6533b7d7fe1ef96bd12678cb

RESEARCH PRODUCT

III: „Relatives Risiko” und „NNT” - anschauliche Maße für binäre Daten

Frank Krummenauer

subject

Contingency tableOphthalmologyVariable (computer science)Relative valuebusiness.industryRelative riskBinary dataStatisticsNumber needed to treatMedicineRisk factorbusinessNumero sign

description

Description of categorial data can often be based on contingency tables. However, percentages appearing in such tables must be meaningful: For most applications, it may be useful to employ factors of causal influence as the row entry variable and relate percentages to sub-groups defined by these row entries ("row percentages"). The comparison of success frequencies (i. e. binary information on "therapy success yes/no") may be based on two therapies, respective success frequencies and their ratio, the relative risk. In addition the success frequencies' difference, the absolute (or excess) risk, can be transformed into the "number needed to treat" (NNT). Many international journals demand this surrogate measure for the short summary of clinical and epidemiological trial reports, since it allows a direct comparison of the results at hand with the results of previous trials. If, for example, prophylactic vaccination strategies are evaluated, the relative risk estimates the order of risk reduction by vaccination, whereas the NNT estimates the number of patients needed to treat or vaccination needed to apply to avoid disease occurrence.

https://doi.org/10.1055/s-2002-35691