Sign Up
Login
Try for Free
Login
The CAGE questionnaire is a four-item, yes-or-no alcohol screen published by John A. Ewing in JAMA in 1984, worded for lifetime experience and built to detect dependence rather than current risky drinking. Primary care, emergency, and behavioral health clinicians use it, and the CAGE-AID variant for drugs, to decide who needs a fuller assessment. This page covers how to document and interpret CAGE results, with a fictional sample note.
Any clinician or trained staff member can ask and score it (verbally or on an intake form); no license, training requirement, or publisher qualification level applies; the treating clinician interprets and documents
Primary care and emergency physicians, nurses, addiction and behavioral health clinicians, trauma-center SBIRT teams, payers and auditors reviewing screening and brief-intervention claims
3 to 6 chart lines (version, time frame, item-level answers and total, current use, follow-up, intervention) · administration about 1 minute
Brief case-finding screen (4 yes-or-no items, one point each, total 0 to 4, lifetime wording; the CAGE-AID asks the same four topics about alcohol or other drugs)
Intake packets and annual visits in primary care, emergency and trauma-center screening, addiction and behavioral health intakes, and any setting that wants a fast dependence-oriented probe; not the preferred routine screen for unhealthy drinking
Ewing 1984 (JAMA) under a Bowles Center permission statement (cite the source; profit-making use needs permission); CAGE-AID Brown and Rounds 1995; USPSTF 2018 and RACGP 2024 name the AUDIT-C for routine screening; described here for documentation, no items reproduced
The CAGE is a four-item alcohol screen developed by John A. Ewing and colleagues at the University of North Carolina at Chapel Hill, first presented at a 1970 meeting in Sydney, validated by Mayfield, McLeod, and Hall in the American Journal of Psychiatry in 1974, and given its canonical exposition by Ewing in JAMA in 1984, which is the citation every reproduction asks for. Each item is answered yes or no and scores one point, so the total runs 0 to 4. The acronym comes from the four topics the items cover: cutting down, annoyance at criticism, guilt, and morning drinking. The items are worded for lifetime experience, so the instrument finds people who have accumulated the consequences of dependence; it asks nothing about how much anyone drinks now. The CAGE-AID (CAGE Adapted to Include Drugs; Brown and Rounds, Wisconsin Medical Journal, 1995) asks the same four topics about alcohol or other drugs together, so its total is neither an alcohol score nor a drug score. Pregnancy-specific descendants, the T-ACE (1989) and the TWEAK (1994), exist because the CAGE performs poorly in prenatal populations.
Three facts carry the documentation. First, a CAGE total is a screening result, never a diagnosis: two or more yes answers is the conventional positive threshold (Ewing 1984), a SAMHSA consensus panel recommended one in primary care to cast a wider net, and no authority maps 0 to 4 onto severity bands or onto DSM-5 alcohol use disorder, which needs its own criteria-based substance use assessment. Second, the time frame decides the meaning: a lifetime positive in a patient abstinent for ten years is a true positive for history and says nothing about current use, and a version reworded to the past 12 months is a modified instrument that must be charted as such. Third, the CAGE and the CAGE-AID are different instruments with different authors, rights holders, and conventions (the University of Washington ADAI overview treats one or more CAGE-AID yes answers as a positive screen), so the version belongs in every entry. For routine adult screening the USPSTF (2018) and Australia's RACGP (2024) now name the AUDIT-C or a single consumption question, because the CAGE detects dependence rather than the full spectrum of unhealthy drinking; the DAST-10 is the structured drug analogue.
The CAGE survives in three places. Primary care intake packets and annual-visit flowsheets still carry it as a legacy item, so family physicians, internists, and nurses chart CAGE results even where the practice's formal screen is now the AUDIT-C or the NIAAA single question. Emergency departments and verified trauma centers, which must screen admitted patients and deliver a documented brief intervention, use the CAGE or the CAGE-AID when a fast verbal probe fits the bedside. Addiction, psychiatric, and behavioral health intakes use it as a dependence-oriented history item beside the fuller substance use assessment, biopsychosocial assessment, or intake note, and the CAGE-AID is the choice when drug involvement is suspected or a program screens for alcohol and drugs in one pass. Outside those uses the neighbors win: the AUDIT-C for current hazardous drinking and quality-measure reporting, the DAST-10 for drug-problem severity, a pregnancy-specific screen in prenatal care, and the CIWA-Ar or COWS when the question is withdrawal severity, which no CAGE result can answer.
