Abstracts

Clinicians are trained to synthesize data and arrive at a diagnosis. We tested the ability of clinicians to arrive at the correct final diagnosis with limited information in patients with seizure like events. In clinical practice patients may be diagnosed with non-epileptic seizures (NES) and epileptic seizures (ES) based on the patients initial appearance. Episodic neurologic events are often difficult to diagnose. Published reports suggest that patients admitted to the epilepsy monitoring unit (EMU) are correctly diagnosed with seizures at a rate of 70-75% before admission. Making a correct diagnosis of seizure-like events is critical. Misdiagnosis impacts quality of life and may lead to iatrogenic complications. Does the appearance and or chief complaint influence the clinicians early impression of the ultimate diagnosis of ES/NES in patients with seizure-like events?
Patients admitted to the Mayo Clinic Scottsdale EMU are digitally photographed. Fifty-nine consecutive patients with seizure like events were evaluated. Photographs and chief complaints were compiled. Blinded raters were instructed to evaluate the photographs and chief complaints independently and then in combination to arrive at a diagnosis of either seizures or NES. Raters were board certified neurologists specializing in Movement Disorders, MS, Stroke, and Epilepsy. No particular method to determine the diagnosis was given to the raters. A non-medical professional was used as a control rater. The final diagnosis in each patient was made after EMU monitoring of typical events. We calculated the proportion of correct diagnoses, inter-rater agreement, sensitivity and specificity.
The sample consisted of 64% female and 36% male patients. 44% had a final diagnosis of seizures and 56% NES. The medical professionals rating the combination of photographs and chief complaints did better than chance, 66% correct, 95% CI (60%-73%), whereas the control rater performed worse than chance, 47% correct. Rating appearance only was 56% correct and complaints only 58% correct. The inter-rater beyond chance agreement, Kappa = 0.50 (moderate). With male patients, the raters are poor at correctly diagnosing NES: 47% correct.
The study emphasizes the ability of some raters based on early initial impressions to correctly diagnose patients referred to an EMU with an accuracy rivaling published reports. While we do not know the methods by which raters made their judgments, the initial encounter may provide clues to the ultimate diagnosis. There continues to be a gender bias against men having NES. Admission to an EMU remains the [quot]gold standard[quot] for diagnosing seizure-like events. Further study is needed to determine the key elements of the initial and subsequent patient evaluation that might allow us more accurate and efficient methods of diagnosing these challenging patients.