Please try the following tools, generated by NEWMEDS:
www.DupCheck.org is a web-based tool to screen for duplicate patients in clinical trials within and across studies, sponsors and therapeutic areas. (Workpackage 10)
DupCheck is a simple tool that can improve compliance, reduce risks of misattributed safety signals and improve efficacy signals in clinical trials. DupCheck does this by screening out patients who are concurrently enrolled in another trial. DupCheck screens across sponsors and indications. Its objective is to help identify patients globally across sponsors and indications who are participating in a concurrent clinical trial or who have recently participated in another trial. These patients would thus not be eligible to participate in a trial and their participation could severely bias study results. In many therapeutic areas even a small number of duplicate patients can lead to a negative or failed trial. In addition enrolling duplicate patients can result in a misattributed serious adverse event.
Pharmacological Imaging and Pattern Recognition Toolbox:
Toolbox for the analysis of brain images for a better classification in the context of drug development (Workpackage 06):
This is a toolbox for the analysis of brain images for a better classification in the context of drug development. The toolbox reads brain images from a number of imaging modalities (fMRI, Structural MRI, PET, Arterial Spin Labeling) in standard formats (Analyze and NIFTI). It then applies a number of Machine Learning algorithms (Support Vector Machines, Gaussian Process Classification, Ordinal Regression) in a cross-validated fashion to achieve predictions at the individual level. The output consists of individual probabilistic predictions, classification accuracy and confusion matrices as well as an illustrative pattern of the brain regions driving the discrimination. This toolbox can be used for a wide variety of applications including drug discrimination, validation of experimental medicine models and patient stratification.
Presentations on this Toolbox can be downloaded here:
Clinical significance calculator for biomarkers of depression treatment outcome (Workpackage 08):
Conventional statistical methods declare how likely a finding is to have occurred by chance (p <0.05). However, this does not automatically mean that a statistically significant finding is also clinically significant. There is no single rule for clinical significance as it depends on the nature of the finding, the phase of illness and the context in which it is applied. To provide such a context, “The clinical significance calculator” is a simple tool to estimate whether a biomarker which predicts the outcome of depression treatment is likely to be clinically significant. Biomarkers are measures taken from blood or other biological material or physiological measurement. Biomarkers include genetic variations, protein levels, and measures of brain activity among others. It is hoped that biomarkers can help predict outcomes in depression and select treatments that are more likely to work for a given individual. This is called personalized medicine. However, for a biomarker to be clinically meaningful it has to predict a difference in outcomes that is not just reliable (which is what statistical significance ensures) but is also large enough to be clinically useful. Using the NICE guidance index of what is clinically significant, i.e. a 3 point improvement on a HAMD score (or its equivalent) this calculator computes the clinical significance to assess whether the effect for a given biomarker is equivalent.