Keith E Muller

University of Florida, Gainesville, Florida, United States

Are you Keith E Muller?

Claim your profile

Publications (161)355.53 Total impact

  • [Show abstract] [Hide abstract]
    ABSTRACT: We used theoretical and simulation-based approaches to study Type I error rates for one-stage and two-stage analytic methods for cluster-randomized designs. The one-stage approach uses the observed data as outcomes and accounts for within-cluster correlation using a general linear mixed model. The two-stage model uses the cluster specific means as the outcomes in a general linear univariate model. We demonstrate analytically that both one-stage and two-stage models achieve exact Type I error rates when cluster sizes are equal. With unbalanced data, an exact size α test does not exist, and Type I error inflation may occur. Via simulation, we compare the Type I error rates for four one-stage and six two-stage hypothesis testing approaches for unbalanced data. With unbalanced data, the two-stage model, weighted by the inverse of the estimated theoretical variance of the cluster means, and with variance constrained to be positive, provided the best Type I error control for studies having at least six clusters per arm. The one-stage model with Kenward-Roger degrees of freedom and unconstrained variance performed well for studies having at least 14 clusters per arm. The popular analytic method of using a one-stage model with denominator degrees of freedom appropriate for balanced data performed poorly for small sample sizes and low intracluster correlation. Because small sample sizes and low intracluster correlation are common features of cluster-randomized trials, the Kenward-Roger method is the preferred one-stage approach. Copyright © 2015 John Wiley & Sons, Ltd. Copyright © 2015 John Wiley & Sons, Ltd.
    Statistics in Medicine 06/2015; DOI:10.1002/sim.6565 · 1.83 Impact Factor
  • Source
    [Show abstract] [Hide abstract]
    ABSTRACT: After conducting a media campaign focusing on the importance of oral and pharyngeal cancer (OPC) examinations, we assessed mechanisms of behavior change among individuals receiving an OPC examination for the first time. We used data from 2 waves of telephone surveys of individuals residing in 36 rural census tracts in northern Florida (n = 806). The second survey occurred after our media intervention. We developed media messages and modes of message delivery with community members via focus groups and intercept interviews. We performed a mediation analysis to examine behavior change mechanisms. Greater exposure to media messages corresponded with heightened concern about OPC. Heightened concern, in turn, predicted receipt of a first-time OPC examination, but only among men. We extended earlier studies by measuring an outcome behavior (receipt of an OPC examination) and demonstrating that the putative mechanism of action (concern about the disease) explained the link between a media intervention and engaging in the target behavior. Improving the quality of media campaigns by engaging community stakeholders in selecting messages and delivery methods is an effective strategy in building public health interventions aimed at changing behaviors. (Am J Public Health. Published online ahead of print May 14, 2015: e1-e8. doi:10.2105/AJPH.2014.302516).
    American Journal of Public Health 05/2015; 105(7):e1-e8. DOI:10.2105/AJPH.2014.302516 · 4.55 Impact Factor
  • Source
    [Show abstract] [Hide abstract]
    ABSTRACT: Researchers seeking to develop complex statistical applications for mobile devices face a common set of difficult implementation issues. In this work, we discuss general solutions to the design challenges. We demonstrate the utility of the solutions for a free mobile application designed to provide power and sample size calculations for univariate, one-way analysis of variance (ANOVA), GLIMMPSE Lite. Our design decisions provide a guide for other scientists seeking to produce statistical software for mobile platforms.
    PLoS ONE 12/2014; 9(12):e102082. DOI:10.1371/journal.pone.0102082 · 3.23 Impact Factor
  • [Show abstract] [Hide abstract]
    ABSTRACT: Medical and health policy decision-makers require improved design and analysis methods for comparative effectiveness research (CER) trials. In CER trials, there may be limited information to guide initial design choices. In general settings, adaptive designs (ADs) have effectively overcome limits on initial information. However, CER trials have fundamental differences from standard clinical trials including population heterogeneity and a vaguer concept of a “minimum clinically meaningful difference”. The objective of this article is to explore the use of a particular form of ADs for comparing treatments within the CER trial context. To achieve this, the authors review the current state of clinical CER. They also identify areas of CER as particularly strong candidates for application of novel AD and illustrate the potential usefulness of the designs and methods for two group comparisons. The authors found that ADs can stabilize power. Furthermore, the designs ensure adequate power for true effects are at least at clinically significant pre-planned effect size, or when variability is larger than expected. The designs allow for sample size savings when the true effect is larger or when variability is smaller than planned. The authors conclude that ADs in CER have great potential to allow trials to successfully and efficiently make important comparisons.
