Document mq0yVkqrmq7kkX4kyeJNqpogO

FILE NAME: Exxon (EXX) DATE: 1958 DOC#: EXX043 DOCUMENT DESCRIPTION: Journal Article - Occupational Cancer - Report Prepared by the Council on Industrial Health ili The Journal OF THE American Medical Association : xe<"V W;i -J '-g?-S j* t ' LaJ-r ' -'*-sSaK,: S f V ' : - t s i i r i ? r. r 8; E dited for the Association U nder the D irection of the Board of T rustees by . AUSTIN SMITH, M.D. VOLUME 166 >at'Maahy-April, 1958 AMERICAN MEICAL ASSOCIATION, CHICAGO 10 COUNCIL ON INDUSTRIAL HEALTH The following report has been prepared by the Committee on Occupational Cancer of the Council on Industrial Health. B. Dixon Holland, M.D., Secretary OCCUPATIONAL CANCER METHODS OF EPIDEMIOLOGICAL STUDY The first form of cancer with an identified causal factor was described nearly two centuries ago. It occurred with unusual frequency among workers in a single occupation and, hence, became known as chimney-sweep's cancer. In spite of the fact that the first knowledge of the etiology of cancer can be traced to Percival Pott's 1 observation on chimney sweeps, w e have no reliable information today from which a reasonable estimate can be made of the proportion of cancers that have some causal rela tionship to an exposure of occupational or indus trial origin. Although a number of specific types of cancer have been associated with occupational exposures since Percival Pott's original studies, there is no reason to believe that these exhaust the possible list of such cancers. In recent years, an increasing amount of scientific literature has appeared on the subject, confusing, at times, both the physician in private practice and the full-time industrial physician. Part of this con fusion arises from the fact that occupational cancer is different from many other well-known occupa tional diseases, more difficult to identify, and, hence, more difficult to study. Experience has demon strated that the identification of occupational or industrial causes of human cancer is made difficult by certain characteristics of cancer itself, as well as by labor mobility. Occupational cancers can rarely be distinguished on morphologic grounds from those due to other causes. Furthermore, several years may elapse between the date a worker is first exposed to an agent capable of causing cancer and the time of development of a tumor to the stage where it can be clinically recognized. From a study of 300 cases of mule-spinner's cancer reported in England and Wales from 1920 to 1928, Henry * found that the time elapsing from ChftJnnim of the Committee on Occupational Cancer is G. Burroughs Mider, M.D., associate director in charge of research, National Cancer Institute^ Bethesda, Md, Members of the Committee are as follows: Harold F. Dorn, PhJX, Biometrics Branch, Division of Research Serv ices, U. $. Public Health Service, Department of Health, Education, and Welfare, National Institutes of Health, Bethesda, Md.; John G. Downing, M.D, Boston; Robert E. Eckardt, M.D., director, Medical Research Division, Esso Research and Engineering Company, Linden, N. J. Mr. Edward A. Lew, Metropolitan Life Insurance Company. New York; W. Harvey Masuigan. Ph.D., Allied C h em ica l a n d D y e Corporation, National Aniline Division, Hopewell, Va.; and John J. Phair, M.D., University of Cincinnati College of Medicine, Cincinnati. the date of starting work as a mule spinner until the date of diagnosis of cancer ranged from 16 to 63 years and that for 67% of the cases this latent period varied between 35 and 54 years. Several of the workers were no longer employed in this occupa tion at the time of diagnosis. The average latent period for mule-spinner's cancer is appreciably longer than that for some other forms of occupa tional cancer. However, a latent period of several years is characteristic of most cancers. For example, the average interval from initial exposure to diag nosis of lung cancer among workers in United States chromate-producing plants was 17.2 years, with a range of 4 to 47 years. * Finally, there is a "normal" or "expected" inci dence of cancer, whereas most occupational dis eases do not "normally" exist. Thus, if a man de velops silicosis, it probably arose as a result of his occupation, since this disease does not have a "nor mal" incidence. Accordingly, it becomes necessary in the study of occupational cancer to establish that the disease occurs either at a higher than nor mal incidence in a given occupation or with a great er relative frequency of site or type, so that there is no doubt about the causal relationship between the disease and the occupation. For this reason, single case reports of silicosis, lead poisoning, or other oc cupational diseases may make significant contribu tions to our understanding of most occupational dis eases, but individual case reports really contribute little to our understanding of occupational cancer. The same problems have arisen in attempts to re late trauma to cancer--and for the same reasons. These difficulties have led to vigorous debates in the scientific literature, which have perhaps been best summarized by Stewart.4 The past three decades have witnessed the birth and large-scale development of experimental meth ods for the study of carcinogenesis. This has per mitted identification of an extensive list of chemical compounds capable of causing cancer in animals in specific experimental situations.5 A number of these compounds are encountered in various manu fa c tu rin g processes. It is not co rrec t to assume from these two facts that the risk of cancer developing 2172 COUNCIL ON INDUSTRIAL HEALTH J.A.M.A., April 26, 1 among industrial workers thereby is either increased or left unchanged. This can be ascertained only by appropriately conducted investigations, and the answer ultimately must be derived from observa tions on workers in specific occupations and indus tries. It is essential, therefore, for the investigator who would study occupational cancer, to have, for com parative purposes, reasonably reliable estimates on the normal incidence of cancer of various types and sites. The collection of such data is a difficult and time-consuming task, requiring almost constant work by a team of trained epidemiologists and stat isticians. Cancer incidence increases with age, shows site-specific variations with sex and color, may be modified by geographic factors, and is being continually altered by the aging of the pop ulation. In addition, certain specific cancers, such as cancer of the stomach in males, may show unpre dictable, unexpected, and unexplained changes in incidence with time. Finally, because of improved diagnostic techniques, greater awareness of cancer, changes in the reporting of disease occasioned by legislative or administrative action, and improved reporting of the cause of death, additional artifacts may be introduced into the calculations. At the present time, the best available rates of normal cancer incidence, prevalence, and relative fre quency are those assembled by Dorn and asso ciates.8 Annual and special reports of the National Office of Vital Statistics provide mortality