Search PubMed⌕ Search

SEARCH · Search PubMed

Results for “Statistical Power”

Search indexed PubMed citations on genomics, clinical trials, systematic reviews and public health. Explore titles, authors and supplied subject terms, then open the PubMed record.

Quote a phrase for an exact phrase match. Source license links do not imply unrestricted reuse.

At least 379 records · Page 21Linked to original sources
PubMed2000

Publication and related bias in meta-analysis: power of statistical tests and prevalence in the literature.

Publication and selection biases in meta-analysis are more likely to affect small studies, which also tend to be of lower methodological quality. This may lead to "small-study effects," where the smaller studies in a meta-analysis show larger treatment effects. Small-study effects may also arise because of between-trial heterogeneity. Statistical tests for small-study effects have been proposed, but their validity has been questioned. A set of typical meta-analyses containing 5, 10, 20, and 30 trials was defined based on the characteristics of 78 published meta-analyses identified in a hand search of eight journals from 1993 to 1997. Simulations were performed to assess the power of a weighted regression method and a rank correlation test in the presence of no bias, moderate bias or severe bias. We based evidence of small-study effects on P < 0.1. The power to detect bias increased with increasing numbers of trials. The rank correlation test was less powerful than the regression method. For example, assuming a control group event rate of 20% and no treatment effect, moderate bias was detected with the regression test in 13.7%, 23.5%, 40.1% and 51.6% of meta-analyses with 5, 10, 20 and 30 trials. The corresponding figures for the correlation test were 8.5%, 14.7%, 20.4% and 26.0%, respectively. Severe bias was detected with the regression method in 23.5%, 56.1%, 88.3% and 95.9% of meta-analyses with 5, 10, 20 and 30 trials, as compared to 11.9%, 31.1%, 45.3% and 65.4% with the correlation test. Similar results were obtained in simulations incorporating moderate treatment effects. However the regression method gave false-positive rates which were too high in some situations (large treatment effects, or few events per trial, or all trials of similar sizes). Using the regression method, evidence of small-study effects was present in 21 (26.9%) of the 78 published meta-analyses. Tests for small-study effects should routinely be performed in meta-analysis. Their power is however limited, particularly for moderate amounts of bias or meta-analyses based on a small number of small studies. When evidence of small-study effects is found, careful consideration should be given to possible explanations for these in the reporting of the meta-analysis.

Bias↗
PubMed1996

A simulation study of cross-validation for selecting an optimal cutpoint in univariate survival analysis.

Continuous measurements are often dichotomized for classification of subjects. This paper evaluates two procedures for determining a best cutpoint for a continuous prognostic factor with right censored outcome data. One procedure selects the cutpoint that minimizes the significance level of a logrank test with comparison of the two groups defined by the cutpoint. This procedure adjusts the significance level for maximal selection. The other procedure uses a cross-validation approach. The latter easily extends to accommodate multiple other prognostic factors. We compare the methods in terms of statistical power and bias in estimation of the true relative risk associated with the prognostic factor. Both procedures produce approximately the correct type I error rate. Use of a maximally selected cutpoint without adjustment of the significance level, however, results in a substantially elevated type I error rate. The cross-validation procedure unbiasedly estimated the relative risk under the null hypothesis while the procedure based on the maximally selected test resulted in an upward bias. When the relative risk for the two groups defined by the covariate and true changepoint was small, the cross-validation procedure provided greater power than the maximally selected test. The cross-validation based estimate of relative risk was unbiased while the procedure based on the maximally selected test produced a biased estimate. As the true relative risk increased, the power of the maximally selected test was about 10 per cent greater than the power obtained using cross-validation. The maximally selected test overestimated the relative risk by about 10 per cent. The cross-validation procedure produced at most 5 per cent underestimation of the true relative risk. Finally, we report the effect of dichotomizing a continuous non-linear relationship between covariate and risk. We compare using a linear proportional hazard model to using models based on optimally selected cutpoints. Our simulation study indicates that we can have a substantial loss of statistical power when we use cutpoint models in cases where there is a continuous relationship between covariate and risk.

Humans↗
PubMed2000

The current state of multiple sclerosis genetic research.

