For four decades, spontaneous reporting has been the main mechanism by which adverse drug reactions are identified after a medicine is released onto the market, and the Australian program has been exceptionally effective. However, spontaneous reporting programs are limited in their ability to identify an association between a drug and an outcome that is common among the users independent of drug use, and they are not sufficiently sensitive to detect a small increase in the risk of certain rare events. In particular, the increasing long-term use of medication for prevention and control of chronic disease in otherwise healthy individuals presents a challenge to which current postmarketing surveillance mechanisms cannot effectively respond. For example, spontaneous reporting cannot detect an increased rate of myocardial infarction associated with hormone replacement therapy, rosiglitazone or rofecoxib; demonstration of these associations requires large, long-term randomised controlled trials.

In recognition of these limitations, the Therapeutic Goods Administration (TGA) recently announced administrative changes that involve a strengthening of pharmacovigilance.1-3 As part of the pre-approval process for a medicine, sponsors will be required to present postmarketing pharmacovigilance and risk minimisation plans. In some cases, the pharmacovigilance plan may simply describe routine passive pharmacovigilance activities. In other cases, where there are important gaps in safety data (eg, groups of potential users for whom documented exposure is inadequate) or where data suggest potentially significant safety issues, the plan will propose studies that specifically address these matters. Studies may use any of the full gamut of epidemiological approaches, including randomised controlled trials, registries of new drug users, and data-linkage and case–control studies.

Risk minimisation plans will outline practical steps that will be undertaken to reduce the risk of known safety hazards during the postmarketing period. Possible steps include a small pack size, second-line use only, and advice given in the product information; these measures were available options in the past, but they will now be listed in a single document. Pharmacovigilance and risk minimisation plans may be updated at any time during the postmarketing period, if a need to do so is identified. In addition, sponsors of medicines that are already registered when the new measures are implemented may also be required to draw up suitable plans on a case-by-case basis.

These changes bring Australia into line with overseas initiatives, particularly those in the United States and Europe, and involve adoption of guidelines prepared by the European Medicines Agency.4,5 The result has been termed a “whole-of-lifecycle” approach, where active investigation of the safety of a medicine may continue for as long as it is in clinical use.1 A newly constituted medicines safety committee will have responsibility for providing advice on pharmacovigilance and risk minimisation plans, and overseeing the spontaneous reporting program.

Medicines prescribed over a long period for disease prevention have typically been approved on the basis of a measured response to a surrogate marker for efficacy, such as improved glucose control, decreased cholesterol level or lowered blood pressure. However, the critical measures of the benefits of such responses to medicines are changes in risk of hard end points, such as myocardial infarction, renal failure and death. These end points are outcomes of the multidimensional effects of medicines — beneficial and adverse — on body systems. Marketing approval of medicines has rarely, if ever, required demonstration of such hypothesised benefits. Under the administrative changes, promising new medicines may still be approved on the basis of surrogate data, but consideration will be given to pharmacovigilance plans that include long-term studies with cohorts that are large enough to determine differences in hard clinical end points and identify any serious adverse effects that may offset benefits.

In 2004, Merck Sharp and Dohme precipitately removed rofecoxib from the market worldwide, because evidence of a significantly higher rate of myocardial infarction than with placebo was an incidental outcome of a randomised controlled trial. There had been prior evidence of this association in the VIGOR (Vioxx Gastrointestinal Outcomes Research) study published in 2000.6 At that time, the manufacturer made a small protocol change to the already commenced APPROVe (Adenomatous Polyp Prevention on Vioxx) study, to allow participants to take low-dose aspirin,7 but apparently took no action to verify the association before 2002.8 Under the new procedures, it will be a routine matter for medicine regulators, including the TGA, to ensure that the sponsor conducts suitable studies to clarify safety concerns such as those raised in the VIGOR study.

Of critical importance to the success of the expanded approach to pharmacovigilance is the level of resourcing at the TGA. Consideration of requirements must include capacity to: evaluate the suitability of pharmacovigilance plans that are included in applications for marketing approval of medicines; review pharmacovigilance plan updates that are submitted by sponsors; identify areas that require further investigation during the lifecycle of a medicine; and prepare reports for the medicines safety committee.

It is disappointing that the initiatives do not include provisions for the TGA to fund, commission or conduct active safety investigations. The US Food and Drug Administration has received a substantial injection of funding for this purpose,4 and Medsafe, the New Zealand medicines authority, is funding product vigilance research through a joint initiative with the Health Research Council of New Zealand.9 Commissioned studies could address questions that sponsors could not reasonably be required to ask, such as those related to genetic markers for hypersensitivity reactions. They could also provide an alternative source of data in cases where the manufacturer chooses not to be transparent about safety matters, as allegedly occurred with rofecoxib10 and cerivastatin.11

As these changes in pharmacovigilance and drug regulation are occurring internationally, in general the TGA will be requiring postmarketing studies in concert with other regulators. In the absence of local initiatives or leadership from the TGA or other federal agencies that encourage the conduct of pharmacovigilance research in this country, Australia may not do much better than to achieve high-level observer status. Linkage of de-identified data from the Pharmaceutical Benefits Scheme and National Cancer Statistics Clearing House, as well as morbidity and mortality data from the Australian Institute of Health and Welfare, would provide a platform for epidemiological studies. Studies using this platform would need to be ones that do not require data on disease severity, concurrent disease, non-subsidised medication or lifestyle factors to control for confounding. While studies using more time-intensive means of data collection should also be encouraged in Australia, the years of inaction with regard to advancing this data linkage cannot be justified.12

The new approach being implemented by the TGA is welcome. Supplementing passive spontaneous reporting with evidence from active pharmacovigilance, in the form of studies designed to address important safety issues associated with medicines, will lead to more timely elucidation of adverse effects, together with reassurance for prescribers and patients. But the failure to include, as an essential part of the proposed changes, a package to ensure that active pharmacovigilance occurs in this country is baffling. This deficiency needs to be addressed urgently, and the matter would benefit from being taken up at a high level of government.