Validation of a phone app for epilepsy diagnosis in India and Nepal

•We devised a diagnostic aid to determine whether episodes of altered consciousness are epileptic or not.•We presented this as a phone app.•We tested it in three populations and compared it to the diagnosis of an experienced clinician.•The app correlated well with clinical diagnosis.•This app could contribute to reducing the epilepsy treatment gap.

PurposeUntreated epilepsy is a major global public health problem with more than 20 million people not being treated for an easily treatable disease. In part this is due to a lack of trained doctors. There are many more non-medical health workers than doctors and they could have an important role in diagnosis and treatment of epilepsy if they had some tools. We have previously described such a tool to distinguish epileptic episodes from other causes of altered consciousness and here present its validation in three new populations.MethodsThe tool was presented as a phone app where the answers to 11 questions provided a probability score which indicated whether episodes might be due to epilepsy or not. It was applied either by non-medical volunteers, health workers, or inexperienced doctors to 132 patients in three separate populations in India and Nepal and compared with the “gold standard” diagnosis of a neurologist with expertise in epilepsy.ResultsThere was good agreement between the app score and the neurologists’ diagnoses (weighted kappa = 75.3%). An app score of 90 or greater had a sensitivity of 88% and a specificity of 100% for diagnosing epilepsy. The app was easy to use with little training and took about 5 min to administer.ConclusionA tool presented as a phone app can be used by non-medical health workers to identify episodes as epileptic or not with good accuracy. It needs to be evaluated more widely but has the potential to play a part in reducing the epilepsy treatment gap.