Samuel Caudill - Academia.edu (original) (raw)

Papers by Samuel Caudill

Chemosphere, Jan 1, 2005

Key methodological aspects are presented for a study of concentrations of polychlorinated dibenzo... more Key methodological aspects are presented for a study of concentrations of polychlorinated dibenzo-p-dioxins (PCDDs), polychlorinated dibenzofurans (PCDFs), polychlorinated biphenyls (PCBs), and organochlorine pesticides in the serum of a sample of the New Zealand population aged 15 years and older. The study took advantage of the sampling frame and sample collection and interview processes of the National Nutrition Study (NNS). An additional blood sample for this organochlorines study was collected by the NNS and questions added to the NNS questionnaire. Serum was obtained from the blood and, based on responses to questions in the questionnaire, samples with possible occupational exposure to organochlorines were excluded. Remaining samples providing at least 2 ml of serum were pooled within 80 strata defined according to geographic area, age group, sex and ethnicity. A minimum number of five individual serum samples was required for pooling within a stratum. Within strata with sufficient samples, two or three pooled samples were created for variance calculation. Eligible for inclusion in the study were 2497 individual serum samples. Sixty strata had sufficient serum samples for pooling and chemical analysis. This was the first study of organochlorine compounds with a national population-based sample. Two factors that made the study feasible deserve emphasis. First, being able to "piggy-back" on another study. Second, pooling of samples to reduce analytic expenses. It is hoped that the methods used in this study will form the basis for other studies investigating organochlorine concentrations in national populations.

JAMA: The Journal of the American Medical Association, 1985

In response to questions about the reliability of the results of screening urine for drugs, we ev... more In response to questions about the reliability of the results of screening urine for drugs, we evaluated the performance of 13 laboratories, which serve a total of 262 methadone treatment facilities, by submitting prereferenced samples through the treatment facilities as patient samples (blind testing). Error rates for the 13 laboratories on samples containing barbiturates, amphetamines, methadone, cocaine, codeine, and morphine ranged from 11% to 94%, 19% to 100%, 0% to 33%, 0% to 100%, 0% to 100%, and 5% to 100%, respectively. Similarly, error rates on samples not containing these drugs (false-positives) ranged from 0% to 6%, 0% to 37%, 0% to 66%, 0% to 6%, 0% to 7%, and 0% to 10%, respectively. These blind tests indicate that greater care is taken with known evaluation samples than with routine samples, laboratories are often unable to detect drugs at concentrations called for by their contracts, and the observed underreporting of drugs may threaten the treatment process. Drug treatment facilities should monitor the performance of their contract laboratories with quality-control samples, preferably through blind testing.

Journal of Nutrition

ABSTRACT

The present study uses gas liquid chromatography (GLC) electron capture detection with packed and... more The present study uses gas liquid chromatography (GLC) electron capture detection with packed and capillary columns to detect polychlorinated biphenyls (PCBs) in serum samples from people living near the electric car repair and maintenance facility of the Southeastern Pennsylvania Transit Authority in Paoli, Pennsylvania. Most of the cohort surveyed had serum patterns similar to patterns for Aroclor 1260 (AR 1260);

Clinical chemistry, Jan 28, 2015

Limited information is available about testosterone concentrations representative of the general ... more Limited information is available about testosterone concentrations representative of the general US population, especially children, women, and non-Hispanic Asians. We obtained nationally representative data for total testosterone (totalT), measured with standardized LC-MS/MS, for the US population age 6 years and older from the 2011-2012 National Health and Nutrition Examination Survey (NHANES). We analyzed 6746 serum samples and calculated the geometric means, distribution percentiles, and covariate-adjusted geometric means by age, sex, and race/ethnicity. The 10th-90th percentiles of totalT values in adults (≤20 years) was 150-698 ng/dL (5.20-24.2 nmol/L) in men, 7.1-49.8 ng/dL (0.25-1.73 nmol/L) in women, and 1.0-9.5 ng/dL (0.04-0.33 nmol/L) in children (6-10 years old). Differences among race/ethnic groups existed in children and men: covariate-adjusted totalT values in non-Hispanic Asians were highest among children (58% compared to non-Hispanic black children) and lowest amon...

