Imprecise exposure assessment and the sample size requirements of case-control studies of residential magnetic field exposure and cancer in adults

A computer program simulating case-control studies is described. It is used to estimate the minimum sample size required and to assess how this is affected by imprecise exposure assessment. In particular, the consequences of neglecting measurements of nonresidential exposure in case-control studies of residentially exposed adults are investigated. According to this model, while the consequent loss of power is not as large as was predicted by algebraic methods, it would be unwise to neglect it when planning a study. © 1995 Wiley-Liss, Inc.     

Prostate cancer association studies: pitfalls and solutions to cancer misclassification in the PSA era

Widespread screening of American men for elevated PSA has changed the characteristics of prostate cancer cases in the U.S. The influence of the changed nature of prostate cancer cases in the PSA era and the need for careful consideration of who is a "case" and who is a "control" on the ability to detect associations of risk factors with prostate cancer in etiologic epidemiologic studies merits discussion. Issue 1: prostate cancer cases diagnosed in the PSA era are enriched with a pool of early lesions, which may differ in etiology, and are deficient in advanced lesions, which are the most likely to be the product of promotion and progression events. By admixing the two types of cases (i.e., imperfect specificity), the associations previously detected using epidemiologic designs when the majority of cases were clinically detected may no longer be apparent in the PSA era when the majority of cases are now detected in the pre-clinical phase. Researchers must now tailor hypotheses such that they are testable using early stage cases or specifically augment the number of advanced cases when testing hypotheses related to extraprostatic growth and progression. Issue 2: even when controls are screened for elevated PSA to rule out the presence of prostate cancer, some proportion of those controls currently will have one or more foci of prostate cancer. The imperfect sensitivity of the PSA test coupled with diagnostic work-up may in part result from (a) lack of PSA elevation in some men with prostate cancer or (b) failure of biopsy to sample the tumor focus in men with elevated PSA. Misclassification of men with undetected prostate cancer as controls usually produces a bias that tends to deflate associations. Given this type of disease misclassification, whether an association still can be statistically detected depends on the extent of misclassification, the magnitude of the true association, the prevalence of the exposure in the true controls, and the sample size, although in general moderate nondifferential misclassification does not lead to profound attenuation. However, under the same scenario attenuation does not occur in cohort or case-cohort studies in which the rate or risk ratio (RR) is calculated. That prostate cancer cases diagnosed in the PSA era are enriched with early stage, minimally invasive disease in our opinion is likely to pose a far more serious obstacle to epidemiologic research on the etiology of clinically important prostate cancer than the issue of inclusion as controls some men who have undiagnosed prostate cancer because of imperfect sensitivity of PSA screening and biopsy sampling error.     

Relative contribution of residential and occupational magnetic field exposure over twenty-four hours among people living close to and far from a power line

This study sought to estimate the relative contribution of exposure to 50 Hz magnetic fields experienced at home, at work/school, or elsewhere to the total exposure over 24 hr. Personal exposure meters were carried by 97 adults and children in the Stockholm area. About half of the subjects lived close (<50 m) to a transmission line and half far (>100 m) away. Spot measurements and calculations for the residential exposure were also made. For subjects living<50 m from the line, the exposure at home contributed about 80% of the total magnetic field exposure, measured in mT-hours. Adults living far away experienced only 38% of the total exposure at home, but children still received 55%. Subjects with low time-weighted average (TWA) exposure both at home and at work spent 84% of their time in fields <0.1 microT, and those with high TWA at both locations spent 69% of their time in fields > or = 0.2 microT. This contrast was diluted if only exposure at one location was considered. For spot measurements and calculations of the residential exposure, both sensitivity and specificity was good. However, the intermediate field exposure category (0.1-0.19 microT) showed poor correlation to the 24 hr personal measurements.     

Underreporting of BMI in adults and its effect on obesity prevalence estimations in the period 1998 to 2001

Objective: To identify the determinants of underreporting BMI and to evaluate the possibilities of using self‐reported data for valid obesity prevalence rate estimations. Research Methods and Procedures: A cross‐sectional monitoring health survey was carried out between 1998 and 2002, and a review of published studies was performed. A total of 1809 men and 1882 women ages 20 to 59 years from The Netherlands were included. Body weight and height were reported and measured. Equations were calculated to estimate individuals’ BMI from reported data. These equations and equations from published studies were applied to the present data to evaluate whether using these equations led to valid estimations of the obesity prevalence rate. Also, size of underestimation of obesity prevalence rate was compared between studies. Results: The prevalence of obesity was underestimated by 26.1% and 30.0% among men and women, respectively, when based on reported data. The most important determinant of underreporting BMI was a high BMI. When equations to calculate individuals’ BMI from reported data were used, the obesity prevalence rate was still underestimated by 12.9% and 8.1% of the “true” obesity prevalence rate among men and women, respectively. The degree of underestimating the obesity prevalence was inconsistent across studies. Applying equations from published studies to the present data led to estimations of the obesity prevalence varying from a 7% overestimation to a 74% underestimation. Discussion: Valuable efforts for monitoring and evaluating prevention and treatment studies require direct measurements of body weight and height.     

