Perspective

Safe Drugs and the Cost of Good Intentions

Hans-Georg Eichler, M.D., Eric Abadie, M.D., June M. Raine, M.D., and Tomas Salmonson, Ph.D.

N Engl J Med 2009; 360:1378-1380April 2, 2009

Article

In recent years, the pendulum of opinion on the health benefits and risks of drugs has swung back from what were often unrealistically high hopes to much grimmer expectations. This shift has resulted from several high-profile incidents related to the drugs' risks and the perception that drug companies downplayed safety findings. Not only has the reputation of the pharmaceutical industry become tarnished as a result, but the performance of drug-regulatory agencies has also come under renewed scrutiny. Some observers have concluded that the entire regulatory system must be revamped because regulators have failed to apply sufficient counterpressure on companies.

Legislators and drug-regulatory agencies in many jurisdictions have taken heed of the calls for a stronger emphasis on drug safety. By 2003, the European Union had instituted a proactive risk-management strategy. In 2005, new European legislation conferred on regulatory agencies the power to require companies to submit, along with their application for marketing authorization, a risk-management plan comprising detailed commitments for postmarketing pharmacovigilance. Other jurisdictions, including the United States, have introduced or are considering similar legal provisions.

These legislative and regulatory changes have been complemented by research initiatives designed to improve safety-evaluation methods. Regulators will probably soon have at their disposal substantially more powerful tools with which to scan the horizon for adverse drug effects, including ever-larger clinical trials, more widespread use of (preplanned) meta-analyses of trials of individual drugs or drug classes, and results from observational studies involving electronic records that link drug-use data to health outcomes for millions of patients. Sophisticated data-mining tools are now available that permit real-time analysis of spontaneous reports of adverse drug reactions.

It would be surprising if these powerful tools did not lead to the detection of new drug-safety signals for both new and well-established drugs. These developments are good news — but there are a number of unintended but foreseeable consequences that will need to be addressed before these advances can be translated into real public health benefits.

First, the number of false safety signals is likely to increase. The interpretation of meta-analyses of adverse drug events is not always straightforward, and observational studies may sometimes produce unreliable findings implying the presence of risks that are later proven to be phantoms. All initial safety signals require further verification to justify regulatory action. That is where the difficulty starts: for example, if the validity of observational findings is questioned, it is usually not because adverse events (especially those in large observational studies) may be attributable to chance, but because of potential bias or confounding.1 Ideally, safety signals would be validated and quantified in randomized, controlled trials, but this approach is often impossible because of the size and duration of the trials that would be required to detect rare adverse drug reactions. So we are often left with very rare but worrisome adverse effects, the risk of which cannot be convincingly verified or refuted without great cost or delay. Such scenarios can result in the loss of public trust and, sometimes, harm to public health from the underuse of potentially lifesaving vaccines and drugs. If regulatory decisions were made on the basis of spurious safety signals, we would end up denying patients — often those with life-threatening conditions — access to drugs even if their overall benefit–risk profiles were favorable.

Second, the push for “safe drugs” means that regulators are coming under increasing pressure to raise the bar and to refuse, restrict, or revoke drug licenses. Some may see such increased stringency as an intended consequence, but would it reflect society's preference? There is a common belief that society is becoming increasingly risk-averse, being less willing to accept risk for a given benefit. But we have not seen convincing evidence of such a change with respect to drugs. Calls for “safe” products can be inappropriately interpreted as demanding zero risk. What if patients and consumers have not become more risk-averse but merely more risk-aware? Media coverage (among other things) can disproportionately increase the public visibility of adverse events, inducing biased judgments.2 If we have merely become more aware of risks as a result of more widespread reporting of the downside of drug treatment, then denying or delaying access to drugs could lead to more, not fewer, disconnects between regulators and their primary stakeholders, patients and health care professionals.

