PubMed Health⌕ Search

Biomedical subjects

G R Howe

Publications and source records attributed to G R Howe.

At least 127 records · Page 7Linked to original sources

Methodological issues in case-control studies: II. Test statistics as measures of efficiency.

The ratio of test statistics has been used to compare the efficiency of matched and unmatched designs, and stratified and pooled analyses for case-control studies. The index has been computed for a wide range of population conditions and it is concluded that (a) pooled analysis is always more efficient than stratified analysis when such pooling leads to a valid estimate of the relative risk, (b) the loss of efficiency by matching in the classic overmatching situation can be substantial, (c) in the confounding situation either design may be more efficient but generally the difference is small. These results add support to the contention that matching is rarely justified in case-control studies.

Analysis of Variance↗

Methodological issues in case-control studies. III--The effect of joint misclassification of risk factors and confounding factors upon estimation and power.

The effect of joint misclassification of polytomous risk factors and polytomous covariates in case-control studies upon estimation and power has been examined using computer simulation and mathematical models. It is concluded that non-differential misclassification of covariates may be as important as non-differential misclassification of exposure variables, and that the resultant bias of the relative risk estimate may be either toward or away from the null. It is pointed out that when a confounding variable is misclassified the calculated p value from the particular study will be incorrect and may either be too large or too small depending on the nature of the confounding.

Clinical Trials as Topic↗

Methodological issues in case-control studies IV: Validity and efficiency of various analysis strategies for continuous variables using the unconditional logistic regression model.

Computer simulation has been used to evaluate the performance of the unconditional logistic regression model when used to analyse continuous data from case-control studies. The size of the bias in the odds ratio estimate introduced by small sample sizes and by the use of the incorrect analysis model has been estimated for various underlying population conditions, together with the power as a measure of efficiency. It is concluded that the model is robust over a wide range of exposures to the risk factor, in the sense that both small sample sizes and use of the incorrect model introduced relatively small biases. The power of the test is again little altered by use of the incorrect analysis model.

Aged↗

Epidemiological assessment of risk to humans from exposure to nitrosamines.

In order for epidemiologists to evaluate the 'nitrosamine hypothesis' it is necessary to develop measures of human exposure to N-nitrosamines - both exogenous and when formed internally through consumption precursors. Dissatisfied with indefinite findings using indirect indices of N-nitrosamine exposure, we have attempted to derive an index based on the known kinetics of N-nitrosamine formation. This has been applied in a case-control study of cerebral tumours and resulted in a suggestive finding of increased risk for exogenous but not total N-nitrosamine exposure. A potential difficulty with our index is doubt as to whether vegetable sources of nitrates and consequent endogenous formation of nitrites indeed result in N-nitrosamine exposure, in view of the possible presence of blocking agents or of vitamin C consumed at the same time as vegetables. Further work is necessary, but we hope to apply the index in a case-control study of gastric cancer currently under analysis.

Dimethylnitrosamine↗

Correlations between cancer incidence rates from the Canadian National Cancer Incidence Reporting System, 1969-78.

Incidence rates for 19 cancers in females and 16 cancers in males have been computed from data reported by 8 Canadian provinces to the National Cancer Incidence Reporting System between 1969 and 1978. The rates, very similar in absolute and relative magnitude to those reported by the U.S. Third National Cancer Survey, have been used to examine patterns of correlation between various cancers within the 8 provinces. There is strong evidence of positive associations between a number of cancers, including a number of associations that have been reported in other similar correlational studies. Correlations that may be of particular interest in suggesting etiologic factors in common include clusters of smoking-related cancers (buccal cavity with pharynx, larynx, lung, and bladder), female sexual cancers (breast, corpus uteri, and ovary), and a group of cancers that have shown correlation in other studies (i.e., cancers of the pancreas and kidney, leukemias, lymphomas, and cancer of the prostate gland). Organs in the gastrointestinal tract (esophagus, stomach, colon, and rectum) anatomically close to each other show a high positive correlation in both females and males, but the further apart the organs are the lower is the correlation; these observations are consistent with other evidence of varying dietary etiologies. Two individual correlations of particular interest are those between female brain tumors and female bladder cancer (two cancers for which little is known of the etiology for a large percentage of them) and those between female breast cancer and female lung cancer. This study, the largest correlational study of incidence data reported to date, demonstrates the utility of such simple correlational analyses.

Canada↗

Food items and food groups as risk factors in a case-control study of diet and colo-rectal cancer.