    Clinical Research and Regulatory Affairs 11/2014; 32(1). DOI:10.3109/10601333.2014.977490
  • Source
    [Show abstract] [Hide abstract]
    ABSTRACT: Objectives: We examined whether health literacy was associated with self-rated oral health status and whether the relationship was mediated by patient-dentist communication and dental care patterns. Methods: We tested a path model with data collected from 2 waves of telephone surveys (baseline, 2009-2010; follow-up, 2011) of individuals residing in 36 rural census tracts in northern Florida (final sample size n = 1799). Results: Higher levels of health literacy were associated with better self-rated oral health status (B = 0.091; P < .001). In addition, higher levels of health literacy were associated with better patient-dentist communication, which in turn corresponded with patterns of regular dental care and better self-rated oral health (B = 0.003; P = .01). Conclusions: Our study showed that, beyond the often-reported effects of gender, race, education, financial status, and access to dental care, it is also important to consider the influence of health literacy and quality of patient-dentist communication on oral health status. Improved patient-dentist communication is needed as an initial step in improving the population's oral health.
    American Journal of Public Health 05/2014; 104(7). DOI:10.2105/AJPH.2014.301930 · 4.55 Impact Factor
  • Sean L Simpson · Lloyd J Edwards · Martin A Styner · Keith E Muller ·
    [Show abstract] [Hide abstract]
    ABSTRACT: Longitudinal imaging studies have moved to the forefront of medical research due to their ability to characterize spatio-temporal features of biological structures across the lifespan. Valid inference in longitudinal imaging requires enough flexibility of the covariance model to allow reasonable fidelity to the true pattern. On the other hand, the existence of computable estimates demands a parsimonious parameterization of the covariance structure. Separable (Kronecker product) covariance models provide one such parameterization in which the spatial and temporal covariances are modeled separately. However, evaluating the validity of this parameterization in high-dimensions remains a challenge. Here we provide a scientifically informed approach to assessing the adequacy of separable (Kronecker product) covariance models when the number of observations is large relative to the number of independent sampling units (sample size). We address both the general case, in which unstructured matrices are considered for each covariance model, and the structured case, which assumes a particular structure for each model. For the structured case, we focus on the situation where the within subject correlation is believed to decrease exponentially in time and space as is common in longitudinal imaging studies. However, the provided framework equally applies to all covariance patterns used within the more general multivariate repeated measures context. Our approach provides useful guidance for high dimension, low sample size data that preclude using standard likelihood based tests. Longitudinal medical imaging data of caudate morphology in schizophrenia illustrates the approaches appeal.
    Journal of Applied Statistics 04/2014; 41(11):2450-2461. DOI:10.1080/02664763.2014.919251 · 0.42 Impact Factor
  • Source
    [Show abstract] [Hide abstract]
    ABSTRACT: Scientists often use a paired comparison of the areas under the receiver operating characteristic curves to decide which continuous cancer screening test has the best diagnostic accuracy. In the paired design, all participants are screened with both tests. Participants with suspicious results or signs and symptoms of disease receive the reference standard test. The remaining participants are classified as non-cases, even though some may have occult disease. The standard analysis includes all study participants, which can create bias in the estimates of diagnostic accuracy since not all participants receive disease status verification. We propose a weighted maximum likelihood bias correction method to reduce decision errors. Using Monte Carlo simulations, we assessed the method's ability to reduce decision errors across a range of disease prevalences, correlations between screening test scores, rates of interval cases and proportions of participants who received the reference standard test. The performance of the method depends on characteristics of the screening tests and the disease and on the percentage of participants who receive the reference standard test. In studies with a large amount of bias in the difference in the full areas under the curves, the bias correction method reduces the Type I error rate and improves power for the correct decision. We demonstrate the method with an application to a hypothetical oral cancer screening study. The bias correction method reduces decision errors for some paired screening trials. In order to determine if bias correction is needed for a specific screening trial, we recommend the investigator conduct a simulation study using our software.
