The language of risk: a question of trust

Risk—its understanding and its communication— can be complex and difficult. This is particularly the case when the risk is unknown and unquantifiable. It involves the probability that a particular hazard will cause a specific effect, and this can be expressed in many different ways: for example, as a percentage (1% risk), as an absolute (1 case/100), as a comparative (high, low), or as some other form of comparison (1 person in your street, the risk of lightning striking, etc.). At the most basic, there is a need to agree on some form of common language to assist in communication with patients or the public.1-3 Communication should of course be seen as a twoway process, and listening to the patient or the public, to their fears and concerns, is as important as the information to be transmitted. To allow better communication, several factors are relevant in determining the risk, and these should be clarified before communication begins. First, is the risk biologically plausible? Could it conceivably happen? Can a hypothesis be developed to account for the potential that a hazard will cause an effect? Second, is the hypothesis testable? Is it possible to measure the hazard and/or effect on the patient or the population? There are two components to this. First, how certain are we that the hazard will cause an effect? Second, what is the level of risk (high or low)? Both of these components are relevant to the communication of risk. Third, what is the impact of the risk? What are the benefits of the procedure? Are the benefits of taking the risk greater than those of not taking it? What are the possible longterm as well as short-term consequences? What would happen if the procedure was not carried out? Fourth, are there ways in which the risk can be reduced, though not necessarily eliminated? From the point of view of risks associated with blood transfusion, the answers to these questions are particularly relevant, as they provide some of the building blocks for the process of communication. By using these factors, a range of scenarios can be developed. At one end of the spectrum is the case of a known hazard that can be measured and whose risks and benefits can be assessed. At the other end would be the case in which, although the hypothesis may be plausible, there is limited or no evidence of a link, the agent cannot be measured, and, thus, estimates of risks cannot be assessed. A hazard such as hepatitis B would equate to the first scenario, and each point listed above can be related to it. It is biologically plausible (indeed it is known) that the virus can be transmitted in blood. It can be measured in blood, and thus an individual donation or a donor population can be screened and estimates of the risk can be made. There are a number of interventions that can be taken to eliminate the risk, and it is possible to work out the benefits for an individual patient, including the cost. Long-term and short-term consequences can be identified. Finally, the individual patient, having listened to the evidence, can make a choice and give consent to the transfusion being given, which is the end point of the whole process. Consent is a central issue in the process of risk communication, and it is discussed in more detail later. An example at the other end of the spectrum would be variant CJD (vCJD). There is little doubt of the biologic plausibility that prions can be transmitted by blood, but it is not clear whether they are thus transmitted. The problem is that prions still cannot be measured routinely, and thus the certainty, level of risk, and impact cannot be assessed. It is still possible, of course, to intervene by taking certain precautions, such as not giving the patient a blood or plasma transfusion. This would be an appropriate use of the precautionary principle, but it will not be possible to know if the risk has been reduced or eliminated, and that intervention may not be for the benefit of the patient. The presentation of evidence at this level of uncertainty is a very difficult matter. In these circumstances, the “expert” may have little to add, as there is “no evidence” on which to base a decision. Such decisions are thus taken as a matter of judgment on the basis of nonscientific evidence. Such decisions should be made as openly and transparently as scientific decisions. There is a third category that falls between the two on the spectrum, in which there may be some evidence of an effect from a particular hazard, but it is not conclusive. This category is best dealt with as a subset of the second category (vCJD). But, first, a word is required about the purpose of the communication of risk.4 Its purpose is to present to the patient or the population information about the risk (the relationship between a hazard and an effect) so that a choice can be made about the use of an intervention. This would involve the giving of information about both the adverse efThe language of risk: a question of trust