Design: Prospective single-centre diagnostic accuracy study
Population: 169 children listed for adenotonsillar surgery for sleep-disordered breathing
Primary outcome: Agreement of video with in-person Brodsky grading
Effect: Sensitivity 70%, specificity 86%; 12% of videos unusable

This prospective single-centre diagnostic accuracy study asked parents of 169 children (median age 4.4) listed for adenotonsillar surgery for sleep-disordered breathing to film the oropharynx with a handheld telemedicine device. Three blinded paediatric ENT surgeons graded tonsil size on the Brodsky scale from the videos.

Agreement between the three graders was strong. Against in-person examination, the videos identified large tonsils (grade 3–4) with 70% sensitivity and 86% specificity. Twelve per cent of videos were unusable to all three graders and nearly a third to at least one, more often in younger children.

Remote triage is attractive where families travel far for clinics. These numbers suggest video can help prioritise, but it misses about three in ten large tonsils and fails outright in one in eight children. It is a supplement to examination, not a replacement for the surgical decision.

Consider parent videos for remote triage of suspected sleep-disordered breathing.
Do not list a child for surgery on video assessment alone.
Give parents simple instructions on lighting and tongue position; poor videos were common in young children.
Bring children with unusable videos to clinic rather than re-filming repeatedly.

Why it matters

Remote first assessment could shorten waits, but its limits need to be clear before it shapes surgical lists.

The statistics, in plain English

Sensitivity of 70% means video missed about 3 in 10 children whose tonsils were large on examination. Specificity of 86% means about 1 in 7 with smaller tonsils were wrongly graded large. A kappa of 0.65 is 'substantial' agreement, not near-perfect.