statistics for the entire country and for various geographic subdivisions.7 Finally, certain states, particularly Connecticut and New York, have assembled simi lar data.8 It is to be hoped that such efforts will be continued and expanded in order that up-to-date, reliable figures will be available for the necessary comparisons, not only to aid occupational cancer studies but also to keep abreast of decade-by-dec ade changes in cancer incidence and distribution. The physician of today, and increasingly in the future, will be called on to provide answers as to whether the increased use of a chemical or the in troduction of a new chemical has increased old hazards or introduced new ones. If the advance ment of the chemical age and the anticipated im provements in standards of living continue, it will be necessary for the physician to keep well in mind the caution so well phrased 10 years ago by Stew a rt4: "It is probable that with the development of chemical industry we will see new chemical can cers and medicine must be on the watch for such developments. But it must refrain from hastily ascribing to industry those tumors whose incidence falls well within the expectations for the popula tion as a whole and from making premature con clusions based on lack of appreciation of statistical methods." While the study of occupational causes of < is complicated, it is not impossible. Since the s of investigation is so broad, a great deal of tinT and energy can easily be spent in fruitless stud unless some preliminary idea can be obtained of| those occupations which appear to be associated! with cancer hazards. The principal support for such ideas or hypotheses is clinical impressions derived ' from the study of individual cases of disease and| the analysis of existing records of morbidity and" mortality. Accordingly, it is believed worthwhile^ to describe various sources of data now available' and to indicate, in summary fashion, certain ap';. proaches and techniques that may prove fruitful. % Official Statistics The official certificate of death required for each;, deceased person in the .United States provides for i the entry of his usual (and probably last) occupa tion and industry. Very little use has been made of this source of information, largely due to uncer tainty concerning the comparability of the entries^ for occupation and industry on death certificates* with corresponding data from the census of popu-T lation. Information on this point soon will be avail-i able from current study of the corresponding en*J; tries for occupation and industry found on the?, death certificate and on the 1950 census schedule for a sample of deaths. ' *i f At best, official mortality statistics provide;a measure of the total effect of an occupation and the general level of living of persons following that occupation. Specific occupational hazards are not - likely to be identified by a study of these records. . Only one entry for occupation is requested, andi frequently either no occupation or an entry of "ret| tired" is reported for persons who have stopped^ working because of old age. Occupations are de- . scribed only in general terms, so that it is not pos- ; sible to identify specific groups of workers who are thought to have been exposed to a suspected health hazard. Furthermore, each industry, and usually each individual company, employs entirely different names for the same job assignment. In spite of their many defects, mortality statistics,; if analyzed with imagination, may provide clues B - f occupational or industrial hazards which can bef* explored by more specific studies. The following examples illustrate the types of studies that ha been made with death certificates. Mortality Rates by Occupation.--The informa-;, tion on cause of death and occupation and industry . on death certificates and census reports can be used i to compute death rates from different forms, of cancer. The most extensive series of death rates3 are those published by the Registrar General of England and Wales. References to these data may be found in an article by Kennaway. A compre hensive study of occupational mortality in the United States has never been made. The National Voi. 166, No. 17 COUNCIL ON INDUSTRIAL HEALTH 2173 Office of-Vital Statistics has initiated a study based upon deaths occurring during 1950, the preliminary results of which have been reported.10 Mortality rates by occupation can be computed for the population of specific localities, for example, a city or county, provided the number of workers is large enough to yield reliable rates. The informa tion needed for such studies is (a) the number of deaths of persons in each of the occupations to be studied, and (b ) the total number of persons in each occupation classification by age and sex. The number of deaths can be obtained from the local or state office of vital statistics; unless special in formation iS available locally, statistics concerning the number of workers by occupation must be ob tained from the U. S. Bureau of the Census. Studies should be done for a period near the date of the census of population in order to avoid the neces sity of preparing-population estimates. A good il lustration of this approach has been reported from Sheffield, England, by Turner and Grace.11 - Mortality Rates for Specific Areas.--The analysis of mortality statistics for cancer of specific sites for local areas, such as a county or group of counties, may be undertaken in orda- to determine whether an unusually high death rate or an unusually large proportion of deaths exists for a specific form of cancer. Available information concerning possible environmental hazards may be studied simulta neously to determine if there are any that appear to be associated with the known facts about can cer mortality. These studies may ;be conducted by systemati cally studying the variation in cancer mortality among the counties of a state, as was done by Mancuso and others12 in Ohio, or-by analyzing mor tality from cancer in an rea where an industry whose workers are thought to be exposed to some cancer hazard is located. Projects of this kind are most.likely to yield useful leads if a large propor tion of workers in the area are employed in a single industry. Mortality rates for total cancer and for cancer of specific sites by age and sex can be computed for a period of several years. If the population under observation is too small to yield reliable age and sex specific death rates, an ex pected number of deaths can be computed by mul tiplying the population of the local area distributed by age and sex by the death rates for some appro priate standard population, for example, the popu lation of the state in which the local area is located. The expected number of deaths is then compared with the observed number. For comparative pur poses, similar computations should be made for other counties that do not have an appreciable number of workers in the industry in question. If the deaths of persons who have been em ployed by a spcifie industry can be identified, a comparison of the causes of death in this group with these of the rest of the population can be made. The study by Hill and others 12 of workers in a factory handling inorganic arsenic compounds is an excellent example of this method. Mortality from Specific Forms of Cancer.