Malloy Kristen, Cardenas Damon, Blackburn August, Whitmire Luke, Cavazos Jose E
Brain Sentinel, San Antonio, TX, United States of America.
Blackburn Statistics, LLC, San Antonio, TX, United States of America.
Epilepsy Behav. 2018 Dec;89:84-88. doi: 10.1016/j.yebeh.2018.09.012. Epub 2018 Oct 31.
There is a high cost associated with recording quality video and electroencephalography (EEG) data in National Association of Epilepsy Center (NAEC) level IV epilepsy monitoring units (EMU). This study considers potential quality measures in EMUs for generalized tonic-clonic (GTC) seizures: types of safety signals, response time, and visibility of patient's limbs for semiology. These quality measures have been summarized across 12 EMUs to estimate response times to GTC seizures and the quality of video data that is captured during admissions.
Video electroencephalographies (vEEGs) from two prospective regulatory studies for the Brain Sentinel device were reviewed. A total of 232 subjects with a history of GTC seizures underwent routine clinical EMU stays. Fifty-four of the study subjects had 96 GTC seizures. The vEEG of events were reviewed for safety signal used, response time, and visibility of patient's limbs.
The average response time from members of the hospital team was 22 s from electrographic generalization (minimum -37 s, maximum 111 s, two no response). For caregivers, average response was 11 s (minimum -15 s, maximum 33 s, 45 not present/no response). In 73% of events, the patient visibility was limited at seizure onset. In 55% of events with limited limb visibility, the visibility was improved (by removing sheets or improving camera angle) >30 s after start of the event. The primary safety signals were as follows: an alert from outside the patient room (54%), button press (23%), hospital team present at seizure start (14%), caregiver vocal alert (6%), and no response (2%).
The average response time of caregivers was twice as fast as the hospital team, underscoring the importance of having a person in the room during onset of a GTC seizure. Diagnostic yield could be improved with more timely removal of patient coverings. It was observed that when patients experienced a GTC seizure, 40% were fully or partially obscured for more than 30 s during the event, compromising the ability of epileptologists to evaluate semiology during seizure onset. Automated seizure alarms may help staff get to patients more quickly and improve diagnostic characterization.
在全国癫痫中心协会（NAEC）四级癫痫监测单元（EMU）中记录高质量视频和脑电图（EEG）数据的成本很高。本研究考虑了EMU中针对全面强直阵挛（GTC）发作的潜在质量指标：安全信号类型、响应时间以及用于癫痫发作症状学分析的患者肢体可见度。对12个EMU的这些质量指标进行了总结，以估计对GTC发作的响应时间以及入院期间捕获的视频数据质量。
回顾了两项针对Brain Sentinel设备的前瞻性监管研究中的视频脑电图（vEEG）。共有232名有GTC发作病史的受试者接受了常规临床EMU住院治疗。其中54名研究对象发生了96次GTC发作。对发作事件的vEEG进行审查，以确定所使用的安全信号、响应时间以及患者肢体的可见度。
医院团队成员从脑电图广泛性放电开始的平均响应时间为22秒（最短-37秒，最长111秒，2次无响应）。对于护理人员，平均响应时间为11秒（最短-15秒，最长33秒，45次未出现/无响应）。在73%的事件中，发作开始时患者的可见度受限。在55%肢体可见度受限的事件中，事件开始>30秒后可见度得到改善（通过掀开床单或改善摄像头角度）。主要安全信号如下：病房外警报（54%）、按钮按压（23%）、发作开始时医院团队在场（14%）、护理人员语音警报（6%）以及无响应（2%）。
护理人员的平均响应时间比医院团队快两倍，这突出了在GTC发作开始时病房内有人员在场的重要性。更及时地掀开患者覆盖物可提高诊断率。据观察，当患者发生GTC发作时，40%的患者在发作期间有超过30秒的时间被完全或部分遮挡，这损害了癫痫专家在发作开始时评估发作症状学的能力。自动癫痫警报可能有助于工作人员更快地赶到患者身边并改善诊断特征描述。
