Solving the dilemma of EEG misinterpretation.

“How Not To Read an EEG,” supplement to this week’s issue ofNeurology®, brings to daylight a “dirty little secret.” Although misinterpretation is concerning with any diagnostic test, EEG misinterpretation seems particularly common. This may ultimately stem from the essence of EEG reading, a pattern recognition skill that cannot be acquired from textbooks, articles, supplements, or guidelines, a process far more enigmatic than appreciating a lesion on anMRI study. To a well-trained electroencephalographer, recognizing an interictal spike is as reproducible and as straightforward as identifying an upgoing toe on neurologic examination, but to the uninitiated, it is witchcraft. As illustrated by Benbadis, the main consequence of EEG misinterpretation is that false-positive reports lead to inappropriate treatments. Outpatients without seizures may be placed on antiepileptic medications. Intensive care unit patients may be treated for nonexistent status epilepticus. Although a cynic might assert that medicine would be better off if this diabolical test had never been invented, those of us who treat epilepsy know that the correctly interpreted EEG adds greatly to accurate diagnosis, classification, and detection of seizures, beyond what is possible from the neurologic history and examination alone, making treatment decisions substantially more effective. Bad EEG may be harmful; good EEG is invaluable. It is possible but cumbersome to reverse the consequences of prior EEGmisinterpretation. Review of prior tracings should be a routine part of epilepsy consultation, because a repeat normal recording does not resolve the issue of a prior false-positive interpretation. This means that the EEG data have to be sent in a format that can be opened without additional proprietary software. In the intensive care unit, clinical context clarifies the import of EEG findings. As Gaspard et al. point out, periodic patterns seen in different clinical contexts may appear somewhat similar; patterns after cardiac arrest may resemble those seen with more reversible processes. This means that the neurointensivist needs to view and understand the EEG findings. The most accurate management decisions are made when EEG findings are correctly integrated into the neurologic dataset. This means that the tracing is interpreted by a trained electroencephalographer informed of clinical context, but is also viewed by the treating neurologist, in the same way that neuroimaging studies are initially interpreted by neuroradiologists, but also examined a second time by neurologists. Fifty years ago, it was understandable that EEG studies were often misinterpreted. Criteria for abnormalities were still being defined and clinical correlates determined. It began to be recognized that many EEG variants, such as wickets and small sharp spikes, were common in normal individuals, and therefore were of no clinical significance. Meaningful fellowship training in clinical neurophysiology was only available in a few centers. We now live in a different world. There exists solid consensus on standards for routine EEG interpretation; controversy remains only for a few of the more complex findings, such as some periodic patterns, seen in patients with critical neurologic illness. Fellowship training is widely available, with oversight and accreditation. Well-trained, certified clinical neurophysiologists are spread throughout the United States. How can EEG misinterpretation still be a major problem? This is because many individuals currently reading EEGs have insufficient training. What would be adequate training? In our opinion, a minimum of 6 months of dedicated training under experienced credentialed electroencephalographers is needed to be able to interpret EEGs independently, with additional time needed to acquire skills in other clinical neurophysiology modalities. Typically, this is not possible during a neurology residency, and can be achieved by an additional residency or fellowship year in an accredited clinical neurophysiology training program. A solution to the problem of misinterpretation would begin with a slow, steady transition to a standard whereby formal EEG interpretation becomes the sole province of those with subspecialty training that includes 6 months’ experience in EEG. The wide dispersion of neurologists with clinical neurophysiology training, as well as the ease of remote access to recordings, makes this possible. This