Statistics in medicine, Jan 30, 2012

The National Centers for Disease Control and Prevention (CDC) provides an ongoing assessment of t... more The National Centers for Disease Control and Prevention (CDC) provides an ongoing assessment of the US population's exposure to environmental chemicals by using biomonitoring in conjunction with CDC's National Health and Nutrition Examination Survey (NHANES). Characterizing the distributions of concentrations of environmental compounds or their metabolites in the US population is a primary objective of CDC's biomonitoring program. Historically, this characterization has been based on individual measurements of these compounds in body fluid or tissue from representative samples of the population. Pooling samples before making analytical measurements can reduce the costs of biomonitoring by reducing the number of analyses. For the first time in NHANES 2005-2006, a weighted pooled-sample design was implemented to facilitate pooling samples before making analytical measurements. This paper describes this design and the estimation method being developed in the National Center...

Clinical chemistry, 1996

Because of the increasing significance of folate nutriture to public health, a "round robin&... more Because of the increasing significance of folate nutriture to public health, a "round robin" interlaboratory comparison study was conducted to assess differences among methods. Twenty research laboratories participated in a 3-day analysis of six serum and six whole-blood pools. Overall means, SDs, and CVs derived from these results were compared within and across method types. Results reported for serum and whole-blood folate demonstrated overall CVs of 27.6% and 35.7%, respectively, across pools and two- to ninefold differences in concentrations between methods, with the greatest variation occurring at critical low folate concentrations. Although results for serum pools were less variable than those for whole-blood pools, substantial intermethod variation still occurred. The overall results underscore the urgent need for developing and validating reference methods for serum and whole-blood folate and for properly characterized reference materials. For evaluating study or ...

BACKGROUND: We evaluated the commutability of a proposed reference material (PRM), with a formula... more BACKGROUND: We evaluated the commutability of a proposed reference material (PRM), with a formula- tion based on dilution of Certified Reference Material 470 (CRM470), for 24 high-sensitivity C-reactive pro- tein (hsCRP) methods. We also investigated whether calibrationbyuseofPRMwaseffectiveinharmonizing results. METHODS: A set of 40 native clinical samples was mea- sured along with PRM and 3 dilutions of PRM. We used weighted

Science of The Total Environment, 1994

We measured the residues of polychlorinated biphenyls (PCBs) in the serum of 23 residents of the ... more We measured the residues of polychlorinated biphenyls (PCBs) in the serum of 23 residents of the New Bedford, Massachusetts, area and from two homogenates each of bluefish and lobsters from the same area. We used congener-specific and total Aroclor quantitative approaches, both of which involved gas chromatography with electron capture detection. Using gas chromatography mass spectrometry (electron ionization mode), we confirmed the presence of PCBs in the combined serum samples and in the aliquots of bluefish and lobsters. In measuring the PCB levels in serum, we found good agreement between the two electron capture detector approaches (r > or = 0.97) when the serum of specific congeners was compared to total Aroclor. We used univariate and multivariate quality control approaches to monitor these analyses. Analytical results for bluefish showed a better agreement between the two techniques than did those for lobsters; however, the small number of samples precluded any statistical comparison. We also measured levels of chlorinated pesticides in the serum samples of two groups of New Bedford residents, those with low PCB levels (< 15 ng/ml) and those with high PCB levels (> or = 15 ng/ml). We found that residents with high PCB levels also tended to have higher levels of hexachlorobenzene (HCB) and 1,1-dichloro-2,2-di-(p-chlorophenyl) ethylene (p,p'-DDE). The higher concentration of all three analytes appears to be influenced by employment in the capacitor industry, by seafood consumption, or both. Using Jaccard measures of similarity and principal component analysis we compared the gas chromatographic patterns of PCBs found in the serum of New Bedford area residents with high serum PCBs with the patterns found in homogenates of lobsters (inclusive of all edible portions except the roe), in homogenates of bluefish fillets taken from local waters, and in serum from goats fed selected technical Aroclors (e.g. Aroclors 1016, 1242, 1254, or 1260). The patterns found in human serum samples were similar to the patterns found in lobster homogenates. Both of these patterns closely resembled patterns found in the serum samples of the goat fed aroclor 1254, as demonstrated by both pattern recognition techniques. In addition, the chromatographic patterns of human serum and of lobsters and bluefish homogenates all indicated the presence of PCBs more characteristic of Aroclors 1016 or 1242.

