The Wald statistic is frequently used to test hypotheses about beta coefficients from multiple linear regression analyses of complex survey data. This statistic requires a consistent estimate of the ...variance-covariance matrix of the regression coefficients. In a survey such as the Second National Health and Nutrition Examination Survey, the sample design limits the researcher to at most 32 degrees of freedom for estimating the variances and covariances with either the balanced half-sample replication or the Taylor series linearization procedure. This article considers the properties of the Wald statistic when the number of beta coefficients approaches the degrees of freedom available from the variance estimation. In this situation, Bonferroni-adjusted t statistics are an attractive alternative.
Abstract
Background: Testicular germ cell tumors (TGCT), the most commonly occurring cancer among young men in the U.S., are thought to be endocrine-related tumors. The relationship between steroid ...hormone levels and development of TGCT, however, has been difficult to study given the rarity of the tumor in the general population and the paucity of pre-diagnostic serum from young men. The U.S. Servicemen's Testicular Tumor Environmental and Endocrine Determinants (STEED) study, however, was able to overcome these obstacles by enrolling military servicemen who had previously donated serum specimens to the Department of Defense Serum Repository. Methods: STEED participants were enrolled between 2002 and 2005. The associations between TGCT risk and log-transformed serum concentrations of testosterone, free testosterone, estradiol, free estradiol, 3α-androstanediol glucuronide (3α-diol-G), an indicator of peripheral androgen action, and sex hormone-binding globulin (SHBG) were examined in 517 case and 790 control participants. Odds ratios (OR) and 95% confidence intervals (CI) were estimated using logistic regression models, adjusting for date of serum donation, age, race/ethnicity, family history of testicular cancer, history of cryptorchidism, height, and body mass index. Results: The analysis found that while testosterone was not significantly associated with TGCT risk, there was an inverse association between 3α-diol-G and risk (OR=0.80, 95%CI=0.63-0.99). In addition, there was a significant inverse association with TGCT and the ratio of 3α-diol-G to testosterone (OR=0.76, 95%CI=0.61-0.93). There was also a significant direct association between estradiol and TGCT risk (OR=1.41, 95%CI=1.06-1.87). The relationships between the free forms of testosterone and estradiol and risk were the same as the relationships between the total forms of each hormone and risk. There was no association between TGCT and SHBG. Conclusion: These results suggest that the hormone metabolism of men who subsequently develop TGCT may vary from that of other men. In men who develop TGCT, more testosterone may be converted to estradiol and less to dihydrotestosterone, the most potent form of androgen, than in men who do not develop TGCT. Further examination of metabolites, both upstream and downstream in the steroid hormone pathway, may help elucidate the mechanism by which these exposures are related to development of TGCT.
Citation Format: {Authors}. {Abstract title} abstract. In: Proceedings of the 103rd Annual Meeting of the American Association for Cancer Research; 2012 Mar 31-Apr 4; Chicago, IL. Philadelphia (PA): AACR; Cancer Res 2012;72(8 Suppl):Abstract nr 4468. doi:1538-7445.AM2012-4468
Abstract
In the United States, known risk factors for hepatocellular carcinoma (HCC) include infection with the hepatitis B virus (HBV), infection with the hepatitis C virus (HCV), excess alcohol ...consumption, diabetes, obesity and several rare metabolic disorders (hemochromatosis, α-1 antitrypsin deficiency, porphyrias). The proportion of risk in the U.S. attributable to each factor separately, and to all factors together, however, is not well understood. To better quantify the risk attributable to these factors, logistic regression analysis was conducted using data from the SEER-Medicare linked database. All persons diagnosed with HCC (n=5,607), with at least 3 pre-diagnostic years of Medicare coverage between 1994 and 2005, were included as cases. A 5% random sample of persons residing in SEER locations (n=190,782) were included as controls. For each risk factor, odds ratios (OR) with their 95% confidence intervals (95%CI), and attributable risks (AR) with their 95%CI were calculated. In addition to calculating overall risks, risks were calculated after stratifying on race/ethnicity, gender and time period of HCC diagnosis. Among all individuals, the AR of all factors together was 63.3%. The AR among males, however, was higher than the AR among females (64.6% vs. 60.7%, respectively). By race/ethnicity, the AR of all factors together was highest among Asians (67.9%). This AR was followed by that among Hispanics (64.9%), then whites (63.3%) and finally, blacks (53.0%). Among specific factors, diabetes had the greatest AR at 34%, followed by alcohol-related conditions at 24.0%, HCV at 20.7%, HBV at 5.7%, metabolic disorders at 3.1% and obesity at 2.0%. These findings differed by gender and race/ethnicity, however. While diabetes had the highest AR among both males and females, the factor with the second highest AR among males was alcohol-related diseases, while among females; the factor with the second highest AR was HCV. The factors with the highest AR among each race/ethnic group were: whites - diabetes (36.4%) and alcohol-related disease (23.5%), blacks - HCV (31.9%) and alcohol-related disease (20.3%), Hispanics - alcohol-related disease (29.9%) and diabetes (28.0%), and Asians - HCV (31.7%) and diabetes (26.5%). These findings indicate that the current increase in incidence of HCC in the U.S. may be fueled by different factors in different racial/ethnic and gender groups. Overall, controlling diabetes might have a greater impact than any other single factor on reducing the incidence of HCC in the U.S.
