Irwin D. J. Bross

Relation between Aortic Atherosclerosis and the Use of Cigarettes and Alcohol: An Autopsy Study

Abstract Examination of 1019 consecutively autopsied white patients who underwent standardized interviews before death, demonstrated no statistically significant relation between aortic atherosclerosis and the use of alcohol. However, a graded and statistically significant increase in the severity of aortic atherosclerosis was found with increasing use of cigarettes measured by both intensity (packages per day) and duration (number of years of smoking). Selective biases in the formation of the study population and faulty collection of historical data have, we believe, been excluded. These findings thus confirm and augment earlier studies of a relation between aortic and coronary atherosclerosis and cigarette smoking.

Spurious effects from an extraneous variable

Abstract Spurious effects from an extraneous variable are a troublesome problem in many areas of research in the biological and behavioral sciences. While investigators have recognized intuitively that there is a relationship between the size of an effect and its chance of being spurious, current textbooks do not contain any explicit statement of this relationship. In this paper one such statement, the Size Rule, is developed. The application of this rule to controversies in the medical area is discussed.

Leukemia in children exposed to multiple risk factors.

Abstract The association between four factors —mothers irradiation before conception, in utero irradiation of the child, previous history of reproductive wastage, and early childhood virus diseases — was studied in children one to four years old. A simple rise in the number of factors did not consistently increase the risk for leukemia. Children with irradiation but without either of the pathologic conditions did not show a greater risk for leukemia, nor did those with the pathologic conditions but without irradiation. However, children exposed to two radiological factors who also had a history of one pathologic event had a significantly increased risk (2.69), as did those with a history of two pathologic factors who had also been exposed to one of the irradiation factors (3.36). The highest risk was 4.64 for children exposed to all four factors. Therefore, only when there was a concordance of the two types of events, irradiation and pathologic, was the risk for leukemia significantly greater.

Aetiological Factors in Mouth Cancer

History has always been one of mans best teachers, a truism which applies also to mans quest to comprehend the nature of disease. We call to mind Snows classical study on cholera through the use of epidemiological techniques. Through his observations a disease was curtailed and the mechanistn of an illness more closely understood at a time when its exact pathogenesis was unestablished. To be learned from this historical study, too, is that one does not need to understand the mechanism of a disease completely in order to accomplish a reduction of its incidence. This point is stressed because it represents the basis of current epidemiological studies in the field of cancer. While many factors may induce cancer and the basic mechanism of cancer is still beyond our reach, we know factors which do play a part in the development of cancer whose removal or reduction will lead to a decrease in incidence of this disease. This is the primary goal of the epidemiologist in his fight against any disease. If a given factor is found to contribute to a disease it is of little importance to the patient whether this factor was of causative or of catalytic nature. What matters to the population is whether the removal of a given factor will reduce the incidence of a given disease. The present investigation on aetiological factors of cancer of the oral cavity was designed to serve towards this end. This report, to be presented in greater detail elsewhere, will outline some of the factors which influence the development of cancer of the mouth and whose control can lead to a drastic reduction in this type of cancer (Wynder, Bross, and Feldman, 1957).

DEVELOPMENT OF LOWER GENITAL CARCINOMAS IN PATIENTS WITH ANAL CARCINOMA: A MORE THAN CASUAL RELATIONSHIP

In a series of 64 anal carcinomas in the female 11 had multiple primaries in the ano‐genital tract. Statistical analysis using control series illustrated that this incidence is more than a casual association of these tumors. In 5 patients the second or third primaries were diagnosed within a period of one year or less (synchronous tumors). In 6 patients the time interval for the appearance of other primaries was of more than one year (metachronous tumors). Of the latter group, 5 had previous x‐ray therapy to the area and the enhancing or primary carcinogenic effect of the ionizing radiation cannot be ruled out. It appears evident, however, that primary tumor in the ano‐genital tract in the female deserves careful investigation and follow‐up if cancer patients are to be kept free of disease.

On the Foundations of Statistical Inference: Discussion

Author(s): L. J. Savage, George Barnard, Jerome Cornfield, Irwin Bross, George E. P. Box, I. J. Good, D. V. Lindley, C. W. Clunies-Ross, John W. Pratt, Howard Levene, Thomas Goldman, A. P. Dempster, Oscar Kempthorne and Allan Birnbaum Source: Journal of the American Statistical Association, Vol. 57, No. 298 (Jun., 1962), pp. 307-326 Published by: Taylor & Francis, Ltd. on behalf of the American Statistical Association Stable URL: http://www.jstor.org/stable/2281641 Accessed: 25-12-2017 20:48 UTC

Taking a Covariable into Account

Abstract A technique is presented for taking a covariable into account when the treatment and response variables are qualitative (i.e., where the familiar technique for the 2×2 contingency table would be used if no adjustment was made for the covariable). The technique is illustrated and exemplified by a clinical trial of chemotherapies for hyaline membrane disease in babies. The scientific rationale for an “exact” statistical test and a simple approximate test (COVAST) is provided. In terms of the mechanics, COVAST is an easy extension of the familiar Sign Test.

A method for evaluation of laxative habits in human subjects

Abstract 1. 1. A method is described for detecting laxative action in normal human subjects and for measuring the potency of laxative agents in constipated patients. 2. 2. Since the subjects record all observations on a daily report card, one physician can apply the method, retaining most of the advantages of the “double-blind” test. Dosage is constant over 1- or 2-week intervals. 3. 3. Both frequency and consistency of stool are measured, for the correlation between these two responses is only +0.5. The subjective categorical judgments of consistency are converted to a single numerical value by the ridit transformation. 4. 4. The laxative potency of an unknown agent emerges from the comparison of its results with those of placebo. The sensitivity of the method is calibrated by a standard laxative, USP fluid extract cascara sagrada in daily doses of 2 c.c. to normal subjects and 4 c.c. to constipated patients. 5. 5. With this method the unknown preparation, a beverage containing an extract of bran, was not distinguished from placebo by a group of 40 normal subjects nor by a group of 20 constipated patients. 6. 6. Dosage-response curves in constipated patients indicate that greatest sensitivity to dosage change lies below 0.5 Gm. daily of powdered USP cascara sagrada.

Rapid Analysis of 2 × 2 Tables

Abstract In many practical applications, especially when a long question schedule is given to two (or more) series of individuals, the computation of a large number of chi-square tests may become burdensome. For certain common types of data the computation can be avoided, or greatly reduced, by the charts presented in this paper. If so desired these charts may be used as a screen for borderline 2 × 2 tables and the occasional borderline case calculated by the usual methods.