    BMC Medical Research Methodology 03/2014; 14(1):37. DOI:10.1186/1471-2288-14-37 · 2.27 Impact Factor
  • [Show abstract] [Hide abstract]
    ABSTRACT: The complexity of system biology means that any metabolic, genetic, or proteomic pathway typically includes so many components (e.g., molecules) that statistical methods specialized for overall testing of high-dimensional and commensurate outcomes are required. While many overall tests have been proposed, very few have power and sample size methods. We develop accurate power and sample size methods and software to facilitate study planning for high-dimensional pathway analysis. With an account of any complex correlation structure between high-dimensional outcomes, the new methods allow power calculation even when the sample size is less than the number of variables. We derive the exact (finite-sample) and approximate non-null distributions of the 'univariate' approach to repeated measures test statistic, as well as power-equivalent scenarios useful to generalize our numerical evaluations. Extensive simulations of group comparisons support the accuracy of the approximations even when the ratio of number of variables to sample size is large. We derive a minimum set of constants and parameters sufficient and practical for power calculation. Using the new methods and specifying the minimum set to determine power for a study of metabolic consequences of vitamin B6 deficiency helps illustrate the practical value of the new results. Free software implementing the power and sample size methods applies to a wide range of designs, including one group pre-intervention and post-intervention comparisons, multiple parallel group comparisons with one-way or factorial designs, and the adjustment and evaluation of covariate effects. Copyright © 2013 John Wiley & Sons, Ltd.
    Statistics in Medicine 02/2014; 33(5). DOI:10.1002/sim.5986 · 1.83 Impact Factor
  • Source
    Sean L Simpson · Lloyd J Edwards · Martin A Styner · Keith E Muller ·
    [Show abstract] [Hide abstract]
    ABSTRACT: Longitudinal imaging studies have moved to the forefront of medical research due to their ability to characterize spatio-temporal features of biological structures across the lifespan. Credible models of the correlations in longitudinal imaging require two or more pattern components. Valid inference requires enough flexibility of the correlation model to allow reasonable fidelity to the true pattern. On the other hand, the existence of computable estimates demands a parsimonious parameterization of the correlation structure. For many one-dimensional spatial or temporal arrays, the linear exponent autoregressive (LEAR) correlation structure meets these two opposing goals in one model. The LEAR structure is a flexible two-parameter correlation model that applies to situations in which the within-subject correlation decreases exponentially in time or space. It allows for an attenuation or acceleration of the exponential decay rate imposed by the commonly used continuous-time AR(1) structure. We propose the Kronecker product LEAR correlation structure for multivariate repeated measures data in which the correlation between measurements for a given subject is induced by two factors (e.g., spatial and temporal dependence). Excellent analytic and numerical properties make the Kronecker product LEAR model a valuable addition to the suite of parsimonious correlation structures for multivariate repeated measures data. Longitudinal medical imaging data of caudate morphology in schizophrenia illustrates the appeal of the Kronecker product LEAR correlation structure.
