Missing files, sparse records and software failure at IBM hamper government effort
to assess feasibility of studying semiconductor worker cancers at Endicott, NY plant
to assess feasibility of studying semiconductor worker cancers at Endicott, NY plant
by Bryant Furlow
EPI NEWS
June 5, 2007—Delays, missing records, and
malfunctioning software at IBM hampered
government scientists’ three-year-long effort to
assess the feasibility of studying cancer rates
among semiconductor workers exposed to solvents
and heavy metals at the IBM plant in Endicott, New
York.
Researchers from the US National Institute for
Occupational Safety and Health (NIOSH), a branch
of the US Centers for Disease Control (CDC),
completed their report in January, but it was not
made public until April.
The NIOSH study was requested by Congressman
Maurice Hinchey (D-NY), and follows a study by
epidemiologist Richard Clapp at Boston University.
Clapp, who had been hired by a law firm
representing former employees, identified significant
elevations in brain, kidney, blood, and skin cancer
mortality among Endicott facility workers. IBM
attempted to block publication of Clapp’s findings,
arguing that the company had not authorized the
disclosure of data from employee death records.
(See the “further reading” box, below.)
The NIOSH study team sought to assess whether
IBM employee and chemical exposure or “industrial
hygiene” records could be used to study cancer
rates and carcinogenic workplace exposures.
But they faced obstacles at almost every step:
missing and incomplete IBM records on who worked
at the facility and when, what chemicals were used in
which departments, delays in IBM’s disclosure of
building blueprints—and even, ironically, software
malfunctions that prevented NIOSH researchers
from accessing
the computer
company’s
chemical expos-
ures database.
Company officials were so uncommunicative that a
former IBM employee had to be sought out by a
NIOSH representative to help decipher nearly 4,000
department and office codes listed in employee files.
“IBM gave us some of the records that we needed to
assess the feasibility of a study among former
workers,” explains Lynne E Pinkerton, the NIOSH
epidemiologist who ran the assessment effort. “They
were concerned about giving us some of the other
records for the study because we could not promise
that we would keep them confidential. To keep
things moving forward, we agreed to review these
records at IBM.”
EPI NEWS
June 5, 2007—Delays, missing records, and
malfunctioning software at IBM hampered
government scientists’ three-year-long effort to
assess the feasibility of studying cancer rates
among semiconductor workers exposed to solvents
and heavy metals at the IBM plant in Endicott, New
York.
Researchers from the US National Institute for
Occupational Safety and Health (NIOSH), a branch
of the US Centers for Disease Control (CDC),
completed their report in January, but it was not
made public until April.
The NIOSH study was requested by Congressman
Maurice Hinchey (D-NY), and follows a study by
epidemiologist Richard Clapp at Boston University.
Clapp, who had been hired by a law firm
representing former employees, identified significant
elevations in brain, kidney, blood, and skin cancer
mortality among Endicott facility workers. IBM
attempted to block publication of Clapp’s findings,
arguing that the company had not authorized the
disclosure of data from employee death records.
(See the “further reading” box, below.)
The NIOSH study team sought to assess whether
IBM employee and chemical exposure or “industrial
hygiene” records could be used to study cancer
rates and carcinogenic workplace exposures.
But they faced obstacles at almost every step:
missing and incomplete IBM records on who worked
at the facility and when, what chemicals were used in
which departments, delays in IBM’s disclosure of
building blueprints—and even, ironically, software
malfunctions that prevented NIOSH researchers
from accessing
the computer
company’s
chemical expos-
ures database.
Company officials were so uncommunicative that a
former IBM employee had to be sought out by a
NIOSH representative to help decipher nearly 4,000
department and office codes listed in employee files.
“IBM gave us some of the records that we needed to
assess the feasibility of a study among former
workers,” explains Lynne E Pinkerton, the NIOSH
epidemiologist who ran the assessment effort. “They
were concerned about giving us some of the other
records for the study because we could not promise
that we would keep them confidential. To keep
things moving forward, we agreed to review these
records at IBM.”
| Further Reading Lynne Pinkerton. 2007. An assessment of the feasibility of a study of cancer among former employees of the IBM facility in Endicott, New York. Final draft report. Industrywide Studies Branch, Division of Surveillance, Hazard Evaluations, and Field Studies, NIOSH. Cincinnati, OH. Dan Ferber. Beset by lawsuits, IBM blocks study that used its data. Science, 2004;304:937-939. Roxanne Nelson. IBM victorious in employee cancer lawsuit. Lancet, 2004;363:788. |
Chemical Exposures Database Software Failed
But when NIOSH representatives arrived at the IBM
facility to review a database of workers' chemical
exposures at Endicott, IBM officials
could not get their software to function.
They later provided the NIOSH team
some data from the malfunctioning
database, but NIOSH scientists were
never given a copy of (or direct access to)
the database.
A NIOSH contractor searched through boxes of
paper files at IBM’s offices, but IBM officials did not
permit him to make copies of any documents.
Instead, he rushed to take notes with pen and paper
from as many files as possible.
But when his notes were compared to the database
records provided by IBM, neither source represented
a complete set of chemical exposure data, NIOSH
found. The paper files had contained records not
disclosed from the database, and the database
records included records not found in the boxes of
paper files. The NIOSH team concluded that IBM had
not conducted regular chemical exposure
assessments in many departments at Endicott.
