Quantification of errors in ordinal outcome scales using shannon entropy: effect on sample size calculations

• Published in: PloS one 2013-07, Vol.8 (7), p.e67754-e67754
• Main Authors: Mandava, Pitchaiah, Krumpelman, Chase S, Shah, Jharna N, White, Donna L, Kent, Thomas A
• Format: Article
• Language: English
• Subjects: Assessments; Bias; Clinical trials; Entropy (Information theory); Errors; Humans; Identification methods; Information theory; Information transfer; Laboratories; Mathematical models; Mathematics; Medical research; Medicine; Models, Statistical; Neurology; Neuroprotective agents; Outcome Assessment, Health Care - standards; Patients; Physics; Randomization; Randomized Controlled Trials as Topic - standards; Sample Size; Size estimation; Social and Behavioral Sciences; Stroke; Stroke - epidemiology; Stroke - therapy; Studies; Transmitters; Uncertainty; Veterans
• ISSN: 1932-6203; 1932-6203
• DOI: 10.1371/journal.pone.0067754

Clinical trial outcomes often involve an ordinal scale of subjective functional assessments but the optimal way to quantify results is not clear. In stroke, the most commonly used scale, the modified Rankin Score (mRS), a range of scores ("Shift") is proposed as superior to dichotomization because of greater information transfer. The influence of known uncertainties in mRS assessment has not been quantified. We hypothesized that errors caused by uncertainties could be quantified by applying information theory. Using Shannon's model, we quantified errors of the "Shift" compared to dichotomized outcomes using published distributions of mRS uncertainties and applied this model to clinical trials. We identified 35 randomized stroke trials that met inclusion criteria. Each trial's mRS distribution was multiplied with the noise distribution from published mRS inter-rater variability to generate an error percentage for "shift" and dichotomized cut-points. For the SAINT I neuroprotectant trial, considered positive by "shift" mRS while the larger follow-up SAINT II trial was negative, we recalculated sample size required if classification uncertainty was taken into account. Considering the full mRS range, error rate was 26.1%±5.31 (Mean±SD). Error rates were lower for all dichotomizations tested using cut-points (e.g. mRS 1; 6.8%±2.89; overall p