INTRODUCTION: Multiple sclerosis (MS) is the most common genetic disease of the nervous system with onset usually in young adulthood. Four genome-wide searches in different Caucasian populations for MS susceptibility loci have been performed, but none reported any linkage at a level that would be regarded as significant according to current criteria. Significant linkage of MS to allelic variants of the major histocompatibility (MHC) locus on chromosome 6p21 has been established although its overall contribution to MS susceptibility has proven difficult to quantify. The objective of this review is not only to provide the reader with an update of MS genetics research, but also to provide a basic knowledge of the techniques being employed to map MS susceptibility genes. The different methodologies are discussed, and specific studies are reviewed in context. METHODS: This review is based on findings from original articles, however, the results of recent candidate gene studies are intended to update previous review articles. RESULTS: There remains no concrete non-MHC locus for MS, although there are enough findings of sufficient interest to warrant further investigation and optimism. Stratification of genome scan data based on MHC class II suggests that it interacts differentially with non-MHC loci and that it contributes moderately to disease susceptibility. Candidate gene studies have continued to return negative and ambiguous results, and follow-up fine mapping of suggestive linkages from the UK genome scan has proven unsuccessful in identifying significant linkages. Genetic analysis of crosses between mouse strains that are differentially susceptible to experimental allergic encephalomyelitis (EAE) has yielded linkages corresponding to putative MS susceptibility loci. However, recent successes in transgenic mice may provide an alternative to EAE, regarded by some as a poor model of MS. CONCLUSION: The first whole genome search for a common human disease was performed over five years ago, and it is now clear, from the lack success in this field, that the genetic complexity of these traits has been underestimated. The genome-wide searches for MS susceptibility genes have suffered from insufficient statistical power, which has probably been compounded by disease and genetic heterogeneity. Studies in isolated populations and better laboratory and clinical definitions of disease are both steps in the right direction to solving these problems. Not withstanding the negative effects of genetic heterogeneity, pooling of resources for meta-analyses may provide the increase in statistical power required for detection of loci that exert a moderate or small effect on disease predisposition.

Animals↗
PubMed1991

Analyzing quality-control trends with moving slope charts.

We have developed and evaluated a new procedure for detecting trends in quality-control measurements and applied it to laboratory data. The method requires the use of sequential or "moving" slope estimates to identify trends. Formulae are derived to estimate the regression error for the moving slope directly from the standard deviation of the analytical measurements obtained during characterization runs. Control limits for the moving slope depend only on this regression error, the span of the slope, and the desired statistical level of control. The moving slope can be plotted with control limits to determine out-of-control points. The statistical power of the moving slope is found to be much greater than that of an often-used test for trends. An example of the use of the moving slope is shown for quality-control measurements for total cholesterol obtained over several years. We conclude that the moving slope procedure has considerably more statistical power than trend rules and that it yields more useful information to the analyst.

Biometry↗
PubMed1995

When you can't ask their names: linking anonymous respondents with the Hogben number.

This article describes a method of linking anonymous subjects with a respondent-generated code using an algorithm based on personal details to produce unique identifiers. It was used to increase confidentiality and statistical power in a year-long work-place health promotion evaluation. Subjects were employees of a large retail chain; 80 per cent were female, and the majority educated to high school level. Of the 385 possible, 81 per cent matched; 67 per cent of the codes were matched on all elements and another 14 per cent were accepted as 'fuzzy' matches. Linking respondents increased the statistical power of the study from an unacceptable 0.4 to an acceptable 0.8. Other research on linking records is briefly discussed, including sample bias and probabilistic matching. This technique is useful when anonymity is likely to raise response rates, but the ideal code could be further sought.

Adult↗
PubMed1987

Useful and extraneous variability in longitudinal assessment of lung function.

Longitudinal measurement is increasingly used to quantify the effects of recent and ongoing influences on lung function, whether treatments of groups of patients, or exposures of working or community populations. The variability in an estimate, eg, mean annual change in FEV1, comes from two sources; variability from true differences in annual change among individuals (called signal), and variability from measurement error (called noise). Signal is useful variability, potentially relatable to explanatory variables, and noise is extraneous. Assuming the variance of true differences remains constant, any increase in noise produces a calculable fall in the proportion of signal in the total observation, which fall we term "signal decay". This is not a function of the number of individuals, which influences rather the statistical power to determine that observed differences are not likely from chance alone. Imprecision in the estimation of individuals' rates of change is a major source of signal decay. Within practical limits, this can be compensated for by increasing the length of the study. Higher rates of subject attrition cause signal decay, in addition to loss of statistical power and susceptibility to survivor bias. Increasing the frequency of testing, within a study of constant length, has little effect on signal and noise, but interval testing protects against secular bias and minimizes data loss from subject attrition.

