SEARCH · Search PubMed

Results for “hypothesis testing”

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.

SourceResearch connectionFrom yearThrough year

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

At least 397 records · Page 22Linked to original sources

PubMed1995

Similarity/equivalence trials for combination vaccines.

In a similarity or "equivalence" trial of a combination vaccine, we wish to show that the combination is sufficiently similar to the separately administered components in some measure of safety, immunogenicity, or efficacy to justify use of the combination. In this setting it is usually required to show similarity in one direction only; specifically, we design the trial to rule out superiority of the separate components by as much as a prespecified quantity theta 0 in an appropriate outcome measure (e.g., a difference or ratio). It is crucial that theta 0 be chosen to be clinically meaningful, so that any difference or ratio less than theta 0 is truly acceptable to clinicians. Estimation is generally more relevant than hypothesis testing in such a trial, and consequently it is natural to consider design and analysis in terms of confidence intervals. The same sample sizes can be obtained, however, from a hypothesis testing approach. The appropriate null hypothesis is that the separate components are superior to the combination by at least theta 0, with rejection of the hypothesis supporting a conclusion of similarity. It is not appropriate to design the trial to test the null hypothesis of no difference, as we would do if we wished to demonstrate superiority of the combination vaccine; failure to reject that hypothesis does not prove similarity, and we might reject the hypothesis when the true difference is unimportant clinically. Further, this inappropriate approach may result in a sample size either larger or smaller than necessary. Sample size formulations are available for various types of comparative measures, e.g., a difference of normally distributed means, a difference or ratio of proportions, and a ratio of hazards.

Clinical Trials as Topic↗

PubMed1986

Mammalian X chromosome inactivation: testing the hypothesis of transcriptional control.

Mammalian X chromosome inactivation is generally considered to be a good example of stable transcriptional repression; however, there has been no satisfactory evidence for transcriptional control. We have made a test of the hypothesis of transcriptional control by Northern blot analysis of RNA from a woman heterozygous for a mutant Hpt allele which shows no detectable transcription of wild-type mRNA. Cells from this Hpt+ Hpt- woman were separated into HPRT+ and HPRT- subpopulations by selection in HAT or thioguanine. The HPRT+ population (in which the Hpt+ is on the active X) transcribed normal Hpt mRNA, while the HPRT- population (in which the Hpt+ allele is on the inactive X) did not. These results provide strong support for the hypothesis of transcriptional control.

Alleles↗

PubMed1997

Estimating Learning Curves of Concept Learning.

In this paper, we describe an approximation method which enables us to study the average generalization performance of learning directly via hypothesis testing inequalities. This unites the learning and the hypothesis testing in a common viewpoint. In particular, we investigate learning curves of a so-called ill-disposed learning algorithm, which can provide useful implications regarding the problem of overfitting from a practical and yet scientific viewpoint. The learning problem is stated in the probably approximately correct (PAC) learning model, but only the average generalization performance is addressed. The resulting bounds are directly related to the number of system weights. The advantages of the theory are that it alleviates the practical pessimism frequently claimed for the results of the VC theory, and provides general insights. We illustrate the results with some numerical simulations. Copyright 1997 Elsevier Science Ltd.

Journal Article↗

PubMed2003

A comparison of parametric versus permutation methods with applications to general and temporal microarray gene expression data.

MOTIVATION: In analyses of microarray data with a design of different biological conditions, ranking genes by their differential 'importance' is often desired so that biologists can focus research on a small subset of genes that are most likely related to the experiment conditions. Permutation methods are often recommended and used, in place of their parametric counterparts, due to the small sample sizes of microarray experiments and possible non-normality of the data. The recommendations, however, are based on classical knowledge in the hypothesis test setting. RESULTS: We explore the relationship between hypothesis testing and gene ranking. We indicate that the permutation method does not provide a metric for the distance between two underlying distributions. In our simulation studies permutation methods tend to be equally or less accurate than parametric methods in ranking genes. This is partially due to the discreteness of the permutation distributions, as well as the non-metric property. In data analysis the variability in ranking genes can be assessed by bootstrap. It turns out that the variability is much lower for permutation than parametric methods, which agrees with the known robustness of permutation methods to individual outliers in the data.

Algorithms↗

PubMed2002

Rehabilitation therapy self-efficacy and functional recovery after hip fracture.

