Covariate study designs

The cohort study offers two additional strategies for analysis — case-cohort and nested case-control study designs the case-cohort comprises all individuals in the cohort with a particular disease and a random sample of unaffected individuals from the cohort the nested case-control design includes cohort members with a particular disease (cases) and unaffected individuals (controls) matched on relevant study covariates ... 

Realistic predichons of study results based on simulations can be made only with realistic simulation models. Three types of models are necessary to mimic real study observations system (drug-disease) models, covariate distribution models, and study execution models. Often, these models can be developed from previous data sets or obtained from literature on compounds with similar indications or mechanisms of action. To closely mimic the case of intended studies for which simulations are performed, the values of the model parameters (both structural and statistical elements) and the design used in the simulation of a proposed trial may be different from those that were originally derived from an analysis of previous data or other literature. Therefore, before using models, their appropriateness as simulation tools must be evaluated to ensure that they capture observed data reasonably well [19-21]. However, in some circumstances, it is not feasible to develop simulation models from prior data or by extrapolation from similar dmgs. In these circumstances, what-if scenarios or sensitivity analyses can be performed to evaluate the impact of the model uncertainty and the study design on the trial outcome [22, 23]. 

Such a step is not required from a statistical point of view, as a result of the role of a properly executed randomization process, but it can be comforting if it supports the clinical relevance of the effect after adjustment for the baseline covariate. If there are specific explanatory factors that are suspected of having an effect on the outcome of interest at the start of a study, it is advisable to incorporate them into the overall study design (for example, through stratified randomization). A brief discussion of this topic has been published by Roberts and Torgerson (1999). The EMEA CPMP has also published a guidance document on baseline covariates (EMEA CPMP 2003). 

Documentation on assumptions should address those assumptions implicit in the pharmacokinetic or pharmacodynamic model and the statistical methodology chosen to evaluate the data. It should also state the assumed sensitivity of the parameters required to define the model relative to the data space being evaluated as well as any preconceived notions regarding biomarkers or surrogate markers evaluated as responses or covariates in the analysis. Hypotheses should be defined based on what was held a priori as true before the study or analysis, what was developed from preliminary or exploratory data analysis, and what would constitute a difference or equivalence in an effect or outcome, hi some instances the criteria for difference as opposed to equivalence can be defined from a statistical viewpoint independent of the actual study design. This approach does not always confer regulatory acceptance, however. 

A covariance study by Li et al. [76] dealt with the comparison of various sampling schemes and accumulation profiles with respect to their usability with covariance processing. The authors divided sparse sampling schemes into the non-uniform and the ti cut-off (CUO) schemes. While common NUS schemes employ—exponentially—increasing A(i spacing as a function of (i, a cut-off design implies the sole acquisition of signals at short <1 times up to a maximum value. The 2D NMR spectrum is then reconstructed with appropriate methods. Various standard, NUS and... 

In study design, the problems of small numbers of subjects and large numbers of potential covariates indicate the need for realistic power estimates, in order to ensure that the numbers are sufficient to enable a deficit to be detected. 

The overall evidence from studies in animals supports the observations of lead neurobehavioral effects in humans. As pointed out by Cory-Slechta (1995), studies in animals have provided a direct measurement of the behavioral process per se, and have done so in the absence of the covariates (e.g., socioeconomic status, parental IQ) known to affect IQ scores in human studies. It is also worth noting that animal studies, in which the experimental design is carefully controlled, have shown that the timing of exposure is crucial, that different neurobehavioral outcomes are affected differently (different thresholds), and that some behavioral alterations last longer than others. 

Note The interpretation of the matrix as the covariance matrix of the errors in x has important applications. The value of any estimate is greatly enhanced if its accuracy is known. is also very useful in initial design and development, as it can be calculated before the estimator is implemented. can be used to study measurement placement and what type and accuracy of information is actually needed. A... 

Such studies can be designed as historic (identification of a pre-existing cohort whose exposure has been identified) or prospective (current exposure ascertainment) with prospective follow-up for occurrence of health events Able to consider multiple exposures and outcomes within single design Can incorporate time-varying covariates ... 

The neurobehavioral test battery used in the 66-month Seychelles study was designed to assess multiple developmental domains (Davidson et al. 1998). The tests were considered to be sufficiently sensitive and accurate to detect neurotoxicity in the presence of a number of statistical covariates. On-site test administration reliability was assessed by an independent scorer, and mean interclass correlations for interscorer reliability were 0.96-0.97 (Davidson et al. 1998). The sample size was determined to be sufficient to detect a 5.7-point difference on any test with a mean (SD) of 100 (16) between low (0-3 ppm) and high (>12 ppm) hair mercury concentration groups for a 2-sided test (A = 0.05 at 80% power). 

---