No law, payer, or author prescribes a CAGE note format; what exists is a permission statement asking for the Ewing 1984 citation, and payer and accreditation rules that ask for a named validated tool, a result, the time spent, and an intervention. What survives review is an entry that names the version, states the time frame actually asked, records the four item-level answers with the total, charts current quantity and frequency beside the screen, labels the result a screen, names the follow-up assessment it triggered, and documents the intervention with its minutes. Each element below carries the pitfall that most often undermines it.
Version: CAGE or CAGE-AID. Chart which instrument was used, in words a later reader cannot mistake: CAGE (original, alcohol only) or CAGE-AID (alcohol or other drugs, including non-medical use of prescription drugs). The two carry different authors, rights holders, validation figures, and positive conventions, and a CAGE-AID yes may concern alcohol on one item and a drug on another, so record which substance produced each positive answer when the patient can say. Pitfall: "CAGE 2/4" when the CAGE-AID was administered, or a CAGE-AID total copied into the problem list as an alcohol result.
Time frame actually asked. The original items ask about lifetime experience. Write "lifetime frame" when the questions were asked as worded, and "modified, past 12 months" (or whatever window was used) when the clinician narrowed them; a narrowed version is a different instrument with different operating characteristics, not the original with a footnote. When lifetime positives refer to a resolved period, date them ("all endorsements precede the patient's sobriety date") and add the current-status line described below. Pitfall: A lifetime positive in a patient sober for years, charted so that the next reader believes the patient is drinking now.
Item-level answers, then the total. Record the four answers in order with the total (C yes, A no, G yes, E no; 2 of 4) rather than the total alone. Which items were endorsed changes the picture: a morning-drinking yes points toward physiological dependence and a withdrawal history, guilt alone often reflects a remote episode, and a single yes is a prompt for follow-up under the SAMHSA wider-net convention rather than a negative. Keep the patient's explanation next to any yes, especially a morning-drinking answer from a shift worker whose "morning" is the end of a night shift and whose drink may have nothing to do with relieving withdrawal. Pitfall: "CAGE positive" or "CAGE 2" with no item pattern, so the entry can be neither interpreted nor reproduced.
Current quantity and frequency, separately. The CAGE has no consumption item, so add the patient's own statement of drinks per typical week, heavy-drinking days, the last drink, and any drug use with route and frequency, on its own line and in their words. This is the line that separates a historical positive from present risk, that Medicare's misuse benefit and the SBIRT codes are built around, and that Canadian and Australian guidance reads against the national weekly limits. Pitfall: A positive screen with no consumption statement, leaving the reader unable to tell current hazardous drinking from remote history.
Interpretation: a screen, in a named population. State the convention applied (positive at two or more yes answers, or at one under the primary-care wider-net convention) and write the conclusion as a screening result: "positive screen; diagnosis not established." Name the population when it changes the reading: at the conventional threshold, sensitivity is low for hazardous drinking in older adults, for women in several primary-care samples, and in pregnancy, so a negative CAGE in those groups rules little out, and a translated version does not import the English threshold's validity. Pitfall: "CAGE 3/4, alcohol use disorder" with no DSM-5 criteria assessed, or "CAGE negative, no alcohol concerns" in a 70-year-old who drinks daily.
Follow-up assessment triggered. A positive screen starts a sequence, so name what happened next and its result: an AUDIT-C or full AUDIT to characterize current drinking, a DAST-10 or TAPS for a drug-related CAGE-AID yes, a DSM-5 criteria-based assessment when a disorder is suspected, and a withdrawal assessment (CIWA-Ar) when current heavy use or a morning-drinking history raises that risk. A negative screen in a patient reporting heavy use also needs a next step, because the CAGE misses risky drinking by design. Pitfall: A positive CAGE with no follow-up instrument or assessment named, or a diagnosis inferred from the screen alone.
