SEARCH · PubMed Health

Results for “bootstrap”

Explore indexed PubMed citations for clinical trials, systematic reviews and public health research. Read source abstracts and follow each citation to its original PubMed record.

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

At least 577 records · Page 32Linked to original sources

PubMed2005

The distribution of genetic parameter estimates and confidence intervals from small disconnected diallels.

The distributions of genetic variance components and their ratios (heritability and type-B genetic correlation) from 105 pairs of six-parent disconnected half-diallels of a breeding population of loblolly pine (Pinus taeda L.) were examined. A series of simulations based on these estimates were carried out to study the coverage accuracy of confidence intervals based on the usual t-method and several other alternative methods. Genetic variance estimates fluctuated greatly from one experiment to another. Both general combining ability variance (sigma(2) (g)) and specific combining ability variance (sigma(2) (s)) had a large positive skewness. For sigma(2) (g) and sigma(2) (s), a skewness-adjusted t-method proposed by Boos and Hughes-Oliver (Am Stat 54:121-128, 2000) provided better upper endpoint confidence intervals than t-intervals, whereas they were similar for the lower endpoint. Bootstrap BCa-intervals (Efron and Tibshirani, An introduction to the bootstrap. Chapman & Hall, London 436 p, 1993) and Hall's transformation methods (Zhou and Gao, Am Stat 54:100-104, 2000) had poor coverages. Coverage accuracy of Fieller's interval endpoint(J R Stat Soc Ser B 16:175-185, 1954) and t-interval endpoint were similar for both h(2) and r(B) for sample sizes n</=10, but for n=30 the Fieller's method is much better.

Analysis of Variance↗

PubMed2003

Statistical analysis of the impact of spectral correlation on observed formation constants from UV-visible spectroscopic measurements.

Information retrieved from UV-visible spectroscopic data by application of a self-modelling factor analysis algorithm showed apparently systematically shifted thermodynamic properties for the same chemical system as a function of spectral slit widths. This empirical observation triggered a systematic investigation into the likely effects of residual and spectral correlation on the numerical results from quantitative spectroscopic investigations. If slit width was a nuisance factor it would reduce the comparability of information evaluated from spectroscopic data. The influence of spectral slit width was investigated by simulation, i.e. by generating and evaluating synthetic spectra with known properties. The simulations showed that increasing spectral correlation may introduce bias into factor analysis evaluations. By evaluation of the complete measurement uncertainty budget using threshold bootstrap target factor (TB CAT) analysis, the apparent shifts are insignificant relative to the total width of the quantity's measurement uncertainty. Increasing the slit widths causes some systematic effects, for example broadening of the registered spectral bands and reduction of spectral noise, because of higher light intensity passing to the detector. Hence, the observed systematic shifts in mean values might be caused by some latent correlation. As a general conclusion, slit width does not affect bias. However, the simulations show that spectral correlation and residual correlation may cause bias. Residual correlation can be taken into account by computer-intensive statistical methods, for example moving block or threshold bootstrap analysis. Spectral correlation is a property of the chemical system under study and cannot be manipulated. As a major result, evidence is given showing that stronger spectral correlation ( r<-0.7) causes non-negligible bias in the evaluated thermodynamic information from such a system.

Journal Article↗

PubMed1996

Confidence intervals for calibration with neural networks.

Based on neural network calibration the confidence intervals of aromaticity determination from infrared reflectance spectra of raw brown coals were estimated by means of the bootstrap method, a simplified Monte Carlo Simulation. The standard deviations and the confidence intervals were estimated to characterise the analysis error.It is shown that confidence intervals of non-linear analysis methods like Back Propagation Neural Networks (BPNN) can be estimated by the bootstrap method. The estimated confidence intervals of the calibration confirm the analysis by BPNN.

Journal Article↗

PubMed2002

JC virus strains indigenous to northeastern Siberians and Canadian Inuits are unique but evolutionally related to those distributed throughout Europe and Mediterranean areas.

