Media Center

Industry Response to the January 2005 Study by Dr. Brad Racette

Posted January 25, 2005

The Racette study does not change our position that there exists no sound epidemiological study linking welding or welding fumes to Parkinson’s disease or Parkinson’s-like movement disorders. The study suffers from inherent bias and methodological flaws that make the results unreliable and not meaningful. For example,

  • There was a clear selection bias. The welder screenings which Dr. Racette uses as the basis for his study were entirely funded by plaintiffs’ attorneys and are being used to gather claimants for lawsuits. Additionally, most of the welders screened were sent to Dr. Racette by plaintiffs’ lawyers. Dr. Racette describes the screenings as “medicolegal,” and has, in sworn testimony, indicated that the screenings were not a “scientific study”. In a deposition conducted in January of 2003, he acknowledged that these screenings could in no way constitute a valid scientific study.
Q: Now the screening that you’re doing down South, that is not a scientifically designed epidemiological study, is it?

A. That is correct.

Q. Is the screening for attorneys for purposes of litigation as far as you know?

A. That is correct.


Q. You do understand that the people you are screening are being sent to you by attorneys.

A. I’m aware of the method of ascertainment.


Q. Is the screening being done down South, is that being funded solely by the plaintiffs’ attorneys?

A. Yes.

  • There was no proper control group. Instead of properly developing a control group which as closely as possible matches that of the welders in his screenings, Dr. Racette compares his results to those found in a 1985 survey of the entire population of Copiah County, Mississippi. The underlying work supporting the Copiah County study was conducted almost 25 years ago, in a different geographic region with different demographics, and using different diagnostic techniques and methods than Dr. Racette employed, thus making comparisons unreliable. In fact, the authors of the Copiah County study themselves said, in 1985, that “it would be risky to assume our results have relevance to other populations,” a caution Dr. Racette apparently ignored. Dr. Racette also makes the erroneous and unsupported conclusion that the prevalence of Parkinson’s disease in a single Mississippi county can be extrapolated to provide an accurate assessment of the prevalence of parkinsonism state-wide in Alabama, when the Copiah County study itself was not designed to ascertain a state-wide rate. Dr. Racette recognized the problems with this approach when he noted in the abstract of his study that testing the validity of his work, “requires duplication on a population based sample that includes a control group to determine the specificity of our methodology.” In the current study, Dr. Racette admits that his findings need to be “confirmed by applying the same screening techniques to a control population.”

  • Dr. Racette failed to properly distinguish between Parkinson’s disease and parkinsonism. In any given population, there will be many more cases of parkinsonism than cases of Parkinson’s disease.  The general parkinsonism category includes all Parkinson’s disease cases, as well as other parkinsonisms, such as multi-system atrophy, progressive supernuclear palsy, dementia with Lewy bodies, as well as secondary parkinsonisms, such as drug-induced and cerebral ischemia.  In addition, there are extrapyramidal disorders such as essential tremor.  Dr. Racette’s methodology does not properly distinguish between these parkinsonisms, and this failing greatly impacts his results because in the Copiah County study, the researchers worked to exclude from their results other parkinsonisms, essential tremor, drug-induced syndromes, as well as extrapyramidal signs. By including these other parkinsonisms in his findings, Dr. Racette artificially inflates the reported risk among welders.

  • Dr. Racette recognizes that his results are speculative and that their validity needs to be confirmed by proper studies. Dr. Racette characterizes his conclusion as merely a “speculation” that welding exposure may increase the prevalence of Parkinson’s disease, and recognizes that his study is not sufficient to validate this speculation. As he says, “[o]f course, this speculation is contingent on applying our methodology to a population-based control group to determine the specificity of our findings.” In fact, all studies to date that have used a population-based control group have found no association between welding and Parkinson’s disease.

  • Dr. Racette’s video screening method did not allow for proper diagnosis. For purposes of his study, Dr. Racette invented a new and unproven approach to diagnosing illness. Instead of meeting with subjects and examining them individually, the standard and accepted method of diagnosing Parkinson’s disease as well as other diseases and afflictions, Dr. Racette chose to perform diagnoses via videotape (screening up to 580 welders in a single day). No evidence exists that Dr. Racette’s unproven video screening method did or could allow effective differentiation between Parkinson’s disease, on the one hand, and other parkinsonisms and extrapyramidal disorders, on the other. This would also inflate artificially the reported risk among welders.

  • Dr. Racette’s extrapolation of his results to all welders in Alabama is questionable. Dr. Racette’s results depend on the assumption that in some way the results of his screening of a population of welders referred by plaintiffs’ lawyers can be extrapolated to all welders in Alabama. This is questionable. Without getting into the technicalities of sampling and coding techniques, obvious questions are raised, for example, by Dr. Racette’s Table 2, where he purports to have screened 60 Alabama welder’s helpers in the 60-69 age range, but assumes that the total number of welder’s helpers in Alabama in that age range was only 35. Something is plainly wrong with these calculations, and they call all the rest into question.

  • There exist critical definitional issues with Dr. Racette’s coding of occupations. In his initial calculations of prevalence among Alabama welders, Dr. Racette compares the findings from his screening of welders (in which he claims that he screened 12% of the state’s welders) to the number of active welders in Alabama in three specified Standard Occupational Codes according to the 2000 Department of Labor Bureau of Labor Statistics Occupational Census (BLS). His method of comparison is improper. This comparison could be valid only if Dr. Racette’s characterizations of his subjects’ occupations matched the BLS’s definition of the relevant occupational codes, as he acknowledges. But we know there is a mismatch: as discussed above, Dr. Racette thinks there are 60 or more welder’s helpers in Alabama in the 60-69 age range, while the BLS thinks there are 35. The article is extremely unclear on the subject of how Dr. Racette coded occupations, but it appears that the problem is that Dr. Racette did not attempt to define occupational categories in the way the BLS defines them; rather, he allowed subjects of his study to self-identify their occupations. This is certain to produce bias in the data, especially given the subjects’ knowledge that they had been selected by lawyers seeking to establish a connection between welding and parkinsonism.

Given all the problems with the study, Dr. Racette’s article is not a sound epidemiological study and in no way changes the scientific landscape. There exists no sound epidemiological study showing an association between welding or welding fume and Parkinson’s disease or Parkinson’s-like movement disorders. On the contrary, the valid studies that exist demonstrate that no such association exists. For example, Dr. Jungson Park recently published studies in the NeuroToxicology Journal and the Industrial Health Journal which found that manganese exposure is not a risk factor for Parkinson’s disease, stating:

As regarding the exposure to hazardous materials, especially Mn, more subjects in the control group than the PD patient group have worked in the occupations with potential exposure to Mn such as welder, smelter, welding rod manufacturer, Mn miner, workers in the iron and steel industries and dry cell battery manufacturers. Furthermore, according to logistic regression in the present study, above occupations with a high potential exposure to Mn showed consistently negative associations with PD after adjusting the confounders such as age, sex and smoking and education level (OR: 0.42, 95% CI 0.22-0.81).

These studies did not have the selection biases and control group issues present in Racette’s paper.

Copyright © 2004 Welding Rod Defense Group. All Rights Reserved.
Important Notice