Intervention, minutes, referral, and record placement. Document the brief intervention (feedback given, the patient's goal, advice offered), the referral made or declined, or the reason no intervention was indicated, and the face-to-face time whenever a time-based code is claimed. Record where the result lives: a primary-care CAGE is an ordinary medical-record entry under HIPAA, while the same result created by a federally assisted Part 2 program is a Part 2 record with its own consent and redisclosure rules. Pitfall: "Counseled on alcohol" with no content, no minutes, and no goal, which supports neither the claim nor the next visit.
CAGE / CAGE-AID DOCUMENTATION BLOCK Date: [ ] Setting: [ ] Clinician: [ ] Encounter: [annual / intake / ED or trauma / prenatal / other] Version: [CAGE, alcohol only / CAGE-AID, alcohol or other drugs] Time frame asked: [lifetime, as worded / modified: state the window] Mode: [asked verbally / self-completed on intake form] Language: [ ] Item-level answers: C [ ] A [ ] G [ ] E [ ] Total: [ ] of 4 (CAGE-AID: substance behind each yes, when the patient can say) Timing of positives: [current / all precede (date or period); explanation] Current use (patient's words): [drinks per typical week; heavy days; last drink; drug use, route, frequency; or none, since (date)] Interpretation: [positive / negative under (convention named); a screen, not a diagnosis; population caveat: older adult, women, pregnancy, translation, or none] Follow-up triggered: [AUDIT-C / AUDIT / DAST-10 / TAPS / DSM-5 criteria assessment / withdrawal assessment; result and date] Intervention: [feedback given; patient's goal; referral made, declined, or not indicated; face-to-face minutes] Plan: [re-screen instrument and date; treatment or counseling; next visit] Record: [code claimed (bare number); storage location; Part 2 status if created by a Part 2 program; consent on file] Clinician signature / credentials: Date:
Free to use and share, no signup. The PDF includes a one-page cheat sheet with element-by-element pitfalls and a pre-sign checklist; the DOCX is the blank documentation block, ready to adapt. Neither reproduces the questionnaire items.
Scenario: a US family medicine annual visit where the practice's intake packet still carries the CAGE; the lifetime-frame screen returns two yes answers, the physician adds a current consumption statement, follows with an AUDIT-C, delivers a brief intervention, and records the time, the claim, and the plan. All details are fictional.
Patient: J.M., 58 · Setting: Family medicine, annual preventive visit · Clinician: A. Delgado, MD · Note date: 09/14/2026
Screen: CAGE, original alcohol-only version, self-completed in English on the practice intake form on 09/14/2026 and reviewed verbally during the visit; lifetime frame as worded. Item-level answers: C yes, A no, G yes, E no; total 2 of 4, positive at the conventional threshold. J.M. places both endorsements in the past two years and describes them as current: two attempts to cut back without lasting change, and regret about drinking more than intended on weekends. No morning drinking and no symptoms on alcohol-free days. No prior CAGE or AUDIT on file.
Current use: In the patient's words: 3 to 4 standard drinks (beer, sometimes bourbon) on 4 or 5 evenings a week, 6 or more on about two Saturdays a month, last drink last night; about 16 to 20 drinks in a typical week, above the NIAAA weekly limit for men. No cannabis, no other drug use, no non-medical use of prescription medication. Denies driving after drinking, injuries, or work problems related to alcohol. Reports poor sleep; blood pressure 146/92 today, higher than at last year's visit.
Interpretation: Positive CAGE in the lifetime frame with both endorsements current, plus a self-reported pattern above recommended limits: consistent with current unhealthy alcohol use. The CAGE is a screen and does not establish a diagnosis; DSM-5 alcohol use disorder criteria were not formally assessed today and will be reviewed at follow-up. No population caveat applies (58-year-old man, English, original wording).