Human polyomavirus JC virus (JCV) isolates around the world are classified into more than 10 geographically distinct genotypes (designated as subtypes). Evolutionary relationships among JCV subtypes were recently examined, and the following pattern of JCV evolution was indicated. The ancestral JCV first divided into three superclusters, designated Types A, B, and C. A split in Type A generated two subtypes, EU-a and -b, containing mainly European and Mediterranean isolates. The split in Type B generated Af 2 (the major African subtype), Bl-c (a minor European subtype), and various Asian subtypes. Type C generated a single subtype (Afl), consisting of isolates derived from western Africa. In this study, JCV isolates prevalent among northeastern Siberians and Canadian Inuits were evaluated in the context of the above-described pattern of JCV evolution. The Siberian/Arctic JCV isolates were classified as belonging mainly to Type A, based on the result of a preliminary phylogenetic analysis. We then examined, using the whole-genome approach, the phylogenetic relationships among worldwide Type A isolates. In neighbor-joining and maximum-likelihood analyses, Type A JCVs worldwide consistently diverged into three subtypes, EU-a, -b, and -c, with high bootstrap probabilities. EU-c was constructed only by northeastern Siberian isolates, derived mainly from Nanais living in the lower Amur River region, and was shown to have been generated by the first split in Type A. Most Siberian/Arctic isolates derived from Chukchis, Koryaks, and Canadian Inuits formed a distinct cluster within the EU-a subtype, with a high bootstrap probability. Based on the present findings, we discuss ancient human migrations, accompanied by Type A JCVs, across Asia and to Arctic areas of North America.

Adult↗

PubMed2002

Estimation of evolutionary parameters with phylogenetic trees.

An important issue in the phylogenetic analysis of nucleotide sequence data using the maximum likelihood (ML) method is the underlying evolutionary model employed. We consider the problem of simultaneously estimating the tree topology and the parameters in the underlying substitution model and of obtaining estimates of the standard errors of these parameter estimates. Given a fixed tree topology and corresponding set of branch lengths, the ML estimates of standard evolutionary model parameters are asymptotically efficient, in the sense that their joint distribution is asymptotically normal with the variance-covariance matrix given by the inverse of the Fisher information matrix. We propose a new estimate of this conditional variance based on estimation of the expected information using a Monte Carlo sampling (MCS) method. Simulations are used to compare this conditional variance estimate to the standard technique of using the observed information under a variety of experimental conditions. In the case in which one wishes to estimate simultaneously the tree and parameters, we provide a bootstrapping approach that can be used in conjunction with the MCS method to estimate the unconditional standard error. The methods developed are applied to a real data set consisting of 30 papillomavirus sequences. This overall method is easily incorporated into standard bootstrapping procedures to allow for proper variance estimation.

Animals↗

PubMed2003

Phylogenetic relationships among JC virus strains in Japanese/Koreans and Native Americans speaking Amerind or Na-Dene.

Many genetic studies using human mtDNA or the Y chromosome have been conducted to elucidate the relationships among the three Native American groups speaking Amerind, Na-Dene, and Eskimo-Aleut. Human polyomavirus JC (JCV) may also help to gain insights into this issue. JCV isolates are classified into more than 10 geographically distinct genotypes (designated subtypes here), which were generated by splits in the three superclusters, Types A, B, and C. A particular subtype of JCV (named MY) belonging to Type B is spread in both Japanese/Koreans and Native Americans speaking Amerind or Na-Dene. In this study, we evaluated the phylogenetic relationships among MY isolates worldwide, using the whole-genome approach, with which a highly reliable phylogeny of JCV isolates can be reconstructed. Thirty-six complete sequences belonging to MY (10 from Japanese/Koreans, 24 from Native Americans, and 2 from others), together with 54 belonging to other subtypes around the world, were aligned and subjected to phylogenetic analysis using the neighbor-joining and maximum-likelihood methods. In the resultant phylogenetic trees, the MY sequences diverged into two Japanese/Korean and five Native American clades with high bootstrap probabilities. Two of the Native American clades contained isolates mainly from Na-Denes and the others contained isolates mainly from Amerinds. The Na-Dene clades were not clustered together, nor were the Amerind clades. In contrast, the two Japanese/Korean clades were clustered at a high bootstrap probability. We concluded that there is no distinction between Amerinds and Na-Denes in terms of indigenous JCVs, although they are linguistically distinguished from each other.

