scispace - formally typeset
Search or ask a question
Author

G. Gordon Brown

Bio: G. Gordon Brown is an academic researcher. The author has contributed to research in topics: Multistage sampling & Logistic regression. The author has an hindex of 1, co-authored 1 publications receiving 352 citations.

Papers
More filters
Journal ArticleDOI
TL;DR: This paper shows how model-adjusted risks, risk differences, and risk ratio estimates can be obtained directly from logistic regression models in the complex sample survey setting to yield population-based inferences.
Abstract: There is increasing interest in estimating and drawing inferences about risk or prevalence ratios and differences instead of odds ratios in the regression setting. Recent publications have shown how the GENMOD procedure in SAS (SAS Institute Inc., Cary, North Carolina) can be used to estimate these parameters in non-population-based studies. In this paper, the authors show how model-adjusted risks, risk differences, and risk ratio estimates can be obtained directly from logistic regression models in the complex sample survey setting to yield population-based inferences. Complex sample survey designs typically involve some combination of weighting, stratification, multistage sampling, clustering, and perhaps finite population adjustments. Point estimates of model-adjusted risks, risk differences, and risk ratios are obtained from average marginal predictions in the fitted logistic regression model. The model can contain both continuous and categorical covariates, as well as interaction terms. The authors use the SUDAAN software package (Research Triangle Institute, Research Triangle Park, North Carolina) to obtain point estimates, standard errors (via linearization or a replication method), confidence intervals, and P values for the parameters and contrasts of interest. Data from the 2006 National Health Interview Survey are used to illustrate these concepts.

395 citations


Cited by
More filters
Journal ArticleDOI
TL;DR: Challenges to incorporation of intersectionality into population health research are identified or expanded upon and have the potential to improve researchers' ability to more specifically document inequalities at varying intersectional positions.

969 citations

Journal ArticleDOI
TL;DR: The latest nationally representative data from the NSCH show that depression, anxiety, and behavioral/conduct problems are prevalent among US children and adolescents and treatment gaps remain.

596 citations

Journal ArticleDOI
TL;DR: Within 4 years of vaccine introduction, the vaccine-type HPV prevalence decreased among females aged 14-19 years despite low vaccine uptake, and the estimated vaccine effectiveness was high.
Abstract: Background. Human papillomavirus (HPV) vaccination was introduced into the routine immunization schedule in the United States in late 2006 for females aged 11 or 12 years with catch-up vaccination recommended for those aged 13-26 years. In 2010 3-dose vaccine coverage was only 32% among 13-17 year-olds. Reduction in the prevalence of HPV types targeted by the quadrivalent vaccine (HPV-6 -11 -16 and -18) will be one of the first measures of vaccine impact. Methods. We analyzed HPV prevalence data from the vaccine era (2007-2010) and the prevaccine era (2003-2006) that were collected during National Health and Nutrition Examination Surveys. HPV prevalence was determined by the Linear Array HPV Assay in cervicovaginal swab samples from females aged 14-59 years; 4150 provided samples in 2003-2006 and 4253 provided samples in 2007-2010. Results. Among females aged 14-19 years the vaccine-type HPV prevalence (HPV-6 -11 -16 or -18) decreased from 11.5% (95% confidence interval [CI] 9.2-14.4) in 2003-2006 to 5.1% (95% CI 3.8-6.6) in 2007-2010 a decline of 56% (95% CI 38-69). Among other age groups the prevalence did not differ significantly between the 2 time periods (P > .05). The vaccine effectiveness of at least 1 dose was 82% (95% CI 53-93). Conclusions. Within 4 years of vaccine introduction the vaccine-type HPV prevalence decreased among females aged 14-19 years despite low vaccine uptake. The estimated vaccine effectiveness was high.

