Evaluation of posturography in the detection of malingering subjects

This study aimed to test the performance of proposed methods for detecting malingering subjects on computerized dynamic posturography using one subject group in three situations (normal, malingering, vestibular weakness).The study design was a prospective, blinded study.The study was conducted at a university hospital.Volunteer subjects aged 20-59 years of age participated.Computerized dynamic posturography was performed under three situations: best effort, faking vestibular weakness, and transient induced vestibular weakness with bilateral simultaneous caloric irrigation.Measured was identification of situation (normal, malingering, induced vestibular weakness) by each of three detection methods: blinded clinical scoring, a set of formulae, and a set of variables (the latter two methods proposed previously by other investigators).Each method performed well. In three-way discrimination, the formulae and clinical scoring each correctly identified approximately 75% of trials. In two-way discrimination (malingering vs. induced vestibular weakness), the best combination of variables slightly outperformed clinical scoring (0.93 vs. 0.88 ROC [receiver operating characteristic] curve area).Computerized dynamic posturography can distinguish malingering in normal subjects from trials performed with best effort or after binaural simultaneous caloric irrigation. The accuracy of blinded clinical scoring was comparable to that of two objective detection methods.