Incomplete health records complicate LGS diagnosis, study finds
Researchers call for standardized processes to improve diagnostic accuracy
Written by |
A study found gaps in electronic medical records that delayed LGS diagnosis. (Photo by iStock)
Incomplete or inconsistent documentation of brain activity, seizure types, and other key symptoms in medical records can make it difficult to accurately identify people with Lennox-Gastaut syndrome (LGS), a rare and severe form of drug-resistant epilepsy, a U.S. study found.
After reviewing electronic health records from 935 people with diagnostic codes for LGS or other forms of drug-resistant epilepsy, doctors specializing in epilepsy initially disagreed on whether patients met the criteria for LGS in about one out of nine cases. A closer review uncovered previously missing information that allowed 45 patients initially classified as not having LGS to be identified as meeting its diagnostic criteria.
The researchers said the findings highlight the need for “standardized terminology, structured clinical documentation, and rigorous multicenter validation processes” to improve diagnostic accuracy, help identify patients for clinical trials, support the development of computer-based tools to identify people with LGS, and improve long-term care.
The study, “Validation of the diagnosis of Lennox–Gastaut syndrome: Challenges and lessons learned,” was published in Epilepsia.
LGS typically begins in early childhood, often between ages 3 and 5. LGS symptoms include seizures that are difficult to control with medication, along with developmental, cognitive, and behavioral problems.
Changing criteria lead to differing diagnoses
Diagnosing LGS involves identifying seizure types, cognitive difficulties, and abnormal patterns of electrical activity in the brain, which can be detected with an electroencephalogram (EEG). But these features can change over time, with some appearing or disappearing as children grow older, making the condition difficult to recognize.
Adding to the challenge, the diagnostic criteria for LGS have changed over the years, leading to differences in how the condition is identified in clinical practice and research. Yet an accurate diagnosis is important for guiding LGS treatment, helping families understand what to expect, and determining eligibility for clinical trials.
To verify LGS diagnoses using established diagnostic criteria, the researchers identified 997 patients whose medical records included diagnostic codes for LGS or other forms of drug-resistant epilepsy between 2016 and 2022 at seven pediatric epilepsy centers across the U.S. After excluding 62 patients who had fewer than two neurological care visits, they reviewed the records of 935 patients to determine who met the criteria for LGS.
The work was part of a larger project to develop computer-based tools that can identify people with LGS from electronic health records.
Researchers gathered information about each patient’s seizures, EEG results, cognitive development, and treatments. Two doctors specializing in epilepsy then independently reviewed each patient’s records to determine whether they met the standard diagnostic criteria for LGS, less restrictive diagnostic criteria used in clinical trials, or neither. A third, senior specialist resolved any disagreements.
The two epilepsy specialists agreed on the initial classification of 830 patients (89%), but disagreed on 105 (11%).
Looking more closely, the researchers found that some patients had been classified as not having LGS because important details were missing or had been inconsistently reported.
These included characteristic EEG findings that were not always captured or clearly documented, seizure descriptions that were too vague to identify specific seizure types, and other missing details, such as the age when seizures first began. Information about learning and developmental difficulties was also often scattered across records from different healthcare providers, making it difficult to piece together a complete picture.
The team went back through the records of 130 patients, searching for additional EEG reports and details about their symptoms. When new information was found, doctors reassessed whether the patients met the criteria for LGS.
This second review changed the classification of 67 patients. Of these, 33 were found to meet the standard diagnostic criteria for LGS, while another 12 met the less restrictive diagnostic criteria. The remaining 22 were excluded because there still wasn’t enough information to determine whether they had LGS.
The most commonly uncovered piece of information was a characteristic EEG pattern called slow spike-and-wave, which was documented in 38 patients (83%) whose LGS classification changed.
Reviewing additional EEG reports was particularly important because slow spike-and-wave activity can appear or disappear over time, meaning it may be present in one recording but not another, the researchers noted.
Additional details about seizure types emerged in 10 patients (22%), while the age when seizures first appeared was clarified in seven (15%). Previously missing evidence of learning or developmental difficulties was also found in three patients (7%).
“This study highlights the challenges of validating the diagnosis of LGS, particularly in extracting information from the [electronic health records] and applying diagnostic criteria across a diverse patient population,” the researchers wrote. “Future work is needed to develop and to implement strategies that facilitate the diagnosis of LGS clinically, to improve patient care, and to improve the identification of patient [groups] for research studies.”
