Analyzing Reliability of Seizure Diagnosis Based on Semiology.

Bo Jin,Han Wu,Jiahui Xu,Jianwei Yan,Yao Ding,Z Irene Wang,Yi Guo,Zhongjin Wang,Chunhong Shen,Zhong Chen,Meiping Ding,Shuang Wang

DOI: https://doi.org/10.1016/j.yebeh.2014.09.057

2014-01-01

Abstract:This study aimed to determine the accuracy of seizure diagnosis by semiological analysis and to assess the factors that affect diagnostic reliability. A total of 150 video clips of seizures from 50 patients (each with three seizures of the same type) were observed by eight epileptologists, 12 neurologists, and 20 physicians (internists). The videos included 37 series of epileptic seizures, eight series of physiologic nonepileptic events (PNEEs), and five series of psychogenic nonepileptic seizures (PNESs). After observing each video, the doctors chose the diagnosis of epileptic seizures or nonepileptic events for the patient; if the latter was chosen, they further chose the diagnosis of PNESs or PNEEs. The overall diagnostic accuracy rate for epileptic seizures and nonepileptic events increased from 0.614 to 0.660 after observations of all three seizures (p < 0.001). The diagnostic sensitivity and specificity of epileptic seizures were 0.770 and 0.808, respectively, for the epileptologists. These values were significantly higher than those for the neurologists (0.660 and 0.699) and physicians (0.588 and 0.658). A wide range of diagnostic accuracy was found across the various seizures types. An accuracy rate of 0.895 for generalized tonic-clonic seizures was the highest, followed by 0.800 for dialeptic seizures and then 0.760 for automotor seizures. The accuracy rates for myoclonic seizures (0.530), hypermotor seizures (0.481), gelastic/dacrystic seizures (0.438), and PNESs (0.430) were poor. The reliability of semiological diagnosis of seizures is greatly affected by the seizure type as well as the doctor's experience. Although the overall reliability is limited, it can be improved by observing more seizures.