Humans↗

PubMed2003

Molecular phylogeny of centrohelid heliozoa, a novel lineage of bikont eukaryotes that arose by ciliary loss.

Recent molecular and cellular evidence indicates that eukaryotes comprise three major lineages: the probably ancestrally uniciliate protozoan phylum Amoebozoa; the ancestrally posteriorly uniciliate opisthokont clade (animals, Choanozoa, and fungi); and a very diverse ancestrally biciliate clade, the bikonts-plants, chromalveolates, and excavate and rhizarian Protozoa. As Heliozoa are the only eukaryote phylum not yet placed on molecular sequence trees, we have sequenced the 18S rRNA genes of three centrohelid heliozoa, Raphidiophrys ambigua, Heterophrys marina, and Chlamydaster sterni, to investigate their phylogenetic position. Phylogenetic analysis by distance and maximum likelihood methods allowing for intersite rate variation and invariable sites confirms that centrohelid heliozoa are a robust clade that does not fall within any other phyla. In particular, they are decisively very distant from the heterokont pedinellid chromists, at one time thought to be related to heliozoa, and lack the unique heterokont signature sequence. They also appear not to be specifically related to either Amoebozoa or Radiolaria, with which they have sometimes been classified, so it is desirable to retain Heliozoa as a separate protozoan phylum. Even though centrohelids have no cilia or centrioles, the centrohelid clade branches among the bikont eukaryotes, but there is no strong bootstrap support for any particular position. Distance trees usually place centrohelids as sisters to haptophytes, whereas parsimony puts them as sisters to red algae, but there is no reason to think that either position is correct; both have very low bootstrap support. Quartet puzzling places them with fairly low support as sisters to the apusozoan zooflagellate Ancyromonas. As Ancyromonas is the only other eukaryote that shares the character combination of flat plate-like mitochondrial cristae and kinetocyst-type extrusomes with centrohelids, this position is biologically plausible, but because of weak support and conflict between trees it might not be correct. Irrespective of their precise position, our trees (together with previous evidence that Chlamydaster sterni has the derived dihydrofolate reductase/thymidylate synthetase gene fusion unique to bikonts) indicate that centrohelid heliozoa are ancestrally derived from a bikont flagellate by the loss of cilia. The centroplast that nucleates their axonemal microtubules is therefore almost certainly homologous with the centrosome of ciliated eukaryotes and should simply be called a centrosome.

Animals↗

PubMed2001

The cytochrome b gene as a phylogenetic marker: the limits of resolution for analyzing relationships among cichlid fishes.

The mitochondrial cytochrome b (cyt-b) gene is widely used in systematic studies to resolve divergences at many taxonomic levels. The present study focuses mainly on the utility of cyt-b as a molecular marker for inferring phylogenetic relationship at various levels within the fish family Cichlidae. A total of 78 taxa were used in the present analysis, representing all the major groups in the family Cichlidae (72 taxa) and other families from the suborders Labroidei and Percoidei. Gene trees obtained from cyt-b are compared to a published total evidence tree derived from previous studies. Minimum evolution trees based on cyt-b data resulted in topologies congruent with all previous analyses. Parsimony analyses downweighting transitions relative to transversions (ts1:tv4) or excluding transitions at third codon positions resulted in more robust bootstrap support for recognized clades than unweighted parsimony. Relative rate tests detected significantly long branches for some taxa (LB taxa) which were composed mainly by dwarf Neotropical cichlids. An improvement of the phylogenetic signal, as shown by the four-cluster likelihood mapping analysis, and higher bootstrap values were obtained by excluding LB taxa. Despite some limitations of cyt-b as a phylogenetic marker, this gene either alone or in combination with other data sets yields a tree that is in agreement with the well-established phylogeny of cichlid fish.