--Since most of the carcinogenic agents of occupational origin that have so far been identified cause cancer of a specific site, several studies have been made in Great Britain, based on death certificates, with a specific type of cancer, for example, cancer of the skin as the cause of death.14 The usual pro cedure has been to review all death certificates filed in a specified area for which population data are available and to compute death rates by occu pation. The areas selected for study have been either the entire country or a specific local area where the existence of a cancer hazard is suspected. Social Security Records.--The records maintained by the Bureau of Old-Age and Survivors Insur ance, Social Security Administration, Department of Health, Education, and Welfare, include the employment history of all workers covered by the act since January, 1937. Information may be ob tained concerning the name and address of each employer, the dates of employment and the in dustrial classification of each employer according to a four-digit code. These records may be used by themselves or in conjunction with official vital statistics for studies of mortality among persons who have worked in specific occupations. For example, one could start with a list of per sons who are known to have been employed in a given occupation or industry. Although detailed information usually will not be given out for per sons who are still living, the date of death and employment history can be obtained for persons who have died. In this way the subsequent history of part or all of the workers who have left the oc cupation can be traced. Alternatively, one could start with a group of persons who have died from a specific form of cancer and obtain their previous work history. A control group of persons who died from causes other than the type of cancer being studied should be investigated simultaneously in order to judge whether the number dying from cancer who had worked in a given occupation was large enough to suggest the existence of an occu pational hazard. Occupational Disease Re])orts.--A survey con ducted by the United States Public Health Service in 1953 found that only 28 states had statutes or regulations requiring physicians to report patients suffering from occupational diseases. No serious attempt is made to promote compliance, so that the regulations often are ignored. A much larger number of occupationally induced diseases are reported in connection with claims for workmen's compensation. About one-half of such diseases are skin disorders. A m ong the rem ain d er, complaints due to physical conditions of the work ing environment, systemic effects of chemicals, and \ / / V 2174 COUNCIL ON INDUSTRIAL HEALTH J.A.M.A., April 26, 1958 infectious diseases are the most common. At best, the claims give only a rough indication of the oc currence of the types of complaints reported. Even where awards have been sustained for occupational cancer, the validity of the alleged causal relation ship frequently fails to satisfy critical and compe tent medical judgment. Neither official reports of occupational diseases nor claims for workmens compensation have yet yielded useful information concerning the incidence of occupational cancer in the United States. In contrast, the Factory Inspection System of Great Britain has been an important source of in formation about occupational cancer. The first med ical inspector was appointed in 1898. As presently organized, the medical staff is responsible for the promulgation and supervision of regulations for various occupations, the investigation of cases of industrial poisonings, the study of new industrial processes that may be injurious to health, and the supervision of the work of physicians who give certificates of fitness for work. A list of lesions that may be caused by known or suspected carcinogenic agents has been estab lished, and these must be reported to the Chief Inspector of Factories. This information is supple mented by inspection of plants where possible carcinogens are produced or handled. Special stud ies have been made of cancers developing among workers engaged in the extraction, processing, or use of shale oil, petroleum products, asbestos, nickel, and dyestuffs. In 1947, for example, Henry 15 published an analysis of 3,753 skin cancers in 2,975 persons reported to the Chief Inspector of Fac tories from 1920 to 1945 inclusive. Pitch, tar, or tarproducts accounted for about 60% of these cases; shale oil, mineral oil, or bitumen caused most of the remainder. Insurance Company Statistics Life Insurance.--Life insurance company follow up studies have the special merit that they usually deal with relatively large numbers of persons who can be automatically traced for long periods of time through the circumstance of their being in sured. A suitable standard of comparison is avail able in the experience of the aggregate body of insured persons of the particular type. Moreover, in the absence of national mortality statistics by socioeconomic classifications, the experience of dif ferent groups of insured persons sheds some light on the relative prevalence of cancer in different socioeconomic classifications. Life insurance companies maintain detailed rec ords of death claims, disability claims, claims for hospitalization, su rg ica l benefits, a n d m e d ica l ex pense benefits, as well as special records for medicoactuarial studies and, occasionally, records of periodic health examinations of their employees also. All of these can be tapped for information regarding the prevalence of cancer. It is important to realize, however, that the records maintained by life insurance companies are prepared primarily for administrative purposes and only incidentally yield information about the prevalence of cancer among insured persons. Generally speaking, from death claim records and records maintained on the insurance in force, it is possible to compute death rates by age, sex, and race for various classes of insured persons. The experience of industrial policy holders of the M et ropolitan Life Insurance Company has been re ported in a number of studies,*8 which show for the years 1911-1955 the mortality rates by age, sex, and race from cancer of the following sites: di gestive organs, female genital organs, breast, buccal cavity, skin, lungs and pleura, bladder, prostate, and other organs. Current information along these lines is published in the Statistical Bulletin of the Metropolitan Life Insurance Company. Because the applicants occupation is usually a major factor in determining the life insurance risk, such information is given on the application for insurance. It is often possible, therefore, to cal culate the cancer mortality among insured persons by industry or occupation. However, occupation is ^ usually available only to the insurance company at time of application for insurance or at time of death, intermediate occupations ordinarily being unavailable. ' The records for group life insurance (that is, insurance issued to groups of employees generally with the cooperation of and contributions from the employer) have been used to calculate can cer death rates for all employees actively at work in particular industries. It should be noted that the cancer death rates for all employees actively at work in an industry so covered may not be sig nificantly higher than average. Yet there may be some processes in that industry in which relatively small numbers of men are employed and which are associated with very high cancer death rates.