Variations in prostate biopsy recommendation and acceptance confound evaluation of risk factors for prostate cancer: Examining race and BMI

BackgroundProstate cancer is ubiquitous in older men; differential screening patterns and variations in biopsy recommendations and acceptance will affect which man is diagnosed and, therefore, evaluation of cancer risk factors. We describe a statistical method to reduce prostate cancer detection bias among African American (n = 3398) and Non-Hispanic White men (n = 22,673) who participated in the Selenium and Vitamin E Cancer Prevention trial (SELECT) and revisit a previously reported association between race, obesity and prostate cancer risk.MethodsFor men with screening values suggesting prostate cancer but in whom biopsy was not performed, the Prostate Cancer Prevention Trial Risk Calculator was used to estimate probability of prostate cancer. Associations of body mass index (BMI) and race with incident prostate cancer were compared for observed versus imputation-enhanced outcomes using incident density ratios.ResultsAccounting for differential biopsy assessment, the previously reported positive linear trend between BMI and prostate cancer in African American men was not observed; no BMI association was found among Non-Hispanic White men.ConclusionsDifferential disease classification among men who may be recommended to undergo and then consider whether to accept a prostate biopsy leads to inaccurate identification of prostate cancer risk factors. Imputing a man’s prostate cancer status reduces detection bias. Covariate adjustment does not address the problem of outcome misclassification. Cohorts evaluating incident prostate cancer should collect longitudinal screening and biopsy data to adjust for this potential bias.     

Characterizing the effect of endocrine disruptors on human health: The role of epidemiological cohorts

Research on endocrine disruptors (EDs) developed from numerous disciplines. In this concert of disciplines, epidemiology is central to inform on the relevance for humans of mechanisms and dose-response functions identified in animals, to characterize the health impact (number of attributable disease cases), the cost associated with ED exposure, and the efficiency of the measures taken to limit exposure. Here, we present epidemiological tools to draw valid inference regarding effects of potential EDs. Epidemiology is generally observational, requiring care to control confounding bias. Many potential EDs have a short biological half-life; approaches relying on repeated biospecimens sampling allow limiting exposure misclassification and the resulting bias. For non-persistent compounds, couple–child cohorts are a central study design. Cohorts can now rely on molecular biology approaches to characterize exposures and intermediate pathways, which corresponds to the advent of molecular epidemiology and allows stronger interactions between epidemiology, toxicology, and molecular epidemiology to characterize the health effects of EDs.     

Relative-risk-estimate bias and loss of power in the Mantel test for trend resulting from the use of magnetic-field point-in-time ("spot") measurements in epidemiological studies based on an ordinal exposure scale.

We assessed the merits of various point-in-time ("spot") measurement protocols in case-control studies based on an ordinal exposure scale. After classifying a number of houses on the basis of prolonged monitoring of the ambient, extremely low frequency (ELF) magnetic field, we determined the probability of misclassification for each "spot" measurement protocol. We calculated the effect of this misclassification on the relative risk estimates and on the Mantel test for trend. We found that classification based on a small group of point-in-time measurements allows an adequate estimate of the relative risk, although the statistical significance of the dose-response gradient may be seriously underestimated. However, the use of automated ambient-field monitors, which results in loss of information on spatial variability, can lead to similar consequences. Therefore, manually collected point-in-time measurements remain a viable option for exposure assessment.     

The Use of Epidemiology in Risk Assessment: Challenges and Opportunities

The assessment of risk from environmental and occupational exposures incorporates and synthesizes data from a variety of scientific disciplines including toxicology and epidemiology. Epidemiological data have offered valuable contributions to the identification of human health hazards, estimation of human exposures, quantification of the exposure–response relation, and characterization of risks to specific target populations including sensitive populations. As with any scientific discipline, there are some uncertainties inherent in these data; however, the best human health risk assessments utilize all available information, characterizing strengths and limitations as appropriate. Human health risk assessors evaluating environmental and occupational exposures have raised concerns about the validity of using epidemiological data for risk assessment due to actual or perceived study limitations. This article highlights three concerns commonly raised during the development of human health risk assessments of environmental and occupational exposures: (a) error in the measurement of exposure, (b) potential confounding, and (c) the interpretation of non-linear or non-monotonic exposure–response data. These issues are often the content of scientific disagreement and debate among the human health risk assessment community, and we explore how these concerns may be contextualized, addressed, and often ameliorated.     

Evaluating Whether a Binary Decision Rule Operates Better Than Chance

We provide an approach to testing whether the accuracy of a binary diagnostic test, which we define as the sum of sensitivity and specificity, is significantly better than chance. We derive an exact confidence interval of size at least 1 — α for the observed accuracy of the test. In addition, we develop tests to compare the accuracy of two such tests applied to the same subjects. These results offer a method for assessing the accuracy of a test at a single test criterion, in contrast to the standard approach of evaluating the total receiver-operating characteristic (ROC) curve for a test.     

Direct and Indirect Classification in Clinical Research

We investigate if the use of a priori knowledge allows an improvement of medical decision making. We compare two frameworks of classification – direct and indirect classification – with respect to different classification errors: differential misclassification, observed misclassification and true misclassification. We analyze general behaviors of the classifiers in an artificial example and furthermore as being interested in the diagnosis of early glaucoma we adapt a simulation model of the optic nerve head. Indirect classifiers outperform direct classifiers in certain parameter situations of a Monte‐Carlo study. In summary, we demonstrate that indirect classification provides a flexible framework to improve diagnostic rules by using explicit a priori knowledge in clinical research.