Dean McKenzie1,2, Christopher Thomas1. 1. Epworth HealthCare, Melbourne, Australia. 2. Department of Statistics, Data Science and Epidemiology, Swinburne University of Technology, Melbourne, Australia.
Abstract
AIM: Relative risks and odds ratios are widely reported in the medical literature, but the latter can be very difficult to understand. We sought to further clarify these important indices. METHODS: We defined both relative risks and odds ratios, then looked at the types of study for which each statistic is suited. We illustrated calculation of relative risks and odds ratios through analysis of tabled data from a recent published longitudinal study, using a 2x2 table, bar charts and R, the open source statistical programming language. Simple rules for when and how to use relative risks and odds ratios are presented. CONCLUSION: Understanding the difference between relative risks and odds ratios and when and how to use them may aid clinical interpretation, dissemination and translation of research findings. This article is protected by copyright. All rights reserved.
AIM: Relative risks and odds ratios are widely reported in the medical literature, but the latter can be very difficult to understand. We sought to further clarify these important indices. METHODS: We defined both relative risks and odds ratios, then looked at the types of study for which each statistic is suited. We illustrated calculation of relative risks and odds ratios through analysis of tabled data from a recent published longitudinal study, using a 2x2 table, bar charts and R, the open source statistical programming language. Simple rules for when and how to use relative risks and odds ratios are presented. CONCLUSION: Understanding the difference between relative risks and odds ratios and when and how to use them may aid clinical interpretation, dissemination and translation of research findings. This article is protected by copyright. All rights reserved.
Keywords:
clinical research; communication; health promotion and prevention; research methods; statistics