- However, it is frequently possible, to determine cancer death'rates for individual plants, and, if these plants concentrate on specific processes, the cancer death rate associated with such processes can be derived. A sttidy covering the period 1927 1934 17 indicated significantly higher than average mortality among felt hat makers, tube, rod, and pipe mill employees, paper manufacture employees, hotel and restaurant workers, shoe and light leather goods producers, and employees of electric and street railways. In recent years a number of corporations have become interested in the mortality from cancer of specific sites among their employees, and such studies have been made in various industries or plants from group life insurance records. The cus tomary procedure has been for the employer to approach the insurance company and request that >r i Vol. 166, No. 17 COUNCIL ON INDUSTRIAL HEALTH 2175 a study be made.. Inasmuch as the relationship be It is important to bear in mind that the records fa,/;i tween the holder of a group life insurance contract maintained by life insurance companies can merely PS and a life insurance company is confidential, a life indicate the fact that cancer mortality has been 1 1 insurance company is not free to publish the re high in certain occupations. It does not by any sults for individual concerns. However, investiga means follow that such cancers are occupational in tions made on an industry-wide basis are not sub origin. Additional investigations must be made to ject to, this restriction. At the present time, about determine whether the association is causal. 36 million persons are insured under group life The records of accident and health insurance insurance policies, and these include most of the companies on operations and treatments for cancer F M employees of the larger companies in the major provide another starting point for investigations f industries. of cancer mortality. However, in such instances an (? Industrial Insurance.--The records for industrial elaborate and costly machinery would be necessary 4*, insurance (that is, insurance issued to individuals to obtain a reasonably complete follow-up. Such a A in small amounts on which premiums are payable follow-up is automatically obtained where a person 1 weekly or monthly, usually collected at the home carries life insurance. The disability claim records of the insured) can be used to compile cancer of life insurance companies can be used as a means death rates according to the occupation of the in of determining the survival and recovery rates sured at the time his policy was issued. It is usu among persons insured under ordinary life insur ally also possible to ascertain the occupation of ance policies who submit disability claims on ac the policyholder at the time of his death. A large count of cancer. Such disability claim records usu f M proportion of the adult males so insured are em ally contain more detailed information as to the A f ployed in industry. About 20 million adult males site and type of cancer than do claim records. are currently insured under industrial policies, and Furthermore, persons insured under ordinary life sx--' these include, chiefly, the lower paid urban wage- insurance policies who submit disability claims on \ :: earners. account of cancer can, as a rule, be traced after 1 m From several studies made in the Metropolitan recovery to determine the rate of recurrence of Life Insurance Company ( based on the occupation cancer. at death), it would appear that the following oc The records of periodic health examinations of cupations have been associated with higher-than- life insurance company employees have a potential s. average death rates from lung cancer: copper value for the investigation of the natural develop smiths and tinsmiths, painters and varnishers, and ment of cancer in its preclinical stages, even though roofers and slaters. Less definite, but suggestive, employment by a life insurance company docs not have been the findings' for such occupations as carry any known occupational cancer hazard. , 1n machinists and mechanics, railway enginemen and `rx- trainmen, proprietors and bartenders of drinking Industrial Records establishments, brick and stone masons, plumbers, The use of public health and insurance company % sctleoathmifnitgtewrsoraknerds. gas fitters, and tailors and other sattatbisetsitcsoninly tihnedicstautedyaroefasocincuwphaticiohnadletcaailnecderstcuadn - / Special medicoactuarial records maintained by ies should be undertaken. These must be under most of the larger life insurance companies on or taken within the industry itself and require the dinary life insurance (that is, insurance issued to careful cooperation of several different groups. individuals in amounts of $1,000 or more) lend Since exposure within industry is far more likely 1 ME *- themselves most readily to the computation of to be greater than outside of industry, it is prob v cancer death rates according to occupation of the able that studies conducted in the primary pro A insured at the time the policy was issued. How ducers of a chemical will be more productive than ever, . the last intercompany study of mortality studies of secondary users. Such investigations are among persons with ordinary life insurance accord difficult and time-consuming to conduct, and un I ing to occupation was made in 1937. This and ear less planning for them is undertaken in advance, lier investigations indicated that cancer mortality they may be impossible to complete. was high among waiters and cooks, hotel keepers, The reason for the lack of such planning in the guards, watchmen and janitors, painters and brick past is found in the history of industrial medicine. and stone masons, and tailors and clothing opera At the turn of the century, industrial medicine was tives. When studies of this kind are undertaken in primarily concerned with the treatment of acci M t the future, they will probably be made under the dental injuries. This state of affairs still exists in . VIi aegis of the Committee on Mortality under Ordi many industrial plants, although the more pro I V nary Insurance and Annuities of the Society of gressive industrial physician is increasingly recog Actuaries. About 60 million persons are currently nizing the importance of preventive and construc V insured under ordinary policies, comprising mostly tive health measures. Intensive safety campaigns middle-class males in nonhazardous