ISSN:1044-3983    

出版商:OVID    

年代:1990

数据来源: OVID

ISSN:1044-3983    

出版商:OVID    

年代:1990

数据来源: OVID

摘要:

This paper reviews the role of statistics in causal inference. Special attention is given to the need for randomization to justify causal inferences from conventional statistics, and the need for random sampling to justify descriptive inferences. In most epidemiologic studies, randomization and random sampling play little or no role in the assembly of study cohorts. I therefore conclude that probabilistic interpretations of conventional statistics are rarely justified, and that such interpretations may encourage misinter pretation of nonrandomized studies. Possible remedies for this problem include deemphasizing inferential statistics in favor of data descriptors, and adopting statistical techniques based on more realistic probability models than those in common use. (Epidemi

ISSN:1044-3983    

出版商:OVID    

年代:1990

数据来源: OVID

摘要:

We examined dietary and other risk factors for renal adenocarcinoma in a case-control study of 203 incident and 207 prevalent cases and 605 neighborhood controls. Using unconditional logistic regression to control selection biases identified in an adjunct study. we tound exposure odds ratios (OR) similar to those from other studies: 1.9 (with a 95% confidence interval (c1) from 0.8 to 4.4) for smoking 2 packs of cigarettes per day; 3.8 (95% CI: 0.78–18) for 3 or more packs per day; 1.7 (95% CI: 0.9–3.2) for women and 1.7 (95% CI: 1.1–2.8) for men in the highest quintile of relative weight (kg); 2.0 (95% C1: 1.3–3.1) for northeastern European ancestry; and 2.3(95% CI: 1.34.1) for history of kidney stones. Incident cases consumed more meats and fewer vegetables than controls: the age-sex-education-adlusted OR for average intake of 85 g (302) of beef per day was 3.4 (95% CI: 1.67.2). Inverse associations were seen for most vegetables. Dietary animal protein, animal fat, and saturated fat, with and without energy adlustment, were weakly associated with disease in unconditional and conditional logistic regressions. Priorhypotheses concerning intake of cholesterol, beta-carotene, preformed vitamin A, and cruciferous vegetables were n o t corrobo

ISSN:1044-3983    

出版商:OVID    

年代:1990

数据来源: OVID

摘要:

Crude estimation of the selection probability ratio (SPR), described previously, was extended to stratified and multivariateestimation and used to assess selection bias in a case-control study of renal adenocarcinoma. I t was shown that the directly pooledestimate of the SPR, using the same weights as the directly pooled estimate of the exposure odds ratio (OR) from the case-controlstudy (assuming the OR and SPR are common to all strata and data are abundant), can be multiplied with the OR to yield anadjusted OR that is free from selection bias. Medical records of 548 interviewed cases were compared with 640 noninterviewedcases, and interviews at 640 controls were compared with mailed questionnaires from 272 (60%) of the noninterviewed controls.Age-sex-adjusted point estimates of SPRs ranged from 0.65 to 1.4. Multivariate estimates from binomial regression ranged from0.34 to 2.0. Higher socioeconomic status and history of renal stones were predictors of participation by both cases and controls.Obesity in women, hypertension, and nonsmoking were predictors in cases only. Heart disease was associated with controlparticipation and case nonparticipation. This study cast doubt on the OR for obesity in women and hypertension in the

ISSN:1044-3983    

出版商:OVID    

年代:1990

数据来源: OVID

摘要:

Cooperation between epidemiology and the laboratory represents a promising perspective for the understanding of the carcinogenicprocess. We describe three collaborative investigations involving the use of markers of internal dose (in particular, hemoglobinadducts) and markers of individual susceptibility in the study of bladder cancer induction by tobacco smoke. These investigationssuggest that (a) hemoglobin adducts formed by aromatic amines are a valid indicator of biologically effective internal dose and(h) genetically determined susceptibility due to metabolic polymorphism may play a role in chemically induced bladder cancer.

