Analyse-it includes ROC curve analysis (CLSI GP10), ROC curve plots, Kappa and Weighted
Kappa (CLSI EP12) to help you determine the diagnostic ability of qualitative and
quantitative diagnostic methods. ROC curve analysis lets you compare methods to
find the best performing test, choose medical decision levels, or compare diagnostic
performance of a new against existing method.
Visualise test performance with ROC curve plot
Plot upto 6 ROC curves to see the diagnostic performance of upto 6 test methods.
From the plot you can see where tests are strongest, and see the medical decision
points that best classify diseased and undiseased subjects.
Describe performance over all decision levels
Calculate true/false-positive & true/false-negative rates, predictive values, likelihood
ratios and cost at each medical decision level.
Choose optimum medical decision levels with decision plots
Decision plots visualise diagnostic ability across all decision levels in terms
of sensitivity & specificity, positive (LR+) & negative (LR-) likelihood
ratios, positive & negative predictive values, or cost (in financial or health
terms).
Compare upto 6 correlated diagnostic tests
DeLong, Delong, Clarke-Pearson is used to compare ROC curves -- the best method
for comparing correlated ROC curves, a less parametric approach than the early 1980's
Hanley & McNeil approach -- in fact, the method now recommend by Hanley &
McNeil themselves. Provides statistical evidence of the best overall test.
Incorporate prevalence and costs
If available, include prevalance of the condition, or costs (in financial or health
terms) of misdiagnosis for more accurate performance evaluation at medical decision
levels.
Describe the performance of a qualitative test
Kappa and Weighted Kappa are included for comparing qualitative agreements between
two judges/raters. Calculates sensitivity/specificity, likelihood ratios, and predictive
values. Equivalent to CLSI EP12-A protocol.
SSL
Live customer support