Little is known about the role of psychological factors in the functional recovery process of hip fracture patients. This study employed a prospective cohort design to test the hypothesis that hospitalized hip fracture patients with greater reported self-efficacy for conducting rehabilitation therapy would have a greater likelihood of recovering to a pre-fracture level of locomotion function six months after the fracture. This hypothesis was tested controlling for pre-fracture level of function and depressive symptoms reported during hospitalization for surgical repair. An original measure of rehabilitation therapy self-efficacy was evaluated prior to hypothesis testing. Study patients were recruited from two hospitals, interviewed during hospitalization and followed up six months later. Patients included in hypothesis test analyses (n = 24) were mostly women (82%) with a mean age of 79 years. Results showed that patients with higher self-efficacy scores had a greater likelihood of locomotion recovery, controlling for pre-fracture locomotion function level (adjusted odds ratio (AOR) = 1.21; 95% confidence interval (CI) = 1.00-1.45; P= 0.05). This positive association between rehabilitation therapy self-efficacy and likelihood of locomotion recovery persisted after adding depressive symptoms (the Center for Epidemiological Studies-depression (CES-D) score) to this logistic regression model (AOR for self-efficacy = 1.18; 95% CI = 0.99-1.42; P= 0.07). It is concluded that rehabilitation therapy self-efficacy is a potentially important psychological factor in helping hip fracture patients recover locomotion functioning.

Activities of Daily Living↗

PubMed1989

An interpretation of cortical maps in echolocating bats.

Target parameters such as reflectivity, range, velocity, and angular position are represented by ordered maps of tuned cortical neurons in insectivorous bats. It is suggested that the response of each neuron in such a map is determined by a hypothesis test conditioned on a particular value of the mapped parameter. The excitation of each neuron is then interpreted as a sample value of a conditional log-likelihood ratio or a log-likelihood function. Interpolation between the samples, which is needed to find the parameter that maximizes the mapped function (e.g., the maximum likelihood parameter estimate), can be accomplished with overlapped tuning curves. An attempt to portray a sharp peak by a weighted sum of relatively broad neuronal tuning curves or interpolation functions results in excitatory center/inhibitory surround behavior. Facilitation or antifacilitation of neurons that are likely to be excited by succeeding observations can be used for sequential detection and tracking. Interpolation and pulse-to-pulse data storage capability are required to explain range jitter sensitivity and to allow for moving target indication in bat sonar. If a cortical map represents an ordered array of hypothesis tests, then many such tests are implemented in parallel when target parameters are unknown. Detection performance is then degraded relative to the idealized situation in which all parameters are specified. Performance in noise may thus appear to be much worse than that of an ideal detector, even if each hypothesis test is optimally implemented.

Animals↗

PubMed1997

Portal vein chemotherapy for colorectal cancer: a meta-analysis of 4000 patients in 10 studies. Liver Infusion Meta-analysis Group.

BACKGROUND: Systemic delivery of cytotoxic drugs yields relatively low doses in the liver, a major site of recurrence for colorectal cancer. Giving chemotherapy by means of continuous portal vein infusion (PVI) into the liver during the immediate postoperative period may improve therapeutic efficacy. PURPOSE: We undertook a meta-analysis to assess the effects on recurrence and survival of administering fluorouracil (5-FU)-based chemotherapy by PVI after colorectal cancer surgery. METHODS: Data on mortality and recurrence were sought for all patients enrolled in randomized trials initiated before 1987 in which a few days (range, 5-7 days) of continuous postoperative PVI of cytotoxic drugs was compared with no further treatment. Data from 10 trials (a promising initial study and nine hypothesis-testing trials) involving about 4000 patients were available for analysis. The main cytotoxic drug in each trial was 5-FU (given with heparin); however, mitomycin C was co-administered in two of the trials. Four trials included an additional control group of patients treated with continuous noncytotoxic PVI of either heparin or urokinase alone; one trial included a second control group of patients treated with continuous systemic infusion of 5-FU. Reported P values are two-sided. RESULTS: Among the 3499 patients randomly assigned to receive either cytotoxic PVI or no further treatment, 1557 deaths are known to have occurred. Survival with and without PVI appeared to be the same for the first 2 years; thereafter, it diverged significantly, with the absolute survival difference (i.e., improvement) associated with PVI at 5 years being 4.7 % (standard deviation [SD] = 1.2 %) (P = .006). When just the nine hypothesis-testing trials were considered, the absolute survival difference was 3.6% (SD = 1.2%) (P = .04). If, ignoring any potential for bias in stage assignment, attention was restricted to patients with Dukes' stage A, B, or C disease (88.3% of the total), the absolute effect on 5-year survival for all 10 trials increased to 6.0% (SD = 1.8%) (P = .001); this estimate remained statistically significant when the initial study was excluded (absolute survival difference = 4.8%; SD = 1.8%; P = .01). In contrast to the highly significant reduction in liver metastases seen in the initial study (79% reduction; SD = 15%; P = .00000007), the reduction found in the nine hypothesis-testing trials was not significant (14% reduction; SD = 10%; P = .2). In the trials with additional control groups, survival appeared to be better with cytotoxic PVI than with noncytotoxic PVI or with systemic cytotoxic drug infusion. CONCLUSIONS: PVI of 5-FU (with or without other cytotoxic drugs) for about 1 week after surgery in patients with colorectal cancer may produce an absolute improvement in 5-year survival of a few percent. Although encouraging, this finding is not statistically secure, and additional evidence from randomized trials involving several thousand more patients is needed.

