Good grief! Can't any story ever be put to rest? When can we say that science has actually answered something. Here's a story about the US FDA trying to decide if more research is needed to see whether or not amalgam fillings for dental caries can cause mercury poisoning. Amalgam is 50% mercury and mercury is an undisputed neuro (and other) toxin. Whether using it in children or pregnant women could cause damage due to mercury poisoning is at issue.
The risk must be small or we'd clearly know that there was one. Here, the benefits of amalgam over other materials (stability, better interaction with the tooth) would have to be weighed against the toxicity risk, if there is any of the latter.
Estimating very small risks--including, by the way, that of dental X-rays--is a very challenging business, not unlike the challenge of estimating genetic risk in GWAS and other wild-goose-chase research. It's hard to estimate risk with acceptable reliability, and perhaps at least as hard to make risk-benefit calculations. The same has been the case with PSA tests for prostate cancer and mammograms to detect early breast cancer, though these two cases may be easier than the dental one because the risks may be somewhat higher and there may be a lot more data.
Testing evolutionary just-so stories about Darwinian adaptive selection in Nature is another instance where theory ideology comes up against the facts, or lack of available facts. Arguments are heated, issues basically never settled.
Things are often complicated by politics (as in climate change), or vested interests, but it doesn't seem to be the case here. Perhaps here we see what happens when scientific technologies improve so much in power that we can even think to ask these kinds of questions.
Whatever the case in this instance, statistical inference on weak, complex probabilistic causal situations will be a major and common aspect of life as well as physical science in the foreseeable future.
No comments:
Post a Comment