Follow-up assessment: AUDIT-C administered verbally in the same visit to characterize current drinking: 7 of 12, positive under the practice protocol for men; item results on the screening flowsheet. Withdrawal risk judged low (no morning drinking, no prior withdrawal, no tremor or sweating on alcohol-free days); CIWA-Ar not indicated. Hepatic panel and GGT added to today's labs.
Brief intervention: Delivered by the physician: feedback on both screen results and where the reported pattern sits against the NIAAA limits, the likely contribution to poor sleep and the rise in blood pressure, and a menu of options. J.M. chose a goal of no more than 2 drinks on any day and two alcohol-free days each week starting this week, and accepted a printed drink-tracking sheet. Declined referral to counseling for now and knows the offer stands. Face-to-face time for the screen review and intervention: 18 minutes.
Plan and record: Nurse phone check in 2 weeks; office visit in 6 weeks to review the tracking sheet, repeat the AUDIT-C, recheck blood pressure, and assess DSM-5 criteria if the pattern persists, with medication options for alcohol use disorder discussed then if criteria are met. Screening and brief intervention claimed under 99408, supported by the tool, result, intervention, and time documented above. Result recorded in the general medical record; this practice is not a Part 2 program, so no Part 2 consent applies. J.M. was told the screen is not a diagnosis and agreed with the plan.
This sample is fictional and for educational purposes. It does not describe a real patient or record; the scores, dates, and details are invented to show documentation structure and are not clinical guidance.
Writing these after every session? BastionGPT drafts complete notes from bullets, dictation, or a transcript.
Generate a note from bulletsUnited States: the instrument is a convention; the rules around it are real. No federal or state LAW names the CAGE, and no law prescribes a screening note format. The USPSTF recommendation of 2018 (grade B; CONVENTION, with quality-measure teeth) covers screening adults 18 and older, including pregnant women, for unhealthy alcohol use, names the AUDIT-C and the NIAAA single question as the most accurate brief instruments, and passes over the CAGE because it detects dependence rather than the full spectrum of unhealthy use; NIAAA's clinician resource points to the same two tools, and the 2018 statement remains current as of September 2026 with no finalized replacement. The USPSTF's 2020 drug-use recommendation (grade B; adults 18 and older; screening by asking, not by testing specimens, and only where care can be offered) evaluated past-year single items, the NIDA Quick Screen, TAPS, and the DAST-10 rather than the CAGE-AID. Under PAYER POLICY, Medicare's alcohol misuse benefit (NCD 210.8, 2011) pays an annual screen under G0442 and up to four counseling sessions a year under G0443 for beneficiaries who misuse alcohol without meeting the decision's dependence criteria, which is the pattern the CAGE is least built to find; the SBIRT codes (99408 and 99409 for commercial plans, G0396 and G0397 for Medicare, H0049 and H0050 where a state Medicaid program adopts them) are time-based and need a named validated tool, the result, the face-to-face time, and, for the intervention codes, the intervention itself, so a bare "CAGE 2/4" supports none of them, and 99408 is not a Medicare substitute for G0442. Accreditation and quality CONVENTIONS ask for the same things: the Joint Commission SUB measure set (2026 specifications manual, as of September 2026) requires alcohol screening with a validated tool by the end of hospital day one for hospitals reporting it; MIPS quality measure 431 (2026) counts systematic screening for unhealthy alcohol use with brief counseling when it is found; and American College of Surgeons trauma verification standards (2022 standards, Q&A current in September 2026) require verified trauma centers to screen at least 80 percent of admitted patients older than 12 and to deliver a documented brief intervention, or a referral where inpatient intervention is unavailable, to at least 80 percent of those who screen positive, with the operative wording for your level in the current Resources for Optimal Care manual. Under LAW, 42 CFR Part 2 as amended by the February 2024 final rule (effective April 16, 2024; compliance date February 16, 2026, now passed, with HIPAA-style breach notification and civil enforcement) attaches to records created by a federally assisted Part 2 program, so a CAGE-AID recorded by such a program is a substance use disorder record with its own consent and redisclosure rules, while an ordinary primary-care CAGE result is a HIPAA record and nothing more; the Part 2 consent page covers the mechanics.