Tobacco Control, 2010

It has been reported that charcoal added to cigarette filters selectively removes many of the mor... more It has been reported that charcoal added to cigarette filters selectively removes many of the more volatile chemicals, but it is not clear to what extent charcoal may reduce the delivery of important less volatile chemical constituents in mainstream cigarette smoke. We analysed machine-derived mainstream smoke deliveries (under three smoking regimens) for variants of a charcoal-filtered cigarette commercially test-marketed in the USA, focusing on selected polycyclic aromatic hydrocarbons (PAHs), phenols and tobacco-specific nitrosamines (TSNAs). While charcoal-containing filters selectively removed lower molecular weight PAHs from mainstream smoke, they did not significantly remove the heavier and more toxic PAHs studied, such as benzo[a]pyrene, a known carcinogen. Likewise, charcoal-containing filters removed phenols and TSNAs from mainstream smoke to differing amounts depending on the compound, filter design and the smoking regimen. The addition of sufficient charcoal to cigarette filters is known to remove many volatile compounds and can potentially reduce deliveries of certain semi-volatile compounds under some machine smoking regimens. Less volatile compounds, with a significant portion in the particulate phase, are less available for selective filtration by charcoal-containing filters than the more volatile compounds that reside predominantly in the gas phase.

Statistics in Medicine, 1995

Using data from a meta-analysis of the effects of oestrogen replacement therapy on the developmen... more Using data from a meta-analysis of the effects of oestrogen replacement therapy on the development of breast cancer, we compared alternative methods for combining dose-response slopes from epidemiological studies. We evaluated issues related both to summarizing data from single studies and to combining results from multiple studies. Findings related to the analysis of individual dose-response studies include: (1) a method of weighing studies that gives greater influence to dose-response slopes that conform to the linear relation of relative risk to duration can lead to large differences in calculated weights as a function of non-linearity; (2) a regression model using a variable-intercept resulted in a mean dose-response slope that increased as much as threefold when compared with the values obtained with a zero-intercept model. When combining results from multiple studies, we found: (1) calculating standard errors of mean dose-response slopes by methods that allow for both among-study and within-study variability (a random-effects type model) gave values different from a method that assumes homogeneity and equal within-study precision (a fixed-effects model); (2) the random-effects model gives mean and standard error results most similar to a bootstrap resampling method as increasing heterogeneity is observed (however, this model could give biased mean estimates compared with the bootstrap method); (3) a components-of-variance model compares favourably with the bootstrap and is easier to apply than the random-effects model. Based on these findings, we recommend the use of methods which incorporate heterogeneity to guard against underestimating the standard error. However, caution is urged because bias in point estimates can occur if extreme heterogeneity is present. Two other observations affect the interpretation of data combined from multiple studies. First, inclusion into a model of quality scores assigned by blinded reviewers had little effect on the mean dose-response slope and its standard error. Second, the number of studies required to achieve desired statistical power, varies with effect size.

Statistics in Medicine, 1995

Inferential statistical methods have traditionally been based on the assumption that one experime... more Inferential statistical methods have traditionally been based on the assumption that one experiment is performed and that interest centres on one or more predetermined hypothesis tests. Exploratory research, on the other hand, often involves multiple hypotheses or repeated investigations under similar or different conditions or both. Several techniques have been proposed to deal with multiple or simultaneous hypothesis testing in single investigations, and procedures to combine observed significance levels for an individual hypothesis test from two or more investigations have been suggested. In this paper we propose a method for identifying important results from multiple statistical tests in multiple investigations. The method is illustrated by using high performance liquid chromatography to identify potential aetiologic contaminants in L-tryptophan samples.

