|
The analysis of risk provides jobs in many fields, but it isn't an end in itself. At least, it shouldn't be. Its purpose is to inform decision making. More often than not, the risk analyst is not the decision maker, so the results of risk analyses — the risk information — must be communicated, and it is the risk analyst who, at least in the first instance, must communicate them.
Communication is crucial in all endeavors involving more than one person. An important component of every professional's competence — alas, an often overlooked and underdeveloped component — is the ability to communicate clearly and fully. For safety professionals, the requirement extends to the communication of risk information, which, for many reasons, can be problematic. A generally poor understanding of the subject of risk is one reason, and the paucity of our risk vocabulary is another.
Guidance given from the earliest stages of education is to design and package information with the intended recipients in mind.
School pupils are taught to tailor the style, length, tone and format of documents to suit their intended readerships, and not to write an essay on the life of a dollar bill in the same way as they would write a report on a laboratory experiment.
There are many ways of designing and packaging risk information. There are also many potential recipients, each of whom may require the information for a different purpose. For example, engineering managers must decide if systems are adequately safe for deployment, and politicians must consider risks when defining policies. Then there is the public, each individual of which must make decisions based on risk information. Should we use mobile telephones, vaccinate our children, buy a house on a flood plane, or evacuate that house at the threat of a storm?
Risk communication is therefore important. We need to choose what information to communicate and how to communicate it. In doing so, we should ask ourselves what assumptions we can reasonably make about the intended recipients before we base our communication on those assumptions. A scientist may flippantly say that a certain carcinogen in a food or drink, taken each day in moderation, increases the average person's risk of cancer over their lifetime by one in a million. A member of the public may reply, "Yes, but is it safe?"
Communication is two-way, and recipients also bear responsibilities. However "good" or "correct" the communicated information, it is subject to interpretation by the person receiving it. Recipients should be conscious of their own requirements and capable of determining whether or not the received information satisfies them. What decision do I need to make? What information do I require in order to be sufficiently informed? Then, when information is received: What does it mean? Is it what I need?
This article is intended to remind safety professionals of the importance of risk communication, and to illustrate how difficult it is to get it right and how easy it is to get it wrong. The appropriateness of safety-critical (and other) decisions depends on how successfully we achieve it. The article identifies a number of factors that influence communication and shows why we should take account of them.
Appropriateness of Information
Risk information may be inappropriate to the decision that needs to be made because it is wrong. Why it is wrong is another issue; it may be known to be wrong and intended to mislead, or it may have been determined wrongly. At the U.K.'s Bristol Royal Infirmary, between April 1990 and August 1993, a surgeon carried out heart operations on 13 infants to correct Atrioventricular Septal Defects (AVSD). Seven died [Refs. 1 and 2]. Yet the surgeon told the parents of the next patient that the risk of mortality as a result of the operation was 20-25%. Where could such a figure have come from? At the time, not only was the death rate of the surgeon's AVSD-operation patients greater than 53%, but also the trend was not encouraging: the last 3 patients had died, as had 5 of the last 6 and 7 of the last 9. The parents approved the operation and the baby died. Yet the surgeon told the parents of the next patient that the chance of success of the operation by him was 80%.
It is likely that the surgeon was not intentionally dishonest, but gave a figure based on overconfidence in his own capability. As a professional, he should have derived the figure from an assessment of his past performance, but it appears that his culture did not include such analysis. Certainly, however, the communicated information did not represent the facts and was inappropriate to the parents' decisions on whether or not to allow their infants to be operated on by the surgeon.
Timing
A friend of mine was suspected of having a liver problem, and his consultant physician recommended a biopsy. My friend agreed, and a hospital appointment was made for the procedure. On the appointed day, he reported to the hospital and underwent preparation. The doctor arrived, made ready to perform the biopsy, and then, at the last minute, requested my friend to sign a permission form. Then, as he was about to do so, she added that there was certain information that she should impart to him. In the procedure, she said, there was a one-in-a-hundred chance of some internal bleeding, a-one-in-a-thousand chance that the bleeding would require intervention, and a one-in-a-million chance that it would be fatal. My friend noted this, signed the permission form, and the biopsy was carried out.