    PLoS ONE 02/2014; 9(2):e88864. DOI:10.1371/journal.pone.0088864 · 3.23 Impact Factor
  • [Show abstract] [Hide abstract]
    ABSTRACT: Participants in trials may be randomized either individually or in groups and may receive their treatment either entirely individually, entirely in groups, or partially individually and partially in groups. This paper concerns cases in which participants receive their treatment either entirely or partially in groups, regardless of how they were randomized. Participants in group-randomized trials are randomized in groups, and participants in individually randomized group treatment trials are individually randomized, but participants in both types of trials receive part or all of their treatment in groups or through common change agents. Participants who receive part or all of their treatment in a group are expected to have positively correlated outcome measurements. This paper addresses a situation that occurs in group-randomized trials and individually randomized group treatment trials-participants receive treatment through more than one group. As motivation, we consider trials in The Childhood Obesity Prevention and Treatment Research Consortium, in which each child participant receives treatment in at least two groups. In simulation studies, we considered several possible analytic approaches over a variety of possible group structures. A mixed model with random effects for both groups provided the only consistent protection against inflated type I error rates and did so at the cost of only moderate loss of power when intraclass correlations were not large. We recommend constraining variance estimates to be positive and using the Kenward-Roger adjustment for degrees of freedom; this combination provided additional power but maintained type I error rates at the nominal level. Copyright © 2014 John Wiley & Sons, Ltd.
    Statistics in Medicine 01/2014; 33(13). DOI:10.1002/sim.6083 · 1.83 Impact Factor
  • [Show abstract] [Hide abstract]
    ABSTRACT: Although drug-eluting stent (DES) compared with bare metal stent (BMS) use reduces in-stent restenosis (ISR) in traditional coronary artery disease, its efficacy in cardiac allograft vasculopathy (CAV) has not been clearly established. CAV is a leading cause of mortality after the first year following cardiac transplantation. CAV treatment options are limited, and DES use has increased significantly in this population. In a retrospective study of heart transplant patients at our institution who underwent percutaneous coronary intervention with a BMS or DES for CAV, we compared baseline characteristics, clinical outcomes, ISR, and target lesion revascularization (TLR). The primary end-point was angiographic ISR assessed by quantitative coronary angiography analyzed as both a binary (≤50% vs. >50%) and continuous variable (follow-up minimal luminal area [MLA]/baseline MLA). Secondary outcomes included TLR and a composite of death, myocardial infarction, heart failure, and retransplantation. In 45 patients with DES, BMS, or both, ISR assessed as a continuous variable was statistically different between the 2 stent groups (follow-up MLA/baseline MLA = 0.796 DES vs. 0.481 BMS; P = 0.0037). There was also a significant difference in ISR (10.8% for DES versus 30.7% for BMS) when assessed as a binary variable. There was no statistically significant difference in TLR or composite cardiovascular outcomes between groups when adjusted for traditional cardiovascular risk factors. ISR assessed as a continuous variable was significantly different between stent groups. However, this did not lead to a difference in TLR or cardiovascular outcomes. This hypothesis-generating finding suggests that patients with CAV may not necessarily need treatment with DES, which can be more costly and carries more potential risk than BMS.
    Journal of Interventional Cardiology 01/2014; 27(1). DOI:10.1111/joic.12088 · 1.18 Impact Factor
  • Source
    Henrietta Logan · Yi Guo · Virginia J Dodd · Keith Muller · Joseph Riley ·
    [Show abstract] [Hide abstract]
    ABSTRACT: The degree of health disparities present in rural communities is of growing concern and is considered "urgent" since rural residents lag behind their urban counterparts in health status. Understanding the prevalence and type of chronic diseases in rural communities is often difficult since Americans living in rural areas are reportedly less likely to have access to quality health care, although there are some exceptions. Data suggest that rural residents are more likely to engage in higher levels of behavioral and health risk-taking than urban residents, and newer evidence suggests that there are differences in health risk behavior within rural subgroups. The objective of this report is to characterize the prevalence of four major and costly chronic diseases (diabetes, cardiovascular disease, cancer, and arthritis) and putative risk factors including depressive symptoms within an understudied rural region of the United States. These four chronic conditions remain among the most common and preventable of health problems across the United States. Using survey data (N = 2526), logistic regression models were used to assess the association of the outcome and risk factors adjusting for age, gender, and race. Key findings are (1) Lower financial security was associated with higher prevalence of cardiovascular disease, arthritis, and diabetes, but not cancer. (2) Higher levels of depressive symptoms were associated with higher prevalence of cardiovascular disease, arthritis, and diabetes. (3) Former or current smoking was associated with higher prevalence of cardiovascular disease and cancer. (4) Blacks reported higher prevalence of diabetes than Whites; Black women were more likely to report diabetes than all other groups; prevalence of diabetes was greater among women with lower education than among women with higher education. (5) Overall, the prevalence of diabetes and arthritis was higher than that reported by Florida and national data. The findings presented in this paper are derived from one of only a few studies examining patterns of chronic disease among residents of both a rural and lower income geographic region. Overall, the prevalence of these conditions compared to the state and nation as a whole is elevated and calls for increased attention and tailored public health interventions.