Adult↗
PubMed2006

A test of the equal environment assumption (EEA) in multivariate twin studies.

In the classic twin design, estimation of genetic and environmental effects is based on the assumption that environmental influences are shared to the same extent by monozygotic and dizygotic twins (equal environment assumption, EEA). We explore the conditions in which the EEA can be tested based on multivariate phenotypic data. We focus on the test whether the correlation between shared environmental factors in dizygotic twins (r(C)) is less than 1. First, model identification was investigated analytically in Maple and Mx. Second, statistical power was examined in Mx. Third, the amount of bias caused by violation of the EEA was evaluated. Finally, applications to empirical data concern spatial ability in adolescents and aggression in children. Bivariate and trivariate models include several instances in which the EEA can be tested. The number of twin pairs that is needed to detect violation of the EEA with a statistical power of .80 (alpha = .05) varied between 508 and 3576 pairs for the situations considered. The bias in parameter estimates, given misspecification, ranged from 5% to 34% for additive genetic effects, and from 4% to 34% for shared environmental effects. Estimates of the nonshared environmental effects were not biased. The EEA was not violated for spatial ability or aggression. Multivariate data provide sufficient information to test the validity of the EEA. The number of twin pairs that is needed is no greater than the number typically available in most twin registries. The analysis of spatial ability and aggression indicated no detectable violation of the EEA.

Aggression↗
PubMed1994

A pedigree series for mapping disease genes in bipolar affective disorder: sampling, assessment, and analytic considerations.

A series of 57 extended pedigrees with high density of bipolar affective disorder is described. Ascertainment and diagnostic procedures are documented and simulation studies to assess statistical power are carried out. The pedigrees, obtained in the US and Israel, are comprised of 1508 adult individuals with best estimate consensus diagnoses (12-71 relatives per pedigree), 490 of whom (including 401 sib pairs) meet criteria for a conservative disease definition (bipolar disorder or recurrent major depression). Cell lines have been established on 1324 of these individuals. Statistical power to detect linkage with lod score analysis, assuming autosomal dominant transmission and highly polymorphic DNA markers, is nearly 100% for alpha (proportion of linked families) = 30%, and 75% for alpha = 20%. This is the largest bipolar pedigree series reported to date; its unique features make it amenable to various gene detection techniques.

Bipolar Disorder↗
PubMed2026

ReMeDy: A Flexible Statistical Framework for Region-Based Detection of DNA Methylation Dysregulation.

Region-based epigenome-wide association studies have demonstrated improved statistical power and biological interpretability compared with probe-wise analyses of DNA methylation data. However, most existing region-based methods characterize methylation dysregulation primarily through changes in mean methylation levels associated with a phenotype of interest. Substantial evidence indicates that phenotype-associated methylation alterations may also manifest through changes in methylation variability or through joint shifts in mean and variability. Despite this, no existing statistical framework jointly models mean-variance methylation changes in a region-based manner. We propose ReMeDy, a flexible statistical framework that uses a hierarchical likelihood approach within a generalized linear model setting to identify differentially methylated regions, variably methylated regions, and regions exhibiting joint differential and variable methylation at a genome-wide scale. Unlike existing models, ReMeDy operates directly on biologically defined co-methylated regions, allowing it to naturally capture spatial correlation inherent in DNA methylation array data, while avoiding reliance on heuristic, user-defined tuning parameters such as smoothing spans and kernel bandwidths that can substantially influence results and introduce subjectivity. Through extensive simulation studies and comprehensive benchmarking against popular models, we demonstrate that ReMeDy maintains false discovery and Type-I error rates at nominal levels while achieving consistently higher statistical power across a wide range of realistic scenarios. Application to population-level DNA methylation data further shows that ReMeDy identifies biologically meaningful regions and pathways implicated in complex human diseases that are not captured by conventional mean-based analyses alone. ReMeDy is implemented as an open-source R package and is freely available at https://github.com/SChatLab/ReMeDy.

