With growing diagnostic and therapeutic advances in neurology, stewardship of limited resources is of increasing importance, particularly in the Canadian public healthcare context. One such limited resource is electroencephalography (EEG), the availability of which is geographically disparate in Canada ^{Reference Niznick, Tang and Kromm1} and affected by labour shortages. ^{Reference Van Winssen, Andrade and Andrade2} Evidence on EEG yield is mixed, and there is a paucity of data reflecting current Canadian practice patterns.

In this study, Patrick and colleagues performed a 6-month retrospective audit of all ambulatory and inpatient short-duration EEGs performed at Kingston General Hospital, evaluating yield (defined as epileptiform and non-epileptiform abnormalities) across settings, requesting providers, indications and patient characteristics. ^{Reference Patrick, Smith and Wang3} The study location allowed the authors to evaluate a large number of EEGs at a single centre with a wide catchment area, facilitating comprehensive assessment across age groups and clinical settings.

The authors found that approximately 40% of EEGs were abnormal and only 15.3% had epileptiform abnormalities. Unsurprisingly, yield varied by clinical indication, with the lowest yield seen in EEGs performed for syncope/presyncope, supporting existing Canadian and UK resource stewardship recommendations. ^4,5 Clinical setting, antiseizure medications/sedatives and age also influenced yield. While pursuit of sleep-deprived EEG did not increase epileptiform yield, the sample size was small. One limitation of note was that latency from event to EEG capture was not recorded; it is possible that this may have influenced diagnostic yield, particularly in outpatients.

It is important to consider that yield does not necessarily equate to usefulness, a distinction that is appropriately discussed in the manuscript. There are circumstances where an abnormal EEG may not impact management (e.g., non-specific slowing, or re-demonstration of abnormalities in a patient with known epilepsy). Conversely, a normal EEG capturing paroxysmal events may influence their characterization. This underscores the role of EEG as an ancillary test that does not replace clinical judgement. Further studies on specific indications are therefore needed to investigate clinical utility.

The authors also evaluated the completeness of paper EEG requisition forms, finding that 56% of submissions were incomplete. These findings guided redesign of a streamlined requisition form, highlighting an excellent example of a data-informed approach to process change.

This study provides valuable insights on EEG utilization in a Canadian healthcare setting. Currently, Choosing Wisely Canada recommendations suggest avoiding routine EEG for syncope with a normal neurologic exam, ⁴ but more specific Canadian guidelines on EEG utilization are lacking. The findings of the present study can therefore guide future resource stewardship efforts.