Animals↗

PubMed2000

RESEARCH: Assessing Uncertainty in Estimates of Nitrogen Loading to Estuaries for Research, Planning, and Risk Assessment.

/ There can be considerable uncertainty associated with calculations of nutrient loading to estuaries from their watersheds, arising from uncertainty in the variables used in the calculation. Analysis of uncertainty is particularly important in the context of planning and management, where such information can be useful in helping make decisions about development in the coastal zone and in risk assessment, where probability of worse-case extremes may be relevant. This fact has been largely ignored when loading calculations have been made, presumably because both uncertainty estimates for the input variables and a standard method were lacking. Parametric (propagation for normal error estimates) and nonparametric methods (bootstrap and enumeration of combinations) to assess the uncertainty in calculated rates of nitrogen loading were compared, based on the propagation of uncertainty observed in the variables used in the calculation. In addition, since such calculations are often based on literature surveys rather than random replicate measurements for the site in question, error propagation was also compared using the uncertainty of the sampled population (e.g., standard deviation) as well as the uncertainty of the mean (e.g., standard error of the mean). Calculations for the predicted nitrogen loading to a shallow estuary (Waquoit Bay, MA) were used as an example. The previously estimated mean loading from the watershed (5,400 ha) to Waquoit Bay (600 ha) was 23,000 kg N yr(-1). The mode of a nonparametric estimate of the probability distribution differed dramatically, equaling only 70% of this mean. Repeated observations were available for only 8 of the 16 variables used in our calculation. We estimated uncertainty in model predictions by treating these as sample replicates. Parametric and nonparametric estimates of the standard error of the mean loading rate were 12-14%. However, since the available data include site-to-site variability, as is often the case, standard error may be an inappropriate measure of confidence. The standard deviations were around 38% of the loading rate. Further, 95% confidence intervals differed between the nonparametric and parametric methods, with those of the nonparametric method arranged asymmetrically around the predicted loading rate. The disparity in magnitude and symmetry of calculated confidence limits argue for careful consideration of the nature of the uncertainty of variables used in chained calculations. This analysis also suggests that a nonparametric method of calculating loading rates using most frequently observed values for variables used in loading calculations may be more appropriate than using mean values. These findings reinforce the importance of including assessment of uncertainty when evaluating nutrient loading rates in research and planning. Risk assessment, which may need to consider relative probability of extreme events in worst-case scenarios, will be in serious error using normal estimates, or even the nonparametric bootstrap. A method such as our enumeration of combinations produces a more reliable distribution of risk.

Journal Article↗

PubMed2002

Population pharmacokinetics of the novel anticancer agent KRN7000.

PURPOSE: KRN7000 is a novel anticancer agent, acting through stimulation of the immune system. The first clinical trial with this agent, which included pharmacokinetic studies, has recently been completed. The aim of the study presented here was to develop a population pharmacokinetic model for KRN7000. METHODS: Plasma concentration-time data were gathered from 24 patients enrolled in a phase I trial in which KRN7000 was administered as a weekly slow injection at doses ranging from 50 to 4800 microg/m(2). These data were used to build a pharmacokinetic model using the nonlinear mixed-effect modeling (NONMEM) program. The model was validated by performance of 200 bootstraps. RESULTS: A three-compartment model with interindividual variability on the central and two peripheral volumes of distribution (V1, V2 and V3) and on clearance (CL) adequately described the data. The final estimates were: V1 2.34 l, V2 2.61 l, V3 2.13 l, and CL 0.130 l/h. Of 24 covariates tested, including both demographic and pathophysiological factors, none showed a significant relationship with the pharmacokinetic parameters obtained. The bootstrap analysis provided parameter estimates within approximately 15% of the original estimates, indicating stability of the model. CONCLUSION: The pharmacokinetic behavior of KRN7000 in the clinical trial could be described by a three-compartment model. Hence, KRN7000 demonstrates linear pharmacokinetics over the investigated dose range. The pharmacokinetics of KRN7000 are not influenced by patient demographic or pathophysiological characteristics.

