3.1 Patients and pathway

There were a total of 123 home videos with a mean 3.4 videos/referral and mode of 1 video/referral. Video duration ranged from 0.2–11.4 min, with a mean of 2.0 min. Video picture quality was mostly considered sufficient for clinical assessment (classified “very good” [19/36] or “good” [16/36]). One referral had videos with poor picture quality as it was taken overnight by a home monitoring system.

35/36 referrals showed videos with the relevant body part displayed. The full body on display, face on display and post event period were shown in over 70% of referrals. Only 22/36 referrals showed interaction with the patient during the event. Of note, a high proportion of referrals (17/36) did not capture the start of the event, which often includes clinically relevant information (Figure 2A). However, a significant proportion of the events happened at predictable times or in clusters so parents/carers were able to capture the beginning of an event.

There was an overall median 10 days for video upload, referral, MDT meeting and issue of a final report (mean 19 days, range 0–123 days; Figure 2B). Most time was spent between initial video referral and MDT meeting (median 5 days). A common contribution to this delay was referring consultants finding it difficult to arrange attendance at the MDT meeting to present cases due to clashing clinical or professional commitments. One referral took 123 days due to such a delay in coordinating with the referring consultant.

Seven referrals had no record of previous EEGs. Nine referrals had previous Video EEG Telemetry (VT) recordings with habitual events recorded in six of these studies. Two VT studies were abnormal without showing seizures, three VT studies confirmed the presence of both epileptic seizures and non-epileptic events, two recorded epileptic seizures only and in one study the child was confirmed to have a movement disorder. After HVR review, repeat telemetry was advised in one child with previously recorded Myoclonic absences; this was to confirm the reported increase in seizure frequency to help guide treatment escalation. In the other eight children with previous VT studies, repeat telemetry was avoided after HVR.

Twenty referrals had previous inter-ictal EEGs, of these seven were normal studies. In one child with a normal EEG a habitual non-epileptic event was recorded during the EEG. After the MDT meeting VT was advised in three children, all with abnormal inter-ictal EEGs. In the first of these VT was advised urgently at the child's regional centre as HVR was consistent with infantile spasms. In the second child VT was undertaken to confirm epilepsy classification as HVR was consistent with a focal seizure. In the third child VT was advised in the future should the frequency increase of current infrequent events seen on HVR, given the background history of lissencephaly. The referring clinician concluded this was not needed after the review in the follow up clinic.

A total of five patients had subsequent EEGs against the consensus of the MDT. Four were for patients that the HVR concluded showed non-epileptic events and no further tests were required. The follow-up EEGs in all four of these cases did not change the diagnosis from that of the HVR MDT. A fifth child had known epilepsy and the HVR MDT thought the episodes were in-keeping with his previous episodes which was confirmed by a subsequent EEG.

3.2 Events

After the MDT meeting the number of referrals classified as non-epileptic events increased from 11 to 22 (Figure 3). These included “transient infant movements”⁸ which comprised head shaking, shuddering attacks, self-gratification or other types of previously reported complex stereotypies. It was helpful to triage these videos as they were sometimes referred urgently as “spasms”, with a few having already had previous EEGs but without habitual event capture. Of those children in whom HVR classified non-epileptic events, three had episodes consistent with a diagnosis of functional neurological disorder (FND). Four of the non-epileptic event referrals were suggestive of other neurological disorders and were followed up by the referring doctor without further neurophysiological tests required.

Ten children were classified as having convincing epileptic events of which two were spasms. This was a similar total proportion to before the meeting, but the number classified as “spasms” decreased from six to two. The first child had two videos that were classified as bilateral convulsive seizures and an absence that was similar in semiology to events previously recorded on video-EEG telemetry (VT). The second had six videos with nocturnal events consistent with frontal lobe hypermotor seizures. Follow-up MRI revealed a structural lesion. The third child's videos revealed events comparable to previously recorded seizures on VT and were classified as myoclonic absence seizures on MDT review. Two children had a mixture of epileptic and non-epileptic events.

The number of referrals classified as uncertain fell from 15 to 2. Of the two for whom consensus MDT diagnosis was not reached, the first referral consisted of very brief eye movements in a child with developmental malformation and delay. A second video was requested but this was also too short at 11 s to arrive at a clinical decision. The second referral showed only one video (captured mid event) not typical of spasms but was classified as uncertain as there were 20 repeated movements. More videos were requested but not provided. Both patients ultimately had a VT study arranged after the MDT.

3.3 Outcomes

Twenty-six patients referred to the service were triaged off the neurophysiology waiting list; of these, six avoided a repeat VT. Five patients were triaged off the neurophysiology waiting list but ultimately later had a further EEG which did not change the HVR diagnosis. Eleven referrals resulted in an attempt to optimize medication, while 25 had no change. Changes to one patient's medication may have been delayed by the MDT HVR as their initial videos were too brief (11 s) and were deemed inconclusive. An EEG 75 days after the first referral ultimately showed atonic and tonic seizures different in semiology to the videos originally sent and treatment was then optimized.

3.4 Feedback from referring consultants

Most consultants (7/10) estimated that they spent over 5 min analyzing the videos before requesting a referral to neurophysiology. Almost all consultants (9/10) reported that HVR improved their understanding of the event. 4/5 consultants responding to the question said that the HVR MDT modulated follow-up investigations; “my patient avoided unnecessary sleep investigations (polysomnography and MSLT) as well as unnecessary further neurophysiological tests e.g. telemetry”. One respondent suggested that the mandatory attendance of the referring doctor at the HVR MDT meeting made the service inaccessible because they did not routinely work on the day it was held.

4.1 Study limitations

A major limitation is imposed by the retrospective nature of the study without a gold-standard investigation offered to all patients independent of the MDT outcome. This means that, for instance, where the MDT determined events as not epileptic in nature there was no routine follow-up VT study to provide confirmation. All cases were followed up clinically by their referring team and review of electronic healthcare records did not provide any examples in which the MDT outcome was altered, including in five patients who eventually had further video-telemetry despite the advice of the MDT that this was not indicated. Follow-up was limited to what could be accessed via electronic healthcare records from our centre; given that this is the regional centre it is reasonable to expect that we cases for whom MDT outcomes were refuted by subsequent clinical progression would have returned. We cannot exclude that some patients might have moved out of the area or sought opinions or diagnosis elsewhere. A further limitation includes the fact that this was a study in an exclusively pediatric patient population. However, other studies have shown video review is reliable in adults^1-5 and therefore suggests a similar HVR service could be adopted for adult patients. As with all MDTs there is no definitive or known optimum method of ensuring a consensus is reached that fully includes the opinions of all the specialists involved. In this case, the meeting chair's decision was the documented decision of the MDT but this was determined after an open discussion.

4.2 Further improvements to the HVR service

To further improve video quality, guidance will be given to parents on what information is most useful to capture on video, such as the whole body, the start and end of the event and interactions with the patient. This may also be paired with an exemplar video.

To improve accessibility, we are allowing the clinical summary to be presented by a clinician who is familiar with the patient if the responsible referring clinician is unavailable. Alternatively, flexibility in meeting timings, for example holding them on different days or times throughout the month, should maximize the opportunities for attendance. There would likely be significantly increased uptake of HVR in other hospitals if it could be coded as a stand-alone diagnostic procedure as suggested by others¹⁴ in order that the activity is appropriately acknowledged and reimbursed where appropriate.