Antibiotics, Antineoplastic↗

PubMed1997

Meta-analysis of cognitive-behavioral treatment studies for bulimia.

A meta-analysis was performed to systematically assess the effect of cognitive-behavioral treatments for bulimia. To protect against past criticisms of meta-analyses, this study focused on well-defined hypotheses with clearly articulated conceptual foundations. Twenty-six studies of the cognitive-behavioral treatment of bulimia were selected through computer searches. Effect sizes were calculated for changes in behavioral outcome measures (25 independent hypothesis tests) and cognitive-attitudinal outcome measures (17 independent hypothesis tests). Additionally, two effect sizes were generated for within and between group comparisons. The analysis revealed an effect size of average r = 0.69 for behavioral outcome measures (average r = 0.64 for between group and average r = 0.74 for within group) and average r = 0.67 for cognitive-attitudinal outcome measures (average r = 0.64 for between group and average r = 0.69 for within group). Follow-up effect sizes were less favorable; however, the diversity of time spans and outcome measures used to calculate follow-up effect sizes limit their utility. Overall, results suggest that the use of a cognitive-behavioral therapy will result in favorable treatment outcomes and implications for future research are discussed.

Adolescent↗

PubMed1998

Fundamental concepts in statistics: elucidation and illustration.

Fundamental concepts in statistics form the cornerstone of scientific inquiry. If we fail to understand fully these fundamental concepts, then the scientific conclusions we reach are more likely to be wrong. This is more than supposition: for 60 years, statisticians have warned that the scientific literature harbors misunderstandings about basic statistical concepts. Original articles published in 1996 by the American Physiological Society's journals fared no better in their handling of basic statistical concepts. In this review, we summarize the two main scientific uses of statistics: hypothesis testing and estimation. Most scientists use statistics solely for hypothesis testing; often, however, estimation is more useful. We also illustrate the concepts of variability and uncertainty, and we demonstrate the essential distinction between statistical significance and scientific importance. An understanding of concepts such as variability, uncertainty, and significance is necessary, but it is not sufficient; we show also that the numerical results of statistical analyses have limitations.

Humans↗

PubMed1993

Clinical trials in adult respiratory distress syndrome.

The need for randomized clinical trials in adult respiratory distress syndrome (ARDS) is now recognized. With this recognition comes the need for researchers to implement proper trial design and for the clinician to be able to interpret results as they apply to clinical practice. The heterogeneity of ARDS as related to etiology, stage, and severity creates the potential for maldistribution of patients in the clinical trial. Likewise, a particular intervention may benefit or not benefit a patient based on these variables. Choosing a minimum of end-points (ideally one) for hypothesis testing is important. The optimal end-point for hypothesis testing pertinent to clinical impact is all-cause mortality at a certain time point (usually 14 or 28 days). Unblinded trials are suboptimal, but necessary, with interventions such as mechanical ventilation treatment modalities. Given these circumstances, treatment protocols should be utilized in both groups. Cooperation of the basic scientist and the clinical scientist is ideal for direction of research in ARDS. Industry funding of ARDS clinical trials is now typical and needed. Under these circumstances, it is important to prevent inappropriate industry influence on trial design, data analysis, data interpretation, and data presentation. The investigator must remain above reproach and the informed consent process must be of the highest standard.

