Measuring expertise in identifying interictal epileptiform discharges

Area of Science:

• Neurology
• Medical Education
• Biostatistics

Background:

• Interictal epileptiform discharges (IEDs) on EEG are crucial for epilepsy diagnosis.
• EEG interpretation varies among readers with different training levels.
• A standardized method to assess IED recognition skills is lacking.

Purpose of the Study:

• To develop and validate a test that quantifies the skill of readers in identifying IEDs.
• To differentiate between expert, experienced non-expert, and novice EEG readers based on IED recognition.
• To establish a reliable method for assessing EEG interpretation proficiency.

Main Methods:

• Developed an online test using 13,262 candidate IEDs, scored by eight fellowship-trained reviewers to create a gold standard.
• Assessed reader performance based on sensitivity, false positive rate, and calibration.
• Utilized a mathematical model to estimate reader skill and threshold, generating receiver operating characteristic (ROC) curves.

Main Results:

• Twenty-nine raters (9 experts, 7 experienced non-experts, 13 novices) completed the test.
• Experts demonstrated superior performance with higher median area under the ROC curve (0.891) compared to experienced non-experts (0.852) and novices (0.735).
• Novices exhibited higher uncertainty (noise level) in IED identification than experienced readers.

Conclusions:

• The developed test effectively distinguishes between expert and non-expert readers based on IED identification ability.
• This assessment can identify and potentially correct variations in IED recognition thresholds among clinicians.
• The findings support the use of standardized testing to improve EEG interpretation accuracy and training.