A Convenient Truth: The Problem With Seizure Under-Reporting and How to Fix It

EEG focal epilepsy
EEG focal epilepsy
Before we begin to rely more heavily on technology for seizure detection, clinician-patient trust needs to improve.

For physicians who manage patients with epilepsy, critical treatment decisions hinge on patient reports of their seizures. Additionally, seizure reporting is essential to the integrity of epilepsy research. Unfortunately, this documentation is often inaccurate, with more than 50% of seizures captured on video-electroencephalography going unreported for both adult and pediatric patients.1

According to a recent paper published in the Lancet Neurology, underreporting is mainly caused by lack of awareness that a seizure has occurred.1 In previous studies, nocturnal seizures and those resulting in impaired consciousness were less likely to be documented.2,3 To address this issue, there has been an increasing focus on the development of electronic tools to aid in seizure detection outside of the clinic. 

These techniques include cerebral measures such as electroencephalography; home-based movement analysis techniques using accelerometry, pressure sensors, and other tools; markers of seizure-related autonomic alterations; behavioral approaches, including currently available smartphone apps that require patients to indicate when a seizure is occurring; and multimodal approaches that combine 2 or more of these techniques.4-7

The authors concluded that, whereas these methods have demonstrated sufficient accuracy in the detection of generalized tonic-clinic seizures, further investigation is needed for the development of techniques to detect other types of seizures.

To further explore issues pertaining to seizure detection and underreporting, Neurology Advisor spoke with Anto Bagić, MD, PhD, FAES, FACNS, professor of neurology and chief of epilepsy at the University of Pittsburgh Medical Center, and director of the UPMC Epilepsy Monitoring Unit, the University of Pittsburgh Comprehensive Epilepsy Center, and the UPMC MEG Epilepsy Program; and Page B. Pennell, MD, professor of neurology at Harvard Medical School and director of research in the epilepsy division at Brigham and Women’s Hospital, who spoke on behalf of the American Epilepsy Society.

Neurology Advisor: What are some of the problems associated with patient underreporting of seizures?

Dr Bagić: Seizure documentation and reporting is the evidence — a factual foundation — for our clinical practice and research in epileptology. Thus, it is hard to overemphasize its importance for everything we do in epilepsy. Simply speaking, accurate reporting is a prerequisite for precise assessment from epidemiology and diagnosis to treatment and prognosis. In practical terms, this affects literally everything we do in the field of epilepsy — whom we diagnose with epilepsy, whom we decide to treat, how we select and dose medications, whom we consider for surgery, who can drive, and who can work.

For instance, because all medications are approved based on reporting — an evidence that is barely 50% reliable — how do you feel about the use of “50% reduction of seizures in 50% of patients” as a threshold for approval? Only when we focus on this issue very thoughtfully, the degree of our routine complacency about this cardinal element of clinical practice becomes evident. And, all of this is when the patients are motivated to report the “truth” — their subjective truth, but what about situations when they may not be disposed accordingly, such as a risk for losing driving privileges or a job?

Accurate documentation and reporting of seizures is of additional social sensitivity in mandatory reporting states such as Pennsylvania. Picture how it feels to have adults crying in your office and telling you that they will lose a job, a house, the ability to support kids in college, if you fulfill your legal obligation and report them. On the other hand, in some published studies, more than half of patients indicated that they would be ready to lie to their doctor to retain a driver’s license. Yet, fallibility is implicit in either action of the doctor, and potential social and societal consequences are serious.

As I tell my patients on the first visit, trust and transparency are the critical prerequisites and foundation of a productive patient-doctor relationship, which is nowhere as important as it is here because we completely rely on patients and families reporting and currently have no objective way to establish the ground truth.

Dr Pennell: Underreporting of seizures by the patient, and even by family members or other close contacts, is extremely common. The more immediate consequences are compromised safety — if the patient reports that he or she is seizure free, then their doctor will provide clearance for them to engage in activities that would be potentially dangerous if a seizure with impaired awareness were to occur. Examples of dangerous activities are climbing heights for work or recreational purposes, scuba diving, operating automatic machinery, and driving. Restriction of driving is such a common and dreadful burden to the patient, but it is the most legislated because of the additional potential to cause harm to others.

Additionally, underreporting of seizures can lead to errant medical treatment decisions. For example, the clinician may decide to wean or lower an antiepileptic drug if he or she thinks the patient has been seizure free for an extended period, or the clinician may not try another antiepileptic drug or device or even surgery if there is the false impression that seizures are better controlled than they are.

Related Articles

As research continues to better understand the underpinnings of sudden unexplained death in epilepsy (SUDEP), fairly consistent risk factors include the number of convulsive seizures and nocturnal seizures. Nocturnal seizures probably have the highest underreporting rate, and therefore a patient may be at higher risk for SUDEP than is recognized by the treatment team.

It is amazing to think that all our treatments for epilepsy are based on changes in seizure rates when underreporting occurs in most patients, to such a high degree, and varies over time even within an individual. Thus, advances in the field of objective seizure detection will greatly help clinical research studies that are aimed at approval of therapies, but also for almost all study designs with other aims, such as prospective observational studies and retrospective studies.