DNA Methylation↗
PubMed1999

Industrialization, electromagnetic fields, and breast cancer risk.

The disparity between the rates of breast cancer in industrialized and less-industrialized regions has led to many hypotheses, including the theory that exposure to light-at-night and/or electromagnetic fields (EMF) may suppress melatonin and that reduced melatonin may increase the risk of breast cancer. In this comprehensive review we consider strengths and weaknesses of more than 35 residential and occupational epidemiologic studies that investigated the association between EMF and breast cancer. Although most of the epidemiologic data do not provide strong support for an association between EMF and breast cancer, because of the limited statistical power as well as the possibility of misclassification and bias present in much of the existing data, it is not possible to rule out a relationship between EMF and breast cancer. We make several specific recommendations for future studies carefully designed to test the melatonin-breast cancer and EMF-breast cancer hypotheses. Future study designs should have sufficient statistical power to detect small to moderate associations; include comprehensive exposure assessments that estimate residential and occupational exposures, including shift work; focus on a relevant time period; control for known breast cancer risks; and pay careful attention to menopausal and estrogen receptor status.

Breast Neoplasms↗
PubMed2004

Nonreplication in genetic studies of complex diseases--lessons learned from studies of osteoporosis and tentative remedies.

Inconsistent results have accumulated in genetic studies of complex diseases/traits over the past decade. Using osteoporosis as an example, we address major potential factors for the nonreplication results and propose some potential remedies. Over the past decade, numerous linkage and association studies have been performed to search for genes predisposing to complex human diseases. However, relatively little success has been achieved, and inconsistent results have accumulated. We argue that those nonreplication results are not unexpected, given the complicated nature of complex diseases and a number of confounding factors. In this article, based on our experience in genetic studies of osteoporosis, we discuss major potential factors for the inconsistent results and propose some potential remedies. We believe that one of the main reasons for this lack of reproducibility is overinterpretation of nominally significant results from studies with insufficient statistical power. We indicate that the power of a study is not only influenced by the sample size, but also by genetic heterogeneity, the extent and degree of linkage disequilibrium (LD) between the markers tested and the causal variants, and the allele frequency differences between them. We also discuss the effects of other confounding factors, including population stratification, phenotype difference, genotype and phenotype quality control, multiple testing, and genuine biological differences. In addition, we note that with low statistical power, even a "replicated" finding is still likely to be a false positive. We believe that with rigorous control of study design and interpretation of different outcomes, inconsistency will be largely reduced, and the chances of successfully revealing genetic components of complex diseases will be greatly improved.

Animals↗
PubMed1997

Unequal randomisation can improve the economic efficiency of clinical trials.

OBJECTIVES: In the majority of clinical trials patients are randomised equally between treatment groups. This approach maximises statistical power for a given total sample size. The objectives of this paper were to determine if, when research costs between treatments differ, it is more economically efficient to randomise additional patients to the cheaper treatment, and how the optimum randomisation ratio can be estimated. METHODS: Estimation of the most economically efficient randomisation ratio for four hypothetical clinical trials using cost-effectiveness analysis. RESULTS: When research costs differ between treatments, and there is no constraint on total sample size, it is always more cost-effective to randomise more patients to the cheaper treatment. For example, a cost ratio between the lesser and more expensive treatment of ten, results in a randomisation ratio of 3.2:1. CONCLUSIONS: Unequal randomisation ratios should be more widely used as this will achieve optimum statistical power for the lowest expenditure of research resources.

Budgets↗
PubMed2004

Effects of censoring on parameter estimates and power in genetic modeling.