Statistics in Medicine, 1989

Using data from the National Health and Nutrition Examination Survey (NHANES II) 1976-1980, we de... more Using data from the National Health and Nutrition Examination Survey (NHANES II) 1976-1980, we demonstrate how cross-sectional total serum cholesterol surveillance data can be used by an individual to assess current and future personal cholesterol risk status. We propose statistical models, based on a person's current measured cholesterol level and the relationship between cross-sectional age and cholesterol percentile estimates, that will allow prediction of future cholesterol levels or the age at which specified cholesterol risk levels will be reached if no cholesterol-altering intervention is taken. These models incorporate the observed variation in the NHANES II data and expected intraperson biological variation and intralaboratory analytical variation. We illustrate the adequacy of the models using data from the longitudinal Framingham Study.

Statistics in Medicine, 2005

Statistics in Medicine, 2011

Pooling samples for analysis was first proposed in the 1940s to reduce analytical measurement cos... more Pooling samples for analysis was first proposed in the 1940s to reduce analytical measurement costs associated with screening World War II recruits for syphilis. Later, it progressed to more complex screening strategies, to population prevalence estimation for discrete quantities, and to population mean estimation for continuous quantities. Recently, pooled samples have also been used to provide efficient alternatives for gene microarray analyses, epidemiologic studies of biomarkers of exposure, and characterization of populations regarding environmental chemical exposures. In this study, we address estimation and bias issues related to using pooled-sample variance information from an auxiliary source to augment pooled-sample variance estimates from the study of interest. The findings are illustrated by using pooled samples from the National Health and Nutrition Examination Survey 2001-2002 to assess exposures to perfluorooctanesulfonate and other polyfluoroalkyl compounds in the U.S. population.

Journal of Exposure Science and Environmental Epidemiology, 2010

Biomonitoring involves the assessment of human or animal populations by measuring organic or biol... more Biomonitoring involves the assessment of human or animal populations by measuring organic or biological compounds or their metabolites in the body fluids or tissues of individuals in those populations. Pooling samples before making analytical measurements can reduce the costs of biomonitoring by reducing the number of analyses. By proper choice of pooled-sample design, population means can be estimated without measuring individual samples. I present a statistical method for characterizing an entire population distribution of such compounds by exploiting the theoretic relationship between interindividual-sample variance and the variation between pooled samples. I use simulation experiments to determine an optimum pooled-sample design as a function of the number of subpopulations and the number of available samples. Using pooled samples to characterize populations is not only more cost-efficient, but also in some cases it can lead to more precise and less biased parameter estimation than that occurs with individual samples.

Journal of Clinical Lipidology, 2008

guidelines to treating multiple lipid abnormalities and greater interest in combination therapy. ... more guidelines to treating multiple lipid abnormalities and greater interest in combination therapy. Purpose: The OCEANS study evaluated the effect of novel proprietary niacin extended-release (NER) and simvastatin combination (NER/S) in patients with mixed dyslipidemia. Methods: Following a 4-week run-in of diet modification and 40 mg/day simvastatin, patients were randomized to an 8-or 12-week NER/S titration scheme to a maximum NER/S dose of 2000 mg/day NER and 40 mg/day simvastatin for 24 weeks. Median changes from screening lipid values, prior to the simvastatin run-in phase, to week 24 in multiple lipid parameters were determined in patients who were lipid treatment-naïve at screening. Results: Following treatment with NER/S (2000/40 mg/d) for 24 weeks, median decreases of 53.0%, 55.9%, and 49.5% in non-HDL-C, LDL-C, and TG, respectively, were observed in treatment-naïve patients (nϭ85) [Figure]. The median increase in HDL-C was 25.0% in these patients. Lastly, a median decrease of 51.4% in the ratio of total cholesterol (Total C) to HDL-C was observed in treatment-naïve patients. Conclusions: In patients with primary Type II or mixed hyperlipidemia, treatment with combination NER/S (2000/40 mg/day) provides substantial improvements in all major lipid parameters associated with the residual risk of cardiovascular events.