Neurology Advisor: As researchers continue to investigate objective measures of seizure detection that may be feasible for patient use, what are some potential ways to address this issue in practice in the meantime?

Dr Bagić: While I see no short cut to meaningful and lasting improvements, raising awareness through sustained, appropriately geared education bears the most potential by far. It is not uncommon for an epileptologist to be told by the patient or family members, “Oh, I never knew that — why have I not heard about this 20 years ago? Doctors always told me that most people with seizures shake, etc.” Naturally, discussing about 40 types of seizures in the context of 25 types of epilepsy would be a counterproductive overdoing, but clarifying that in fact most people with epilepsy don’t shake would be a very valuable start.

Because “nonshaking” seizures are missed much more often, an appropriate awareness of this may be a big necessary step toward improving self-reporting. Appropriate engagement of the family and social circle is a necessary step with all possible ramifications. Encouraging the use of currently available assistive tools and options discussed in the Lancet article — as a supplement but not a replacement for being watchful — would be the next best step. Additionally, in my opinion, building trust and transparency has no alternative.

Dr Pennell: At this point, addressing seizure detection in practice is challenging. Embrace (by Empatica) was very recently cleared by the US Food and Drug Administration for seizure detection.8 However, as the article highlights, this device as well as other devices have the best sensitivity for convulsive, generalized tonic-clonic seizures. Other seizure types, such as focal impaired (complex partial) seizures and absence seizures, are less likely to be detected with high sensitivity.

Additionally, there can still be many false-positive alarms with all seizure detection devices. This puts the caregivers in a vulnerable situation for second guessing and guilt if they did not buy the device and something tragic happens, such as injury from a seizure or SUDEP, or if the device was used but they could not respond fast enough or in a way that changed the outcome. The outcomes are often not preventable even with immediate detection, and thus it could give the patient and family a false sense of security as well. The false-positive rates can also lead to unnecessary interventions initially and alarm fatigue later.

Neurology Advisor: What should be the focus of future research in this area?

Dr Bagić: An attempt to obtain an automatic yet specific comprehensive behavioral reading through a multimodal approach based on the ongoing lines of research discussed in the Lancet article — that in the future may be supplemented with the algorithms for a reliable automatic facial “reading” — remains the best foreseeable hope. However, it is only a surrogate for a plausible and convenient practical solution for a befitting route to the source — the brain. Seizures are fundamentally an electrical phenomenon, and accessing the brain’s electrical activity driving seizures noninvasively with sufficient “resolution” (ie, sensitivity and specificity) remains the most important and most challenging direction.

Many creative and innovative attempts have already been made, including wearables, but none have solved multiple existing technical obstacles well enough and combined the results within a single solution yet. An interesting but even more technically challenging approach would be to have a convenient method to continually monitor magnetic fields associated with normal and abnormal brain activity, similar to an approach implemented in routinely used magnetoencephalography that currently does not exist as a wearable option.

Dr Pennell: There is more work to be done on better multimodal technologies with high sensitivity and specificity for all seizure types, how to differentiate them, and the alarm and response options. We need to consider patient acceptance, but we also need thoughtful studies of implementation practices and the impact on patients, caregivers, and clinicians.


  1. Elger CE, Hoppe C. Diagnostic challenges in epilepsy: seizure under-reporting and seizure detection.Lancet Neurol. 2018;17(3):279-288.
  2. Kerling F, Mueller S, Pauli E, Stefan H. When do patients forget their seizures? An electroclinical study.Epilepsy Behav. 2006;9(2):281-285.
  3. Detyniecki K, Blumenfeld H. Consciousness of seizures and consciousness during seizures: are they related?Epilepsy Behav. 2014;30:6-9.
  4. Bialer M, Johannessen SI, Levy RH, et al. Seizure detection and neuromodulation: a summary of data presented at the XIII conference on new antiepileptic drug and devices (EILAT XIII).Epilepsy Res. 2017;130:27-36.
  5. Van de Vel A, Cuppens K, Bonroy B, et al. Non-EEG seizure detection systems and potential SUDEP prevention: state of the art: review and update. Seizure. 2016;41:141-153.
  6. van Andel J, Thijs RD, de Weerd A, Arends J, Leijten F. Non-EEG based ambulatory seizure detection designed for home use: what is available and how will it influence epilepsy care?Epilepsy Behav. 2016;57(pt A):82-89.
  7. Jory C, Shankar R, Coker D, McLean B, Hanna J, Newman C. Safe and sound? A systematic literature review of seizure detection methods for personal use.Seizure. 2016;36:4-15.
  8. PR Newswire. Embrace by Empatica is the world’s first smart watch to be cleared by FDA for use in neurology. https://www.prnewswire.com/news-releases/embrace-by-empatica-is-the-worlds-first-smart-watch-to-be-cleared-by-fda-for-use-in-neurology-300593398.html. February 5, 2018. Accessed March 5, 2018.