    BMC Public Health 10/2013; 13(1):906. DOI:10.1186/1471-2458-13-906 · 2.26 Impact Factor
  • Source
    [Show abstract] [Hide abstract]
    ABSTRACT: GLIMMPSE is a free, web-based software tool that calculates power and sample size for the general linear multivariate model with Gaussian errors ( GLIMMPSE provides a user-friendly interface for the computation of power and sample size. We consider models with fixed predictors, and models with fixed predictors and a single Gaussian covariate. Validation experiments demonstrate that GLIMMPSE matches the accuracy of previously published results, and performs well against simulations. We provide several online tutorials based on research in head and neck cancer. The tutorials demonstrate the use of GLIMMPSE to calculate power and sample size.
    Journal of statistical software 09/2013; 54(10). DOI:10.18637/jss.v054.i10 · 3.80 Impact Factor
  • Source
    [Show abstract] [Hide abstract]
    ABSTRACT: Suboptimal vitamin B-6 status, as reflected by low plasma pyridoxal 5'-phosphate (PLP) concentration, is associated with increased risk of vascular disease. PLP plays many roles, including in one-carbon metabolism for the acquisition and transfer of carbon units and in the transsulfuration pathway. PLP also serves as a coenzyme in the catabolism of tryptophan. We hypothesize that the pattern of these metabolites can provide information reflecting the functional impact of marginal vitamin B-6 deficiency. We report here the concentration of major constituents of one-carbon metabolic processes and the tryptophan catabolic pathway in plasma from 23 healthy men and women before and after a 28-d controlled dietary vitamin B-6 restriction (<0.35 mg/d). liquid chromatography-tandem mass spectrometry analysis of the compounds relevant to one-carbon metabolism showed that vitamin B-6 restriction yielded increased cystathionine (53% pre- and 76% postprandial; P < 0.0001) and serine (12% preprandial; P < 0.05), and lower creatine (40% pre- and postprandial; P < 0.0001), creatinine (9% postprandial; P < 0.05), and dimethylglycine (16% postprandial; P < 0.05) relative to the vitamin B-6-adequate state. In the tryptophan pathway, vitamin B-6 restriction yielded lower kynurenic acid (22% pre- and 20% postprandial; P < 0.01) and higher 3-hydroxykynurenine (39% pre- and 34% postprandial; P < 0.01). Multivariate ANOVA analysis showed a significant global effect of vitamin B-6 restriction and multilevel partial least squares-discriminant analysis supported this conclusion. Thus, plasma concentrations of creatine, cystathionine, kynurenic acid, and 3-hydroxykynurenine jointly reveal effects of vitamin B-6 restriction on the profiles of one-carbon and tryptophan metabolites and serve as biomarkers of functional effects of marginal vitamin B-6 deficiency.
    Journal of Nutrition 08/2013; 143(11). DOI:10.3945/jn.113.180588 · 3.88 Impact Factor
  • Source
    Yi Guo · Henrietta L Logan · Deborah H Glueck · Keith E Muller ·
    [Show abstract] [Hide abstract]
    ABSTRACT: Many researchers favor repeated measures designs because they allow the detection of within-person change over time and typically have higher statistical power than cross-sectional designs. However, the plethora of inputs needed for repeated measures designs can make sample size selection, a critical step in designing a successful study, difficult. Using a dental pain study as a driving example, we provide guidance for selecting an appropriate sample size for testing a time by treatment interaction for studies with repeated measures. We describe how to (1) gather the required inputs for the sample size calculation, (2) choose appropriate software to perform the calculation, and (3) address practical considerations such as missing data, multiple aims, and continuous covariates.