occupations. and much safety engineering have reduced indus- - y / JUILm S H p tS 'i- 217G COUNCIL ON INDUSTRIAL HEALTH J.A.M.A., April 26, 1958 trial accidents to the point approaching the infini tesimal. Thus, only 3% or less of absenteeism in industry is due to in-plant accidents. The increas ing importance of preventive1-constructive mcusiues m m dustna l medicine is attested bv the recent rec ognition granted occi pational medicine as a section ot the American board of Preventise Medicine, Suspicion for the existence of an occupational cancer max also arise from the clinical judgment of tlie physician Main' medical advances have .uiseii out of the "clinical lmiu lies" of well Pained and obseivnnt pin.Acinus. It is necessaiv onlv to point out that the clinical hunch can do no moie than arouse suspicion, it neither proves noi dispiove.s the existence of oeenpational canecr H o w ever, u h e n the clinical hunch serves as tin1 basis of careful, detailed study, it max lead to significant advances m our knowledge. Such clinical bundles may arise out of a plnsician's being inipiessed by tin1 num ber of cases he sec's in a iclatixelx i estrieted group, b\ coiollarx clinical findings in in dividual cases, or even In inquiries from the work ers themselves, the nnsxveis to which he mav not knoxv b u t yet feel obligated to develop The1 in creasing potential significance of occupational can cer aiising from the1i lcreased cpiantitv and diversi fication of chemicals used m or m ade In mdustrx' place's an increasing rexponsibilitv on the industrial phvsician, full or part-time W hen a suspicion lias arisen, fiom studies of 'public health or insurance statistics as outlined aliox'e1, by clinical judgment, or from an unusual location of a cancer, such as scrotum or bladder, that an occupational cancer mav exist in a given industry, subsequent studies follow a pattern which it mav lie helpful to outline1 The first step ;s to obtain all of the earner case's of that in d u stn for a jiciiod of time (perhaps 5 or 10 years) and to anaK/.c them for oxen-;.11 incidence in the industry. It is recognized immediately that the first requisite is that some1 system of cancel i('porting within the mdustiy is neeessarc Let in assume, for ('sample, that m a gix'e'n industry 50 case's of cancer have been reported oxer the* past iO ve-aix If this oc curred m an industn in which there was an aver age animal emploxtv census of 200, tins xxould u'present an incidence of 2.500 case's per 100.000 per year, which would lie distinctly abnormal. If this incidence happened in oi.lx one year, it could well be a statistical ,,rtifacl, hence, it is important to have tlie study encompass more than one vear On the* other hand, i: this occurred in an industry m which the average annual census was 2 000. tins would represent an incidence of 250 pei 100.000 per year, which would lie just about the anticipated rate If it happened in an indnstiy with an average annual census of 20,000, this would repic'sent an incidence of 25 per 100,000 per year. Such a result would clearly be so ranch below the expected that it should be immediately evident that only about 10% of the cases are actually being reported with the system used. Such a result should lead to a critical evaluation of the effectiveness of the report ing system. More detailed studies would be futile in this stage unless the cancer being studied is of an unusual site, such as the scrotum or bladder. L et u.x assume, however, tha t an incidence of 250 per 100,000 per year has been determined. The next step would be to analyze the cases by sex. Tims, if it were found tha t 40 of the 50 cases ocenrred in men and that about 80% of the employees were male and 20% were female, the incidence rates of tlie sexes would be approximate! / the same. On the other hand, if 40 of the 50 cases occurred in men in an industry in which only .20% of the population were men, it would certainly be sug gestive of two situations: first, that the incidence in men was suggestively high ( i. e., 1,000 per 100, 000 per y e a r ' and second, tha t the incidence in x\ omen inordinately low (i. e,, CO per 100,000 per year). Both situations would warrant further study in an attempt to understand the causes. The next step in the procedure would be to pull out the medical record of each of the 50 eases re ported and study it in great detail. The importance of an adequate medical record, containing a maxi mum of information and documented by originals or copies of laboratory reports, x-ray reports, biopsy reports, summaries of hospitalizations, autopsy rec ords. and death certificates (when death has been tin1 result), now becomes apparent. In looking to the future, every physician, and particularly every industrial physician, can understand the importance of tlie maintenance of adequate industrial medical records. It is apparent that if the industrial medical leeord is inadequate the promise of rewarding re sults from the detailed study of them is indeed remote. However, after the medical records have been found adequate, detailed studies of them will now permit the proper classification of cases. For in stance, such studies may reveal that cases reported as lung or liver cancer are, in reality, metastases oi cancers of different primary sites. They should, therefore, be reclassified as to the primary site of origin. In addition, it may be found that certain conditions initially believed to be cancer and re ported as such have been found on detailed study to be some disease other than cancer. The initial repent on such eases may never have been changed and tlie eases carried erroneously as cancer cases. Such detailed study, therefore, may lead not only to an alteration ot the total cases reported but also to changes in the distribution of cancer cases in organ systems. This detailed study of the medical record also permits all cases to be classified by age at apparent onset of the disease. This study may be rewarding Vol. 166, No.. 17 COUNCIL ON INDUSTRIAL HEALTH 2177 because if it is found that the average age at onset of cancer, either in general or in site-specific situa tions, is significantly lower than in the general population this itself may be justification for pur suing the studies further. The next step in the procedure is to obtain an occupational history on the individuals with cancer. This is not an easy task, since the individual him self, if still living, may not remember all of the various occupations he has been involved in, even within one industry. At this point, the personnel and/or payroll office may be of great assistance to the investigator. Because it is necessary for in dustry to know the cost of its finished goods, cost accounting records are usually maintained. Since different jobs may demand different rates of pay, it is frequently possible by reviewing such records to follow a man's career within an industry'in quite specific detail, at least, as far as different depart i ments and the major classifications within each of these departments are concerrted. A study of such occupational records may then provide a further clue