THE ONTARIO FACT-FINDING PANEL DISCUSS INTERPRETATION ISSUES IN HEALTH STUDIES

Epidemiological or health studies of human beings exposed to a particular chemical or physical irritant of some nature are extremely unlikely not to have some major issues. These include having a large enough group of people to make a statistically valid comparison. They include having a group of individuals exposed to the item of concern, such as TCDD, and then being able to compare their health to a relatively equal group of individuals who have not been exposed. By a relatively equal group we mean of similar age, similar genetic background , similar diet, similar overall health originally etc..

Other issues in this study would include the impossibility of having perfect knowledge. How many of us can remember an exposure say to an on the job solvent that occurred twenty-five years ago? Yes if you collapsed at the time due to an acute exposure you probably would remember. Otherwise not. How many of us have been exposed to a chemical and not even known it? This exposure could be via air, food or water. Therefore determining who is in the control group allegedly not having been exposed to the chemical of concern is a difficulty.

The Panel discuss at length problems with exposure assessment. "Even the most comprehensive and high-quality epidemiological studies typically have monitoring information on exposure for only a few points in time for some of the study population, which must be used to represent the exposure history of the entire population. Thus for most of the timeframe covered by any epidemiological study, exposures must be estimated or extrapolated, from a relatively small amount of monitoring data...". Thus exposure misclassification can lead to either false negative conclusions or to false positive conclusions regarding associations between TCDD and health.

The Panel then states "Clearly, exposure misclassification can and probably does have a powerful effect in nearly all observational epidemiology studies. What's more, it typically drives relative risk estimates towards the null.". Also disease misclassification can and does occur. Current medical records are considered for example a much better indication of cancer presence than death certificates listing the cause of death.

Confounding is a huge issue which can only be somewhat mitigated in these studies. Obesity and smoking for example are risk factors for many diseases. Suggesting that a persons' cancer is caused by exposure to a single large exposure to TCDD may not reflect a lifetime of smaller exposures to TCDD through foods and inhalation. Similarily suggesting that exposure to TCDD through annual road spraying wasn't large enough to cause their cancer probably is ignoring their previous exposures to TCDD by eating fish downstream of Uniroyal Chemical or even through exposure from spraying of nearby Ontario Hydro power lines. To date I have seen no suggestion that this study has considered multiple exposures to 2,3,7,8 TCDD and or 2,4,5-T from many other known sources in our environment. In other words it is not accurate to separate out Ontario government causes of TCDD exposure and suggest that they haven't added to citizens toxic burden.

Another well recognized confounding variable is known as the "healthy worker effect". In essence the Panel are focusing on workers exposed to 2,4,5-T and TCDD and comparing them to the general population. However the forestry and Hydro workers as a group are probably healthier (at least initially) than the general population in order for them to be working full time. Also the general population includes the very elderly who are more likely to have serious other health issues.

The Fact-Finding Panel concluded that as much as possible the studies they reviewed all considered possible limitations of epidemiological studies. While that is mildly comforting that is all it is. While I'm not suggesting any flaws specific to the Panel's work, I am suggesting that the whole field of risk assessment is fraught with information difficulties, assumptions and extrapolations. Perhaps the mathematical models and calculations are the best we've got. Perhaps the best we've got is significantly underestimating the effects of Ontario's poisoning of the environment. Perhaps not. It was still incredibly stupid for an incredibly long length of time.