    BMC Medical Research Methodology 07/2013; 13(1):100. DOI:10.1186/1471-2288-13-100 · 2.27 Impact Factor
  • Source
    [Show abstract] [Hide abstract]
    ABSTRACT: Marginal deficiency of vitamin B-6 is common among segments of the population worldwide. Because pyridoxal 5'-phosphate (PLP) serves as a coenzyme in the metabolism of amino acids, carbohydrates, organic acids, and neurotransmitters, as well as in aspects of one-carbon metabolism, vitamin B-6 deficiency could have many effects. Healthy men and women (age: 20-40 y; n = 23) were fed a 2-day controlled, nutritionally adequate diet followed by a 28-day low-vitamin B-6 diet (<0.5 mg/d) to induce marginal deficiency, as reflected by a decline of plasma PLP from 52.6±14.1 (mean ± SD) to 21.5±4.6 nmol/L (P<0.0001) and increased cystathionine from 131±65 to 199±56 nmol/L (P<0.001). Fasting plasma samples obtained before and after vitamin B6 restriction were analyzed by (1)H-NMR with and without filtration and by targeted quantitative analysis by mass spectrometry (MS). Multilevel partial least squares-discriminant analysis and S-plots of NMR spectra showed that NMR is effective in classifying samples according to vitamin B-6 status and identified discriminating features. NMR spectral features of selected metabolites indicated that vitamin B-6 restriction significantly increased the ratios of glutamine/glutamate and 2-oxoglutarate/glutamate (P<0.001) and tended to increase concentrations of acetate, pyruvate, and trimethylamine-N-oxide (adjusted P<0.05). Tandem MS showed significantly greater plasma proline after vitamin B-6 restriction (adjusted P<0.05), but there were no effects on the profile of 14 other amino acids and 45 acylcarnitines. These findings demonstrate that marginal vitamin B-6 deficiency has widespread metabolic perturbations and illustrate the utility of metabolomics in evaluating complex effects of altered vitamin B-6 intake.
    PLoS ONE 06/2013; 8(6):e63544. DOI:10.1371/journal.pone.0063544 · 3.23 Impact Factor
  • [Show abstract] [Hide abstract]
    ABSTRACT: The aim of this study was to examine risk factors for reduced mouth or throat cancer (MTC) knowledge using a sample of rural North Floridian adults. Telephone interviews were conducted across rural census tracts throughout North Florida in 2009-2010, using a survey adapted for cultural appropriateness. The sample consisted of 2,393 individuals (1,059 males and 1,334 females; 1,681 whites and 712 blacks). Only 9% of the study respondents indicated they had not heard of MTC; however, only 12% endorsed knowing "a lot." Higher education levels and health literacy indicated they had more MTC knowledge. Among female participants, whites had more knowledge than blacks (OR = 1.9). Among black participants, males had more knowledge than females (OR = 1.7). Conversely, greater concern with MTC was associated with lower education levels, health literacy, and financial status, but higher depression scores. Awareness that excessive sun exposure is a risk factor for MTC was lower than for earlier studies using more urban samples. This study adds to the literature on MTC knowledge and concern because this sample was drawn exclusively from rural populations in North Florida, a group with the highest MTC morbidity and mortality. An unanticipated finding was that blacks were more concerned than their white rural counterparts. This study was also the first to report that depression was associated with increased concern about MTC. The goal is to persuade at-risk groups to obtain MTC screenings with the goal of reducing disparities in MTC whenever they occur.
    The Journal of Rural Health 06/2013; 29(3):294-303. DOI:10.1111/jrh.12003 · 1.45 Impact Factor
  • Yueh-Yun Chi · Keith E. Muller ·
    [Show abstract] [Hide abstract]
    ABSTRACT: Medical images and genetic assays typically generate data with more variables than subjects. Scientists may use a two-step approach for testing hypotheses about Gaussian mean vectors. In the first step, principal components analysis (PCA) selects a set of sample components fewer in number than the sample size. In the second step, applying classical multivariate analysis of variance (MANOVA) methods to the reduced set of variables provides the desired hypothesis tests. Simulation results presented here indicate that success of the PCA in the first step requires nearly all variation to occur in population components far fewer in number than the number of subjects. In the second step, multivariate tests fail to attain reasonable power except in restrictive, favorable cases. The results encourage using other approaches discussed in the article to provide dependable hypothesis testing with high dimension, low sample size data (HDLSS).
