Estimates of table cell frequencies (raw data) from effect size measures
This tool provides estimates of table cell frequencies (raw data) based on effect size measures and on their 95% confidence intervals. Such data is frequently needed in the context of meta-anlaysis. The following notation is used for result presentation:
a: Number of exposed/treated with outcome/event
b: Number of exposed/treated with no outcome/event
c: Number of unexposed/control with outcome/event
d: Number of unexposed/control with no outcome/event
Input values
Number of events
Results
Estimates of raw data
Each row corresponds to a possible solution (set of values meeting the criteria of the input values). If no values are provided, please recheck your input values.
Notes
Please note that when there is no information on the number of events/participants developing the outcome of interest, a range of values (rather than a single value) is provided for each cell. This reflects uncertainty with effect size measures - for example, both a risk ratio of 1.45 and a risk ratio of 1.54 may be simply reported as '1.5'.
When there is no information on the number of events, two result sets (solutions) are presented when estimating raw data from risk differences and odds ratio, as such estimations involve applying the quadratic formula. Please choose the result set/solution that more probably applies to your data.
Risk differences and the respective confidence interval bounds MUST NOT be entered as percentages, but rather as proportions.
This interactive tool was developed by Bernardo Sousa-Pinto (Department of Community Medicine, Information and Health Decision Sciences; Faculty of Medicine, University of Porto, Porto, Portugal) and Carlo Di Pietrantonj (Regional Epidemiology Unit SeREMI, Local Health Unit Alessandria- ASL AL, Alessandria, Italy). For more information, please check the
estimraw R package
. Contact: bernardo@med.up.pt