Conference Coverage

Researchers mine free-text diary entries for seizure cluster insights


 

REPORTING FROM AES 2019

– Free-text diary entries by patients with epilepsy are a “largely untapped” source of information about the frequency and treatment of seizure clusters, researchers said at the annual meeting of the American Epilepsy Society. In addition, patients may describe other clinically relevant concerns such as tiredness, depression, head injury, or seizures while driving, researchers said.

Joyce A. Cramer, a clinical research consultant affiliated with Yale University, New Haven, Conn. Jake Remaly/MDedge News

Joyce A. Cramer

To examine how seizure clusters are reflected in the electronic diaries of patients with epilepsy, Joyce A. Cramer, a clinical research consultant and colleagues examined data from EpiDiary, a set of mobile and Web-based apps designed to help patients with epilepsy manage their medications and record their symptoms. EpiDiary prompts patients to indicate whether they were seizure free, had a seizure, or had a seizure cluster on a given day. Patients also have the ability to enter free-text notes.

“This was the first-ever review of the unstructured, free-text notes,” Ms. Cramer said.

Investigators used lexical analysis to identify free-text comments that potentially were about seizure clusters, based on the use of words such as “lots,” “many,” or “repeat.” Researchers reviewed every flagged comment to confirm whether it pertained to a seizure cluster. They defined a cluster as two or more seizures on a calendar day.

An algorithm flagged 5,955 entries by 1,839 users. Clinician review confirmed that 2,645 of the flagged comments (44.4%) pertained to seizure clusters. Of the confirmed clusters, 512 (19.4%) were found only through the free-text notes and had not been documented through structured data elements such as seizure cluster check-boxes or seizure counts.

“Extra medicine was taken for clusters by 553 users on 3,818 days,” the researchers reported. “This was 30.1% of all users and 56.5% of those commenting on clusters.” In some instances, patients named specific medications, including lorazepam, clonazepam, midazolam, clobazam, rectal diazepam, other diazepam, and clorazepate.

Free-text diary entries could help researchers study various topics. The authors highlighted examples of entries that “contained other clinically relevant information,” including the following:

  • Massive ongoing cluster with about 20% apneic events.
  • My constant question seems to be: HOW can I function in life when just small outings bring about this incessant tiredness?
  • Started feeling like I was having an aura and pulled over.
  • Thought about suicide for the first time in a while.

Interpretations of the seizure cluster data are limited, the researchers noted. The algorithm might have missed some free-text comments that were about seizure clusters. And in some instances, researchers used words such as “puffs” to identify seizures when a connection to seizures was not entirely clear. In addition, patients may have used a definition of cluster that was different from the definition used by the investigators.

UCB Pharma and Irody, the company that owns EpiDiary, funded the study. Irody’s founder and president was a coauthor, and another author holds stock or options in Irody. Ms. Cramer consults for Irody, UCB, and other pharmaceutical companies.

SOURCE: Fisher RS et al. AES 2019. Abstract 1.424.

Next Article: