Document QXjMVX251LxEY6mL3Ynrn933v

Letters LA'^ wr * S. coot- headers are invited to submit tetters for publication in this department. Submit to lr~ vmg ft Tabershaw, MO., Editor, PO, Box 2169, Rockville, MO 20652, Letters should be typewritten, double spaced and should be design*fed "fin Publication". Vinyl Chloride and Its Polymers Letter-to-the-Editor. -- This is in reference to the article by Tabershaw and Gaffey: Mortality Study of Workers in the Manufacture of Vinyl Chloride and its Polymers, IOM 16, August 1974. I find this article to be quite interesting as well as intriguing. Although it appears that the methodology employed could withstand a great deal of scrutiny, I am not at all convinced regarding the in terpretation of the results. For example, sentence one of paragraph two of the abstract is a state ment of fact. i.e. as long as readers do not get confused regarding a "study population" and the "population at risk"; the latter being the one on which inferences are made and presumably the population of importance -- in most epidemiological efforts. Very interesting, but what can we make of the statement? Not much, really, for the real inference to the "population at risk" can be found under (2) m the 3rd paragraph of the ab stract. The last sentence of the abstract suggests that vinyl chloride may be associated with cancer of multiple sites. Nonsense; no such a suggestion exists. In the same light, vmyl chloride may not be associated with cancer of multiple sites. These high SMR's may well be a consequence of sampling variability and in fact, the authors were unable to infer that significant excesses occur in the risk population. The discussion on page 513 presents rationale in support of the conclusions found in the abstract. White the authors intuitive judgment may well be correct the results are inconclusive. Using statistical justification as support for thenconclusions is most certainly invalid. Robert B Reger Chief. Statistics Branch NIOSH-ALFORD Morgantown, WV 265C5 ftepif Mttrt I. agM't f Letter Reger's first comment on our paper makes a distinction between "population exposed to risk" and "study population," with which we agree. In every epidemiological study of which we are aware, the study population is. usually by design, a subgroup of the total population exposed to risk. Unless that subgroup is chosen by some kind of probability sampling, which is rarely the case, generalization from the study population to the population at risk is not a purely statistical process. It is mediated mon sennsce# data bearir generalizati this comm The seco elusions oi on which statistically responses points, we most relev The first this case ' Mint cc plete; V single t How effective is your hearing conservation program? Your audiometer is the Pleats tend me more information on Greaon-Stadler audiometer*. final authority. NAME. FIRM/INSTITUTION. ADDRESS__________ CITY. .STATE - .ZIP. Mail this coupon to QRASON-STADLER COMPANY. SB WlNTHROP STREET. CONCORD, MASS., 017*2 15 R E-&E-I-V-E-P No matter how thorough your firm`s noise reduction or protection program, the only way you can verify that it is effective is to keep reliable records of the hearing threshold of all employees involved. This means audiometric measurements performed with accurately calibrated and reliable audiometers such as the Grason-Stadler Automatic 1703 or the Manual 1707. Grason-Stadler A ON COMPANY 96 Wlnthrop St.. Concord. MaiMChusott* 01742. T*l:$17/ 369-3767. TWX: 710-347-1050,_________________ j Tt W le to 0 Pi ti- 772 JUN181980 Journal of Shirley D. Conner ucc 041605 S the author, ktw COrrec, We U*n, >0,t for the,, V inva|l(j 'Her Branch 1RD ' WV 24*,; Ks ur pap,., between and "study ' agme In f which 'ulation ts_ up of the ^k. Unlw, tie kind of rarely the he study at risk is ess. It is mediated by a certain amount of com mon sense, and the existence of other ^ta bearing on the plausibility of the f,neralization. Our investigation shares n;4 common defect. The second comment attacks our con clusions on the ground that the SMR's no which they are based are not eatistically significant. Of the several ^ponses which might be made to this points, we believe the following two are rost relevant. The first is that the evidence at hand in his case is a complex of facts; (a) malignant neoplasms do in fact occur , are often than expected, while other causes of death, with trivial exceptions, do not; (b) the deaths from malignancies which cause this excess are generally in the high exposure groups, and the more stringent the criterion for high exposure the greater the concentration of excess deaths; (c) for one site, digestive can cers, the observations agree with what is known from other evidence. We can measure the statistical significance of each SMR, but we know of no way to put a numerical measure on the significance of the above pattern. The second is that the interpretation of a non-significant, as opposed to a significant, SMR is not as simple as Reger implies. There is ample evidence that most employed groups have a lower mortality than the general population, so that there is a question as to what the "expected" SMR should really be. There is no problem when the SMR is significantly greater than 100, but when it is not, one enters muddy waters, statistically speaking. We still believe that we are justified in what we consider to be a temperate conclusion from data which are ad mittedly less than overwhelming. Irving R. Tabershaw, M D William R. Gar.ey, PhD. JOM BACK ISSUES Mint condition issues of the JOURNAL OF OCCUPATIONAL MEDICINE are available for Vol. 14 (1972) com plete; Vol. 15 (1973) with the exception of numbers 3, 4 and 9; and complete sets of the current year. Price $3 single copy orders. Complete year sets $34. JOM - BACK ISSUES Mayo Publications P.O. Box 247 Downers Grove, IL 60515 ------------------- (w)-------------------V MEDICAL DIRECTOR he ty. juction i verify of the ormed meters or the 1:017/ The Bettis Atomic Power laboratory, a division of Westinghouse Electric Corporation, on* of the nation's leading design and development centers of nuclear reac tors, Ins an immedist* opening tor * Division Medical Director. Position requires a full range of occupational health ac tivities including ptrtodc examinations oriented to control of radiation exposure, occupational disability care, health counseling end preventive services. Background in radiation medicine a plus. United States citizenship is required. This is a major career opportunity offering an attractive salary and benefits. If interested, please send resume in confidence to: Daniel T. Masta, Suporvltnr Professional Employment, OepL JO 11-74 Westinghouse 8ETTIS ATOMIC POWER LABORATORY opwiM hr the Atomic Entr Commiisori by Watio^moM Electric Corporatise 7,0, Sm . Wilt MiffNo, To, 1S1ZZ An Equtl Opportunity Employer Journal of Occupational Medicine/Vol. IE, No. 12/December 1974 773 UCC 041606