INTIRNAL
caPR eaPO N O CN ci

U N IO N C A R B ID E C O R P O R A T IO N

To CN*rr»o)

OlV<«tQO

LOCOCtOo

Dr. Carol Stack
Chemical Manufacturers Association
2501 M Street, MW
Suite 200
Washington, DC 20037

OLD

n iG G E B U R V R O A D , P A N B U R V ,

CT OSS 17

November 4, 1983
Qngintftirtg Otfit-

Cooy co

EHA Phase II Proposal
on Vinyl Chloride Study
(UCC South Charleston and
Texas City)
Dear Carolt
As a follow-up to our phone conversation of October 31, 1983 I shall set
my conanents on the EHA Phase IIProposal to paper. In order of concern they
are as follows>
1.

I do not feel that persons previously classified as exposed should be
considered *unexposed" on the basis of available informati n in
1983. The EHA proposal contains a suggestion to analyze the data two
ways, using the original classification as well as the revised
classification. The substantial number of reclassified individuals
for some companies is definitely a cause for concern, however, I do
not feel this will be resolved simply by re-analyzing the data and
contrasting
it
with
results
obtained
using
the original
classification.
I feel that this will only create problems in
interpretation if the results are different.
union Carbide's plant in west Virginia was affected by this
reclassification problem. By searching the memories of those who
were involved in selecting the original cohort for the 1972 study, I
as not convinced that they did not have good reason to assume that
persons selected for the exposed group were indeed exposed. It is
very likely that they had data available to them which is no longer
to be found. While it is true that we cannot confirm the exposure
status on the basis of records available to us today, it is entirely
possible and even likely that if we were to start de novo in creating
an exposed cohort, this cohort would contain some individuals not
included in the original study.
Therefore, I conclude that only one of two approaches is justified
under the circumstances) 1) analyte the original cohort using the
original classification criteria, and 2) start de novo and
independently identifying an exposure cohort using records available
today. Under the EHA proposed approach, individuals included in the
original cohorts could only be excluded) there is no provision for
the inclusion of employees found to be inadvertently ommitted from
the orignial cohort. This could lead to substantial bias.

UCC

014879

 2 -

2.

The coding of contributory causes from the death certificate is a
worthwhile endeavor. All death certificates should be coded in this
l u m t f, not just those obtained in the Phase II study.

3.

is the national death index (NOI) e reliable source for det reining
individuals known to be *alivea? is it valid to assume that if the
person does not appear as a death after 15 months have elapsed, that
this person must be alive? I state these simply as questions as I do
not have the necessary information to provide any answers.

4.

I would prefer to see deaths which occurred during the time period of
the 8th and the 9th revisions of the International Classification of
Diseases used in preference to coding back to the 7th revision. The
■Monson" program allows for the conversion of codes from the 8th to
the 7th or from the 7th to the 8th. It is always preferable to be as
up to date as is practicable.

I am sorry that I was unable to attend the last meeting.
to seeing you at the next one!

I look forward

Sincerely,

Susan G. Austin, Se.O.
Corporate Director of Epidemiology
SGA/pmb
ccs

Mr.
Mr.
Mr.
Dr.
Mr.
Dr.

J.
I.
T.
T.
N.
0.

H. Barrett
1. Greenberg»
S. Lawrence
A , Lincoln
L. Wheeler
Wong

ucc

014880