Athens Insomnia Scale (AIS)

Both are brief, validated insomnia severity scales; the choice depends on context and availability. AIS (8 items, 0–24) is based on ICD-10 criteria and more common in Europe; ISI (7 items, 0–28) is more common in North America and research. AIS references the past month; ISI references the past 2 weeks. Both have good sensitivity/specificity for insomnia. For comparability across studies, decide a priori and use consistently; they are not interchangeable (different score ranges and timeframes).

AIS ≥6 indicates probable insomnia disorder based on ICD-10 criteria with ~89% sensitivity and ~86% specificity. However, the scale is not diagnostic; a score ≥6 warrants further clinical evaluation to confirm insomnia disorder diagnosis (symptom duration ≥3 months, daytime impairment, exclusion of sleep apnea). A patient with AIS = 6–7 may have mild, situational sleep disturbance or early insomnia, not yet meeting full clinical criteria. Combine with clinical interview and sleep history.

Item 4 is subjective self-report of estimated total sleep duration (e.g., '5 hours per night'). It is prone to recall bias and perception bias; patients often underestimate or overestimate sleep duration. If sleep duration estimate on AIS is very low (<5 hours) or discrepant from what the patient reports elsewhere, consider asking about a sleep diary (1–2 week log of sleep times) or actigraphy (wristband recording). Objective measurement may clarify whether the patient truly has shortened sleep or perceives sleep as shorter due to poor quality.

AIS is designed for individuals with regular sleep schedules and may not be appropriate for shift workers, who have forced irregular timing and may score high even without insomnia disorder. For shift workers, assess whether insomnia symptoms occur when attempting to sleep during their scheduled sleep window (despite circadian misalignment) versus reflecting circadian desynchrony alone. Consider the Circadian Rhythm Sleep Disorder assessment or specialized shift worker measures. Standard AIS cutoffs may overestimate insomnia in this population.

A reduction of ≥4–5 points from baseline or a change from above-threshold (≥6) to subclinical (<6) indicates treatment response. For example, a patient with baseline AIS = 14 (moderate) improving to AIS = 6–8 shows meaningful change. Ideally, achieve AIS <6 (no insomnia). In CBT-I trials, typical responders show 50% reduction from baseline within 8 weeks. Monitor every 2–4 weeks during active treatment to detect early response or non-response.

AIS was developed for ICD-10 criteria but applies closely to DSM-5 insomnia disorder as well (both require sleep initiation/maintenance difficulty plus daytime impairment for ≥3 months). An AIS ≥6 is consistent with DSM-5 insomnia in most cases. However, AIS does not explicitly ask about symptom duration (≥3 months) or exclusion of substances/sleep disorders. For DSM-5 diagnosis, combine AIS with structured diagnostic interview to confirm timing and rule-outs. AIS is a severity measure, not a diagnostic instrument per se.