safety issues, especially with regard to the timing of publication. Editors differed about how they would initiate such interaction. While many would be willing to liaise, with the authors' permission, with a regulatory body or funding agency, the pharmaceutical industry was still distrusted in many quarters. Editors might also feel anxious about working with the more strident consumer groups.
Without more openness and trust, and willingness between the players in pharmacovigilance to work together, the communication of drug safety was not going to be improved to any extent, Dr McNamee thought. - Contributed.