ARE you concerned about the health risks of living near a waste incinerator? If a giant new plant is about to be built near your home, the answer may well be yes. In the UK, a government consultation exercise on waste management, launched this week, is likely to lead to a large increase in the amount that gets burnt, sparking fears of huge incinerators springing up across the nation and spewing noxious pollutants over neighbourhoods.
The protests are already under way. Plans to build Europe鈥檚 biggest incinerator on the outskirts of Peterborough in the English midlands were rejected last month by local councillors, who talked of residents becoming 鈥済uinea pigs鈥�. Similar plans for a plant near Norwich are also on hold, with one local politician describing it as a 鈥渨eapon of potential mass pollution鈥�.
With the battle lines drawn, it鈥檚 a safe bet that both sides will turn to science in the expectation of getting definitive answers. They can expect to be disappointed. As with many disputes over the health consequences of new technologies 鈥� from overhead power lines and cellphones to nuclear installations 鈥� the science already shows that the chances of serious harm to anyone鈥檚 health are small. What science will never be able to show is that the risk is zero.
Advertisement
If the debate over incinerators is not to turn into another of those rows that rattle on for years, both sides need to accept some basic truths about the limitations of science. No one doubts that incineration can produce some pretty nasty stuff, notably dioxins, acid gases and heavy-metal residue. Science can tell us whether incineration plants can do an adequate job of protecting us from any meaningful threat, and it has already done so: the risk from incinerators is too small to worry about. Yet many people are still not satisfied. They demand proof that there will be no risk.
The reason 鈥渮ero risk鈥� is a nonsensical notion lies in the statistical methods epidemiologists use to assess threats to public health. These methods have scored some spectacular successes over the years, and have saved countless lives by tracing effects back to their causes: linking lung cancer with smoking, for example.
Proving causality, however, is a notoriously tricky business. The most reliable way to gauge the level of risk would be to recruit vast numbers of people, choose half of them at random and expose them to the alleged health risk. Then if there is something to worry about, it will be pretty obvious: the exposed group will have significantly higher levels of ill-health.
The problem with this approach is clear. Deliberately exposing people even to a suspected health risk is ethically unacceptable. Even if it were not, such studies would take decades. Instead, epidemiologists use 鈥渃ase-control鈥� studies, in which people known to suffer from the alleged health effect are compared with a control group free of the effect. Comparing the numbers of people in each group who were exposed to the supposed threat then reveals their relative level of risk.
The trouble with case-control studies is that they compel epidemiologists to use non-random samples, which offer no protection against one group being subtly but systematically different from the other. Such studies are also retrospective, so researchers must rely on people鈥檚 guesstimates of how they were exposed to the risk factor. If the health effect is rare, investigators are further handicapped by having only small numbers to analyse.
鈥淣o one has convincingly explained why overhead power lines should cause childhood leukaemia鈥�
Epidemiologists are all too familiar with these limitations, and have made heroic efforts to get around them. Even so, it鈥檚 generally agreed 鈥� by the World Health Organization, among others 鈥� that studies finding no more than an approximate doubling in relative risk should be treated with scepticism. Such an increase could easily be the result of unknown biases, confounding effects or sheer fluke.
As it happens, this is just the level of relative risk typically found in studies of notorious health scares such as those linking childhood leukaemias with overhead power lines and cellphones with cancer. Another reason not to take these findings very seriously is their lack of plausibility: no one has found a convincing explanation of why, for example, overhead power lines should cause leukaemia in children.
When case-control studies repeatedly produce modest evidence of implausible health threats, they are telling us it鈥檚 pointless doing more of the same. This is precisely the state of play with the issue of waste incineration. The studies so far have uncovered inconsistent and implausible evidence for a low level of risk. It is only natural that people living near huge incineration plants want more assurance 鈥� but epidemiology simply cannot provide it. The type of study needed is financially, logistically and ethically out of the question.
There are, of course, genuine reasons to be concerned about incinerators. They are noisy, they damage air quality and they increase traffic levels. You don鈥檛 need subtle statistical analysis to detect the roar of a waste-laden truck.