The very concept of risk is tricky. To you and me, it means pretty much what the dictionary says…the probability that something bad might happen. And to us, the part of that definition that most influences how worried a risk makes us feel is the subjective ‘bad’ part, more than the objectively quantifiable likelihood. We are emotional survivalists more than we are probabilistic rationalists.

But when it comes to government risk management policy making, shouldn’t we want the probabilities to matter more and the emotions less? The goal of risk policy, after all, is to maximize protection of public and environmental health in a world of limited resources. Doesn’t it seem like the best way to do that is to be more objective, more evidence-based, more rational, to focus on the facts, not our feelings?

A lot of governments try to do just that, using what is called a risk-based approach to policymaking about threats to public health, which basically just means - get all the facts so you can objectively figure out what’s best. But as appealing as that rational approach may seem, policy makers don’t operate in some ideal technocratic vacuum of pure reason. Risk managers have to respond to public passions too, even when our fears exceed what the evidence warrants, or fall short of what the evidence warns. So they also use what’s known as a hazard-based approach, which basically means taking action against a potential threat as soon as there is a reasonable hint of a possible hazard, and people are worried, way before all the facts are in.

This conflict between a risk-based and hazard-based approach to risk management is much more than a semantic issue relevant only to technocrats and wonks. We depend on government to protect us from what we can’t protect ourselves from as individuals, and when those risk managers make these decisions, your health and mine are at stake. So to help illuminate this hidden but important policy question, here are the basics about these two approaches to risk management, and then a look at how the battle over which approach is best is currently playing out in one controversial issue, the fight over the regulation of Bisphenol A (BPA).

The RISK-BASED approach starts with this formula:

Risk (the probability of a negative consequence) = Hazard X Exposure.

Important details about Hazard include…

  • Is the ‘thing’ (a product, an ingredient, a process, a behavior) hazardous?
  • How hazardous? (severity)
  • In what ways (what does it do to us?)
  • At what dose? (This applies to things we ingest or inhale or get on our skin.)
  • Is it particularly hazardous to select people/subpopulations?

…while critical details about Exposure include…

  • Are we exposed at all, since something that’s hazardous isn’t a risk until we’re actually exposed. (Think of a poisonous snake locked safely in a cage.)
  • To how much? (For chemicals this is about dose. For behaviors, like drunk driving, it’s about frequency.)
  • In what ways?
  • At what age? (Materials or behaviors that are hazardous for young kids may not be for adults).

With all those critical details, you can make a reasonably well-informed risk-based policy, or what industry argues is policy based on ‘sound science’

But suppose the early evidence suggests that some product or process or behavior might be a hazard. Suppose we just don’t know yet how hazardous it might be, or at what dose, or to which subpopulations, and we know next to nothing about the exposure side. But there are some plausible signs of danger (e.g. initial lab tests on a chemical, or early studies that suggest playing violent video games may be bad for kids.) We don’t want the government to sit around, in many cases for years, until they have the rest of the facts, before acting to protect us, right?

That’s where the HAZARD-BASED approach comes in. It triggers action if there is a hint of hazard, before the full details about hazard are known, or any of the exposure details are even considered. Better safe than sorry, right? Environmentalists call this the Precautionary Principle

With most risks, the default for most governments is the hazard-based approach….make sure it’s safe before allowing it…the way the U.S. Food And Drug Administration does with drugs, for example. But as safe as it sounds, this precautionary approach has problems. If we’re too careful and ban everything suspected of being hazardous, we could lose out on the benefits of products that turn out to be safe, like life-saving drugs for patients with only months to live, or processes, like some types of genetically modified food.

But the seemingly more rational risk-based approach has problems too. It’s more fully informed and rational, and allows society to enjoy the benefits of many products (antibacterial nano silver particles in food storage containers to kill germs and prevent spoilage) and processes (genetically modified crops). But approving products or processes or behaviors (driving while using a cell phone) before all the hazard and exposure facts are known means that if later research shows that the early hints of hazard were right, we’ve exposed ourselves to harm, some of which may be irreversible, while we were waiting for all the ‘sound science’ to come in. Consider how a fully risk-based ‘get all the answers first’ approach would make it harder to respond now to climate change, since we’re still developing some critical scientific details.

Now, what does this look like in the real world? BPA is a great example. Some environmental scientists are passionately convinced that the hazard evidence is convincing enough that we should act. they have been very effective at stirring media and public concern. In response, applying a hazard based approach, Canada has banned BPA from baby bottles and sippy cups. The U.S. has too (only after manufacturers and retailers pulled those products from the market), and several countries around the world have done the same, or are thinking about it.

But at the level of proof that a risk-based approach requires, most government policy makers are not persuaded that BPA is dangerous. They say that at the levels to which we are exposed to BPA, it’s safe. Relying on a risk-based approach that takes into account not just hazard but also exposure, governments have rebuffed environmentalist calls to ban the stuff. Including the Canadians!

Of their hazard-based ban on baby bottles, in 2010 they said;

"Our science indicated that bisphenol A may (my emphasis) be harmful to both human health and the environment and we were the first country to take bold action in the interest of Canadians."

Of their risk-based approach to the same risk, in 2012 they said;

"Based on the overall weight of evidence…Health Canada's Food Directorate continues to conclude that current dietary exposure to BPA through food packaging uses is not expected to pose a health risk to the general population, including newborns and young children.” (italics are my emphasis)

Now, consider the implications of this ambivalence to public health. The ban on baby bottles and sippy cups, a hazard-based response to public concern, ignores the ‘exposure’ part of the RISK = HAZARD X EXPOSURE formula. The evidence about BPA suggests that if there is a risk, the greatest potential hazard is damage to the health of the developing fetus, and fetal exposure comes from what pregnant mothers eat and drink. Moms don’t use baby bottles and sippy cups! The hazard-based precautionary ‘reassure a frightened public’ response to BPA fails to protect the very population which the evidence suggests is most at risk. OOPS!

But the risk-based approach most governments are taking to BPA could be dangerous too. If the fears about BPA turn out to be true, but we continue to permit its use while we do all the research necessary to fully understand hazard and exposure, we could be harming (during fetal development) a whole generation of children, and possibly genetically threatening their offspring as well. OOPS!

BPA illustrates how most governments use both approaches to risk management. But it also illustrates how many governments are now moving toward a more hazard-based approach, particularly with industrial chemicals that are mostly controlled under the more risk-based ‘sound science’ method. This is especially true of governments with a legacy of not having been precautionary enough about risks like Mad Cow disease and blood transfusions and thalidomide (the morning sickness medicine that caused serious birth defects in the kids of mothers who used it in Canada and many European countries, where despite early hints of hazard the drug was approved). Many observers feel that these governments, including the EU and Japan, are now adopting hazard-based risk management less to actually protect public health and more as a reaction to public mistrust caused by their historic failures.

Like I said, risk is a tricky business for anybody trying to analyze things objectively, as we want our policy makers to do. They have to account for scientific complexities and uncertainties, weigh difficult tradeoffs between harms and benefits and public values, respond to intense public emotions, and deal with pressure from advocates and vested interests lobbying and campaigning and pressuring for one outcome or another. We are at serious risk because of the subjective nature of the way all these factors shape the risk management policy making we depend on to protect us. But in a democracy there is no way around this mess, because of the very subjective nature of ‘risk’ itself.

Image: Environmental Health and Safety at UF.