Paradigm Shift in Pharmacovigilance

Joanna Haas, Vice President of Pharmacovigilance at Genzyme, delves into the driving forces behind the changing nature of safety risk management, including the dissolution between pre- and post-marketing safety activities and the sources and usage of safety information.

The pharmaceutical industry is experiencing a paradigm shift in the assessment and management of patient-related safety issues. This shift is a consequence of a new appreciation for the importance of subtle safety issues not identifiable by traditional safety monitoring techniques.

As part of these changes, pre- and post-marketing safety, once viewed as separate, must now be viewed as a continuum of safety activities. The availability of electronic health records provides new sources of safety information that can be explored for evidence of potential safety concerns. A discordant framework for interpreting potential safety signals, communicating risks and making appropriate decisions about reducing risks has led to public distrust and uncertainty. The management of safety risks is a central concern for the pharmaceutical industry, as safety issues must be evaluated and managed throughout the life of each product.
 
Shift in paradigm
Thalidomide, sold as a sedative in many countries in the mid-1950s, led to the dramatic and virtually pathognomonic congenital abnormality phycomelia, characterized by underdeveloped, flipper-like limbs. Thalidomide had not been commercialized in the US due to the resistance of Frances Kelsey, a reviewer for the FDA; as a result, the US was largely spared the epidemic of thalidomide births. For over 45 years, the method for preventing serious drug safety problems due to pharmaceuticals was modeled on the “thalidomide experience.”

In this context, an alert practitioner recognizes a serious adverse drug reaction.  A system of surveillance, collection, evaluation and reporting then funnels the observations of clinicians into a central repository where patterns of abnormalities can be recognized. A certain level of triage permits rapid sorting of the most important information. The worst reactions (serious) are communicated quickly, particularly if they are fatal or life-threatening. Similarly, new (previously unappreciated) adverse reactions are identified as more important than those that are already known. Hence, the processing of individual safety case reports, communicated by a medical practitioner to a pharmaceutical company and by the pharmaceutical company to the FDA, forms the backbone of the post-marketing safety system for pharmaceutical products. This process continues to be valuable for significant and distinctive conditions that can be readily associated with specific drug products.

More subtle safety issues

The “thalidomide model” works and continues to work for problems that share the characteristics of thalidomide and congenital abnormalities: a distinctive clinical picture, a low background rate in the absence of the suspect drug and a temporal relationship that favors identification of a causal association between exposure and adverse experience. However, not all safety issues have these characteristics, and when one or more is absent, individual case reports have limited value in suggesting a link, let alone establishing one, between a drug and a side effect.

The relationship between Vioxx and heart disease provides an excellent illustration of the limits of individual case reporting. Cardiovascular disease is the leading cause of mortality and morbidity in the general population, and cardiovascular events related to Cox-2 inhibitors are in no way different from other cardiovascular events. A modest increase in the background rate of such events in patients exposed to a drug would not be evident, particularly if the risk increased with long-term exposure to the product. Individual case reporting is not designed to identify this kind of safety issue.

Even more difficult to interpret is a purported adverse drug reaction that mimics the natural history of the underlying disease being treated. There are two hallmark examples of this sort of safety problem: asthma mortality in relation to drugs used to treat asthma and suicide in relation to antidepressant medications. Since the 1960s, concerns have been raised about a potential increase in the risk of death among asthmatic patients using beta adrenergic receptor agonists to treat bronchospasm due to asthma. Acute asthma death in that setting would be clinically indistinguishable from death due to a severe asthma attack in the absence of treatment, and the use of potent drugs is most common in patients at the highest risk for asthma deaths.  Individual case reports would not be able to separate cause and confounder.

Similar difficulties complicate the evaluation of a possible relationship between treatment with antidepressants and suicidal ideation among adolescents prescribed antidepressants. In this situation, the very individuals considered most at risk for the adverse outcome (suicide or suicidal behavior) are those most likely to be prescribed drugs, as opposed to other forms of therapy. Case reports of suicidal ideation in patients on SSRIs could not be readily interpreted because the adverse experience of concern is part of the natural history of the indication for which the drug is prescribed. This phenomenon, in which the side effect is also a manifestation of the condition being treated, is referred to as “confounding by indication.”

