Document Z8pOG43p7zBkE2rxwvLVNYO08

A. U. A. ARCHIVES OF WOUSTRI.4L Hh.it.TH Limiting the investigation to cases coming to eutffiy, as has been frequently done in earlier studies, still lurther restricts its use in generalization. The problem with which we are concerned is whether asbestos miners experience more lung cancer than does the general population. The answer necessitates the collection of reliable infor mation on asbestos miners as a group, as well as on the general population. It scents advisable to discuss the differ ences between the epidemiological approach and that toed in the studies which have been reported to date. - A very important consideration is the fact that long cancer, in spite of its increasing numbers, is stiU a disease of low incidence; that is, in a given population not many persons will centric:' this particular disease. This fan requires that large samples or groups must be stud ied to provide meaningful results. Recognizing the difficulty of obtaining such large samples, most earlier writers deviated from the epidemiological method and sought to circumvent the requirement of observing well persons by (I) compar ing the rrlcihv frequency of cancer in vari ous 'sites: (2) comparing the reiative frequency of cancer in a group of hospital ized patients; (3) comparing the relative frequency of cancer in a group of cases coming to autopsy. Attempting to compare tw<> leqnitation .groups. Inditing only at the relative fre quency of canter in various body sites, may result in finding a higher percentage frelative frequency) in one of the groups, when, in fact, the mertetity rate of cancer at a particular organ is exactly the same in both groups. This is- very clearly dem ons:rated in the excellent article by Dora* The 'mortality rate from a particular cause ' is the Trite measure of comparison. It is apparent char selected gnwps such **' hospitalized patients or autopsy cases may not be in any way representative of a larger group, and chu in dealing with such samples, the observer may easily find --- cases_pf (..given disease.than would be found in another group of the same size, bin representative of the general population. It is true that investigation of cases from such a sample can furnish information valuable for research, but the use of this information in drawing generalizations is necessarily restricted, ft is the obligation of both the investigator and of those who read his report to make proper comparisons and to draw only those conclusions which are valid and justified. A good statistical study of cases of cancer of the hutg occur ring in a group uf autopsies can lead to a proper inference concerning (he frequency of lung cancer among cases coming to autopsy, but only to socb cases. For in formation from such a study to be pro jected ta somflarger'group, it is'neccssary that the autopsies represent a good sample cl that larger group To assume that soeh is the case in any particular series is dan gerous and likely to be false. There it some danger that the figures reported by some juthors may be miscon strued as applying to asbestos workers or even asbestos miners, when, in fact, (he authors in question do not make this gen eralization. nor can the generalization be nude for the reasons staled. Close study of the reports reveals that the percentages cited relate only to the group of autoppes covered by the particular investigation. The present study, in contrast to the earlier works, has been planned to utilize the epidemiological method. A well-defined group of asbestos miners has been estab lished in such a war that ic constitutes a guod sample of the whole population uf asbrsth* miners in Quebec. Data for all members or this group have been collected anil analyzed. Those concerning hmg can ter have received most careful considera tion/ Details of the methods employed win he set forth bier, but the type of approach is constricted to. permit of fair comparisons and valid generalizations. Collection and Analysis of Data A preHminair survey c! mwiccs uf information in Ftbruaty, 1956, involved &3S .................................. Vet. 17. /. W* /W S'cfj.