Adult↗

PubMed2002

A population analysis of the pharmacokinetics of Cremophor EL using nonlinear mixed-effect modelling.

PURPOSE: The purpose of this study was to develop a population pharmacokinetic model for Cremophor EL used as a formulation vehicle for paclitaxel. METHODS: Plasma concentration-time data from 70 patients (85 courses) treated with paclitaxel dissolved in Cremophor EL were used. The nonlinear mixed-effect modelling (NONMEM) program was used for the population pharmacokinetic analysis. The influence of patient characteristics on the pharmacokinetics of Cremophor EL was determined. The stability of the final model was evaluated using bootstrapping. RESULTS: The data were optimally fitted to a three-compartment model with Michaelis-Menten elimination from the central compartment. The following pharmacokinetic parameters were estimated: volume of the central compartment (V1=2.59 l), volumes of two peripheral compartments (V2=1.81 l, V3=1.61 l), intercompartmental clearance between central and peripheral compartments (Q12=1.44 l/h, Q13=0.155 l/h), maximal elimination rate (Vmax=0.193 ml/h), and concentration at half Vmax (Km=0.122 ml/l). Interindividual variability of the pharmacokinetic parameters was quantified for V1 (25%), V2 (36%) and Vmax (31%). Residual variability consisted of a combined additional (0.095 ml/l) and proportional error (7%). Gender, body surface area and performance status according to the World Health Organization were significantly correlated with V1, V2 and Vmax, respectively ( P<0.0001). The median parameter estimates of 1000 bootstrap samples were in accordance with those obtained with the original data set, indicating the validity of the population model. CONCLUSIONS: The population model was able to adequately describe the pharmacokinetic parameters and influence of covariates on the pharmacokinetics of Cremophor EL. This model can be used when studying the relationship between the pharmacokinetics and toxicity of Cremophor EL, and the drug's influence on the pharmacokinetics of paclitaxel.

Adult↗

PubMed2002

Population pharmacokinetic and limited sampling models for carboplatin administered in high-dose combination regimens with peripheral blood stem cell support.

OBJECTIVE: By means of a nonlinear mixed effect modeling technique, a population pharmacokinetic (PK) model was developed to evaluate the effects of a variety of covariates on clearance and other pharmacokinetic parameters of ultrafilterable carboplatin administered in high-dose combination regimens with peripheral blood stem cell support. In addition, single-sample and two-sample limited sampling models (LSMs) were derived to estimate carboplatin's AUC that could be used in the design of drug dosing regimens. METHODS: A total of 44 female patients with advanced ovarian cancer participated in two phase I studies. All 44 patients received a high-dose carboplatin chemotherapy with other anticancer drugs. A population PK model was applied to the plasma concentration-time data of ultrafilterable carboplatin using the NONMEM and Xpose computer programs. The Xpose program utilized a general additive modeling technique to identify significant patient covariates and PK parameter relationships. The resultant PK model was validated using a bootstrap method. Stepwise linear regression analyses were used to develop LSMs based on the correlation between carboplatin's AUC and plasma concentrations. RESULTS: The best structural covariate-free model for high-dose carboplatin was a linear two-compartment model with an exponential error model to account for intersubject variability and a CCV error model to account for intrasubject variability. Subsequently, a final covariate model for clearance (l/min) was obtained as follows: TVCL=0.101+0011*(WT-62.35)-0.0658*(SCR-0.65) where WT is body weight (kg) and SCR is serum creatinine (mg/dl). Both WT and SCR were found to significantly influence carboplatin's total clearance. It was determined that the best single-sample LSM was AUC(LSM)=0.553*C(240min) ( r=0.998). CONCLUSION: Both a population PK model and a LSM for high-dose carboplatin were developed following its administration in combination chemotherapeutic regimens with peripheral blood stem cell support. In both cases, the models performed well when analyzed in the context of the retrospective and bootstrap analyses. Prospective studies in ovarian cancer patients should be conducted to further tailor the current models.