Assessment of risk in these situations cannot be based on the clinical exercise of evaluating a series of individual case reports. The dissection of cause and effect can be complex and requires other approaches that may include interventional studies, such as large simple safety studies, or observational studies, particularly those using pharmacoepidemiological designs. Such studies have not been part of the standard pharmaceutical armamentarium. They require specialized expertise and suitable data sets and can be costly and time consuming. 

Pre- and post-marketing safety

After the first product approval, in the first indication, in the first market, traditional distinctions between pre- and post-marketing safety begin to fade. The ICH Pharmacovigilance Planning document (E2E) provides a bridge between the Common Technical Document (CTC) and subsequent safety information collected while the product is on the market. In this setting, it is impossible to exclude safety information or to discount it because of its source; this applies to pre-clinical and clinical data as well as to post-marketing safety information. This continuum of safety information leads to a life-cycle approach to the management of safety issues and puts a premium on building good communication among various parts of the organization with respect to flow and interpretation of information and recommendations for action. The implications for the organization are significant.

Computerization of relevant safety information

Computerization of medical and safety information is accelerating the move away from dependence on individual safety case reports as the primary source of safety information. For decades, clinical development has also adopted the “thalidomide model” of individual case reporting, as established by US and international regulatory reporting systems.  In the clinical development setting, the use of the electronic case report form (eCRF) fundamentally changes the nature of safety monitoring. Expedited reporting of serious cases is no longer the only way to identify safety problems within a clinical trial. Non-serious information accrues in parallel, and provision must be made for its review to identify any relevant safety data. In addition to risk assessment, the idea of risk-minimization, while originally developed as a post-marketing concept, can and should be applied in the clinical development setting.

It is, however, in the post-approval context that the prospective impact of information technology is greatest. Large quantities of information potentially relevant to patient safety are collected by a number of systems for a variety of purposes. For example, health claims data are available for insured patients in several systems. Electronic medical records supporting direct health care delivery contain information relevant to drug safety, even though they are collected for quite different purposes. While there are obstacles, the main problems involved in utilizing these systems of information have to do with how such systems will be supported, used and interpreted – in short, how these systems will be governed.

A variety of forces – including the extension of Medicare to cover drugs and the rising costs of medical care overall – appear to be converging to overcome the problems of creating an information system able to provide data on drug safety. Who will have access to these resources? How will research questions be formulated and results presented and evaluated? Unlike the individual safety case report, these data would, for the most part, be out of the hands of the pharmaceutical industry.

Discordant framework for interpretation

Analysis of safety information is open to more dispute and interpretation than is the analysis of efficacy data (although when one begins to speak of effectiveness in the post-marketing setting, many of the same issues emerge).  This is because safety issues are rarely resolved with the help of a well-defined endpoint and a clear decision-making paradigm. Add to this the emotional language of risk and the difficulty that patients have in evaluating the risks of drugs and the process of interpreting findings and communicating results is fraught with potential disagreement.

Pharmaceutical companies have traditionally provided most of the information on which safety assessments are based. This may change to the extent that data from large external databases is increasingly important in assessing risks. Nonetheless, overall, the pharmaceutical sponsor will generally have the deepest understanding of its product and of the strengths and limits of the information on which assessments rest. However, the sponsor also has potential financial conflicts of interest. Other stakeholders in the regulatory process have potential conflicts of interest as well, but these are often less visible or at least less readily quantifiable.

Managing risks throughout the life of a product

These shifts in thinking about safety alter the management of known and potential human safety risks during clinical development and after commercialization. Activities that may previously have been delayed as post-marketing commitments may need to be anticipated, perhaps even tested, during clinical development. The effect is to extend risk assessment, risk management and risk minimization to an earlier point in clinical development. Organizational flexibility, increased collaboration and reduced territoriality are required to accommodate these changes.

About the contributor

Dr. Joanna F. Haas joined Genzyme in 2000 after spending 15 years working in clinical development and pharmacovigilance for leading pharmaceutical companies in France, Germany, Austria and the US. She is a recognized leader in the management of human safety risks and has participated in industry and government collaborations in the US and Europe to optimize pharmacovigilance regulations and standards.