Canada and Australia: no requirement, and reference limits a positive screen should be documented against. In Canada, no federal or provincial LAW or payer names the CAGE. Canada's Guidance on Alcohol and Health (Canadian Centre on Substance Use and Addiction, January 2023; CONVENTION) replaced the 2011 low-risk drinking guidelines with a continuum: 2 or fewer standard drinks a week is low risk, 3 to 6 is moderate, 7 or more is increasingly high, a Canadian standard drink holds 13.45 g of ethanol, and there is no known safe amount in pregnancy or when trying to conceive. CAMH's guidance for professionals asks for a recent alcohol history (drinks per week and the most in a day) and a validated tool such as the AUDIT-C or AUDIT, with assessment rather than diagnosis by screen after a positive result. A current, instrument-specific position from the Canadian Task Force on Preventive Health Care or the College of Family Physicians of Canada could not be confirmed for this page, so none is claimed. In Australia, the RACGP preventive-activities guidance (alcohol recommendation dated June 28, 2024; CONVENTION) recommends screening adults 18 and older, including in pregnancy, every two years and names the AUDIT-C; the NHMRC guidelines (December 2020) set the reference limits at no more than 10 standard drinks a week and 4 on any one day, with a standard drink at 10 g of ethanol and no alcohol under 18, in pregnancy, or while planning a pregnancy; no MBS item and no state law names the CAGE. Neither instrument carries a qualification level or purchase restriction: any clinician may ask the four questions, and the boundary is attribution and non-commercial use, covered in the next paragraph. In every jurisdiction a translated CAGE is charted with its language and the convention actually applied, because a Spanish primary-care study found a study-optimal threshold of one and a Brazilian Portuguese pregnancy study found the conventional threshold failed its construct-validity test.
The CAGE questionnaire is the work of John A. Ewing, published in JAMA in 1984, and is made available under a Bowles Center for Alcohol Studies (University of North Carolina at Chapel Hill) permission statement, dated February 2004, that asks users to cite Ewing 1984 and reserves any profit-making use for negotiated permission; a 2014 Occupational Medicine questionnaire review states instead that reproduction requires permission from JAMA, and no current source resolves that conflict or grants blanket commercial, EHR-vendor, or public-web rights, so this page does not call the CAGE public domain, and a commercial product should clear rights before reproducing it. The CAGE-AID is copyrighted through the Wisconsin Medical Journal (Brown and Rounds, 1995) and is conventionally reproduced with the journal's permission and the authors' consent for non-commercial clinical use. BastionGPT is not affiliated with, or endorsed by, any of these publishers. This page reproduces no test items, stimuli, norms, or scoring materials.
The numbers behind these errors are specific. A diagnostic meta-analysis of 10 studies and 4,562 patients found pooled sensitivity at the conventional threshold of 0.87 in inpatients, 0.71 in primary care, and 0.60 in ambulatory patients and concluded the CAGE has limited value for screening at that threshold (Aertgeerts, Buntinx, and Kester, 2004); a 2007 review put average sensitivity at 0.71 and specificity at 0.90 for abuse or dependence and reported sensitivity of 0.38 in one primary-care sample of white women (Dhalla and Kopec); in older primary-care patients the conventional threshold identified 14 percent of hazardous drinkers with 97 percent specificity (Adams, Barry, and Fleming, JAMA, 1996), and in adults over 60 moving from two yes answers to one shifted sensitivity and specificity from 70 and 91 percent to 86 and 78 percent (Buchsbaum and colleagues, 1992). The 93 percent sensitivity that consumer pages repeat traces to the 1974 psychiatric-inpatient validation (366 admissions, 39 percent alcoholic), not to primary care. No published audit counts how often the version or the time frame is missing from CAGE chart entries; the SBIRT and Medicare documentation rules described under compliance considerations are what turn those omissions into a claim problem. The BastionGPT Clinical Advisory Board sees the same errors most often in CAGE documentation reviews:
BastionGPT is specifically trained, tuned, and clinically tested on substance use assessments and screening documentation.