The contributions of food items and food groups as risk factors in a previously reported case-control study of diet and colo-rectal cancer have been analyzed. The study included 348 patients with colon cancer, 194 with rectal cancer, 542 neighbourhood controls individually matched to the cases on the basis of age and sex and a second control series of 535 surgical hospital controls frequency matched to the cases. For colon cancer, as in the previous analysis, the major risk factor was saturated fat, individual food items or groups failing to make a significant contribution to the risk. In particular there was no protective effect of dietary fibre and, for cruciferous vegetables, only a minor protective effect in females. No individual cruciferous vegetable made an important contribution to this effect. For rectal cancer, on the other hand, a significant effect of saturated fat, independent of other food items or groups, was only found for females in the highest consumption category. For males, consumption of eggs, beef and veal significantly increased risk but not consumption of pork, while for females, there was a non-significant increase in risk with consumption of eggs, no increased risk with consumption of beef or veal and a significantly increased risk with consumption of pork. There was no protective effect of dietary fibre or of cruciferous vegetables for rectal cancer, but in females, there was a significantly increased risk for consumption of beer, though this was somewhat reduced when controlled for consumption of saturated fat. There was no indication of an effect of alcohol in either sex or of beer in males. Thus, these results confirm the previous report in showing a significant effect of saturated fat in increasing risk of colon cancer but suggest a contribution of meats to risk of rectal cancer.

Canada↗

Reliability of computerized versus manual death searches in a study of the health of Eldorado uranium workers.

An epidemiological follow-up study of 16,000 uranium mine and refinery employees has made use of computerized techniques for searching a national death file. The accuracy of this computerized matching has been compared with that of corresponding manual searches based on one-eighth of the worker file. The national death file--Canadian Mortality Data Base--at Statistics Canada includes coded causes of death for all deaths back to 1950. The machine search was carried out using a generalized record linkage system based upon a probabilistic approach. The machine was more successful than the manual searchers and was also less likely to yield false linkages with death records not related to the study population. In both approaches accuracy was strongly dependent on the amount of personal identifying information available on the records being linked.

Canada↗

Methodological issues in case-control studies: validity and power of various design/analysis strategies.

Computer stimulations have been used to estimate the efficiency, as measured by the statistical power, of various combinations of design and analysis strategies for case-control studies. Conditions under which the various forms of analysis yield consistent relative risk estimators are derived for the general model. The results indicate that the loss of efficiency resulting from the use of a less than optimum design or analysis strategy in many real life situations is small. Practical considerations are of more importance than theoretical statistical ones in deciding upon appropriate strategies. It is concluded that matching is rarely, if ever, justified in most case-control studies of chronic diseases.

Adult↗

A follow-up study of a ten-percent sample of the Canadian labor force. I. Cancer mortality in males, 1965-73.

A system is described for monitoring the mortality experience of an approximately 10% sample of the Canadian Labor Force with the use of computerized record linkage to the Canadian mortality data base. Results are presented for the 4,203 cancer deaths that occurred between 1965 and 1973 in 415,201 males with known occupational histories from 1965 to 1969. A number of previously suggested associations between certain occupations and cancers have been confirmed in these results, and a number of new associations also have been observed. The system provides a powerful tool for both generating and testing hypotheses, and this power will increase as further mortality experience is accumulated by the cohort.

Age Factors↗

Cancer mortality (1965-77) in relation to diesel fume and coal exposure in a cohort of retired railway workers.

A cohort study of 43,826 male pensioners of the Canadian National Railway Company was conducted. The cause of death of 17,838 pensioners who died between 1965 and 1977 was ascertained by computerized record linkage to the Canadian national mortality data base. The main finding was an elevated risk of lung cancer for those employed in occupations involving exposure to diesel fumes and coal dust, with highly significant dose-response relationships observed. That such association may be due in part to smoking cannot be excluded; but in view of the widespread exposure to diesel fumes, the finding warrants further investigation. The present study also demonstrated the utility and feasibility of large-scale occupational cohort studies conducted with the use of computerized record linkage to national mortality records.

Aged↗

Confidence interval estimation for the ratio of simple and standardized rates in cohort studies.