ISSN:1044-3983    

出版商:OVID    

年代:1990

数据来源: OVID

摘要:

To assess the incidence of and risk factors tor human immunodeficiency virus (HIV) infection among HIV-negative intravenousdrug abusers. we studied a cohort ot intravenous drug abusers recruited from drug treatment centers in Milan and northern Italy.We enrolled 933 subjects between January 1, 1987 and April 15, 1989 and we followed 460 subjects with one or more visits fora mean fdlaw-up duration of 10.4 months. The incidence rate of HIV infection was 7.4 per 100 person-years, equivalent to aone-year risk of 7.3%. Relative risk was higher in subjects who had been using intravenous drugs for less than 2 years (RR = 2.3).In a case-control analysis, recent trequent syringe sharing was the behavioral variable most strongly associated with HIV infection,with the highest risk in subjects sharing often (OR = 6.1, 90% CI = 2.614.7). We found no association with the use of cocainein addition to heroin nor with sexual habits. Among biologic variables, relative risks were increased in individuals whoseT4-lyrnphocyte count was lower than 1,000 at first visit (RR = 8.5, 90% CI = 2.9–24.3) or who were carrying HBsAg

ISSN:1044-3983    

出版商:OVID    

年代:1990

数据来源: OVID

摘要:

A case-control analysis d Parkinsons disease and smoking habits was conducted in a cohort of 50,002 men who attended Harvard College (Cambridge, MA) o r the University of Pennsylvania (Philadelphia, PA) between 1916 and 1950 and were followed in adulthood tor morbidity and mortality data. Ninety-six cases of Parkinsons disease were identified from responses to a mailed yxionnaire, addresaed in 1976 to the University of Pennsylvania alumni and in 1977–1978 to the Harvard alumni. Four controls trom the same population were selected for each case. Even though no statistically significant association was found, the results support a negative association between cigarette smoking and Parkinsuns disease.

ISSN:1044-3983    

出版商:OVID    

年代:1990

数据来源: OVID

摘要:

Recent epidemiologic evidence indicates an association between fat distribution and many diseases. To assess the validity of circumference measurements obtained by self-report, the authors analyzed data from 123 men aged 4C75 years and 140 women aged 41–65 years, drawn from two large ongoing prospective studies. On mailed questionnaires, subjects were asked to measure and record their weight and waist and hip circumferences. These data were compared with standardized measurements taken approximately six months apart by technicians who visited participants at their homes. Crude Pearson correlations between self-reported waist circumferences and the average of two technician-measured waist circumferences were 0.95 for men and 0.89 for women. Similar correlations for hip measurements were 0.88 for men and 0.84 for women, and for waist-to-hip ratios, 0.69 for men and 0.70 for women. After adjusting for age and body mass index (kg/m2), correlations for waist-to-hip ratios were 0.55 for men and 0.58 for women. Correlations became stronger after correcting for random within-person variability from daily or seasonal fluctuations. Self-reported and measured weights were highly correlated: 0.97 for men and 0.97 for women. Self-reported waist, hip, and weight measurements appear reasonably valid. The moderate degree of measurement error for the ratio of self-reported waist and hip circumferences, however, implies that previously reported associations based on self-report of these measures may have been appreciably underestimated. (Epidemiology 1990; 1:46&473)

ISSN:1044-3983    

出版商:OVID    

年代:1990

数据来源: OVID

摘要:

We reviewed published studies on the association of age at natural menopause and cigarette smoking. All demonstrated an earlier median o r mean age of menopause among smokers; for current smokers vs noncurrent smokers the difference ranged from 0.8 to1.7 years. For studies that presented suitable data, we computed prevalence odds ratios of menopause for current smokers vs never-smokers, current smokers vs noncurrent smokers, and former smokers vs never-smokers. The Mantel-Haenszel summary odds ratiosand 95% confidence intervals for these contrasts were: 1.9 (1.7–2.2), 1.7 (1.5–1.9), and 1.3 (1.C-1.7). Studies that presenteddata on amount of cigarettes smoked per day demonstrated dose-response trends when analyzed using the Mantel-Haenszelextension. The consistency of results across studies, the persistence of the effect when age and other covariates were considered,and the dose-response relation all support the hypothesis that smoking Increases the risk of early menopause. (Epidemiology

ISSN:1044-3983    

出版商:OVID    

年代:1990

数据来源: OVID