Employee Records Missing, Incomplete
Chemical exposure records were not the only
problem. Employee work assignment records were
equally disorganized and incomplete. Nearly a
quarter of the personnel files evaluated by NIOSH
had inconsistent dates of hire. Files were in the
wrong folders, according to the NIOSH report, and
IBM provided only year-end employee work
assignments.
“We estimate that ‘year-end’ personnel files, on
average, miss approximately 21% of the
departments in which an employee worked,” the
report states.
Worse yet, entire boxes of documents were simply
missing. “We're clearly missing personnel records for
workers who quit working for IBM prior to 1965,”
notes Pinkerton.
But when NIOSH representatives arrived at the IBM
facility to review a database of workers' chemical
exposures at Endicott, IBM officials
could not get their software to function.
They later provided the NIOSH team
some data from the malfunctioning
database, but NIOSH scientists were
never given a copy of (or direct access to)
the database.
A NIOSH contractor searched through boxes of
paper files at IBM’s offices, but IBM officials did not
permit him to make copies of any documents.
Instead, he rushed to take notes with pen and paper
from as many files as possible.
But when his notes were compared to the database
records provided by IBM, neither source represented
a complete set of chemical exposure data, NIOSH
found. The paper files had contained records not
disclosed from the database, and the database
records included records not found in the boxes of
paper files. The NIOSH team concluded that IBM had
not conducted regular chemical exposure
assessments in many departments at Endicott.
Employee Records Missing, Incomplete
Chemical exposure records were not the only
problem. Employee work assignment records were
equally disorganized and incomplete. Nearly a
quarter of the personnel files evaluated by NIOSH
had inconsistent dates of hire. Files were in the
wrong folders, according to the NIOSH report, and
IBM provided only year-end employee work
assignments.
“We estimate that ‘year-end’ personnel files, on
average, miss approximately 21% of the
departments in which an employee worked,” the
report states.
Worse yet, entire boxes of documents were simply
missing. “We're clearly missing personnel records for
workers who quit working for IBM prior to 1965,”
notes Pinkerton.
Boxes of documents have gone missing, say government investigators
Occupational Health
Chemical Exposure Records a 'Major Concern'
Asked about the NIOSH report, IBM is unapologetic. “We
cooperated with NIOSH during its feasibility assessment,”
IBM spokesman Ari Fishkind told epiNews by email. “It
concluded that there is sufficient information available if
a study was conducted. As always, we remain committed
to working with the agency.”
Pinkerton agrees that a study of employee cancer rates
is possible—assuming IBM cooperates.
“Obtaining the exposure records would be a major
concern,” she concedes.
The NIOSH team estimates that at least 6.7 percent of
the former employees for whom IBM provided records—
some 1,881 people—had a “high potential” for exposure
to chemical solutions used to etch, plate and laminate
circuit boards.
Little Hope for Definitive Answers
It remains unclear what, if any, conclusions could be
drawn from a full-blown NIOSH study.
“There are not enough data to estimate [chemical]
exposures quantitatively,” Pinkerton points out. “It may
be possible in the case of some specific chemicals or
groups of chemicals to assign qualitative levels of
exposure—high versus low.”
Even if NIOSH does confirm elevated cancer rates, she
cautions, it may be impossible to determine whether
workers’ cancers are due to chemicals they encountered
at IBM. At best, Pinkerton believes, NIOSH will be able to
compare employee cancer rates to those of the general
public.
“There are enough data to determine if former IBM
employees have a higher rate of cancer than the general
population,” she explains, but “the lack of exposure data
will limit the study's ability to determine whether the
cancers are due to workplace exposures.”
Asked about the NIOSH report, IBM is unapologetic. “We
cooperated with NIOSH during its feasibility assessment,”
IBM spokesman Ari Fishkind told epiNews by email. “It
concluded that there is sufficient information available if
a study was conducted. As always, we remain committed
to working with the agency.”
Pinkerton agrees that a study of employee cancer rates
is possible—assuming IBM cooperates.
“Obtaining the exposure records would be a major
concern,” she concedes.
The NIOSH team estimates that at least 6.7 percent of
the former employees for whom IBM provided records—
some 1,881 people—had a “high potential” for exposure
to chemical solutions used to etch, plate and laminate
circuit boards.
Little Hope for Definitive Answers
It remains unclear what, if any, conclusions could be
drawn from a full-blown NIOSH study.
“There are not enough data to estimate [chemical]
exposures quantitatively,” Pinkerton points out. “It may
be possible in the case of some specific chemicals or
groups of chemicals to assign qualitative levels of
exposure—high versus low.”
Even if NIOSH does confirm elevated cancer rates, she
cautions, it may be impossible to determine whether
workers’ cancers are due to chemicals they encountered
at IBM. At best, Pinkerton believes, NIOSH will be able to
compare employee cancer rates to those of the general
public.
“There are enough data to determine if former IBM
employees have a higher rate of cancer than the general
population,” she explains, but “the lack of exposure data
will limit the study's ability to determine whether the
cancers are due to workplace exposures.”
Fact Sheet
|
'The lack of chemical exposure data will
limit our ability to determine whether the
cancers are due to workplace exposures'
limit our ability to determine whether the
cancers are due to workplace exposures'