Genetic and environmental influences on variance in phenotypic traits may be estimated with normal theory Maximum Likelihood (ML). However, when the assumption of multivariate normality is not met, this method may result in biased parameter estimates and incorrect likelihood ratio tests. We simulated multivariate normal distributed twin data under the assumption of three different genetic models. Genetic model fitting was performed in six data sets: multivariate normal data, discrete uncensored data, censored data, square root transformed censored data, normal scores of censored data, and categorical data. Estimates were obtained with normal theory ML (data sets 1-5) and with categorical data analysis (data set 6). Statistical power was examined by fitting reduced models to the data. When fitting an ACE model to censored data, an unbiased estimate of the additive genetic effect was obtained. However, the common environmental effect was underestimated and the unique environmental effect was overestimated. Transformations did not remove this bias. When fitting an ADE model, the additive genetic effect was underestimated while the dominant and unique environmental effects were overestimated. In all models, the correct parameter estimates were recovered with categorical data analysis. However, with categorical data analysis, the statistical power decreased. The analysis of L-shaped distributed data with normal theory ML results in biased parameter estimates. Unbiased parameter estimates are obtained with categorical data analysis, but the power decreases.

Computer Simulation↗
PubMed2006

G-protein beta3 subunit gene variant is unlikely to have a significant influence on serum uric acid level in Japanese workers.

The C825T variant of the G-protein beta3 subunit (GNB3) gene has attracted renewed attention as a candidate gene for obesity, hypertension and hyperuricemia. The main role of G-protein is to translate signals from the cell surface into a cellular response. The 825T allele is associated with a splice variant of GNB3 protein and enhanced G-protein activation. We examined the relationship between this variant and the risk of hyperuricemia in Japanese workers. The study subjects were 1,452 men and 1,169 women selected from 3,834 men and 2,591 women in 1997. On the basis of common clinical criteria, hyperuricemia I was defined as serum uric acid >or= 7.0 mg/dl in men and 6.0 mg/dl in women or taking antihyperuricemic medication. The hyperuricemia I group consisted of 186 men and 20 women and its control of 1,266 men and 1,149 women. Hyperuricemia II was defined as serum uric acid > 5.7 mg/dl (median) in men and 3.9 mg/dl (median) in women or taking antihyperuricemic medication. The hyperuricemic II group consisted of 684 men and 570 women and its control of 768 men and 599 women. To replicate previous significant results in young Caucasian men, we selected these criteria because the authors of the study in young Caucasian men adopted the median in their subjects as a cut-off. The statistical power was estimated as 99% based on the significant results in Caucasians. Genotype and allele distributions in men and women with hyperuricemia I and II were not significantly different from those in the corresponding control groups. Logistic regression analysis on hyperuricemia I and II, and multiple regression on serum uric acid level demonstrated no significant effect of the C825T genotype. Despite the sufficient statistical power, this study could not demonstrate the significant influence of C825T on hyperuricemia or serum uric acid. The targeting of this polymorphism is unlikely to be beneficial in the prevention of hyperuricemia in the general Japanese population.

Cross-Sectional Studies↗
PubMed2005

Exposure assessment implications for the design and implementation of the National Children's Study.

Examining the influence of environmental exposures on various health indices is a critical component of the planned National Children's Study (NCS). An ideal strategy for the exposure monitoring component of the NCS is to measure indoor and outdoor concentrations and personal exposures of children to a variety of pollutants, including ambient particulate and gaseous pollutants, biologic agents, persistent organics, nonpersistent organics (e.g., pesticides), inorganic chemicals (e.g., metals), and others. However, because of the large sample size of the study (approximately 100,000 children), it is not feasible to assess every possible exposure of each child. We envision that cost-effective strategies for gathering the necessary exposure-related information with minimum burden to participants, such as broad administration of product-use questionnaires and diaries, would likely be considered in designing the exposure component of the NCS. In general a biologic (e.g., blood, urine, hair, saliva) measure could be the dosimeter of choice for many of the persistent and for some of the nonpersistent organic pollutants. Biologic specimens, such as blood, can also indicate long-term internal dose to various metals, including lead and mercury. Environmental measures, on the other hand, provide pathway/source-specific exposure estimates to many of the environmental agents, including those where biologic measurements are not currently feasible (e.g., for particulate matter and for some gaseous criteria pollutants). However, these may be burdensome and costly to either collect or analyze and may not actually indicate the absorbed dose. Thus, an important technical and logistical challenge for the NCS is to develop an appropriate study design with adequate statistical power that will permit detection of exposure-related health effects, based on an optimum set of exposure measurement methods. We anticipate that low-cost, low-burden methods such as questionnaires and screening type assessments of environmental and biologic samples could be employed, when exposures at different critical life stages of vulnerability can be reliably estimated by these simpler methods. However, when reliability and statistical power considerations dictate the need for collecting more specific exposure information, more extensive environmental, biologic, and personal exposure measurements should be obtained from various "validation" subsets of the NCS population that include children who are in different life stages. This strategy of differential exposure measurement design may allow the exposure-response relationships to be tested on the whole cohort by incorporating the information on the relationship between different types of exposure measures (i.e., ranging from simple to more complex) derived from the detailed validation subsamples.