Adult↗

PubMed2000

Evidence for multiple species within the endoparasitic dinoflagellate Amoebophrya ceratii as based on 18S rRNA gene-sequence analysis.

Parasitism within the group of dinoflagellates is a widespread phenomenon. Whether the parasitic dinoflagellates exhibit specificity in their infection is not well known, but this possibility has become an important issue in the development of biological control of harmful algal blooms. The 18S rDNA sequences from the parasite Amoebophrya sp. and its dinoflagellate host Dinophysis norvegica were determined and compared with the published sequence of Amoebophrya sp. infecting Gymnodinium sanguineum and other dinoflagellates. The results showed that the sequence from the parasite within D. norvegica was clustered with that of the one from G. sanguineum with 100% bootstrap support in a maximum-likelihood analysis. The observed identity between these two sequences was 93%, which indicates that they are not identical species. The two sequences from Amoebophrya sp. were deeply branched within the group of dinoflagellate sequences and represent the earliest diverging dinoflagellates. The sequence from the parasite Parvilucifera infectans, also infecting D. norvegica, was not closely related to the Amoebophrya sp. sequences. The sequence from D. norvegica appeared as a sister group to a cluster containing Prorocentrum lima and Alexandrium spp. without significant bootstrap support. The data presented herein support the hypothesis that A. ceratii comprises more than one species, and this opens the possibility that infections of harmful algal species might involve more than one Amoebophrya species.

Animals↗

PubMed2001

Random amplified polymorphic DNA analysis applied to acute otitis media caused by penicillin non-susceptible Streptococcus pneumoniae.

The spread of penicillin non-susceptible Streptococcus pneumoniae (PNSSP) is an emerging problem for the treatment of acute otitis media (AOM). Attendance of children at day care centers, as well as the spread of PNSSP, is a risk factor for AOM. The status of the spread of PNSSP during the acute infection phase of AOM has not been evaluated. We examined the clonality of samples from seven children in a day care center who simultaneously developed AOM caused by PNSSP. The seven isolates from the children, and six control samples were grouped by serotyping, by determining resistance to antimicrobial agents, and by genotyping, carried out by sequencer-based random amplified polymorphic DNA (RAPD), and validated by bootstrap analysis. There was no evidence to indicate the direct dissemination of PNSSP among these patients in the day care center, although the simultaneous occurrence of PNSSP AOM had initially suggested a clonal outbreak. The possible presence of a common ancestral strain suggested the importance of surveillance during the carrier state. The result of RAPD genotyping was highly reproducible, as validated by the high bootstrap score. The use of an automated sequencer, in combination with a careful choice of primers, and commercially established kits, played a significant role in the reproducibility of the studies.

Child↗

PubMed2004

Molecular phylogeny of monocotyledons inferred from combined analysis of plastid matK and rbcL gene sequences.

Using matK and rbcL sequences (3,269 bp in total) from 113 genera of 45 families, we conducted a combined analysis to contribute to the understanding of major evolutionary relationships in the monocotyledons. Trees resulting from the parsimony analysis are similar to those generated by earlier single or multiple gene analyses, but their strict consensus tree provides much better resolution of relationships among major clades. We find that Acorus (Acorales) is a sister group to the rest of the monocots, which receives 100% bootstrap support. A clade comprising Alismatales is diverged as the next branch, followed successively by Petrosaviaceae, the Dioscoreales-Pandanales clade, Liliales, Asparagales and commelinoids. All of these clades are strongly supported (with more than 90% bootstrap support). The sister-group relationship is also strongly supported between Alismatales and the remaining monocots (except for Acorus) (100%), between Petrosaviaceae and the remaining monocots (except for Acorus and Alismatales) (100%), between the clade comprising Dioscoreales and Pandanales and the clade comprising Liliales, Asparagales and commelinoids (87%), and between Liliales and the Asparagales-commelinoids clade (89%). Only the sister-group relationship between Asparagales and commelinoids is weakly supported (68%). Results also support the inclusion of Petrosaviaceae in its own order Petrosaviales, Nartheciaceae in Dioscoreales and Hanguanaceae in Commelinales.

