Version 2018-09-17, ruwen.boehm@pharmakologie.uni-kiel.de
The calculations of the pharmacovigilance measurements of disproportionality have been successfully validated against R/epiR.
A free online statistical calculator for a contingency table that has 2 rows and 2 columns for the two categorial data to be analyzed. This software is especially designed for pharmacovigilance data but might be useful for other data as well.
Capabilities:
rates,
Chi-squared values corrected as recommended by Yates (chisq),
basic statistical measurements of disproportionality for like RRR, ROR, PRR and diagnostic statistic's measurements like Likelihood Ratios (LR+,LR-) and Pos/Neg Predictive Values
optional correction for real-world background,
(e.g., prevalence data of drug usage or event/disease occurrence in the common population),under-reportingand confounders/tied values for number of records.
These measurements of disproportionality are calculated as RRR=DE*N/(D*E), PRR=(DE/D)/(dE/d) and ROR=DE*de/(De*dE). A value of 1 is considered normal background noise. The confidence interval can be estimated using Chi-squared with Yates' correction chisq > 4 or lower bound of the 95% confidence interval (CI) of RRR, PRR or ROR, e.g. for s = sqr( 1/DE + 1/De + 1/dE + 1/de ) for ROR with CI = e ^ ( ln ROR ± 1,96s ).