Chemosphere, Jan 1, 2005

Key methodological aspects are presented for a study of concentrations of polychlorinated dibenzo... more Key methodological aspects are presented for a study of concentrations of polychlorinated dibenzo-p-dioxins (PCDDs), polychlorinated dibenzofurans (PCDFs), polychlorinated biphenyls (PCBs), and organochlorine pesticides in the serum of a sample of the New Zealand population aged 15 years and older. The study took advantage of the sampling frame and sample collection and interview processes of the National Nutrition Study (NNS). An additional blood sample for this organochlorines study was collected by the NNS and questions added to the NNS questionnaire. Serum was obtained from the blood and, based on responses to questions in the questionnaire, samples with possible occupational exposure to organochlorines were excluded. Remaining samples providing at least 2 ml of serum were pooled within 80 strata defined according to geographic area, age group, sex and ethnicity. A minimum number of five individual serum samples was required for pooling within a stratum. Within strata with sufficient samples, two or three pooled samples were created for variance calculation. Eligible for inclusion in the study were 2497 individual serum samples. Sixty strata had sufficient serum samples for pooling and chemical analysis. This was the first study of organochlorine compounds with a national population-based sample. Two factors that made the study feasible deserve emphasis. First, being able to "piggy-back" on another study. Second, pooling of samples to reduce analytic expenses. It is hoped that the methods used in this study will form the basis for other studies investigating organochlorine concentrations in national populations.

JAMA: The Journal of the American Medical Association, 1985

In response to questions about the reliability of the results of screening urine for drugs, we ev... more In response to questions about the reliability of the results of screening urine for drugs, we evaluated the performance of 13 laboratories, which serve a total of 262 methadone treatment facilities, by submitting prereferenced samples through the treatment facilities as patient samples (blind testing). Error rates for the 13 laboratories on samples containing barbiturates, amphetamines, methadone, cocaine, codeine, and morphine ranged from 11% to 94%, 19% to 100%, 0% to 33%, 0% to 100%, 0% to 100%, and 5% to 100%, respectively. Similarly, error rates on samples not containing these drugs (false-positives) ranged from 0% to 6%, 0% to 37%, 0% to 66%, 0% to 6%, 0% to 7%, and 0% to 10%, respectively. These blind tests indicate that greater care is taken with known evaluation samples than with routine samples, laboratories are often unable to detect drugs at concentrations called for by their contracts, and the observed underreporting of drugs may threaten the treatment process. Drug treatment facilities should monitor the performance of their contract laboratories with quality-control samples, preferably through blind testing.

Journal of Nutrition

ABSTRACT

The present study uses gas liquid chromatography (GLC) electron capture detection with packed and... more The present study uses gas liquid chromatography (GLC) electron capture detection with packed and capillary columns to detect polychlorinated biphenyls (PCBs) in serum samples from people living near the electric car repair and maintenance facility of the Southeastern Pennsylvania Transit Authority in Paoli, Pennsylvania. Most of the cohort surveyed had serum patterns similar to patterns for Aroclor 1260 (AR 1260);

Clinical chemistry, Jan 28, 2015

Limited information is available about testosterone concentrations representative of the general ... more Limited information is available about testosterone concentrations representative of the general US population, especially children, women, and non-Hispanic Asians. We obtained nationally representative data for total testosterone (totalT), measured with standardized LC-MS/MS, for the US population age 6 years and older from the 2011-2012 National Health and Nutrition Examination Survey (NHANES). We analyzed 6746 serum samples and calculated the geometric means, distribution percentiles, and covariate-adjusted geometric means by age, sex, and race/ethnicity. The 10th-90th percentiles of totalT values in adults (≤20 years) was 150-698 ng/dL (5.20-24.2 nmol/L) in men, 7.1-49.8 ng/dL (0.25-1.73 nmol/L) in women, and 1.0-9.5 ng/dL (0.04-0.33 nmol/L) in children (6-10 years old). Differences among race/ethnic groups existed in children and men: covariate-adjusted totalT values in non-Hispanic Asians were highest among children (58% compared to non-Hispanic black children) and lowest amon...