for the continuation of these studies. Thus, if an inordinate number of cancers appear among one group of workers or if cancers of a specific site all occur in men with a long service in one depart ment, further detailed studies again may be war ranted. At this stage, the numerator of the incidence equation is available, but the denominator is not. This equation is as follows: Incidence (cases per 100,000 per year) = total number of cases for n years -- population at risk X n X 100,000. Let us assume that, at this point, examination of the rec ords shows that there are 10 cases of lung cancer among gadget wreckers. Further review of the records discloses that among these 10 persons the minimum service as a gadget wrecker is nine years. Two more men have had 10 years of such service, one 11, two 13, one 15, one 18, one 23, and one 25 years. It would then be extremely helpful to the physician if he were able to obtain from the per I sonnel department the number of men who have worked nine years or longer as a gadget wrecker. This figure would supply him with the denomi nator in the above equation. Thus, if our study en i. compasses 10 years, ending Dec. 31, 1955, and 100 men have worked nine years or more as a gadget wrecker by that date, the equation becomes as follows: Incidence = 10=100X10X100,000=1,000 I per 100,000 per year. This incidence of lung can f- cer is some 30 times the expected incidence in men of all ages and about 14 times the expected inci t dence in men over 45 years of age. > rf It might appear that the study is now complete, ns. but actually i t .should proceed .further. The next $ step would-be to obtain from the personnel depart ment the names of all men who have ever, regard less of the duration of such work, worked as a gadget wrecker. Then the medical records of each of these men should be reviewed in order to insure that no cases of cancer have been overlooked. Ex perience in occupational cancer studies has shown that this review may frequently uncover proved cancer cases among these medical records, which, for one reason or another, may have escaped re porting. It is conceivable that such studies will reveal cases with an exposure time which is less than the previous minimum uncovered in the re ported cases and, hence, may require revision of the population at risk. This type of analysis of medical records will insure that no cases have been missed. The collection of data on total employment in the department will permit analysis of cancer data in a variety of ways, most of which will give negative results which only serve to establish more firmly the significance of the positive results in the more selected groups. Thus, analysis of incidence rates for all men employed one day or longer, 5 years or longer, 9 years or longer, and 15 years or longer may show no statistically significant in creased incidence until employment has exceeded 9 years, and an increasingly significant increase the more employment has extended beyond 9 years. It is obvious from what has been said that when the incidence rates have been determined they should be analyzed for statistical significance in comparison with the expected rates in the general population. This analysis should generally be per formed by someone well acquainted with statistics as applied to biological problems. This is important because a four-fold or five-fold increase in inci dence, which may appear on the surface to be sig nificant, may really not have any significance sta tistically because of the small number of cases. In addition, consideration should be given to the possibility that a difference in the characteristics of the industrial population and the general popu lation with which it is compared may account for the observed difference in incidence. The incidence of cancer of specific sites is known to vary by age, sex, race, and geographical region, so that the pos sible effect of these factors should be investigated in the interpretation of differences in incidence rates. Finally, when these rates have been determined, one last step may be worthwhile. This is to go to the foreman of the operation involved and inquire about the existence of shift books. A careful study of such shift books may reveal a pattern of occu pation by subdivisions in the department which cannot be determined from the records maintained by the personnel and/or payroll office. It may be found that, although 100 men work in a depart ment, perhaps only one-half or one-quarter of them are involved in one particular operation of the de- i / K si' ) 11 V or / i i i J I \ Of X U 'AT, "S l V )f A s 1 1 f M Jl/jI IV V/ ;i ; : ? f Mi 1? I lVf 2178 COUNCIL ON INDUSTRIAL HEALTH J.A.M.A., April 26, 1958 partment and that all of the specific cancer cases are to be found among those men performing this specific operation. This final study may be of great importance from two standpoints: first, it may in crease the significance of the incidence rates, and second, it may serve to pinpoint the exposures and, hence, lead to the easier detection of the carcino genic agent and, thus, to its ultimate control. It is necessary to point out that, in small indus tries with a heavy labor turnover and also in cer tain industries with problems peculiar to the in dustry, such as the construction industry, industrial medical records may be quite incomplete or even lacking. In such cases, obviously, the above tech niques cannot be applied. Other methods, such as the use of social security records or high-caliber union medical services, may need to be developed for statistical, epidemiological studies of this na ture. In these cases, again, the first requisite will be the collection and periodic analysis of adequate medical records. From what has been said above, it can be seen that occupational cancer studies even within an in dustry can be laborious, time-consuming proce dures. There are no real shortcuts to such adequate studies, and attempts to perform "quickie" or illconceived studies may lead to very misleading re sults. Adequate studies require the services of a phy sician who is interested in such studies and trained well enough to determine the adequacy of medi cal records. He must also have enough familiarity with the industry being studied in order that occu pational histories will have some meaning to him. He must, in the course of such studies, enlist the support of the personnel and/or payroll depart ments, a competent biostatistieian, the foreman, and other groups within the industry, including, particularly, engineers and chemists who are fa miliar with the operations being studied. Finally, he needs, for comparative purposes, age, sex. and site-specific incidence rates for the general popula tion and for segments of the population more com parable to those he is studying, if his group differs in any significant way from the general population. The above comments outline certain require ments for the future which will greatly assist in the performance of occupational cancer studies. First, and most important, are adequate industrial medical records. Second is a mechanism of record ing and