    Communication in Statistics- Simulation and Computation 05/2013; 42(5). DOI:10.1080/03610918.2012.659819 · 0.33 Impact Factor
  • Source
    Matthew J Gribbin · Yueh-Yun Chi · Paul W Stewart · Keith E Muller ·
    [Show abstract] [Hide abstract]
    ABSTRACT: Using covariance or mean estimates from previous data introduces randomness into each power value in a power curve. Creating confidence intervals about the power estimates improves study planning by allowing scientists to account for the uncertainty in the power estimates. Driving examples arise in many imaging applications. We use both analytical and Monte Carlo simulation methods. Our analytical derivations apply to power for tests with the univariate approach to repeated measures (UNIREP). Approximate confidence intervals and regions for power based on an estimated covariance matrix and fixed means are described. Extensive simulations are used to examine the properties of the approximations. Closed-form expressions are given for approximate power and confidence intervals and regions. Monte Carlo simulations support the accuracy of the approximations for practical ranges of sample size, rank of the design matrix, error degrees of freedom, and the amount of deviation from sphericity. The new methods provide accurate coverage probabilities for all four UNIREP tests, even for small sample sizes. Accuracy is higher for higher power values than for lower power values, making the methods especially useful in practical research conditions. The new techniques allow the plotting of power confidence regions around an estimated power curve, an approach that has been well received by researchers. Free software makes the new methods readily available. The new techniques allow a convenient way to account for the uncertainty of using an estimated covariance matrix in choosing a sample size for a repeated measures ANOVA design. Medical imaging and many other types of healthcare research often use repeated measures ANOVA.
    BMC Medical Research Methodology 04/2013; 13(1):57. DOI:10.1186/1471-2288-13-57 · 2.27 Impact Factor
  • Source
    [Show abstract] [Hide abstract]
    ABSTRACT: BACKGROUND: Oral and pharyngeal cancer is a serious health threat that goes unnoticed by most people. Increasing screenings for oral and pharyngeal cancer is essential to achieving early detection when the disease is most treatable. PURPOSE: We tested the effectiveness of a media campaign designed to increase intentions to seek an oral and pharyngeal cancer screening. We further examined whether concern and knowledge of oral and pharyngeal cancer mediated screening intentions. METHODS: Participants in the intervention condition received messages on posters, handheld fans, pamphlets, and magnets displayed on the sides of cars or trucks. Participants in the intervention and comparison conditions (N = 1,790) were surveyed prior to and after the intervention. RESULTS: Intervention participants reported greater intentions to seek free oral and pharyngeal cancer screenings. Concern about oral and pharyngeal cancer partially mediated the effect whereas knowledge did not. CONCLUSIONS: Our media campaign successfully increased screening intentions by heightening concerns.
    Annals of Behavioral Medicine 03/2013; 46(1). DOI:10.1007/s12160-013-9480-z · 4.20 Impact Factor

Publication Stats

3k Citations
355.53 Total Impact Points


  • 2007-2015
    • University of Florida
      • • Department of Health Outcomes and Policy
      • • Department of Biostatistics
      • • College of Medicine
      Gainesville, Florida, United States
  • 2014
    • University of Colorado
      • Department of Biostatistics and Informatics
      Denver, Colorado, United States
  • 1981-2006
    • University of North Carolina at Chapel Hill
      • • Department of Biostatistics
      • • Department of Radiology
      • • Department of Medicine
      • • Department of Biomedical Engineering
      North Carolina, United States
  • 2005
    • University of North Carolina at Charlotte
      Charlotte, North Carolina, United States
  • 1988
    • North Carolina Central University
      • School of Education
      Durham, NC, United States
  • 1984
    • United States Environmental Protection Agency
      Washington, Washington, D.C., United States