Base Sequence↗

PubMed2002

Phylogeny of Saururaceae based on mitochondrial matR gene sequence data.

DNA sequences of matR gene from three species of Saururaceae and the selected outgroups, Chloranthus holostegius and Zippelia begoniaefolia, are reported. All DNA sequences of six species in four genera of Saururaceae and the two outgroups are analyzed on PAUP 4.0 8b to reconstruct the phylogeny. A single matR gene tree is generated from parsimony, distance, and likelihood analyses, respectively. The three trees with the same topology are slightly different in bootstrapping support for some clades. The result indicates that Saururaceae is monophyletic. Anemopsis is sister to Houttuynia, and the two genera form the first diverging lineage of the family. The sister group relationship between Saururus and Gymnotheca is also supported by a relatively high bootstrap value. The result is different from all the former phylogenetic opinions on Saururaceae based on morphology, but it is supported by the evolution of flower-bract stalk in Saururaceae. In addition, some characteristics of the matR gene are analyzed. The MatR gene is a relatively better tool to reconstruct the molecular clock because the base substitution bias greatly decreases in the gene.

Journal Article↗

PubMed2004

Estimating the expectation of the log-likelihood with censored data for estimator selection.

A criterion for choosing an estimator in a family of semi-parametric estimators from incomplete data is proposed. This criterion is the expected observed log-likelihood (ELL). Adapted versions of this criterion in case of censored data and in presence of explanatory variables are exhibited. We show that likelihood cross-validation (LCV) is an estimator of ELL and we exhibit three bootstrap estimators. A simulation study considering both families of kernel and penalized likelihood estimators of the hazard function (indexed on a smoothing parameter) demonstrates good results of LCV and a bootstrap estimator called ELL(bboot). We apply the ELL(bboot) criterion to compare the kernel and penalized likelihood estimators to estimate the risk of developing dementia for women using data from a large cohort study.

Bias↗

PubMed2005

A nonparametric comparison of conditional distributions with nonnegligible cure fractions.

Survival data with nonnegligible cure fractions are commonly encountered in clinical cancer clinical research. Recently, several authors (e.g. Kuk and Chen, Biometrika 79 (1992) 531; Maller and Zhou, Journal of Applied Probability, 30 (1993) 602; Peng and Dear, Biometrics, 56 (2000) 237; Sy and Taylor, Biometrics 56 (2000) 227) have proposed to use semiparametric cure models to analyze such data. Much of the existing work has been emphasized on cure detections and regression techniques. In contrast, this project focuses on the hypothesis testing in the presence of a cure fraction. Specifically, our interest lies in detecting whether there exists survival differences among noncured patients between treatment arms. For this purpose, we investigate the use of a modified Cramér-von Mises statistic for two-sample survival comparisons within the framework of cure models. Such a test has been studied by Tamura et al., (Statistics in Medicine 19, 2000, 2169) using bootstrap procedure. We will focus on developing asymptotic theory and convergent algorithms in this paper. We show that the limiting distributions of the Cramér-von Mises statistic under the null hypothesis can be represented by stochastic integrals and a weighted noncentral chi-squares. Both representations lead to concrete numerical schemes for computing the limiting distributions. The algorithms can be easily implemented for data analysis and significantly reduce computing time compared to the bootstrap approach. For illustrative purposes, we apply the proposed test to a published clinical trial.

Algorithms↗