Third, simply putting more emphasis on drug safety may well fail to buy back trust in drugs. Indeed, it is difficult to conceive how increasing the profile of adverse-effect findings and related communications and adding to the media coverage would reassure patients, consumers, and health care professionals. In fact, we may instead witness an upward spiral of risk awareness, in which better pharmacovigilance tools will detect more safety signals (some true and some false), which will draw more attention to the downside of medicines and, in turn, give rise to a call for greater emphasis on safety and better pharmacovigilance tools, and so on.

We believe that such a negative scenario is preventable, if the more intense surveillance of drug safety is complemented by additional measures. For instance, regulators need to refine their methods of assessing benefit–risk balances and switch from “implicit” to “explicit” decision making — that is, to an approach involving explicit descriptions not only of all decision criteria and interpretations of data but also valuations, such as the weighting factors for potential treatment outcomes.

Ideally, regulators should also shift from the use of qualitative statements to quantitative descriptions of the size of the net health benefits. We believe that such refined assessments would provide greater transparency and consistency to the decision-making process. This improvement would provide an opportunity to systematically incorporate patients' values and preferences into decisions and create a better communication tool, enabling regulators to publicly justify decisions.

Drug regulators should also lead the way in developing a consensus on what is considered a tolerable level of risk. Experts in other areas of risk regulation, such as the regulation of carcinogenic residues in food or of nuclear power, have long recognized that calls for zero tolerance of risk are untenable and have instead established a consensus definition of “negligible” and “tolerable” risk levels. There is evidence that patients are willing to accept a finite level of risk for a given benefit. For example, a recent survey of patients with multiple sclerosis found that a majority would “definitely” or “probably” use a treatment that was “significantly more effective than currently available drugs,” even with a 1-in-1000 chance of a fatal side effect.3 This information has been factored into a quantitative benefit–risk analysis of natalizumab4 — work that offers an interesting case study for regulatory decision making about drugs with a low probability of serious adverse effects. We believe that the public health would be better served if patients' views of acceptable risk levels were incorporated into benefit–risk assessments.

At the European Medicines Agency (EMEA), we no longer use terms like “ensuring drug safety” in public communications, instead striving to ensure a “positive benefit–risk profile” — a phrase implying the concept of tolerability of risk. Indeed, we consider the debate over “safe drugs” to be unhelpful, since words trigger expectations — and every drug is unsafe under some conditions. We also believe that regulators should begin to communicate about benefits and risks, rather than just risks — by presenting, in quantifiable terms that patients can understand, the benefits of taking a drug and the odds of experiencing adverse effects. Currently, companies, though required to communicate risks, naturally focus on promoting their drugs' benefits, whereas regulators emphasize the risks — leaving patients and health care providers searching for balanced information. Historically, regulators shied away from communication about benefits for fear of being perceived as doing the industry's marketing, and certainly more detailed emphasis by regulators on drugs' health benefits will draw some negative responses. But the alternative seems to be a lack of comprehensive and balanced information on drugs, progression of the spiral of risk awareness, and further erosion of the public trust in drug treatments.

In addition to shifts in regulatory behavior, some changes are needed in physicians' prescribing practices with respect to off-label use of drugs.5 Prescribers need to be fully conversant with a drug's license and need to discuss with their patients the probable risks and benefits of an off-label use. We also foresee a role for third-party payers, who need to carefully evaluate when to fund off-label drug treatment.

As the early-20th-century satirist Kurt Tucholsky wryly noted, “The opposite of `good' is `good intention.'” Although uncritical calls for “ensuring drug safety” are probably well meant, they may be doing harm as well as good. If, however, we make the most of emerging pharmacovigilance tools and new legislative provisions, we should be able not only to refine benefit–risk assessments but also to maximize the public health benefit of new medical treatments.

The views expressed in this article are the personal views of the authors and do not necessarily represent the position of EMEA or any other regulatory agency.

Drs. Eichler, Abadie, Raine, and Salmonson report being full-time employees of drug-regulatory agencies. No other potential conflict of interest relevant to this article was reported.

