A pediatric epilepsy diagnostic tool for use in resource-limited settings: A pilot study

It is estimated that nearly 80% of the 50 million people affected with epilepsy globally live in regions where specialist care and diagnostic tests are scarce and care is often delivered through a primary health provider with limited training. To improve diagnostic accuracy of the history and physical examination, we developed and piloted a questionnaire to discriminate between focal versus generalized epilepsy, with the future goal to guide medication choices.

Through literature review and retrospective chart review of 75 children with epilepsy at Boston Children's Hospital, a 15-item questionnaire was developed. Simple motor seizures were excluded for the purposes of this questionnaire. The questionnaire was then translated in local dialects and prospectively validated at Muhimbili National Hospital in Dar Es Salaam, Tanzania, and University Teaching Hospital in Lusaka, Zambia. Children 6months-18years of age with suspected or active epilepsy were identified, and a nonphysician administered the questionnaire to the patient's caregiver. Next, each patient was evaluated by a pediatric neurologist blinded to the questionnaire results, and together with locally obtained but remotely interpreted EEG, an electroclinical diagnosis was made. The questionnaire data were compared with this clinical gold standard.

A total of 59 children participated: 28 from Tanzania and 31 from Zambia. Sixteen patients were excluded: 5 were excluded because of incomplete data, and 11 did not meet criteria for epilepsy based on initial screening questions. Of the remaining 43 patients, 28 had focal or multifocal epilepsy (65%), and 15 (35%) had generalized epilepsy. The questionnaire had a sensitivity of 78% and positive predictive value of 81.5%. Data were analyzed using a Rasch model, testing the questionnaire's internal consistency, reliability, and its discriminative validity in classifying focal versus generalized epilepsy against an electroclinical diagnosis. The mean epilepsy score for focal epilepsy was 0.084 logits compared with -1.147 logits for generalized epilepsy, demonstrating a large effect size [F (1, 41)=13.490, p<0.001].

Our questionnaire provides a straightforward method to improve diagnostic accuracy, and could assist in bridging the diagnostic gap in pediatric epilepsy in resource-limited settings. This tool was specifically designed to be easily implemented by any healthcare provider. This pilot study prompts broader prospective validation in additional settings for further refinement, and for performance assessment of impact on provider's practice, ability to guide medication choices, and ultimately improve treatment outcomes in resource-limited regions.