Q&As Concerning the Previous Studies

Why did the company launch an epidemiological investigation of brain cancer at Spring House?

Rohm and Haas decided to investigate whether there might be an excess of brain cancer at Spring House in 2001 after a diagnosis of brain cancer in a young research chemist, which came on the heels of a recently completed study of brain cancer conducted by BP Amoco at its research facility in Chicago. The company reviewed available records to identify cases of brain cancer among the 4,800 Spring House employees who had worked at the facility since its opening in 1963.

The company commenced a case-control study in June 2002, designed to look at those cases which had been identified to determine if there might be any workplace association. The case-control study was done first to determine whether or not a link or correlation between brain cancer and any workplace chemicals could be found. Thereafter, the company further sought to determine through a cohort mortality study whether the number of brain cancer deaths at Spring House were higher than what would be expected in the U.S. general population.

How does the company determine if a study is needed?

Several factors may prompt the company to conduct a study, including concern by employees or former employees, information from outside sources such as cancer registries, or other evidence that indicates there is an excess or undue elevation in a particular disease or illness. Since 1972, Rohm and Haas has conducted 12 epidemiological studies.

What were the criteria for including someone in the 2002 case-control study?

The 2002 case-control study included all known cases of brain cancer among employees whose employment records indicated that they had been assigned at some point to the Spring House facility.  The study focused on primary malignant brain cancers (i.e., the cancer did not originate from another cancer elsewhere in the body), but any brain cancers identified were presumed to be primary unless the study team could confirm otherwise. The study team identified 12 cases of brain cancer and 3 benign brain tumors.

Work histories were taken from Human Resources and Epidemiology databases. Records included the job title and department for each employee. Building locations for all job/department combinations were not recorded in these databases but were determined by interviews of long-term employees and by reviewing other records, including business reports, scientific notebooks and technical reports.  

What is a “work history?”

A work history is generated for each employee by Human Resources. It contains demographic information and employment information – date of hire, date of separation, job title, job level, department code, site code. Older work histories at ROH recorded all adjustments to pay.  Although ROH work histories include department and/or business information, and occasionally a lab number, they do not include building information. For the case control study, a subject’s building assignment was determined from his business and department assignment in his work history, historical records (building occupancy permit dates), and employee and retiree interviews. Spring House employees can request a copy of their work history from the Spring House Human Resources representative. 

What was the design and methodology of the case-control study?

The 2002 study was a matched analysis case-control study, which includes matching the 12 potential cases of brain cancer with a group of four controls for each case, selected on the basis of age, race, gender and date of hire. For each case and control, a work (occupational) and personal (non-occupational) exposure analysis was constructed based on a review of the individual's work history, projects, job title, department and chemicals used. These exposure analyses were constructed based on a review by medical and health professionals of research notebooks, B-reports and technical reports, as well as interviews of current or former colleagues where appropriate.

Non-occupational exposures were included where known or indicated in medical records, such as smoking, alcohol use, head trauma, diabetes, epilepsy, etc. Exposures were given weights based on length of time and proximity to the chemicals (i.e., author, researcher, both, etc.). The same was done for each case and control as to which buildings each may have worked in, such as whether the person ever worked in a particular building and for what duration, categorized by whether he or she worked in the building ever, less than or equal to the median time or above the median time (in years).

Several analyses were done on the data to determine what, if any, associations could be drawn between the chemicals an individual could have been exposed to and brain cancer. As part of the analysis, 82 chemicals were identified as having at least one mention in an animal or human study associating it with an occurrence of brain tumor.  From these, the team identified 36 categories of chemicals that include more than 20,000 chemicals for which at least one study participant had a potential exposure. The study also evaluated 10 different buildings where the study subjects had worked, as well as 21 non-occupational or self-reported exposures.

In the case-control study, how were cases assigned to a category of chemicals? 

In the case control study, if a person was exposed to at least one of the chemicals included in the category, the person was considered as exposed to that category. Persons could be exposed to multiple chemicals in their time at Spring House and could be represented in more than one chemical category.  

Since many of the chemists moved among buildings with job changes and lab relocations, how were cases identified to a certain building? Has there been any weighting to time spent in particular buildings, or for temporary assignments to other buildings?

Work histories contain site locations, departments, and job titles or positions held throughout the employee’s employment with the company. Job changes were captured in employees work history. Job changes may include a new department, job position and/or change of site location. A study participant’s time in a department is accounted for only if his work history was changed to reflect the assignment.  

Work histories do not contain building location. For the case-control study, an employee’s building location was determined for each department and job position combination identified in the employee’s work history. The building location was determined from a variety of sources, including the Research SAS database, building occupancy permits, interviews, notebooks and B-Reports, and Department Groups associated with specific buildings.

In the case control study, exposure to buildings was weighted depending on the length of time spent in a building. Three levels of analyses for exposure to building location were performed when at least 2 subjects had data for the same building, and one of the two had to be a case. The first level of analysis of building location was a yes or no, if a case or control ever worked in a building. The second level of analyses was looking at the length of time or duration spent in a building. The case or control either did not work in a building, worked in a building less than or equal to the median, or worked in the building greater than the median. Lastly, cases and controls were assigned to a building in which they worked the longest.