The catastrophic misinterpretation model [Behav. Res. Ther. 24 (1986) 461-470] proposes that panic attacks result from misinterpretation of interoceptive stimuli as precursors to physical or psychological emergency. Inconclusive evidence for the model may be partly explained by limitations of the questionnaires developed to measure catastrophic misinterpretation. For example, the Body Sensations Interpretation Questionnaire (BSIQ) is unable to clarify whether anxiety-related interpretations of ambiguous interoceptive stimuli represent catastrophic misinterpretations or responses masking feared outcomes (e.g., heart failure). Additionally, it lacks items relating to several DSM-IV criteria for panic, thereby limiting content validity. Reliability is also potentially compromised due to experimenter-coding of participant-generated responses. A modified form of the BSIQ was developed to address these limitations and evaluated with non-anxious controls (n=34) and people with panic disorder (n=38). The revised questionnaire demonstrated good to excellent internal consistency, inter-rater reliability, and construct validity and is a useful development of the BSIQ.