Statistics in medicine, Jan 30, 2012

The National Centers for Disease Control and Prevention (CDC) provides an ongoing assessment of t... more The National Centers for Disease Control and Prevention (CDC) provides an ongoing assessment of the US population's exposure to environmental chemicals by using biomonitoring in conjunction with CDC's National Health and Nutrition Examination Survey (NHANES). Characterizing the distributions of concentrations of environmental compounds or their metabolites in the US population is a primary objective of CDC's biomonitoring program. Historically, this characterization has been based on individual measurements of these compounds in body fluid or tissue from representative samples of the population. Pooling samples before making analytical measurements can reduce the costs of biomonitoring by reducing the number of analyses. For the first time in NHANES 2005-2006, a weighted pooled-sample design was implemented to facilitate pooling samples before making analytical measurements. This paper describes this design and the estimation method being developed in the National Center...

Clinical chemistry, 1996

Because of the increasing significance of folate nutriture to public health, a "round robin&... more Because of the increasing significance of folate nutriture to public health, a "round robin" interlaboratory comparison study was conducted to assess differences among methods. Twenty research laboratories participated in a 3-day analysis of six serum and six whole-blood pools. Overall means, SDs, and CVs derived from these results were compared within and across method types. Results reported for serum and whole-blood folate demonstrated overall CVs of 27.6% and 35.7%, respectively, across pools and two- to ninefold differences in concentrations between methods, with the greatest variation occurring at critical low folate concentrations. Although results for serum pools were less variable than those for whole-blood pools, substantial intermethod variation still occurred. The overall results underscore the urgent need for developing and validating reference methods for serum and whole-blood folate and for properly characterized reference materials. For evaluating study or ...

BACKGROUND: We evaluated the commutability of a proposed reference material (PRM), with a formula... more BACKGROUND: We evaluated the commutability of a proposed reference material (PRM), with a formula- tion based on dilution of Certified Reference Material 470 (CRM470), for 24 high-sensitivity C-reactive pro- tein (hsCRP) methods. We also investigated whether calibrationbyuseofPRMwaseffectiveinharmonizing results. METHODS: A set of 40 native clinical samples was mea- sured along with PRM and 3 dilutions of PRM. We used weighted

Science of The Total Environment, 1994

We measured the residues of polychlorinated biphenyls (PCBs) in the serum of 23 residents of the ... more We measured the residues of polychlorinated biphenyls (PCBs) in the serum of 23 residents of the New Bedford, Massachusetts, area and from two homogenates each of bluefish and lobsters from the same area. We used congener-specific and total Aroclor quantitative approaches, both of which involved gas chromatography with electron capture detection. Using gas chromatography mass spectrometry (electron ionization mode), we confirmed the presence of PCBs in the combined serum samples and in the aliquots of bluefish and lobsters. In measuring the PCB levels in serum, we found good agreement between the two electron capture detector approaches (r > or = 0.97) when the serum of specific congeners was compared to total Aroclor. We used univariate and multivariate quality control approaches to monitor these analyses. Analytical results for bluefish showed a better agreement between the two techniques than did those for lobsters; however, the small number of samples precluded any statistical comparison. We also measured levels of chlorinated pesticides in the serum samples of two groups of New Bedford residents, those with low PCB levels (< 15 ng/ml) and those with high PCB levels (> or = 15 ng/ml). We found that residents with high PCB levels also tended to have higher levels of hexachlorobenzene (HCB) and 1,1-dichloro-2,2-di-(p-chlorophenyl) ethylene (p,p'-DDE). The higher concentration of all three analytes appears to be influenced by employment in the capacitor industry, by seafood consumption, or both. Using Jaccard measures of similarity and principal component analysis we compared the gas chromatographic patterns of PCBs found in the serum of New Bedford area residents with high serum PCBs with the patterns found in homogenates of lobsters (inclusive of all edible portions except the roe), in homogenates of bluefish fillets taken from local waters, and in serum from goats fed selected technical Aroclors (e.g. Aroclors 1016, 1242, 1254, or 1260). The patterns found in human serum samples were similar to the patterns found in lobster homogenates. Both of these patterns closely resembled patterns found in the serum samples of the goat fed aroclor 1254, as demonstrated by both pattern recognition techniques. In addition, the chromatographic patterns of human serum and of lobsters and bluefish homogenates all indicated the presence of PCBs more characteristic of Aroclors 1016 or 1242.

