Insomnia Severity Index Documentation: Scoring & Sample Note

Both versions are authorized, and the chart must say which. The current distributor catalog calls the last month the standard interval and expressly accepts a two-week interval, especially for shorter treatment studies and CBT-I monitoring, provided the same interval is used at baseline, post-treatment, and follow-up; the 2011 psychometric paper used the one-month window while most CBT-I clinical forms use two weeks, which is why so many pages present two weeks as the only ISI. Patient, clinician, and significant-other versions exist and give complementary information, and the ISI-3 is a separate three-item derivative with its own two-week and one-month forms. Name the respondent version, the recall interval, the language and country version, and the mode at every administration, and never compare a two-week result with a one-month result as if they were the same measurement.

Whichever rule your episode names, because the numbers have different provenances. The six-point figure is a minimally important difference estimated from an eszopiclone trial of 828 adults using quality-of-life, fatigue, and productivity anchors, whose authors called for study of generalizability. The eight-point-or-more reduction is the responder convention from Morin's CBT and medication trial, and a current score below 8 is that trial's operational remission rule; both reappear in the FDA summaries for digital CBT-I products, and one dossier used a six-point threshold in its prespecified analysis and eight points post hoc. The 2011 study found a mean change of about minus 8.4 for moderate improvement rated by an independent evaluator, supporting evidence rather than a rule; a five-point figure published in 2024 applies to a modified six-item score, not the standard total; and a 2024 review of 81 trials found within-person thresholds ranging from 3 to 8 points. So write "minus 6, exceeds the six-point MID estimate; does not meet this episode's eight-point response rule; above the below-8 remission threshold," and never "clinically significant" without a source.

Both results and the discordance, then the examination of it. Check diary completeness and timing, sleep opportunity and schedule, arithmetic, clock-checking, medications, adherence to the sleep window and stimulus control, acute stress, napping, and whether the ISI recall window and the diary week actually overlap. Then ask what drove the ISI change: less distress, greater satisfaction, better daytime function, less perceived interference, or true nocturnal improvement. Moderate convergence with diaries and weak convergence with polysomnography are expected because the ISI measures perceived severity and impact (in the 2011 clinical sample the total correlated about .59 with diary wake time and mostly not at all with polysomnography), so do not erase a real reduction in distress because sleep efficiency has not normalized, and do not claim sleep-continuity recovery when the diary is still impaired. A defensible sentence: "ISI improved by six points, driven clinically by lower distress and daytime interference; diary sleep efficiency and nocturnal wakefulness remain impaired, so behavioral treatment continues."

Call it operational ISI remission under the below-8 convention, and say so in those words. Then separately state whether the patient currently meets interview-based insomnia criteria, whether diary sleep and daytime function have recovered, what residual symptoms remain, whether medication or another active treatment is maintaining the result, and whether OSA, a circadian disorder, mood symptoms, or another condition is still active. The below-8 rule comes from treatment-trial conventions and the digital-CBT-I regulatory summaries; it is not a stand-alone diagnostic determination, and "cured" is not a chart word. At completion, record the diary, functional recovery, residual symptoms, comorbidity status, and the relapse plan with recontact thresholds beside the classification.

It depends on who you are and what you are building. The ISI is copyright of Charles M. Morin and distributed by Mapi Research Trust: an individual clinician, physician, student, or non-funded academic can register for direct access and use the authorized paper form under its terms at no fee; funded academic projects, healthcare organizations, commercial users, and IT companies submit a project request, and a licence and fees may apply. Electronic versions are treated as derivative works, so embedding the wording, response controls, or scoring in an EHR flowsheet, patient portal, app, or digital therapeutic requires its own permission process with screenshot review, and a free public calculator hosted by a commercial entity is commercial and electronic use whatever it charges; the web-based ISI was validated in 2011, but validation is not permission. Direct access is not public domain: no retyping, redistribution, modification, translation, or online posting without authorization, and self-translation by a bilingual clinician is not authorized. Charting "ISI 12/28, two-week patient version" is documentation of a result, not reproduction. Terms as stated on the distributor's catalog in June 2026; no public fee schedule is published.

Guidelines name it as an example; no payer reviewed requires it. The AASM's 2021 guideline strongly recommends multicomponent CBT-I and treats response and remission as critical outcomes without naming an instrument; its 2025 quality measures name the ISI and PSQI as acceptable questionnaires and expect a clinician's global summary rather than a filed form; the 2025 VA/DoD guideline names the ISI as an example screening instrument that identifies who needs further evaluation. No national Medicare source names the ISI as required for 96127 or adopts its thresholds as coverage rules, so billing turns on payer, practitioner, setting, medical necessity, and documentation, and a score in the note does not itself create separate payment. Prescription digital CBT-I products (Somryst, cleared 2020; Sleepio, cleared 2024) used the ISI in their FDA dossiers, and Medicare covers qualifying digital mental health treatment devices when ordered, used incident to professional services, and incorporated into an ongoing plan of care under codes G0552 through G0554, with no ISI threshold; the clearance confers no licence and no mandate. In Australia, Better Access requires an outcome measure of the clinician's choosing (the ISI is not named), and RACGP guidance recommends the ISI with a diary; in Canada, no federal rule names it.

Yes for insomnia severity and treatment monitoring, no for diagnosing or grading OSA: document apnea testing, respiratory severity, PAP or other treatment, adherence, residual sleepiness, and driving or occupational safety separately, and note that a randomized trial of 145 patients with comorbid insomnia and OSA found CBT-I delivered before CPAP improved insomnia symptoms, CPAP acceptance, and average use. The instruments answer different questions: the ISI measures retrospective perceived insomnia severity, dissatisfaction, interference, visibility, and distress; the PSQI measures broader past-month sleep quality and disturbance across duration, efficiency, disturbances, medication use, and daytime dysfunction; the Epworth Sleepiness Scale measures propensity to doze and belongs in the sleepiness and safety picture; the STOP-Bang estimates apnea risk; and the prospective diary supplies the nightly variables that prescribe and adjust CBT-I. A sophisticated insomnia record uses the ISI plus the diary, adds the Epworth when sleepiness matters and the PSQI only when broad sleep quality is the construct, and verifies each instrument's own permissions.

Yes. Give it the facts (version and recall, total, baseline and date, the criterion your episode uses, the diary summary, adherence, medications, comorbid sleep disorders, and the plan) and it drafts the full entry: the score with its band as a label, the change against the named criterion, the diary as a separate stream, the reconciling interpretation, and the plan with the next administration dated, ready for your review. It can also check a finished note for a band charted as the diagnosis, an unnamed version or recall, six and eight points cited as one rule, below 8 written as cured, and a diary that is missing or overruled, and it never needs the instrument's items to do so. BastionGPT is HIPAA-compliant with a signed BAA on every plan, and your data is never used to train models.