Computer simulation has been used to compare four methods for calculating confidence intervals for simple rate ratios estimated from cohort studies. The method proposed by Cornfield (1956. In Proceedings of the Third Berkeley Symposium on Mathematical Statistics and Probability. Vol. IV, 135-148) for interval estimation for the odds ratio in a 2 X 2 table (adapted to the ratio of two Poisson variables) gives reasonable and conservative results, and is recommended in view of its relative computational simplicity. Three methods are compared for interval estimation for the ratio of standardized rates, such as the standardized morbidity or mortality ratio, when the data are stratified by some confounding factor. The adapted Cornfield method again is the recommended procedure. An examination of the use of a summary chi square test with and without continuity correction indicates that the test size closest to the nominal is given by excluding such a correction when the test statistic is computed.

Computers↗

Evaluation of a self-administered dietary questionnaire for use in a cohort study.

A self-administered diet questionnaire designed for use in a cohort study has been assessed in comparison with a detailed quantitative diet history previously used in and validated for case-control studies. One hundred fifty-eight women aged 40 to 59 were asked to complete the self-administered questionnaire and 123 returned it by mail. Of these women, 50 were interviewed at home using the detailed diet history. Estimates of intake of major nutrients other than fat and fatty acids are comparable from the two methods although in general, estimated intake from the self-administered questionnaire was higher than the diet history. The difference in intake of fat and fatty acids was largely but not completely accounted for by differences in amounts of added fat recorded. It is concluded that when applied to large numbers of women in a cohort study, the self-administered questionnaire is feasible to administer and will be sufficiently accurate in estimating nutrient intake as compared to estimates from a detailed, interview-type diet history.

Adult↗

Breast cancer mortality following fluoroscopic irradiation in a cohort of tuberculosis patients.

A study has been conducted to determine the mortality experience from 1950-1977 of a cohort of women treated for tuberculosis in Canadian sanatoria between 1930 and 1952. Approximately 50 percent of these women received substantial breast tissue doses of fluoroscopic irradiation in conjunction with their treatment by artificial pneumothorax. A preliminary analysis of 23572 women known alive at the beginning of 1950 has shown a highly significant breast cancer mortality risk for those women exposed to such radiation. There is evidence of decreasing effect with increasing age at first exposure, and no increase in risk is observed until ten years after first exposure.

Breast Neoplasms↗

Dietary factors in relation to the etiology of colorectal cancer.

A case control study of 542 cases of colorectal cancer, 542 neighborhood controls, and 535 hospital controls has been conducted in two areas of Canada. The consumption of 13 nutrients has been estimated individually for all study members using an extensive diet history questionnaire. Highly significant dose-response relationships are observed between saturated fat intake and increased risk of both colon and rectal cancer for both sexes. In contrast no increase in risk was observed for intake of any level of polyunsaturated fat. The data are thus supportive of the hypothesis that dietary saturated fats may be of etiologic significance in cancers of the large bowel.

Colonic Neoplasms↗

The effect of hysterectomies and screening on mortality from cancer of the uterus in Canada.

The trends in mortality from cancer of the uterine cervix, cancer of the endometrium and all uterus in Canada and the ten Canadian Provinces from 1951-53 to 1974-76 at ages 30-64 have been re-evaluated in relation to screening for cancer of the cervix in 1966 and 1971 and hysterectomies for non-malignant conditions from 1969 to 1976. By means of series of mathematical models, the effect of different prior extrapolations of the numbers of hysterectomies performed by age and province has enabled the potential impact of hysterectomies on the trends of mortality from uterine cancer to be evaluated by relating deaths to "uteri at risk" rather than "women at risk" irrespective of the presence of an intact uterus. It has been found that the numbers of hysterectomies performed have little impact on the fall in mortality certified to cancer of the cervix or to all uterine cancer, but in the decade 1964-66 to 1974-66 they convert an apparent fall in mortality from cancer of the endometrium to stability. Re-evaluation of the fall in mortality from 1960-62 to 1970-72 in relation to the intensity of screening shows little impact of the hysterectomies performed on the significant correlation reported previously. However, after extension of the analysis to 1974-76 the correlation disappears. This finding may indicate a more limited potential for the application of screening, as practised in Canada, to reduce mortality from cancer of the cervix, than has been anticipated previously.

Adolescent↗

Estimated benefits and risks of screening for breast cancer.

A controlled randomized trial of breast cancer screening has been initiated in Canada. This paper presents an analysis of the possible benefit from screening relative to the possible radiation risk from mammography for those women who will be screened in the trial. It shows that with modern low-dose mammography, even when a conservative estimate of possible reduction in mortality due to early detection is applied to the data, the estimated benefit substantially exceeds any possible hazard.

Adult↗