Citation Format: {Authors}. {Abstract title} abstract. In: Proceedings of the 101st Annual Meeting of the American Association for Cancer Research; 2010 Apr 17-21; Washington, DC. Philadelphia (PA): AACR; Cancer Res 2010;70(8 Suppl):Abstract nr 1816.
This paper outlines an approach for the design and analysis of randomized controlled trials investigating community-based interventions for behavioral change aimed at health promotion. The approach ...is illustrated using the Community Intervention Trial for Smoking Cessation (COMMIT), conducted from 1988 to 1993, involving 11 pairs of communities in North America, matched on geographic location, size, and sociodemographic factors. The situation discussed is when assignment to intervention is done at the community level; for COMMIT, the very nature of the intervention required this. The number of communities as a key determinant of the statistical power of the trial. The use of matched pairs of communities can achieve a gain in statistical efficiency. Randomization is used to obtain an unbiased assessment of the intervention effect; randomization also provides the basis for statistical analysis. Permutation tests (and corresponding test-based confidence intervals), using community as the unit of analysis, follow directly from the randomization distribution. Within this framework, individual-level covariates can be used for imputation of missing values and for adjusting analyses of intervention effect.
Representative risk estimation is fundamental to clinical decision-making. However, risks are often estimated from non-representative epidemiologic studies, which usually underrepresent minorities. ..."Model-based" methods use population registries to improve externally validity of risk estimation but assume hazard ratios (HR) are generalizable from samples to the target finite population. "Pseudoweighting" methods improve representativeness of studies by using an external probability-based survey as the reference, but the resulting estimators can be biased due to propensity model misspecification or inefficient due to variable pseudoweights or small sample sizes of minorities in the cohort and/or survey. We propose a two-step pseudoweighting procedure that poststratifies the event rates among age/race/sex strata in the pseudoweighted cohort to the population rates to produce efficient and robust pure risk estimation (i.e., a cause-specific absolute risk in the absence of competing events). For developing an all-cause mortality risk model representative for the US, our findings suggest that HRs for minorities are not generalizable, and that surveys can have inadequate numbers of events for minorities. Poststratification on event rates is crucial for obtaining reliable risk estimation for minority subgroups.
Relations between maternal anthropometric status during pregnancy and infant feeding practices and growth from birth through the first 6 mo of life were examined in a cohort of 351 Israeli ...mother-infant pairs of North African descent. Maternal weight, height, and triceps skinfold thicknesses were determined at 6 and 9 mo of pregnancy, while infants’ weights and lengths were measured at birth and at 1, 2, 3, and 6 mo of age with concurrent collection of age-specific maternal-reported infant feeding data. On the basis of multiple-linear-regression analysis that adjusted for potential covariates, mean maternal weight at the first prenatal visit and at 6 and 9 mo of pregnancy were positively associated with birth length (P for trend in all cases < 0.0001) and with linear growth between birth and 1, 3, and 6 mo of age. Maternal skinfold thickness at 9 mo of pregnancy and maternal height were also significantly associated with birth length. Moreover, maternal height, weight, and skinfold thickness at 6 and 9 mo of pregnancy were positively associated with mean birth weight. After adjustment for morbidity in the past month and other covariates, infants breast-fed exclusively had greater attained weight and weight gain in the first 3 mo compared with infants who were bottle-fed exclusively, breast-fed and bottle-fed, or solid-fed exclusively. These findings underscore the need for programs that improve the nutritional status of women before, during, and after pregnancy, and encourage exclusive breast-feeding of infants for at least the first 3 mo of life.
The associations between maternal anthropometric measures during pregnancy and infant feeding practices and growth in the first 6 months of life were investigated in a cohort of 351 mother-infant pairs of North African (Morocco, Tunisia, Algeria, or Libya) descent but currently residing in Israel’s Negev region. Mean maternal weight at the first prenatal visit and at 6 and 9 months of pregnancy was significantly and positively associated with birth length and with linear growth between birth and 1, 3, and 6 months of age. Maternal skinfold thickness at 9 months of pregnancy and maternal height also were significantly associated with birth length, while maternal height, weight, and skinfold thickness at 6 and 9 months of pregnancy were positively associated with mean birth weight. The rate of exclusive breast feeding declined from 34% at 1 month to 18% at 2 months to 6% at 3 months. After adjustment for covariates such as morbidity in the preceding month, maternal anthropometric status, and socioeconomic factors, infants who were exclusively breast-fed had greater attained weight and weight gain in the first 3 months of life than their counterparts who were exclusively bottle-fed, breast- and bottle-fed, or solid-fed. These findings underscore the importance of programs that improve the nutritional status of women before, during, and after pregnancy and promote exclusive breast feeding for at least the first 3 months.