Drop the Likelihood Ratio

Summary

Objective: To present a geometric ruler to determine predictive values of binary diagnostic test results from prevalence, sensitivity and specificity without calculation.

Methods: On logarithmic scale the relationships between pre-/post-test odds and the likelihood ratio as well as sensitivity, specificity and the likelihood ratio appear as simple linear (additive) equations.

Results: Each of these additive equations can be solved geometrically in form of a ruler with three scales. By amalgamation we devise a novel non-electronic tool which omits the intermediate step of likelihood ratio determination.

Conclusions: We propose a simple geometric method to aid in interpretation of diagnostic test results for both practical and educational purposes.

Keywords

Bayes’ rule - Caraguel-Vanderstichel nomo-gram - ruler - diagnostic test

• References

• 1 Fagan TJ. Letter: Nomogram for Bayes theorem. N Engl J Med 1975; 293: 257.
• 2 Hellmich M, Lehmacher W. A ruler for interpreting diagnostic test results. Methods Inf Med 2005; 44: 124-126.
• 3 Caraguel CG, Vanderstichel R. The two-step Fagan’s nomogram: ad hoc interpretation of a diagnostic test result without calculation. Evid Based Med 2013; 18: 125-128.
• 4 Deeks J, Altman D. Diagnostic tests 4: Likelihood ratios. Br Med J 2004; 329: 168-190.
• 5 Warner E, Messersmith H, Causer P. et al. Systematic review: using magnetic resonance imaging to screen woman at high risk for breast cancer. Ann Intern Med 2008; 148: 671-679.
• 6 Benish WA. The use of information graphs to evaluate and compare diagnostic tests. Methods Inf Med 2002; 41 (02) 114-118.
• 7 Hughes G. Information graphs for epidemiological applications of the Kullback-Leibler divergence. Methods Inf Med 2014; 53 (01) IV-VI.
• 8 Ughi E, Jassó J. Physical and virtual worlds in teaching mathematics: possibilities for an effective cooperation?. Proceedings of the Fourth Congress of the European Society for Research in Mathematics Education (CERME 4, 2005). Available at. http://www.mathematik.uni-dortmund.de/~erme/CERME4/index.php.
• 9 Carley S, Dosman S, Jones SR, Harrison M. Simple nomograms to calculate sample size in diagnostic studies. EMJ 2005; 22 (03) 180-181.