Assessing electrodiagnostic skills among residents and fellows: Relationships between workplace-based assessments using the Electromyography Direct Observation Tool and other measures of trainee performance

Introduction/Aims: Graduate medical education programs must ensure residents and fellows acquire skills needed for independent practice. Workplace-based observational assessments are informative but can be time- and resource-intensive. In this study we sought to gather “relations-to-other-variables” validity evidence for scores generated by the Electromyography Direct Observation Tool (EMG-DOT) to inform its use as a measure of electrodiagnostic skill acquisition. Methods: Scores on multiple assessments were compiled by trainees during Clinical Neurophysiology and Electromyography rotations at a large US academic medical center. Relationships between workplace-based EMG-DOT scores (n = 298) and scores on a prerequisite simulated patient exercise, patient experience surveys (n = 199), end-of-rotation evaluations (n = 301), and an American Association of Neuromuscular & Electrodiagnostic Medicine (AANEM) self-assessment examination were assessed using Pearson correlations. Results: Among 23 trainees, EMG-DOT scores assigned by physician raters correlated positively with end-of-rotation evaluations (r = 0.63, P =.001), but EMG-DOT scores assigned by technician raters did not (r = 0.10, P =.663). When physician and technician ratings were combined, higher EMG-DOT scores correlated with better patient experience survey scores (r = 0.42, P =.047), but not with simulated patient or AANEM self-assessment examination scores. Discussion: End-of-rotation evaluations can provide valid assessments of trainee performance when completed by individuals with ample opportunities to directly observe trainees. Inclusion of observational assessments by technicians and patients provides a more comprehensive view of trainee performance. Workplace- and classroom-based assessments provide complementary information about trainee performance, reflecting underlying differences in types of skills measured.