Pharmaceutical risk communication: Sources of uncertainty and legal tools of uncertainty management

Abstract

The purpose of this article is to differentiate between the different functions of pharmaceutical risk communication and make explicit their different legal natures and implications. Risk communication has been generally categorised as a warning act, which is performed in order to prevent or minimise risk. Risk analysis has also drawn attention to the role played by information in reducing uncertainty about risk. Both approaches focus on the safety aspects of communication and its role in harm reduction. However, there are cases where a risk cannot be avoided or uncertainty reduced, for example the side effects associated with pharmaceutical products or when a decision about drug approval or withdrawal has to be made on the basis of available evidence. In these instances, risk communication has other purposes than preventing risk or reducing uncertainty. This article analyses the legal instruments which have been developed to control and manage the risks related to drugs (such as the notion of ‘development risk’ or ‘residual risk’) and relates them to different kinds of uncertainty. These are conceptualised as epistemic, ecological, metric, ethical and stochastic, depending on their nature. By referring to this taxonomy, different functions of pharmaceutical risk communication are identified and connected with the legal tools of uncertainty management.

Keywords:

• risk communication
• risk management
• risk epistemology
• precautionary principle
• residual risk
• development risk
• well-founded suspicion

Notes

 1. Drugs are products under constant testing (Arzneimittel sind Produkte in Dauererprobung) declared the Health Minister Dr. Focke in the ministerial statement for the provision of the German Medicines Act (Arzneimittelgesetz 1976): Scheu 2003, p. 701.

 2. See Osimani et al. for a Bayesian formalisation of this principle.

 3. The problem of underreporting is mentioned by several authors: Di Fabio 1993, p. 116, Schönhöfer 1993, p. 111, Meyboom et al. 1997, p. 383, Hartmann-Besche 1998, p. 124, Räpple 1998, p. 109, Talbot and Nilsson 1998, p. 428. Consistency is flawed and biased towards underreporting also because of contingent factors: fear of possibly incurring in legal responsibility, and time constraints in the ambulatory workflow.

 4. A statistical methodology which can help with the integration of data coming from heterogeneous sources is the Bayesian paradigm: see the proposal advanced by Waller and Evans 2003, and author & Russo, forthcoming.

 5. See also Räpple 1998, pp. 50–57.

 6. The distinction avoidable-unavoidable is used in the legal literature also in a different sense. This is when a risk can be avoided through intervention or precaution. This is related to measures which can be taken in order to minimise or prevent unexpected risks (development risk) or known residual risks. Among these measures, communication plays a fundamental role, which however differs from the function it has in the declaration of residual risk. This distinction will be treated in the last part of the paper.

 7. Residual risk may have a different threshold in drug approval and drug prescription. In the decision concerning drug marketing, the benchmark for comparing benefit and risk refers to the relevant population. Instead in the single prescription, the doctor should compare the individual risk faced by the single consumer to the expected therapeutic outcome on the basis of his specific health profile and disease condition. Therefore residual risk splits into product and therapeutic residual risk.

 8. See Hughes et al. 2007 and Temple 2007 for an example of comparison between benefits and risks of terfenadine, chlorpherinamine and loratidine.

 9. The risk of contact allergy to neomycin sulfate seems for instance to vary with age and increase with the number of additional positive reactions to other standard series allergens (Menezes de Pádua et al. 2005).

10. ‘Ethnic group may act as a marker for underlying genetic or environmental differences in this susceptibility’ (McDowell et al. 2006, pp. 1, 4). Relevant cases with regard to this point are the recent approval of isosorbide dinitrate plus hydralazine (BiDil) limited to the use in black patients (McDowell et al. 2006, p. 1).

11. A meta-analysis on the literature devoted to the investigation of the relationship between ethnicity and ADRs to cardiovascular drugs has for instance delivered the result that the relative risk of angio-oedema from ACE (angiotensin converting enzyme) inhibitors in black patients is 3.0 compared to non-black patients (C.I. 95%: 2.5 to 3.7); the relative risk of cough in concomitance with ACE inhibitors is 2.7 when compared between East-Asians and white patients (C.I. 95%: 1.6 to 4.5); and the relative risk of intracranial haemorrhage with thrombolytic therapy is 1.5 (C.I. 95%: 1.2 to 1.9). (McDowell et al. 2006, pp. 2–3). See also Henry et al. 1996 for gastrointestinal reactions to NSADs.

12. In the US, the theory of the learned intermediary relates precisely to the inadequacy of product information as a basis for individual prognosis: the doctor intermediates between the pharmaceutical firm and the drug consumer by selecting the appropriate product on the basis of an individual risk–benefit assessment and is also supposed to be in the best position for selecting the appropriate information for the consumer about his personal risk–benefit prognosis. See Ferguson 1992 for discussion. See also Calman 1996 and more recently, Author's, 2007, chapter 6, for a general discussion on the problems related to communicating probability of risks in these settings.

13. This is the topic of so called ‘sensitivity analysis', the branch of decision theory devoted to the estimation of information value conditional on its capacity to change the decision.

14. Different speech acts with the same illocutionary point can be distinguished on their turn through ‘operations’ on the seven components constituting the illocutionary force. For example reporting differs from asserting because of the different propositional condition that the event reported be either in the past or in the present, but not in the future (Searle and Vanderveken 1989, p. 187); requesting differs from ordering for the position of authority of the speaker (mode of achievement) (Searle and Vanderveken 1989, p. 201); promising differs from threatening for the presupposition (preparatory condition) that the state of affair described by the propositional content is good for the hearer, whereas in the threat this is bad (Searle and Vanderveken 1989, p. 203).

15. Author's 2007 presents a distinction of the responsibility spheres (patient, doctor, pharmaceutical firm) concerning drug damage in relation to information duties.