See how clinicians use it day to day on the AI therapy notes page.
Many BastionGPT users report saving more than 90 minutes per day on documentation.
HIPAA-compliant with a signed BAA on every plan. Your data is never used to train models. BastionGPT drafts, you review and sign.
Each of the four items is answered yes or no and each yes scores one point, so the total runs 0 to 4. Two or more yes answers is the conventional positive threshold (Ewing, JAMA, 1984; the 1974 validation found a threshold of two or three useful and four alone too insensitive). A SAMHSA consensus panel for primary care recommended lowering the threshold to one yes to cast a wider net, and in adults over 60 that move raised sensitivity from 70 to 86 percent while dropping specificity from 91 to 78 percent, so a single yes is a prompt for follow-up rather than a negative. There are no severity bands: pages that grade 2 as moderate and 3 as severe are inventing a scale no CAGE publication supports, and no score establishes DSM-5 alcohol use disorder, which needs its own criteria-based assessment. The CAGE-AID follows its own convention, with the University of Washington ADAI overview treating one or more yes answers as a positive screen. Chart the convention you applied and the item pattern, never the total alone.
Under the lifetime frame, yes: the items ask about experience ever, so remote endorsements never expire and a lifetime CAGE can stay high indefinitely after recovery. That is a true positive for history and no evidence of current use. Chart it so the two cannot be confused: "CAGE, original alcohol-only version, lifetime frame; C yes, A yes, G yes, E no; 3 of 4. All endorsements precede the patient's sobriety date; reports no alcohol in nine years; current consumption screen negative. Lifetime result documents remote alcohol history and does not establish current drinking." Put the current status and its duration on its own line, keep the fuller history in the substance use assessment or a relapse prevention plan rather than in the screen line, and if you narrowed the questions to the past 12 months instead, chart the result as a modified instrument, because the operating characteristics change with the window.
The CAGE (Ewing, 1984) asks its four items about alcohol only. The CAGE-AID, CAGE Adapted to Include Drugs (Brown and Rounds, Wisconsin Medical Journal, 1995), asks the same four topics about alcohol or other drugs together, including non-medical use of prescription drugs. That single change means a CAGE-AID total is not substance-specific: two yes answers can reflect alcohol on one item and cannabis on another, so the chart needs the substance behind each positive when the patient can say. The instruments also differ in authorship, rights holders, validation (in the original primary-care study of 124 patients the conjoint version was more sensitive and less specific than the alcohol-only CAGE), and positive convention, with the CAGE-AID overview treating one or more yes answers as a positive screen. Name the version in every entry, never carry a CAGE-AID result into the problem list as an alcohol finding, and follow a drug-related yes with the DAST-10 or TAPS so the drug side of the result gets its own assessment.
For routine screening of adults for unhealthy alcohol use, the AUDIT-C or the NIAAA single question. The USPSTF's 2018 recommendation names those two among the most accurate brief instruments and passes over the CAGE because it detects dependence rather than the full spectrum of unhealthy use; NIAAA's clinician resource points to the same two; Australia's RACGP (2024) names the AUDIT-C for two-yearly screening; and CAMH's Canadian guidance asks for a recent consumption history plus the AUDIT-C or AUDIT. Medicare's misuse benefit and MIPS measure 431 are built around current unhealthy drinking short of dependence, which the CAGE is least designed to find. The CAGE keeps a role as a fast dependence-oriented probe, a legacy intake item, and historical case-finding information, and a positive CAGE still deserves a consumption screen so the note shows current risk rather than history alone.