Tobacco Control, 2010

It has been reported that charcoal added to cigarette filters selectively removes many of the mor... more It has been reported that charcoal added to cigarette filters selectively removes many of the more volatile chemicals, but it is not clear to what extent charcoal may reduce the delivery of important less volatile chemical constituents in mainstream cigarette smoke. We analysed machine-derived mainstream smoke deliveries (under three smoking regimens) for variants of a charcoal-filtered cigarette commercially test-marketed in the USA, focusing on selected polycyclic aromatic hydrocarbons (PAHs), phenols and tobacco-specific nitrosamines (TSNAs). While charcoal-containing filters selectively removed lower molecular weight PAHs from mainstream smoke, they did not significantly remove the heavier and more toxic PAHs studied, such as benzo[a]pyrene, a known carcinogen. Likewise, charcoal-containing filters removed phenols and TSNAs from mainstream smoke to differing amounts depending on the compound, filter design and the smoking regimen. The addition of sufficient charcoal to cigarette filters is known to remove many volatile compounds and can potentially reduce deliveries of certain semi-volatile compounds under some machine smoking regimens. Less volatile compounds, with a significant portion in the particulate phase, are less available for selective filtration by charcoal-containing filters than the more volatile compounds that reside predominantly in the gas phase.

Statistics in Medicine, 1995

Using data from a meta-analysis of the effects of oestrogen replacement therapy on the developmen... more Using data from a meta-analysis of the effects of oestrogen replacement therapy on the development of breast cancer, we compared alternative methods for combining dose-response slopes from epidemiological studies. We evaluated issues related both to summarizing data from single studies and to combining results from multiple studies. Findings related to the analysis of individual dose-response studies include: (1) a method of weighing studies that gives greater influence to dose-response slopes that conform to the linear relation of relative risk to duration can lead to large differences in calculated weights as a function of non-linearity; (2) a regression model using a variable-intercept resulted in a mean dose-response slope that increased as much as threefold when compared with the values obtained with a zero-intercept model. When combining results from multiple studies, we found: (1) calculating standard errors of mean dose-response slopes by methods that allow for both among-study and within-study variability (a random-effects type model) gave values different from a method that assumes homogeneity and equal within-study precision (a fixed-effects model); (2) the random-effects model gives mean and standard error results most similar to a bootstrap resampling method as increasing heterogeneity is observed (however, this model could give biased mean estimates compared with the bootstrap method); (3) a components-of-variance model compares favourably with the bootstrap and is easier to apply than the random-effects model. Based on these findings, we recommend the use of methods which incorporate heterogeneity to guard against underestimating the standard error. However, caution is urged because bias in point estimates can occur if extreme heterogeneity is present. Two other observations affect the interpretation of data combined from multiple studies. First, inclusion into a model of quality scores assigned by blinded reviewers had little effect on the mean dose-response slope and its standard error. Second, the number of studies required to achieve desired statistical power, varies with effect size.

Statistics in Medicine, 1995

Inferential statistical methods have traditionally been based on the assumption that one experime... more Inferential statistical methods have traditionally been based on the assumption that one experiment is performed and that interest centres on one or more predetermined hypothesis tests. Exploratory research, on the other hand, often involves multiple hypotheses or repeated investigations under similar or different conditions or both. Several techniques have been proposed to deal with multiple or simultaneous hypothesis testing in single investigations, and procedures to combine observed significance levels for an individual hypothesis test from two or more investigations have been suggested. In this paper we propose a method for identifying important results from multiple statistical tests in multiple investigations. The method is illustrated by using high performance liquid chromatography to identify potential aetiologic contaminants in L-tryptophan samples.