reporting disease which is capable of peri odic analysis. This suggests McBee sort cards and IBM or other mechanical methods of analyzing data. Third is a mechanism of recording occupa tions and the occupational exposures of various segm ents of th e w orking force. These occupational data must be capable of mechanical analysis so that a periodic running survey of occupational ex posures, including their length, may be correlated with the medical data. If industry and industrial physicians become cognizant of the demands which, will be made of them in the future and plan a t . this time for the recording of data so that they can be analyzed, it will be possible for them to become aware of any significant shifts in the inci dence of disease, including occupational cancer, in the working force. The earlier such shifts are rec ognized, the sooner can their causes be studied and understood. This in turn will lead to the con trol of significant exposures and the genuine frui tion of industrial medicine, whose confines lie within the preventive-constructive framework of medicine. These comments, although true of occu pational cancer, are equally valid for all occupa tional disease. The methodology can be applied equally successfully by the industrial physician, the private physician, the public health physician, and the union physician, but, because of the need for assistance from other groups within industry, it will probably be most successfully applied by the industrial physician.'It will become important for industry to know that its engineering controls are keeping pace with its technological developments, and this can most effectively be done through stud ies comparable to those outlined above. Summary In the majority of cases, occupational cancer cannot be distinguished, morphologically, from cancer not due to occupation. As a result, the diag nosis of occupational cancer is more difficult than the diagnosis of many other occupational diseases. The suspicion for the existence of an occupa tional cancer hazard may arise from clinical ob servations, from public health statistics, or from the statistics available from insurance companies. The value and limitations of each of these must be realized. Failure to realize the limitations of these methods may lead to misleading conclusions. Failure to appreciate their value may result in overlooking important leads for further detailed studies. \ " ' In modern industry, with the tremendous expan sion in the manufacture and use of chemicals, some of which may be carcinogenic, the responsibilities of the medical profession for detecting new occu- _ pational cancers have been increased. The first requisite in such studies is the maintenance of adequate medical records by competently trained physicians. In addition, it will be necessary to main tain adequate employment and occupational rec ords. The correlation of cancer with occupation must be made from these records with a full ap preciation of the natural history of nonoccupational cancer, such as its age, sex, and site distribution. Specific increases in age, sex, and site-specific cancer incidence above normal incidences for the same age, sex, and site represent the only reliable Vol. 166, N.o. 17 COUNCIL ON INDUSTRIAL HEALTH 2179 evidence for the existence of an occupational can cer hazard. The importance of normal incidence rates in the general population, to serve as a base line control, is therefore self-evident. Further, a careful methodology for occupational cancer studies from industrial medical and occupational records . is required. ' The important aspect of occupational cancer studies lies in the fact that an understanding of the disease will permit its control and, ultimately, its complete prevention. With modern engineering methods and control and protective devices, pre vention can be achieved at reasonable cost. References 1. Pott, P.: Chirurgicaf Works, vol. 2, Dublin, James Williams, 1778, p. 403. 2. Henry, S. A'.:. Cancer of Scrotum in Relation to Oc cupation, London, Oxford University Press, 1946. ` 3. Machle, W., and Gregorius, F.: Cancer of Respiratory System in United States Chromate-Producing Industry, Pub. Health Rep. 83*1114-1127 (Aug. 27) 1948. Health of Workers in Chromate-Producing Industry, Publication 192, Federal Security Agency, U. S. Public Health Service, 1953. 4. Stewart, F. W.: Occupational and Post-traumatic Can cer, Bull. New York Acad. Med. 23:145-162 (March) 1947. 5. Shubik, P., and Hartwell, J. L.: Survey of Compounds Which Have Been Tested for Carcinogenic Activity, Supple ment 1, Department, of Health, Education, and Welfare, U. S. Public Health Service, to be published. . 6. Dorn, H. F., and others: Morbidity from Cancer in United States: I. Variation in Incidence by Age, Sex, Race, Marital Status, and Geographic Region, Public Health Mono graph' 29, Department of Health, Education, and Welfare, U. S, Public Health Service, 1955, pp. 1-121. 7. Vital Statistics of the United States, two volumes, an , nual, Department of Health, Education, and Welfare, U. S. Public Health Service. Vital Statistics: Special Reports, vol. 43, no 11, Department of Health, Education, and Welfare, U. S. Public Health Service, 1956. 8. Griswold, M. H.; Wilder, C. S.; Cutler, S. J.; and Pollock, E. S.: Cancer in Connecticut, 1935-1951, Hartfort, Conn., Connecticut State Department of Health, 1955. Annual Report, Bureau of Cancer Control, New York State Department of Health. 9. Kennaway, E. L.: Data Relating to Cancer in Publica tion of Ceneral Register Office, Brit. J. Cancer -1:158-172 (June) 1950. 10. Moriyama, I. M., and Guralnick, L.: Occupational and Social Class Differences in Mortality in Trends and Differentials in Mortality, in 1956 Conference of Milbank Memorial Fund, New York, Milbank Memorial Fund, 1956. 11. Turner, H. Nl., and Grace, H. G.: Investigation into Cancer Mortality Among Males in Certain Sheffield Trades, J. Ilyg. 33:90-103 (Jan.) 1938. 12. Mancuso, T. F.; McFarland, E. \1.; and Porterfield, J. D.: Distribution of Cancer Mortality in Ohio, Am. J. Public Health -13:58-70 (Jan.) 1955. 13. Hill, A. B., and others: Studies in Incidence of Can cer in Factory Handling Inorganic Compounds of Arsenic: I. Mortality Experience in Factory, Brit. J. Indust. Med. 3:1-6 (Jan.) 1948. 14. Henry, S. A., and Irvine, E. D.: Cancer of Scrotum in Blackburn Registration District, 1837-1929, J. Hyg. 38: 310-340 (July) 1936. Atkin, M., and others: Mortality from Cancer of Skin and Lip in Certain Occupations, Brit. J. Cancer 3:1-15 (March) 1949. Kennaway, E. L., and Ken naway, N. M.: Further Study of Incidence of Cancer of Lung and Larynx, ibid. 1:260-298 (Sept.) 1947. 15. Henry, S. A.: Occupational Cutaneous Cancer Attrib utable to Certain Chemicals in Industry, Brit. M. Bull. 4: 389-401, 1947. 16. Twenty-five Years of Health Progress, New York, Metropolitan Life Insurance Company, 1937. Health Prog ress 1936-1945, New York, Metropolitan Lile Insurance Company, 1948. Dublin, L. I., and Spiegelman, M.: Health Progress Among Industrial Policyholders 1946-1950, in Transactions, Society of Actuaries, vol. 3, 1951. Lew, E. A., and Spiegelman, M.: Mortality of Industrial Policyholders P950-1955, ibid., vol. 9, 1957. 