ISSN : 2146-3123
E-ISSN : 2146-3131

Assessment of Relative Risk Measurement Comparing with Odds Ratio, Attributable Risk and Number Needed to Treat
Necdet Süt 1, Mustafa Şenocak 2
1Biostatistics Editor, Balkan Medical Journal Department of Biostatistics and Medical Informatics, Trakya University Faculty of Medicine, Edirne, Turkey
2İstanbul Üniversitesi Cerrahpaşa Tıp Fakültesi Biyoistatistik Anabilim Dalı, İstanbul
Pages : 213-221


Objectives: It was aimed to clarify the theoretical aspects of relative risk (RR) and other risk measurements (odds ratio “OR”, attributable risk “AR” and number needed to treat “NNT”), to examine the associations between RR and others, to refer the limitations of RR in medical decision making and to put forward some solution suggestions.

Study Design: All possible RR, OR, AR, NNT values and their confidence intervals were calculated in a sample of 200 subjects. In addition, an example practice was performed based on real clinical data.

Results: The value of RR was the same in different clinical conditions. The change in absolute risk measurements such as AR and NNT was parallel with clinical change rate.

Conclusion: Interpretation of RR causes problem since it can take the same value in different clinical conditions. Therefore, in addition to RR, considering the AR measurement which is parallel with clinical change rate, may be useful while interpreting the RR values.

Keywords : Controlled clinical trials; odds ratio; risk assesment
Viewed : 3061
Downloaded : 827