Source Information

Dr. Eichler is the senior medical officer of EMEA, London. Dr. Abadie is the chair of the Committee for Medicinal Products for Human Use (CHMP) of EMEA and scientific advisor to the General Directorate at the Agence Française de Sécurité Sanitaire des Produits de Santé, Saint-Denis, France. Dr. Raine is the chair of the Pharmacovigilance Working Party of CHMP and director of vigilance and risk management of medicines at the Medicines and Healthcare Products Regulatory Agency, London. Dr. Salmonson is the vice chair of CHMP and senior expert at Läkemedelsverket (the Medical Products Agency), Uppsala, Sweden.

References

References

1. 1

   Vandenbroucke JP. Observational research, randomised trials, and two views of medical science. PLoS Med 2008;5:e67-e67
   CrossRef | Web of Science | Medline

2. 2

   Fischhoff B. Risk perception and communication. In: Detels R, Beaglehole R, Lansang MA, Gulliford M, eds. Oxford textbook of public health. 5th ed. Oxford, England: Oxford University Press (in press).
   

3. 3

   Calfee J. A representative survey of M.S. patients on attitudes towards the benefits and risks of drug therapy. Washington, DC: AEI-Brookings Joint Center for Regulatory Studies, 2006.
   

4. 4

   Thompson JP, Noyes K, Dorsey ER, Schwid SR, Holloway RG. Quantitative risk-benefit analysis of natalizumab. Neurology 2008;71:357-364
   CrossRef | Web of Science | Medline

5. 5

   Gazarian M, Kelly M, McPhee JR, Graudins LV, Ward RL, Campbell TJ. Off-label use of medicines: consensus recommendations for evaluating appropriateness. Med J Aust 2006;185:544-548
   Web of Science | Medline

Citing Articles (9)

Citing Articles

1. 1

   Zhong Yuan, Bennett Levitan, Jesse A. Berlin. (2011) Benefit-risk assessment: to quantify or not to quantify, that is the question. Pharmacoepidemiology and Drug Safety 20:6, 653-656
   CrossRef

2. 2

   Hans C. Ebbers, Aukje K. Mantel-Teeuwisse, Ellen H.M. Moors, Huub Schellekens, Hubert G. Leufkens. (2011) Todayʼs Challenges in Pharmacovigilance. Drug Safety 34:4, 273-287
   CrossRef

3. 3

   Enrique Ginzburg, Franck Dujardin. (2011) Physicians’ perceptions of the definition of major bleeding in major orthopedic surgery: results of an international survey. Journal of Thrombosis and Thrombolysis 31:2, 188-195
   CrossRef

4. 4

   P M Coplan, R A Noel, B S Levitan, J Ferguson, F Mussen. (2011) Development of a Framework for Enhancing the Transparency, Reproducibility and Communication of the Benefit–Risk Balance of Medicines. Clinical Pharmacology & Therapeutics 89:2, 312-315
   CrossRef

5. 5

   Erik Tambuyzer. (2010) Towards a framework for personalized healthcare: lessons learned from the field of rare diseases. Personalized Medicine 7:5, 569-586
   CrossRef

6. 6

   Serena Frau, Maria Font Pous, Maria Rosa Luppino, Anita Conforti. (2010) Risk Management Plans: are they a tool for improving drug safety?. European Journal of Clinical Pharmacology 66:8, 785-790
   CrossRef

7. 7

   Deborah A. Wing, Barbara Powers, Durlin Hickok. (2010) U.S. Food and Drug Administration Drug Approval. Obstetrics & Gynecology 115:4, 825-833
   CrossRef

8. 8

   Rebecca Dresser, Joel Frader. (2009) Off-Label Prescribing: A Call for Heightened Professional and Government Oversight. The Journal of Law, Medicine & Ethics 37:3, 476-486
   CrossRef

9. 9

   J Vincent. (2009) World Health and Bioterrorism. Clinical Pharmacology & Therapeutics 85:6, 561-565
   CrossRef