17. Lew, E. A.: Use of Life Insurance Company Records for Cancer Studies, A. M. A. Arch. Indust. Hyg. 5:198-203 (M arch) 1952. ' , - . V' 1 __ BLOOD SUGAR DETERMINATIONS.--Blood sugar in the majority of clinical laboratories is determined by obsolete analytical methods which yield false values. Instead of the actual glucose content of the blood, these values repre- ' sent glucose equivalents of the combined copper reducing power of glucose plus ' .\ , other substances--such as glutathione, ergothioneine, uric acid, amino acids--which ' . are entirely unrelated to blood sugar. The reducing power of these substances varies , ' . - with the several reagents employed, so that the sugar values are distorted to variable ~ ! . - / -degrees. Thus the non-sugar reducing matter adds to the actual sugar content 10-15 - ' '' mg. percent in the Hagedom-Jensen method, 15-25 mg. per cent in the original ' ' / : Folin-Wu method 25-35 mg. per cent by the old Shaffer-Hartmen method. In con- L ; : sequence, different laboratories report fpr the same sugar concentrations a variety - U t; of figures which allow no basis for comparison. This situation has l>een, and still is, - T; .J > '. ,'the source of much confusion and misunderstandings. For 28 years now we have --been using methods which give the true, correct sugar values. Since these are at ! > i-s ,'i least as simple as the methods which yield false values, there is no justification what- v\- 'i ' ever for the inertia that keeps false values in circulation.-M iehael Somogyi, Ph.D., *o> Sugar Tolerance Tests: Their Relation to Diabetes and the Nature of Diabetes, Pro : ceedings of the Forty-fifth Annual Meeting of the Medical Section of the American . Lie Convention in N ew York, 1957. ,i - / Vol. 167, No. 6 EXAMINATIONS AND LICENSURE 757 r man of the department should accompany each plication. Requests for accommodation should made to Dr. Lalich at the same time. Meeting on Electroencephalography in Atlantic City.--The 12th annual meeting of the American Electroencephalographic Society will be held June 12-15 at the Hotel Traymore, Atlantic City, N. J. Dr. Robert S. Dow will give the presidential ad dress the morning of June 14 and will serve as chairman for a symposium on "Electrophysiology of the Cerebellum," which will include the follow ing topics and participants: Introduction and Historical Survey, Dr. Dow. Cerebellar Responses to the Stimulation of Afferent Connec tions, Dr. Ferdinando A. Morin, Detroit. Electrical Activity of the Cerebellar Cortex and Nuclei and How It May Be Modified, John M. Brookhart, Ph.D., Port land, Ore. Efferent Discharges Resulting from Electrical Stimulation of the Cerebellar Cortex and Nuclei, Ray S. Snider, Ph.D., Chicago. Significance of Eleetrophysiologieal Studies in Relation to the Localization of Function in the Cerebellum, Dr. Wil liam W. Chambers and James M. Sprague, Ph.D., Phila delphia. Relation of the Cerebellum to Sensory Mechanisms, Dr. Dow. Twenty-five papers are scheduled and a sympo sium on EEG of the "Premature, Newborn and Young Infant" is planned for June 15. Speaker at the annual banquet (June 14, 7 p. m.) will be Donald B. Lindsley, Ph.D., who will discuss "Lindslev's Folly: Reminiscences About the Early Days of EEG During the Depression Era (1930-1940)." For information write Dr. Jerome K. Merlis, Secre tary, American Electroencephalographic Society, EEG Laboratory, University Hospital, Baltimore 1, Md. Hunterian Lecture at the Royal College of Sur geons (1952), was author of "The Thyroid: A Physiological, Pathological, Clinical, and Surgical Study," and was attached to the Royal Marsden Hospital in a research capacity (since 1956). He was awarded a travelling fellowship by the Kellogg Foundation and by the International College of Surgeons, spending a year in the U. S. lecturing on diseases of the thyroid. CORRECTIONS Occupational Cancer.--In the report to the Council on Industrial Health in the April 26, 1958, issue of The Journal, Page 2177, column 1, second new para graph, the equation starting on line 3 should read as follows: "Incidence (cases per 100,000 per year) r= [total number of cases for n years -l (popula tion at risk X n)] X 100,000." The equation start ing on the fifth line from the end of the same para x graph should read: "Incidence = [10 (100 X 10)1 100,000 rrz 1,000 per 100,000 per year." Aero Medical Association.--in the Washington News in Tim Journal, Feb. 22, 1958 issue, front advertising page 27, and April 5, 1958 issue, front advertising page 27, it was stated that the Aero Medical Association was "now nine years old" and was "formed ten years ago," respectively. These statements were erroneous. The Aero Medical Association was founded in 1929. The 30th anni versary will be observed at the association's annual meeting in Los Angeles, April 27-29, 1959. EXAMINATIONS AND FOREIGN International Meeting of Gynecologists and Obste tricians.--The International Professional Associ ation of Gynecologists and Obstetricians will meet in Brussels, Belgium, under the presidency of Dr. Francisco Luque Beltran, of Madrid, July 18-19. Two main subjects to be discussed are: "Criteria and Processes of Specialization in Gynecology and Obstetrics" and "Economic Classification of Fees for Obstetric and Gynecologic Functions According to the Hierarchy of Values." Registration fee is 500 Belgian francs to be deposited to the account of l'Association Professionnelle des Obsttriciens et Gyncologues Belges: Dr. Stocq, 23 Drve du Caporal Uucle, Banque de la Socit Gnrale de Belgique, Brussels, Belgium. Information can be requested from the General Secretary of the Con gress, Dr. Geeraert, 211, Avenue Louise, Brussels, Belgium. DEATHS IN OTHER COUNTRIES Tobias Levitt.--Formerly surgical first assistant it the thyroid clinic, New End Hospital, Hampstead, England, Tobias Levitt, F. R. C. S., died in Charing Cross Hospital Feb. 22. He was elected to give a LICENSURE EDUCATIONAL COUNCIL FOR FOREIGN MEDICAL GRADUATES, INC. Educational Council for Foreign Medical Graduates, Inc.: The American medical qualification examination to be given henceforth twice a year for foreign medical grad uates. Medical Schools' in the United States and Foreign Countries, Sept. 23. Final date for filing application is June 23. Executive Director, Dr. Dean F. Smiley, 1710 Orrington Ave., Evanston, Illinois. NATIONAL BOARD OF MEDICAL EXAMINERS N a t io n a l B o a r d o f M k d ic a l E x a m in l u s : Written, Parts I and 1. All medical centers in the United States and Can ada, June 17-18. Sec., Dr. John F. Hubbard, 133 S. 36th St., Philadelphia. BOARDS OF MEDICAL EXAMINERS A l a b a m a : Examination. Montgomery, June 17-19. Sec., Dr D. G. Gill, State Office Building, Montgomery 4. A k iz o n a : Examination. Phoenix, July 16-18. Reciprocity. Phoenix, July 19. See., Dr. Thomas N. Bate, 826 Security Bldg., Phoenix. C a l if o r n ia : Written Examination. San Francisco, June 16 19; Los Angeles, August 18-21; Sacramento, Oct. 20-23. Oral Examination. San Francisco, June 14; Los Angeles,