There are reasons to doubt the figures given, but the issue here is timing. The information was provided too late to be of help in making the decision to undergo the biopsy, for my friend had made that decision long before, when discussing the matter with the consultant physician. For risk information to be relevant, it must be timely. Its dispatcher should be working to a predetermined decision-making schedule, which should have been set by the decision maker.
Late information is a problem in all management situations. It introduces an extra stage of decision making, for it places managers in the position of having to judge whether to make the required decision in the absence of necessary information or to delay it until the information is available. This extra stage can introduce significant risks, for both uninformed and delayed decisions carry potentially calamitous consequences. Frequently, managers in such situations merely curse their luck, but they could do more than that. The added risk can, in many circumstances, be obviated by the decision maker, planning in advance for the necessary information to be available on time.
We need not only to carry out risk analysis, but also to do so with the knowledge of its purpose and its proper schedule.
To Inform or to Reassure?
The purpose of a medical practitioner providing risk information to a patient is, surely, to inform the patient's decision making. It was with this in mind that I engaged in discussion with a medical acquaintance. Referring to the cases of the Bristol Royal Infirmary and my friend's liver biopsy, I asked how medical professionals could justify such substandard risk communication when potential life-and-death decisions depended on it. "You don't believe that that's what the information is for, do you?" asked my acquaintance. "Of course," I said. "What else?" "It's not to inform," she said, "it's to reassure."
I don't believe that the biopsy information given to my friend was to reassure him. Rather, it was provided to satisfy a legal requirement for patients to make "informed decisions." But the AVSD-operation information certainly gives the impression of being intended for reassurance. And that's how it seems to be in many industry sectors. The use of misleading information in the selling (or mis-selling) of financial products, such as insurance policies, pensions and loans, shows that the problem is not confined to medicine. Yet, the dependence of the public on "professionals" is near absolute. The alternative to accepting the information provided as true is for every member of the public to seek a second opinion every time professional advice is required. It is important for us to strive to be both objective and honest in providing risk information.
Misleading Information
It is also important to be honest, for risk information can be communicated to give a false impression. Huff [Ref. 3] retells a story of statistical "proof" provided by the U.S. Air Force that jet flying was safer than flying in conventional aircraft. The proof consisted of figures showing that the death rate, in fatalities per 100,000 flying hours, was higher in ordinary planes than in jets. As the Air Force jets in question carried only one or two persons, while the conventional aircraft carried many passengers, more were at risk per flying hour in the latter. Huff points out that if genuine information had been the aim, deaths would have been presented per passenger-hour or passenger-mile, rather than per plane-hour.
Averages
Risk figures are often given in averages, particularly to the public. Yet, average figures can be misleading. The physician who considers the average patient rather than the one before him is likely to be far off the mark in both diagnosis and treatment.
In civil aviation, it is reported that there is about one fatal crash per million flying hours. This may be useful as a reference point against which to monitor the airworthiness of new aircraft prior to licensing them, but it does not help the public to decide which planes, routes or airlines to use or avoid. As crashes occur more frequently at takeoff and landing than during cruising, numbers of crashes per flight would make it apparent that someone who makes a few long flights is exposed to a smaller risk than someone who travels a shorter total distance on many flights.
Road-transport risk is often presented in terms of an average mortality rate per year, but this fails to inform decisions that ordinary people may wish to make. It does not distinguish between a driver's chance of being killed, killing another road user, and being in an accident in which a passenger is killed. It does not facilitate a decision on what time of the day or night it is safest or least safe to venture out as a driver or pedestrian. It does not inform on what type of person is most likely to have an accident or which types of roads have the highest and lowest accident rates. To propose an average risk figure is to suggest that there is a common risk to all drivers regardless of their gender or age, the time of day, or their geographical location. This is not the case.
next page »
|
