Wing Video Tape

["Otto G. Raabe" <ograabe@ucdavis.edu>]

February 13, 1998
Davis, CA

This week I viewed the video tape made by Steve Wing (University of North
Carolina) of his study of multiple myeloma incidence at 4 DOE facilities. I
have heard that the DOE is making this tape available to employees at these
facilities. As you recall that study did not find any radiation influence
on the incidence of multiple myeloma! However, Steve Wing contends that if
you separate out those persons whose exposure occurred after age 45 you
find an effect. Such a selective grouping of the data invalidates the
significance, if any, of the findings. I took notes about the values
presented in the video, and they are presented below. I may have made some
transcription mistakes, but I believe these are the values.

                 <0.01 Sv          0.01-0.05 Sv          >0.05 Sv
Myeloma Cases       83                  5                  10
Controls           341                 31                  19
Odds Ratio         1.0                0.76                4.34
95% CI Range                       (0.26-2.21)         (1.46-2.90)

Well, you can see that there is something wrong with the odds ratio at
>0.05 Sv since it is larger than the stated confidence interval. This
obvious error definitely was present in the video.

I ran a Fisher's Exact Test using the statistics software INSTAT. For the
>0.05 Sv group the actual odds ratio is 2.16 (CI 0.97-4.83) from Woolf's
approximation, and the two-sided p value is 0.092, which is not
significant. Unless there is some scientific basis for knowing that an
increase in myeloma should be associated with these small radiation
exposures, the two-sided test should be used. This is supported by the
reduced number of cases in the 0.01-0.05 Sv group with an odds ratio of
0.66 (0.25-1.76).(The one-sided value is 0.052, which is also not
significant.) There is NO basis in these data for claiming that
occupational exposures caused multiple myeloma.

Wing forces a linear model through these data to claim a significant
positive slope (a radiation induction of myeloma). However, the net of the
cases among >0.01 Sv showed no effect! This study actually shows little to
support the contention that small occupational exposures to ionizing
radiation cause multiple myeloma. It is further invalidated by its
inability to consider the numerous confounding factors that could affect
the occurrence of multiple myeloma.

Otto     
		*****************************************************
		Prof. Otto G. Raabe, Ph.D., CHP
                [President, Health Physics Society, 1997-1998]
		Institute of Toxicology & Environmental Health (ITEH)
		     (Street address: Old Davis Road)
		University of California, Davis, CA 95616
		Phone: 530-752-7754  FAX: 530-758-6140 [NEW AREA CODE]
		E-mail ograabe@ucdavis.edu