Abstracts

Rationale: Misinterpreted EEGs are a commonly recognized problem leading to unnecessary treatment and adverse outcomes. Insufficient training during neurology residency may contribute to this problem. Methods: Neurology residents rotating through the Boston VA Healthcare System used a novel computerized EEG instruction program, developed on-site and known as the Modular Real-time EEG Education Guide (MR EEG). Through this program, the residents enter their interpretations of full-length EEGs and receive immediate feedback on how closely their interpretations compare to the previously entered reports by consensus of three epileptologists. The residents completed up to 25 EEGs during their three-week rotation. The first and last five EEGs in these series were matched in abnormalities and difficulty level as decided by the three epileptologists. Correct resident classification of these EEGs as normal or abnormal was compared between these two sets of five EEGs and used as a direct outcome measure for improvement in EEG interpretation ability. These five EEGs included one normal, one with focal slowing, one with generalized slowing, one with focal epileptiform discharges, and one with generalized epileptiform discharges. In addition, an EEG knowledge based multiple choice pre and post-test were administered to these residents. With a different cohort of residents, we have previously shown significant improvements in similar EEG knowledge based tests following usage of the MR EEG program. Results: In total, 29 residents began the program of which 24 completed the pre and post-tests (11 PGY-3 and 13 PGY-4) and 15 completed all 25 EEGs (7 PGY-3 and 8 PGY-4). The average resident score on the knowledge test improved from 57.2% to 73.1% (p < 0.001). There was a significant difference in the baseline pre-test comparing the PGY-3 (43.6%) to PGY-4 (68.7%) residents although both classes showed significant improvements when evaluated independently. Comparing the subject matched EEG 1-5 and 21-25, there was a non-significant trend toward improvement within the PGY-3 class from 60% of EEGs correctly classified as normal or abnormal to 74.4% (p=0.13). This was not seen when evaluating the PGY-4 class (75.9% to 75%, p=0.91). Conclusions: This is the first study, to our knowledge, to directly assess improvement in residents' ability to interpret EEG following a teaching intervention. We illustrate that this computerized independent study program may be effective in improving real-world outcomes for residents who have not previously had significant exposure to EEG (PGY-3). The PGY-4 class had previously been exposed to a teaching set of different EEGs with similar topics using MR EEG during their PGY-3 year. Their higher baseline scores may indicate a retention of knowledge from the previous year's EEG teaching or benefits from other clinical training. Overall, we believe that MR EEG, or a similar independent study module with computerized feedback may be a useful adjunct to standardized EEG teaching. Funding: This was an unfunded study.