535 citations

Journal ArticleDOI
TL;DR: Marginal standardization is the appropriate method when making inference to the overall population, and prediction at the means should not be used with binary confounders.
Abstract: Background: We review three common methods to estimate predicted probabilities following confounder-adjusted logistic regression: marginal standardization (predicted probabilities summed to a weighted average reflecting the confounder distribution in the target population); prediction at the modes (conditional predicted probabilities calculated by setting each confounder to its modal value); and prediction at the means (predicted probabilities calculated by setting each confounder to its mean value). That each method corresponds to a different target population is underappreciated in practice. Specifically, prediction at the means is often incorrectly interpreted as estimating average probabilities for the overall study population, and furthermore yields nonsensical estimates in the presence of dichotomous confounders. Default commands in popular statistical software packages often lead to inadvertent misapplication of prediction at the means. Methods: Using an applied example, we demonstrate discrepancies in predicted probabilities across these methods, discuss implications for interpretation and provide syntax for SAS and Stata. Results: Marginal standardization allows inference to the total population from which data are drawn. Prediction at the modes or means allows inference only to the relevant stratum of observations. With dichotomous confounders, prediction at the means corresponds to a stratum that does not include any real-life observations. Conclusions: Marginal standardization is the appropriate method when making inference to the overall population. Other methods should be used with caution, and prediction at the means should not be used with binary confounders. Stata, but not SAS, incorporates simple methods for marginal standardization.

533 citations

Journal ArticleDOI
17 Nov 2015-JAMA
TL;DR: Both the incidence of early-stage prostate cancer and rates of PSA screening have declined and coincide with 2012 USPSTF recommendation to omit PSA screened men from routine primary care for men.
Abstract: Importance Prostate cancer incidence in men 75 years and older substantially decreased following the 2008 US Preventive Services Task Force (USPSTF) recommendation against prostate-specific antigen (PSA)–based screening for this age group. It is unknown whether incidence has changed since the USPSTF recommendation against screening for all men in May 2012. Objective To examine recent changes in stage-specific prostate cancer incidence and PSA screening rates following the 2008 and 2012 USPSTF recommendations. Design and Settings Ecologic study of age-standardized prostate cancer incidence (newly diagnosed cases/100 000 men aged ≥50 years) by stage from 2005 through 2012 using data from 18 population-based Surveillance, Epidemiology, and End Results (SEER) registries and PSA screening rate in the past year among men 50 years and older without a history of prostate cancer who responded to the 2005 (n = 4580), 2008 (n = 3476), 2010 (n = 4157), and 2013 (n = 6172) National Health Interview Survey (NHIS). Exposures The USPSTF recommendations to omit PSA-based screening for average-risk men. Main Outcomes and Measures Prostate cancer incidence and incidence ratios (IRs) comparing consecutive years from 2005 through 2012 by age (≥50, 50-74, and ≥75 years) and SEER summary stage categorized as local/regional or distant and PSA screening rate and rate ratios (SRRs) comparing successive survey years by age. Results Prostate cancer incidence per 100 000 in men 50 years and older (N = 446 009 in SEER areas) was 534.9 in 2005, 540.8 in 2008, 505.0 in 2010, and 416.2 in 2012; rates began decreasing in 2008 and the largest decrease occurred between 2011 and 2012, from 498.3 (99% CI, 492.8-503.9) to 416.2 (99% CI, 411.2-421.2). The number of men 50 years and older diagnosed with prostate cancer nationwide declined by 33 519, from 213 562 men in 2011 to 180 043 men in 2012. Declines in incidence since 2008 were confined to local/regional-stage disease and were similar across age and race/ethnicity groups. The percentage of men 50 years and older reporting PSA screening in the past 12 months was 36.9% in 2005, 40.6% in 2008, 37.8% in 2010, and 30.8% in 2013. In relative terms, screening rates increased by 10% (SRR, 1.10; 99% CI, 1.01-1.21) between 2005 and 2008 and then decreased by 18% (SRR, 0.82; 99% CI, 0.75-0.89) between 2010 and 2013. Similar screening patterns were found in age subgroups 50 to 74 years and 75 years and older. Conclusions and Relevance Both the incidence of early-stage prostate cancer and rates of PSA screening have declined and coincide with 2012 USPSTF recommendation to omit PSA screening from routine primary care for men. Longer follow-up is needed to see whether these decreases are associated with trends in mortality.

386 citations