With the help of quantitative EEG (qEEG) software to graphically display complex data, nurses can be trained to spot nonconvulsive seizures (NCS) in neurologic intensive care patients and potentially shorten the time to treatment, a new study by researchers at Duke University Medical Center shows.
NCS occur commonly among neuro-ICU patients, and may cause further injury to already damaged brains—and lead to worse outcomes—according to neurologist Christa Swisher, MD, the lead investigator. The results of the study were published in Neurocritical Care.
Although more than 90% of seizures in critically ill hospital patients are nonconvulsive, NCS by definition have no clear physical manifestation. They can be detected by continuously monitoring a patient’s EEG, but that approach is obviously not practical.
Quantitative EEG software provides a time-compressed, simplified graphical display calculated from raw EEG data. Board-certified neurophysiologists receive extensive training in interpreting EEGs and qEEG, but due to their workloads, they read the data only every six hours on average.
“Many studies have demonstrated that a delay in recognition and therefore a delay in treatment results in seizures being difficult to control, and in the meantime they cause more brain injury. The longer that they seize, the worse they do,” Swisher says.
In the study, neurologists gave neuro-ICU nurses a brief training on reading qEEG displays tailored to each patient. With the wide variety of conditions resulting in-patient admissions to the neuro-ICU, patients’ EEGs show significant variation. So Jennifer H. Kang, MD, a neurology resident at Duke and first author of the study, reviewed each patient’s qEEG information with a nurse in order to point out the patterns that indicate a seizure and contrast them with artifacts such as muscle movements that can stand out from background.
Tailored Training Sessions Key to Effective Use of EEG
The nurses then checked the displays in hourly blocks for evidence of NCS. “After tailored training sessions, the nurses were able to pick up recurrent seizures with a very good sensitivity and a low false-positive rate,” Swisher says. They detected seizures with a sensitivity of 85.1% and a specificity of 89.9% compared with the reading of conventional continuous EEG data by two neurophysiologists.
Swisher says that it would be feasible to incorporate this information into clinical practice because the nurses could alert the neurocritical care providers to periods of concern, who could discuss the findings with the neurophysiology team to decide whether seizures are in fact present. The neurocritical care team would then decide whether the patient needed to be treated with intravenous anti-seizure medications that can be “loaded” to get them to therapeutic levels very quickly.
“The study demonstrates that it is very reasonable to train nurses on qEEG screening with a tailored and simplified training session,” Swisher says. She compares the process to training nurses to review bedside telemetry and alert the medical team when worrisome changes in a patient’s EKG are detected.
Because qEEG software is readily available, “this approach could change the way patients are managed not only at academic institutions but also at private and community hospitals, to provide better treatment and better outcomes,” Swisher says.