TEST FOR TREATMENT EFFECT BASED ON BINARY DATA WITH RANDOM SAMPLE SIZES

1990 ◽  
Vol 32 (1) ◽  
pp. 53-70 ◽  
Author(s):  
Jun Shao ◽  
Shein-Chung Chow
Keyword(s):  
Random Sample ◽  
Treatment Effect ◽  
Binary Data ◽  
Sample Sizes

scholarly journals Optimal Allocation of Interviews to Baseline and Endline Surveys in Place-Based Randomized Trials and Quasi-Experiments

2018 ◽  
Vol 42 (4) ◽  
pp. 391-422 ◽  
Author(s):  
Donald P. Green ◽  
Winston Lin ◽  
Claudia Gerber
Keyword(s):  
Random Sample ◽  
Treatment Effect ◽  
Randomized Trials ◽  
Optimal Allocation ◽  
Average Treatment Effect ◽  
Baseline Survey ◽  
Baseline Measure ◽  
Panel Surveys ◽  
Planning Stage ◽  
Average Treatment

Background: Many place-based randomized trials and quasi-experiments use a pair of cross-section surveys, rather than panel surveys, to estimate the average treatment effect of an intervention. In these studies, a random sample of individuals in each geographic cluster is selected for a baseline (preintervention) survey, and an independent random sample is selected for an endline (postintervention) survey. Objective: This design raises the question, given a fixed budget, how should a researcher allocate resources between the baseline and endline surveys to maximize the precision of the estimated average treatment effect? Results: We formalize this allocation problem and show that although the optimal share of interviews allocated to the baseline survey is always less than one-half, it is an increasing function of the total number of interviews per cluster, the cluster-level correlation between the baseline measure and the endline outcome, and the intracluster correlation coefficient. An example using multicountry survey data from Africa illustrates how the optimal allocation formulas can be combined with data to inform decisions at the planning stage. Another example uses data from a digital political advertising experiment in Texas to explore how precision would have varied with alternative allocations.

scholarly journals Approaches for Structural Investigations of Binary Data Using Confirmatory Factor Models

2018 ◽  
Vol 7 (6) ◽  
pp. 68
Author(s):  
Karl Schweizer ◽  
Siegbert Reiß ◽  
Stefan Troche
Keyword(s):  
Sample Size ◽  
Binary Data ◽  
Factor Models ◽  
The Other ◽  
Simulation Studies ◽  
Sample Sizes ◽  
Structural Investigations ◽  
Ml Estimation ◽  
Confirmatory Factor ◽  
The Relationship

An investigation of the suitability of threshold-based and threshold-free approaches for structural investigations of binary data is reported. Both approaches implicitly establish a relationship between binary data following the binomial distribution on one hand and continuous random variables assuming a normal distribution on the other hand. In two simulation studies we investigated: whether the fit results confirm the establishment of such a relationship, whether the differences between correct and incorrect models are retained and to what degree the sample size influences the results. Both approaches proved to establish the relationship. Using the threshold-free approach it was achieved by customary ML estimation whereas robust ML estimation was necessary in the threshold-based approach. Discrimination between correct and incorrect models was observed for both approaches. Larger CFI differences were found for the threshold-free approach than for the threshold-based approach. Dependency on sample size characterized the threshold-based approach but not the threshold-free approach. The threshold-based approach tended to perform better in large sample sizes, while the threshold-free approach performed better in smaller sample sizes.

scholarly journals ANOVA with random sample sizes: An application to a Brazilian database on cancer registries

2013 ◽  
Author(s):  
Célia Nunes ◽  
Gilberto Capistrano ◽  
Dário Ferreira ◽  
Sandra S. Ferreira
Keyword(s):  
Random Sample ◽  
Cancer Registries ◽  
Sample Sizes

Experience of implementing a novel random sampling BICR audit for investigator (INV)-assessed progression-free survival (PFS) in the PALOMA-3 trial.

2017 ◽  
Vol 35 (15_suppl) ◽  
pp. 1058-1058
Author(s):  
Xin Huang ◽  
Ke Zhang ◽  
Nicholas C. Turner ◽  
Cynthia Huang Bartlett ◽  
Carla Giorgetti ◽  
...  
Keyword(s):  
Random Sample ◽  
Primary Endpoint ◽  
Treatment Effect ◽  
Random Sampling ◽  
Randomized Clinical Trials ◽  
Metastatic Breast ◽  
Progression Free Survival ◽  
Primary Objective ◽  
Third Party ◽  
Double Blind

