Assessment of the validity of masking in electrodiagnostic research.

OBJECTIVES: Performed by thousands of specialists and researched for >60 yrs, electrodiagnostic testing rates poorly in evidence-based guidelines in part because there are no valid masked research studies. This study assesses whether the electrodiagnostic examination can be adequately masked.

DESIGN: Prospective, controlled, masked trial. A total of 150 persons aged 55-80 yrs, thought clinically to be asymptomatic (n = 32), with low back pain (n = 46), or with spinal stenosis (n = 63), underwent needle electromyography and nerve conduction studies by specialists in physical medicine and rehabilitation who were not permitted to ask any clinical information or to perform a formal examination. Examiners rated potential bias and recorded clues that might bias the examination.

RESULTS: Examiners noted potential for bias in conversation in 17 and due to pain behavior in one. Half of these potential unmaskings pointed away from the correct clinical diagnosis, leaving potential for bias in <6% of the subjects. The clinical diagnosis had no relationship to electrodiagnostician rating of pain behavior in response to testing or repositioning, to limb atrophy, and to arthritis (P > 0.05 all instances).

CONCLUSIONS: The electrodiagnostic examination can be successfully masked. This information will help future research in electrodiagnostics meet the criteria of evidence-based medicine.