Animals↗

PubMed1997

Testing the hypothesis of the multidimensional model of anorexia nervosa in adolescents.

This study statistically tested six hypothesized risk factors of the model for anorexia nervosa. Forty-three adolescents with anorexia nervosa and 85 controls were administered the EAT, EDI, and FES. In addition, 43 parents of anorexics and 85 parents of controls completed the Family History Data Sheet, the FES, and the Perfect Child Questionnaire. Three of six hypothesized risk factors were confirmed: family history of depression, feelings of ineffectiveness, and poor interceptive awareness. Log-linear analysis indicated that the hierarchical model that best fit the data had significant two-way interactions with anorexia nervosa, G2 (11, N = 128) = 65.87, p < .001. In addition, alcohol and drug abuse or dependence figured prominently in the family history of patients with anorexia nervosa. The multidimensional model for anorexia nervosa holds up as an exploratory model of this condition in the adolescent age group.

Adolescent↗

PubMed2004

Testing the hypothesis of a recombinant origin of the SARS-associated coronavirus.

The origin of severe acute respiratory syndrome-associated corona-virus (SARS-CoV) is still a matter of speculation, although more than one year has passed since the onset of the SARS outbreak. In this study, we implemented a 3-step strategy to test the intriguing hypothesis that SARS-CoV might have been derived from a recombinant virus. First, we blasted the whole SARS-CoV genome against a virus database to search viruses of interest. Second, we employed 7 recombination detection techniques well documented in successfully detecting recombination events to explore the presence of recombination in SARS-CoV genome. Finally, we conducted phylogenetic analyses to further explore whether recombination has indeed occurred in the course of coronaviruses history predating the emergence of SARS-CoV. Surprisingly, we found that 7 putative recombination regions, located in Replicase 1ab and Spike protein, exist between SARS-CoV and other 6 coronaviruses: porcine epidemic diarrhea virus (PEDV), transmissible gastroenteritis virus (TGEV), bovine coronavirus (BCoV), human coronavirus 229E (HCoV), murine hepatitis virus (MHV), and avian infectious bronchitis virus (IBV). Thus, our analyses substantiate the presence of recombination events in history that led to the SARS-CoV genome. Like the other coronaviruses used in the analysis, SARS-CoV is also a mosaic structure.

Computational Biology↗

PubMed2001

Problem-solving abilities in unipolar depressed patients: comparison of performance on the modified version of the Wisconsin and the California sorting tests.

Problem solving relies on such abilities as decision-making, planning, initiation and hypothesis testing. Although problem-solving deficits have been consistently reported in depression, the specific nature of these deficits is not fully elucidated. In order to assess and isolate cognitive processes underlying problem-solving impairments in depression, depressed patients and normal controls were evaluated with the modified version of the Wisconsin Card Sorting Test (WCST) and the California Card Sorting Test (CCST). The California Card Sorting Test, unlike the modified WCST, provides several different measures of concept generation, concept identification and concept execution. Compared with controls, depressed patients did not show any deficits on all the measures of the modified WCST. In contrast, depressed patients evidenced mild impairment on the CCST with a specific deficit on concept generation but no major problems in concept identification and concept execution. The deficit in concept generation may be rooted in multiple factors such as hypothesis-testing deficits, a loss of cognitive flexibility and a conservative style of response. Since a positive relation between problem-solving deficits and the mean duration of the depressive episode was observed, problem-solving abilities might be predictive of poorer outcome in patients with unipolar affective disorders.

Adult↗

PubMed2001

A three-outcome design for phase II clinical trials.

The goal of a phase II trial is to make a preliminary determination regarding the activity and tolerability of a new treatment and thus to determine whether the treatment warrants further study in the phase III setting. Phase II clinical trials are typically designed in the hypothesis testing framework with two possible outcomes, either reject the null hypothesis H(0) or reject the alternative hypothesis H(a), based on the observed activity level. However, in cases where the observed activity is "borderline," the decision regarding the future of the agent is not as clear as the prespecified hypothesis test would indicate. In this paper we propose an alternative design that allows for three outcomes: reject H(0), reject H(a), or reject neither. We describe the theoretical properties of this design and illustrate it with several examples. We focus on the clinical implications of the three-outcome design. Control Clin Trials 2001;22:117-125

Clinical Trials, Phase II as Topic↗

PubMed2001

Optimum receivers for pattern recognition in the presence of Gaussian noise with unknown statistics.

