The IMMP is both pro-active and non-interven-tional, which offers a number of advantages. It is best suited for drugs newly released in order to gain early information, but can be used for the study of signals generated by spontaneous reporting (Coulter and Edwards, 1990). In the context of the IMMP, a much higher reporting rate is achieved with PEM methodology than with intensified spontaneous reporting. At times PEM is used only in certain regions and the programme then relies on intensified spontaneous reporting for the other regions. The information is gained in the "real world" of normal clinical practice where the use of drugs is subject to many influences not present in the artificial world of controlled clinical trials, e.g. the presence of diseases other than the one under study with the trial drug and the use of multiple drugs concurrently. It provides data that are not collected in the standard spontaneous reporting schemes, e.g. rates, but the two methods are complementary in pharmacovigilance. When a drug has more than one indication, this information is collected for the whole cohort and rates are calculated for each indication. The demographics of the cohort of patients prescribed the drugs are routinely established, as are the doses used. Risk can be measured and risk factors identified. Observations can be made on the full range of patients, some not usually included in clinical trials —the young, the old, the pregnant. Some of these patients may be outside the age range for approved use. With longitudinal monitoring, trends in prescribing can be seen and there is a greater chance of identifying longer term outcomes or events such as pregnancy or death. Reasons for cessation of therapy provide information routinely on deaths, compliance and efficacy. Death rates can be calculated.
The IMMP is also highly adaptable. It has been used successfully to study drugs used long term for the treatment of chronic diseases and for drugs used intermittently, e.g. for the relief of migraine. It has also proved to be a good method for the study of the adverse effects of intra-uterine devices. The questionnaires can incorporate requests for specific information for particular drugs, e.g. severity of asthma, tests of liver function. The cohorts can be used for investigation of signals and there is potential for their use in the identification of reactions of long latency using record linkage.
The IMMP is not suitable for the study of rare reactions where very large cohorts are required to provide adequate statistical power. In these circumstances the method is not cost effective. Neither is it suited for drugs that are unlikely to be used widely, particularly if their use is restricted to small specialist groups or hospital inpatients.
The full safety profile of new drugs just released is unknown and it is to the benefit of all concerned to get good information on safety as quickly as possible. There is scope for PEM methodology to be used much more widely and pooled information from several centres would provide added value. Countries with a large population who feel unable to mount a national scheme, could use the method regionally. NZ has a population of only 3.8 million, but the IMMP has been able to produce results that have made a significant contribution to pharmacovigilance. Countries struggling to obtain a "worthwhile" number of reports from their spontaneous reporting programmes could switch resources to PEM studies of a few selected drugs.
Experience in NZ has shown that PEM stimulates awareness and interest in drug safety. An immediate effect of the IMMP was a doubling of the number of reports in the spontaneous reporting programme. The cost-benefit ratio of getting very good information on a small number of drugs important in a particular society would seem to be more favourable than getting very little information on a wide range of drugs. The cost of doing this is not high. Most of the IMMP work has been done on a budget of less than US$150 000 per annum providing for a half time physician and a small group of support staff. The use of modern information technology and simple epidemiologi-cal methods facilitates this cost effectiveness.
Was this article helpful?