Telemedicine and Neurology: When Is Virtual Neurologic Examination Appropriate?

High rates of agreement in detecting an abnormality on the virtual and in-person neurologic examination may support the validity of the virtual neurologic examination depending on the clinical scenario, according to study findings published in Neurology: Clinical Practice.

Study authors conducted a real-world retrospective chart review of patients who were examined virtually and in-person within 4 months at outpatient general neurology and neuromuscular clinics from 2 tertiary academic care centers during the COVID-19 pandemic to assess the utility of virtual vs in-person neurologic examination in arriving at an accurate localization and diagnosis. Adults who underwent an initial consultation using videoconferencing between March 2020 and June 2021 and who underwent a subsequent in-person visit with a bedside neurologic examination within 4 months of the initial virtual consultation were eligible for inclusion. The primary outcome was overall agreement between virtual and in-person neurologic examination, defined as the presence or absence of any abnormality on both examinations. To test overall agreement between the 2 modes of examination, the Cohen kappa coefficient was calculated. Fourfold contingency tables were used to calculate sensitivity and specificity.

A total of 81 patients (mean age, 48; women, 61%) were included in the study, of whom 41 (50.6%) were seen in a neuromuscular clinic and 40 (49.4%) were seen in a rapid general neurology clinic.

Overall, there was decent agreement between the findings on the virtual and in-person neurologic examinations with regard to the detection of an abnormality or a normal examination (agreement, 64%; P =.003).

Three (3.7%) patients had an abnormality detected on the virtual exam that was not found on the in-person exam, while 26 (32%) patients had normal virtual exams and had an abnormality on their in-person exam. In 6 (23%) cases, the findings were incidental and unrelated to the referral reason.

The specificity and sensitivity of the virtual exam for identifying an abnormality on examination was 86% and 53%, respectively, when using the in-person exam as the gold standard.

The highest agreement between findings on the virtual and in-person exams was found for patients who presented with cranial nerve symptoms (86%), weakness (75%), and pain (75%).

Substantial agreement between the 2 exam types was found for gait abnormalities, while moderate agreement was found for extraocular movements, facial weakness, dysarthria, fasciculation, and lower limb weakness. Fair agreement was found for bulk, upper limb weakness, and sensation, while no agreement was found for hypokinetic or hyperkinetic movements and cerebellar signs.

A large proportion of cases demonstrated a consistent localization and diagnosis (44% and 57%, respectively) after the virtual and in-person assessments. While localization changed in 12 (15%) cases, diagnosis changed in 11 (14%) cases.

There was a refinement (ie, an adjustment in the clinician’s impression without a complete change in localization or diagnosis) in localization and diagnosis in 33 (41%) and 24 (30%) cases, respectively, between visits. Of those who had a change in localization and/or diagnosis, only 3 patients had a documented difference in their virtual and in-person exams.

Study limitations include the small sample size, retrospective data collection method, variation in the delivery of in-person neurologic examinations, and certain unavoidable confounders.

“Future work should be pursued to determine factors that influence the level of agreement between the VNE [virtual neurologic examination] and IPNE [in-person neurologic examination] (e.g., video quality and presenting symptoms), including prospectively designed studies,” the study authors concluded.