Clinical sign profile likelihood ratios for bovine spongiform encephalopathy suspects.

This study has described a method for generating the probability of a bovine spongiform encephalopathy (BSE) suspect being a true positive BSE case. A weighting equivalent to the clinical sign frequencies recorded in histopathologically confirmed BSE cases was assigned to 14 clinical observations and a clinical profile score for the case was generated by the summation of the weightings. This method was applied to 50 histopathologically confirmed (true positive cases) BSE suspects and 50 histopathologically unconfirmed (false positive cases) BSE suspects. The profile scores for the true positive BSE suspect cases were statistically significantly higher than the profile scores of the false positive BSE suspect cases (P = 0.0014) using a Mann-Whitney U non-parametric test. The mean profile score for the true positive cases was 944 and for the false positive scores was 879. Likelihood ratios for BSE suspects with different clinical profile scores were computed using different clinical profile score cut off point and ranges. A BSE suspect with profile score of 727 or above (the lowest cut off point) and 1037 or above (the highest cut off point) had likelihood ratios for being BSE of 1. 09 and 4.00 respectively. The likelihood ratios for a BSE suspect being a true positive BSE case in the profile score ranges 627-866, 867-966, 967-1036 and 1037-1067 were 0.35, 1.06, 1.30 and 4.0 respectively. Further investigations or a revisit may be justified in animals with a low probability of being BSE. Copyright 2000 Harcourt Publishers Ltd.