Not on the score alone. The SBIRT codes (99408 and 99409, G0396 and G0397, and H0049 and H0050 where a state Medicaid program adopts them) and Medicare's G0442 and G0443 are PAYER POLICY: they need a named validated instrument, the result, the face-to-face time for the time-based codes, and, for the intervention codes, the brief intervention itself documented with its content. The CAGE is a validated instrument, so a note that names it, records the item-level result and the time frame, adds the consumption statement, and documents the intervention with its minutes can support a claim; "CAGE 2/4, counseled" cannot. Two further cautions: Medicare's benefit under NCD 210.8 targets misuse that does not meet its dependence criteria, so an AUDIT-C or single-question result is the better documentation fit, and 99408 is not a Medicare substitute for G0442. Medicaid rules are state-specific, so check the state's SBIRT policy before choosing a code.
Only if the record is created or received by a federally assisted Part 2 program. Part 2 follows the source of the record, not the word alcohol: an ordinary primary-care CAGE result is a HIPAA record and nothing more, while a CAGE-AID recorded at intake in a Part 2 program is a substance use disorder record with Part 2 consent and redisclosure rules. Information a patient conveys orally to a clinician outside a Part 2 program, who then documents it independently, does not become a Part 2 record for that reason alone. The February 2024 final rule (effective April 16, 2024; compliance date February 16, 2026, now passed) aligned Part 2 with HIPAA on breach notification and enforcement and allows a single consent for treatment, payment, and health care operations, while keeping special protection for SUD counseling notes. Document where the result lives and, in a Part 2 program, which consent is on file; the Part 2 consent page covers the forms and the workflow.
With the population caveat written into the interpretation. At the conventional threshold the CAGE identified 14 percent of hazardous drinkers among older primary-care patients (Adams, Barry, and Fleming, 1996) and 48 percent of older medical outpatients with abuse or dependence (Jones and colleagues, 1993); in predominantly white women sensitivity ran 38 to 50 percent across the studies Bradley and colleagues reviewed in 1998; and in pregnancy the instrument performs poorly enough that the T-ACE (1989) and TWEAK (1994) were built to replace it, with ACOG endorsing the T-ACE. A negative CAGE in these groups therefore rules little out, and the note should say so and pair the screen with a consumption statement or an AUDIT-C. In prenatal care, document why a pregnancy-specific screen was used instead (the clinical question is current exposure, not lifetime dependence history), record that instrument's result, and note the counseling that no amount of alcohol is known to be safe in pregnancy. In older adults, interpret a single yes seriously under the wider-net convention rather than waiting for two.
The four items cover, at the topic level, cutting down, annoyance at criticism, guilt, and morning drinking; that is where the acronym comes from. This page does not print or paraphrase the questions, because the rights position is contested: the Bowles Center for Alcohol Studies permission statement (2004) asks users to cite Ewing 1984 and requires negotiated permission for any profit-making use, a 2014 questionnaire review in Occupational Medicine states that reproduction requires JAMA's permission, and no current source grants blanket commercial, EHR-vendor, or public-web rights. For a non-commercial clinical form, the permission statement supports use with the citation, NIAAA and SAMHSA reproduce the instrument in free clinical materials, and the JAMA article is the source of record. For a commercial product, an ad-supported web tool, or an EHR vendor build, clear rights first. The CAGE-AID sits under separate copyright through the Wisconsin Medical Journal (1995) and is reproduced in toolkits with the journal's permission and the authors' consent.
Yes. Give it the facts (version, time frame as asked, the four item-level answers, the patient's consumption statement, any follow-up instrument and result, the intervention delivered and its minutes, and where the result is stored) and it drafts the full entry: the screen result with the convention named, current use on its own line, a screening-level interpretation with the population caveat, the follow-up, the intervention, the plan, and the record placement, ready for your review. It can also check a finished note for a missing version, a lifetime positive with no time frame or current-status line, a total without the item pattern, no consumption statement, a diagnosis inferred from the screen, or an intervention with no content or minutes. BastionGPT is HIPAA-compliant with a signed BAA on every plan, and your data is never used to train models.
The instrument facts and compliance claims on this page trace to these sources, last verified September 2026:
Educational content, not legal or billing advice. Sample notes are fictional. Follow your organization's policies and your board, payer, and jurisdiction requirements.