Census Tract Epidemiology Reports are Hazardous to Your [mental] Health

Census tract mortality studies aimed at confirming or denying a facility chemical emission threat to health repeat the same error over and over and over.

These are death certificate-based mortality studies. Quantitative extrapolation of mortality risk from agents such as asbestos demonstrate that large excess risk would not be observed by the method, and that excess risk observed by the method would not be due to the emissions.

Possibly mesothelioma cases could be observed in such a study. [Over much of the period, and possibly now, there is no ICD code for mesothelioma, and therefore death certificates would not be coded for it in state death registries.] Possibly, x-ray evidence of pleural plaques in a sample of residents could observe an effect of the facility. Probably a soil sampling study could detect the presence of asbestos contamination. But census tract mortality would not find a problem.

OSHA estimates a 6 per 100 lifetime risk of cancer [mostly lung cancer] from a 45 year working career exposed to 0.1 fiber/cc of asbestos. Thus, 4.5 fiber-years. Which is approximate doubling from population risk for males, and more than doubling for females. Possibly a statistically significant result could be observed at a 2 per 100 lifetime attributable risk at an excess of 30% if the expected deaths from lung cancer were upwards of 100, which translates into expected death from all causes upwards of 2000 in this cohort.

The actual ATSDR report found 276 lung cancer deaths among men, a rate ratio 0f 1.07, with an upper confidence interval of 1.20 [that is, results rule out with 95% confidence a risk as high as 1.20]. Combining races would tend to reduce the SIR based on the expected cases, since the lung cancer mortality rate among blacks is higher than whites.

For mesotheliomas, there were 6 observed vs. about 10 expected; while the SIR was .6 or so, the UCI was 1.23. [Unclear how the investigators got mesothelioma diagnosis from death certficates.]

So, contrary to the stated conclusion, the ATSDR study found an increase in lung cancer, not statistically significant, but only suceeded in ruling out a 20% excess which translates to a lifetime risk of 1.2 per 100. This translates back to a .07 fiber-year exposure.

There is some reassurance that the mesotheliomas were not in excess.


Second report finds minimal cancer risk

Friday, July 08, 2005

By DARRYL R. ISHERWOOD Staff Writer

HAMILTON - State health officials have determined that townshipwide, residents face no increased risk of cancer or other disease from the operation of the W.R. Grace and Co. Zonolite plant, which processed asbestos-laden ore for more than 40 years. ]

But state officials admit that the recent study, conducted on the entire population of Hamilton, does not present as valuable a picture of the health risks as a more narrowly focused study completed in conjunction with federal health officials in March.

This latest report, issued by the state Department of Health and Senior Services late last month, confirms what federal health regulators said three months ago: despite years of processing vermiculite contaminated with tremolite asbestos, the former insulation plant on Industrial Drive did not significantly raise the risk of cancer and other asbestos-related illnesses among township residents.

"(The) results do not indicate that asbestos-related outcomes for Hamilton as a whole have been impacted by potential asbestos exposure from the Zonolite/W.R. Grace facility," said the report, which was issued to the township's health officer.

Michael Berry, the health department scientist [sic] who conducted the study, said that while the report was good news, it was not as accurate as the study done in March. That study looked at Trenton, Lawrence and Hamilton residents who lived within one mile of the plant.

"The problem with the municipalwide study is Hamilton is a very large town and though there may have been significant exposures in the nearby neighborhoods, we certainly had no reason to believe that all or most of Hamilton was affected," Berry said.

Berry said because of Hamilton's size - the township covers 40 square miles - studying health effects of residents who live miles from the plant served little purpose. But in order to be consistent with a national effort to study the health risks in communities where the vermiculite was shipped, Hamilton was included in the state's report.

Another limitation of the study, Berry said, is that it focused only on Hamilton mailing addresses, but there are many areas of the township that have Trenton addresses, which could skew the sample.

That possibility would have been addressed in the March study, which did not limit itself to Hamilton residents, but rather to the people who lived closest to the plant.

Migration is also a weakness of both the recent study and the one completed in March, he said. Residents who moved and were diagnosed with an asbestos-related disease in a different area of the state or country would not fall within either report.

There could be a migration both into and out of a study area," Berry said. "But we hope it would average out."
The study was conducted using cancer and health data from seven New Jersey municipalities that all reportedly received vermiculite ore from a mine in Libby, Mont., that was later found to be contaminated with the deadly form of asbestos.
The report included cancer statistics from 1986 through 1995 and mortality rates from asbestos-related diseases from 1979 through 1998.