Statistics in Medicine, 1989

Using data from the National Health and Nutrition Examination Survey (NHANES II) 1976-1980, we de... more Using data from the National Health and Nutrition Examination Survey (NHANES II) 1976-1980, we demonstrate how cross-sectional total serum cholesterol surveillance data can be used by an individual to assess current and future personal cholesterol risk status. We propose statistical models, based on a person's current measured cholesterol level and the relationship between cross-sectional age and cholesterol percentile estimates, that will allow prediction of future cholesterol levels or the age at which specified cholesterol risk levels will be reached if no cholesterol-altering intervention is taken. These models incorporate the observed variation in the NHANES II data and expected intraperson biological variation and intralaboratory analytical variation. We illustrate the adequacy of the models using data from the longitudinal Framingham Study.

Statistics in Medicine, 2005

Statistics in Medicine, 2011

Pooling samples for analysis was first proposed in the 1940s to reduce analytical measurement cos... more Pooling samples for analysis was first proposed in the 1940s to reduce analytical measurement costs associated with screening World War II recruits for syphilis. Later, it progressed to more complex screening strategies, to population prevalence estimation for discrete quantities, and to population mean estimation for continuous quantities. Recently, pooled samples have also been used to provide efficient alternatives for gene microarray analyses, epidemiologic studies of biomarkers of exposure, and characterization of populations regarding environmental chemical exposures. In this study, we address estimation and bias issues related to using pooled-sample variance information from an auxiliary source to augment pooled-sample variance estimates from the study of interest. The findings are illustrated by using pooled samples from the National Health and Nutrition Examination Survey 2001-2002 to assess exposures to perfluorooctanesulfonate and other polyfluoroalkyl compounds in the U.S. population.

Journal of Exposure Science and Environmental Epidemiology, 2010

Biomonitoring involves the assessment of human or animal populations by measuring organic or biol... more Biomonitoring involves the assessment of human or animal populations by measuring organic or biological compounds or their metabolites in the body fluids or tissues of individuals in those populations. Pooling samples before making analytical measurements can reduce the costs of biomonitoring by reducing the number of analyses. By proper choice of pooled-sample design, population means can be estimated without measuring individual samples. I present a statistical method for characterizing an entire population distribution of such compounds by exploiting the theoretic relationship between interindividual-sample variance and the variation between pooled samples. I use simulation experiments to determine an optimum pooled-sample design as a function of the number of subpopulations and the number of available samples. Using pooled samples to characterize populations is not only more cost-efficient, but also in some cases it can lead to more precise and less biased parameter estimation than that occurs with individual samples.

Journal of Clinical Lipidology, 2008

guidelines to treating multiple lipid abnormalities and greater interest in combination therapy. ... more guidelines to treating multiple lipid abnormalities and greater interest in combination therapy. Purpose: The OCEANS study evaluated the effect of novel proprietary niacin extended-release (NER) and simvastatin combination (NER/S) in patients with mixed dyslipidemia. Methods: Following a 4-week run-in of diet modification and 40 mg/day simvastatin, patients were randomized to an 8-or 12-week NER/S titration scheme to a maximum NER/S dose of 2000 mg/day NER and 40 mg/day simvastatin for 24 weeks. Median changes from screening lipid values, prior to the simvastatin run-in phase, to week 24 in multiple lipid parameters were determined in patients who were lipid treatment-naïve at screening. Results: Following treatment with NER/S (2000/40 mg/d) for 24 weeks, median decreases of 53.0%, 55.9%, and 49.5% in non-HDL-C, LDL-C, and TG, respectively, were observed in treatment-naïve patients (nϭ85) [Figure]. The median increase in HDL-C was 25.0% in these patients. Lastly, a median decrease of 51.4% in the ratio of total cholesterol (Total C) to HDL-C was observed in treatment-naïve patients. Conclusions: In patients with primary Type II or mixed hyperlipidemia, treatment with combination NER/S (2000/40 mg/day) provides substantial improvements in all major lipid parameters associated with the residual risk of cardiovascular events.