William Tatum, DOWilliam Tatum, DO
Results from the OSmartViE study demonstrate that smartphone videos of outpatients being evaluated for epilepsy can be predictive of video electroencephalogram (EEG) diagnosis and ultimately add value, when combined with medical history and physical exam, to the diagnostic process.

The prospective, masked, diagnostic accuracy study included smartphone videos from 44 nonconsecutive epilepsy clinic outpatients (age ≥18 years, 70.5% women). The videos were recorded during evaluation process prior to EEG monitoring. The videos were evaluated by physicians from 8 different academic epilepsy centers.

Videos submitted were from events that could have or could not have been epileptic seizures. Board certified experts in epilepsy and clinical neurophysiologists reviewed the videos and made the final diagnosis based on video EEG monitoring (VEM).

Ultimately, 44 videos were uploaded and underwent 530 video reviews performed by19 reviewers (10 epilepsy experts and 9 senior neurology residents), with a mean of 6.6 experts per video and 5.5 residents per video.

Using video EEG, reviewers diagnosed 11 epileptic seizures, 30 psychogenic nonepileptic attacks, and 3 physiologic nonepileptic events.

Following review of final smartphone videos, the expert epilepsy reviewers accurately predicted a video EEG diagnosis of epileptic seizures 89.1% of the time (95% CI, 84.2%-92.9%), with a specificity of 93.3% (95% CI, 88.3%-96.6%). Notably, accuracy of diagnoses for epileptic and psychogenic nonepileptic attacks were lower for residents, “suggesting experience is important when using smartphone videos alone without EEG to predict a diagnosis,” study author William O. Tatum, DO, director of the Comprehensive Epilepsy Center at Mayo Clinic and professor of neurology at Mayo Clinic College of Medicine & Health Sciences, told NeurologyLive.

The rate of correct diagnoses increased from 78.6% to 95.2% when medical history and physical examination results were combined with smartphone videos, increasing the odds of receiving a correct diagnosis over 5 times compared with using history and physical exam alone (95% CI, 1.01-54.3; P =.02).  Notably, a correct diagnosis of 25% of the smartphone videos, composed solely of psychogenic nonepileptic attacks, was achieved by 100% of the reviewers.

“Overall, convulsive episodes (9 correct vs 2 incorrect/unknown reviews; P  <.03) and psychogenic nonepileptic attacks (148 correct vs 13 incorrect/unknown; P  <.001) were significantly more likely to be correctly diagnosed from video review,” Tatum and colleagues wrote.

“Misdiagnosis has significant downfalls including use of antiseizure medication, loss of driving privileges, social isolation, and financial implications,” Tatum told NeurologyLive. “Therefore, with a high degree of accuracy and inter-rater reliability in patients with nonepileptic conditions, encouraging patients to capture a video as adjunctive information to a complete and detailed history and physical examination is what we currently recommend.”

Some limitations of using smartphone videos as a diagnostic tool include the degree of digital sophistication, preserved patient consciousness and motor abilities, and costs to maintain electronic platforms and wireless connection with established video transfer privacy.

“Expert evaluation of smartphone videos reflects a modern mobile health tool and is a useful adjunct to HP [history and physical exam]. The ability to identify patients without epilepsy before inpatient VEM could triage and realign resources to patients for whom the need is high (ie, surgery),” Tatum and colleagues concluded.
REFERENCE:
Tatum WO, Hirsch LJ, Gelfand MA, et al. Assessment of the predictive value of outpatient smartphone videos for diagnosis of epileptic seizures. JAMA Neurol. Published online January 21, 2020. doi:10.1001/jamaneurol.2019.4785.