We develop algorithms to detect a known pattern or a reference signal in the presence of additive, disjoint background, and multiplicative white Gaussian noise with unknown statistics. The presence of three different types of noise processes with unknown statistics presents difficulties in estimating the unknown parameters. The standard methods such as expected-maximization-type algorithms are iterative, and in the framework of hypothesis testing they are time-consuming, because corresponding to each hypothesis one must estimate a set of parameters. Other standard methods such as setting the gradient of the likelihood function with respect to the unknown parameters will lead to a nonlinear system of equations that do not have a closed-form solution and require iterative methods. We develop an approach to overcome these handicaps and derive algorithms to detect a known object. We present new methods to estimate unknown parameters within the framework of hypothesis testing. The methods that we present are direct and provide closed-form estimates of the unknown parameters. Computer simulations are used to show that for the images tested, the receivers that we have designed perform better than existing receivers.

Journal Article↗

PubMed2006

Body cooling and the diving capabilities of muskrats (Ondatra zibethicus): a test of the adaptive hypothermia hypothesis.

We tested the hypothesis that immersion hypothermia enhances the diving capabilities of adult and juvenile muskrats by reducing rates of oxygen consumption (V O2). Declines in abdominal body temperature (T(b)) comparable to those observed in nature (0.5-3.5 degrees C) were induced by pre-chilling animals in 6 degrees C water. Pre-chilling did not reduce diving V O2 of any animal tested in 10 degrees C or 30 degrees C water, irrespective of the nature of the dive. Most behavioural indices of dive performance, including average and cumulative dive times, were unaffected by T(b) reduction in adults, but depressed in hypothermic juveniles (200-400 g). Hypothermia reduced diving heart rate only on short (<25s) dives (16% reduction, P=0.01), but did not affect the temporal onset of diving bradycardia. Post-immersion V O2 was higher for pre-chilled than for normothermic muskrats, but the difference became insignificant on longer (>90 s) dives. Our findings suggest that the mild hypothermia experienced by muskrats in nature has minimal effect on diving and post-immersion metabolic costs, and thus has little impact on the dive performance of this northern semi-aquatic mammal.

Acclimatization↗

PubMed1994

Do nonclinical checkers exhibit deficits in cognitive control? Tests of an inhibitory control hypothesis.

We tested the hypothesis that persons who engage in compulsive checking may do so to compensate for cognitive errors produced by deficient inhibitory control. In two experiments, undergraduates were classified by scores on the MOCI checking subscale as checkers or noncheckers. On self-report measures, checkers were significantly more depressed, more anxious, more prone to cognitive slips, and more likely to engage in obsessive-compulsive behaviors. However, checkers performed similarly to noncheckers on laboratory tests of inhibitory control of cognition. Checkers and noncheckers were equally able to (1) ignore distractors in a selective attention task, (2) suppress inappropriate word meanings in a sentence comprehension task, and (3) inhibit retrieval of to-be-forgotten items in a memory task. These results suggest that compulsive checking does not arise from failures of inhibitory control of cognition.

Adult↗

PubMed1999

Advances in clinical trials in the twentieth century.

This article considers the rise of the randomized clinical trial during the twentieth century. Before such development could begin, probability and statistics needed to merge. Sir RA Fisher introduced randomization in the 1920s and, beginning in the 1930s and 1940s, randomized clinical trials in humans were being performed by using the statistical-hypothesis-testing paradigm. Randomization gave unbiased comparisons and a way to perform hypothesis testing without model assumptions. To preserve the benefits of randomization, a type of analysis called intent-to-treat analysis is appropriate. Needed development has occurred and is occurring in refining ethical standards, monitoring trials of serious irreversible endpoints while preserving type-I error, and instituting independent data- and safety-monitoring boards. Recent methodology has also been concerned with the appropriateness of using surrogate endpoints. A current area of debate is the appropriateness of using Bayesian statistical methods in this context.

Bayes Theorem↗