Sign Up
Login
Try for Free
Login
The MADRS (Montgomery-Åsberg Depression Rating Scale) is a clinician-rated, ten-item depression severity scale, published in 1979 and designed to be sensitive to change, with a total of 0 to 60. