Safety in the EMU: Reaching Consensus

Abstract

Long-term video-EEG monitoring in the epilepsy monitor-ing unit (EMU), while generally a safe procedure (1, 2), can be associated with adverse events. The majority of EMUs have experienced falls, status epilepticus, and postictal psychosis during monitoring (3); although rare, fatalities and near-fatalities have also occurred in the EMU (2, 4). Even patients with psychogenic nonepileptic seizures are prone to adverse events, usually falls, at a significant rate (5). Provocative proce-dures such as antiepileptic drug (AED) withdrawal and sleep deprivation increase the risk for adverse events, particularly if there is a history of generalized tonic-clonic seizures (GTCs) as a proportion of them will inevitably experience difficult to control seizures that may require intubation and treatment in the intensive care unit (ICU). Such risks should be understood, prior to monitoring, by both the treating physician as well as the patient. Nonetheless, because video-EEG monitoring is an elective procedure for which possible morbidities are predict-able and potentially systematically preventable, safety in the EMU must always be the primary goal of any admission plan.Despite potential hazards and acknowledged importance of safety measures, there is a wide variation in practice with regard to drug withdrawal, seizure observation, and rescue protocols (6) in the EMU and a clear lack of consensus between epilepsy centers. Implementation of the safety measures may be costly, especially when additional personnel are involved. Automated safety alerts may miss critical events owing to the high false-positive rate that are ignored and can lead to delay in responding to seizures (7).Because of the importance of this issue, the American Epilepsy Society formed a workgroup to search for evidence and best practices regarding safety measures for patients ad-mitted to the EMU. The first product of this effort was a survey published by Shafer et al. in 2011, identifying the extent to which adverse events occurred in the EMU, including falls, sta-tus epilepticus, postictal psychosis, and pulling of implanted electrodes (3).In their most recent article, “A consensus-based approach to patient safety in epilepsy monitoring units: Recommenda-tions for preferred practices” (8), Shafer et al. build upon these findings to establish a set of consensus practice recommenda-tions for enhancing patient safety in the EMU. First, a set of statements regarding safety were developed by four work-groups in the key areas: seizure observation, seizure provoca-tion, management of acute seizures, and activity/environment. Because the authors found that literature searches revealed lack of evidence, expert consensus was sought using the Delphi methodology. This methodology consists of a series of iterative questionnaires and anonymous feedback (9). A set of statements from the workgroups were consolidated by a screening committee, evaluated by a small group of indepen-dent experts, and then further revised. The resulting state-ments were then submitted to the Delphi process; they were sent in an e-mailed survey to a select group of American Epi-lepsy Society (AES) members, an expert group, and workgroup members. Respondents rated each of the safety statements on a scale between one (completely agree) and nine (completely disagree). A second survey was sent, with each item’s mean and spread; the participants were asked to reevaluate them. Items where rating showed a strong agreement (at least eight) were accepted. These exhaustive efforts generated a set of 30 safety recommendations with strong consensus.Some of the seemingly more important recommenda-tions did not reach sufficient consensus to be included. For example, the question of whether continuous observation is needed for all patients all the time did not reach consensus. Given the relatively stringent criteria needed for a recommen-dation to reach acceptable consensus, a lack of consensus on a particular recommendation should not necessarily indicate that the safety measures need not be implemented. For example, the use of a pulse oximeter did not reach consensus, but it may nevertheless be considered because of ease of implementation, low cost, and effectiveness as an adjunctive alarm device.Will adherence to these safety measures translate to improvement in safety outcomes? There is some encouraging data to suggest that it will. After instituting EMU safety process improvements that included staff education, expanded EMU staffing, fall prevention protocols, and enhanced communica-tion hand-offs, one large EMU was able to achieve significant decrease in missed seizures and a small, though statistically insignificant, decrease in rate of falls (10). Further studies are definitely warranted to determine the extent of improved outcomes as a result of these recommendations.