Adolescent↗
PubMed2004

Accuracy and power of statistical methods for detecting adaptive evolution in protein coding sequences and for identifying positively selected sites.

The parsimony method of Suzuki and Gojobori (1999) and the maximum likelihood method developed from the work of Nielsen and Yang (1998) are two widely used methods for detecting positive selection in homologous protein coding sequences. Both methods consider an excess of nonsynonymous (replacement) substitutions as evidence for positive selection. Previously published simulation studies comparing the performance of the two methods show contradictory results. Here we conduct a more thorough simulation study to cover and extend the parameter space used in previous studies. We also reanalyzed an HLA data set that was previously proposed to cause problems when analyzed using the maximum likelihood method. Our new simulations and a reanalysis of the HLA data demonstrate that the maximum likelihood method has good power and accuracy in detecting positive selection over a wide range of parameter values. Previous studies reporting poor performance of the method appear to be due to numerical problems in the optimization algorithms and did not reflect the true performance of the method. The parsimony method has a very low rate of false positives but very little power for detecting positive selection or identifying positively selected sites.

Algorithms↗
PubMed2005

Low plasma coenzyme Q10 levels as an independent prognostic factor for melanoma progression.

BACKGROUND: Abnormally low plasma levels of coenzyme Q10 (CoQ10) have been found in patients with cancer of the breast, lung, or pancreas. OBJECTIVE: A prospective study of patients with melanoma was conducted to assess the usefulness of CoQ10 plasma levels in predicting the risk of metastasis and the duration of the metastasis-free interval. METHODS: Between January 1997 and August 2004, plasma CoQ10 levels were measured with high-performance liquid chromatography in 117 consecutive melanoma patients without clinical or instrumental evidence of metastasis according to American Joint Committee on Cancer criteria and in 125 matched volunteers without clinically suspect pigmented lesions. Patients taking CoQ10 or cholesterol-lowering medications and those with a diagnosis of diabetes mellitus were excluded from the study. Multiple statistical methods were used to evaluate differences between patients and control subjects and between patients who did (32.5%) and did not (67.5%) develop metastases during follow-up. RESULTS: CoQ10 levels were significantly lower in patients than in control subjects (t test: P < .0001) and in patients who developed metastases than in the metastasis-free subgroup (t test: P < .0001). Logistic regression analysis indicated that plasma CoQ10 levels were a significant predictor of metastasis (P = .0013). The odds ratio for metastatic disease in patients with CoQ10 levels that were less than 0.6 mg/L (the low-end value of the range measured in a normal population) was 7.9, and the metastasis-free interval was almost double in patients with CoQ10 levels 0.6 mg/L or higher (Kaplan-Meier analysis: P < .001). LIMITATIONS: A study with a larger sample, which is currently being recruited, and a longer follow-up will doubtlessly increase the statistical power and enable survival statistics to be obtained. CONCLUSIONS: Analysis of our findings suggests that baseline plasma CoQ10 levels are a powerful and independent prognostic factor that can be used to estimate the risk for melanoma progression.

Adult↗
PubMed1997

Vocational outcome of intervention for low-back pain.

Practical management guidelines for occupational health physicians are needed for the individual support of employees with low-back pain. In this study the level of evidence regarding the efficacy of intervention with vocational outcome parameters was assessed. In a systematic literature search, 40 randomized clinical trials on different types of intervention were retrieved. Their internal validity and statistical power criteria were assessed. The randomization procedure, blinding of patients, and sample size were problematic in most studies. For patients with acute low-back pain limited or moderate evidence was found for the efficacy of no bed rest, a short period of bed rest, and spinal manipulation. For chronic patients limited evidence was found for the efficacy of antidepressants. For the other types of intervention, studies with sufficient statistical power were lacking. Such studies are needed before more-detailed evidence-based guidelines can be formulated for occupational health care.

Absenteeism↗