Personalized care of paediatric drug-resistant epilepsy in Africa: A single-centre pilot study utilizing mobile health and genetic testing.

The study was a small pilot with only 39 completers, limiting statistical power and generalizability. Heart rate data was excluded because the wearable device did not record it consistently. Engagement with the technology dropped significantly over time, with only about 60% of participants still using the device at six months, which may bias the later data. Clinicians were aware of the app's use, which may have influenced how they recorded seizures in clinic notes. The genetic panel had a low diagnostic yield (7.5%) compared to literature, possibly due to the specific panel design or the high frequency of non-genetic etiologies in this cohort. Selection bias was possible because inclusion depended on parental consent rather than clinical phenotype.

Two authors are employees or shareholders of Aparito, the company that helped design the mobile application used in the study. The work was funded by the Newton Fund, South African Medical Research Council, and the National Research Foundation of South Africa.

Do not conclude that the app accurately reflects true seizure frequency compared to clinic records. The lower counts in the app may reflect under-reporting by caregivers, recall bias in clinic records, or technical issues, and the study design cannot definitively separate these factors.