1058 Background: PFS has frequently been used as a primary endpoint for evaluating efficacy of anticancer therapies in randomized clinical trials. Given high correlation between INV and independent (BICR) assessments with respect to the relative treatment effect, a pre-planned BICR audit of INV progression assessment in a random subgroup of patients (pts) instead of a BICR review of all progression assessments can be an acceptable approach to verify the INV assessments and to evaluate the potential bias in INV PFS results. Methods: PALOMA-3 was a randomized, double blind, placebo (PCB) controlled, Ph 3 study with the primary objective of demonstrating the superiority of palbociclib (PAL) + fulvestrant (F) over PCB + F in women with HR+, HER2- metastatic breast cancer (MBC). The primary endpoint was INV assessed PFS. BICR assessment of PFS was performed with the use of a novel audit approach involving a random sample–based BICR to verify if the INV assessed PFS was accurate. A third-party core imaging laboratory performed the blinded review for a randomly selected subgroup of pts (~40%). NIH and PhRMA methods were used to evaluate the potential for bias in the INV PFS results. Results: PAL + F improved PFS in patients with HR+, HER2- MBC. The observed INV HR was 0.46 (95% CI: 0.36, 0.59; stratified 1-sided p < 0.0001) in favor of PAL + F. The median PFS was 9.5 mo (95% CI: 9.2, 11.0) in the PAL + F arm and 4.6 mo (95% CI: 3.5, 5.6) in the PCB + F arm (Lancet Oncol. 2016; 17: 425–39). The estimated HR of the complete BICR data incorporating the information from the complete INV assessed PFS and the random sample audited BICR subgroup was 0.33 with the upper bound of the 1-sided 95% CI of 0.47. The results confirmed the INV assessed treatment effect and detected no INV bias in favor of PAL + F. Conclusions: PALOMA-3 is the first registrational trial to use a BICR audit and has received positive reviews from regulatory agencies. The experience of implementing the random sampling BICR audit in PALOMA-3 demonstrates that this approach can be used for randomized, double blind oncology trials with solid tumors where INV assessed PFS is the primary endpoint and a large treatment effect is targeted. Sponsor: Pfizer. Clinical trial information: NCT01942135.

Bimonthly assessment of magnetization transfer magnetic resonance imaging parameters in multiple sclerosis: a 14-month, multicentre, follow-up study

2010 ◽  
Vol 16 (3) ◽  
pp. 325-331 ◽  
Author(s):  
S. Mesaros ◽  
MA Rocca ◽  
MP Sormani ◽  
P. Valsasina ◽  
C. Markowitz ◽  
...  
Keyword(s):  
Multiple Sclerosis ◽  
Treatment Effect ◽  
Repeated Measures ◽  
Magnetization Transfer ◽  
Random Effect ◽  
Lesion Volume ◽  
Sample Sizes ◽  
Double Blind ◽  
Over Time

This study was performed to assess the temporal evolution of damage within lesions and the normal-appearing white matter, measured using frequent magnetization transfer (MT) MRI, in relapsing—remitting multiple sclerosis (RRMS). The relationship of MT ratio (MTR) changes with measures of lesion burden, and the sample sizes needed to demonstrate a treatment effect on MTR metrics in placebo-controlled MS trials were also investigated. Bimonthly brain conventional and MT MRI scans were acquired from 42 patients with RRMS enrolled in the placebo arm of a 14-month, double-blind trial. Longitudinal MRI changes were evaluated using a random effect linear model accounting for repeated measures, and adjusted for centre effects. The Expanded Disability Status Scale (EDSS) score remained stable over the study period. A weak, but not statistically significant, decrease over time was detected for normal-appearing brain tissue (NABT) average MTR (—0.02% per visit; p = 0.14), and MTR peak height (—0.15 per visit; p = 0.17), while average lesion MTR showed a significant decrease over the study period (—0.07% per visit; p = 0.03). At each visit, all MTR variables were significantly correlated with T2 lesion volume (LV) (average coefficients of correlation ranging from —0.54 to —0.28, and p-values from <0.001 to 0.02). At each visit, NABT average MTR was also significantly correlated with T1-hypointense LV (average coefficient of correlation = —0.57, p < 0.001). The estimation of the sample sizes required to demonstrate a reduction of average lesion MTR (the only parameter with a significant decrease over the follow-up) ranged from 101 to 154 patients to detect a treatment effect of 50% in a 1-year trial with a power of 90%. The steady correlation observed between conventional and MT MRI measures over time supports the hypothesis of axonal degeneration of fibres passing through focal lesions as one of the factors contributing to the overall MS burden.

scholarly journals Synthetic observations from deep generative models and binary omics data with limited sample size

2020 ◽  
Author(s):  
Jens Nußberger ◽  
Frederic Boesel ◽  
Stefan Lenz ◽  
Harald Binder ◽  
Moritz Hess
Keyword(s):  
Sample Size ◽  
Binary Data ◽  
Generative Models ◽  
Small Sample ◽  
Generative Adversarial Networks ◽  
Omics Data ◽  
Sample Sizes ◽  
Odds Ratios ◽  
Limited Sample ◽  
Limited Sample Size

AbstractDeep generative models can be trained to represent the joint distribution of data, such as measurements of single nucleotide polymorphisms (SNPs) from several individuals. Subsequently, synthetic observations are obtained by drawing from this distribution. This has been shown to be useful for several tasks, such as removal of noise, imputation, for better understanding underlying patterns, or even exchanging data under privacy constraints. Yet, it is still unclear how well these approaches work with limited sample size. We investigate such settings specifically for binary data, e.g., as relevant when considering SNP measurements, and evaluate three frequently employed generative modeling approaches, variational autoencoders (VAEs), deep Boltzmann machines (DBMs) and generative adversarial networks (GANs). This includes conditional approaches, such as when considering gene expression conditional on SNPs. Recovery of pair-wise odds ratios is considered as a primary performance criterion. For simulated as well as real SNP data, we observe that DBMs generally can recover structure for up to 100 variables with as little as 500 observations, with a tendency of over-estimating odds ratios when not carefully tuned. VAEs generally get the direction and relative strength of pairwise relations right, yet with considerable under-estimation of odds ratios. GANs provide stable results only with larger sample sizes and strong pair-wise relations in the data. Taken together, DBMs and VAEs (in contrast to GANs) appear to be well suited for binary omics data, even at rather small sample sizes. This opens the way for many potential applications where synthetic observations from omics data might be useful.

Sign in / Sign up

Export Citation Format

Share Document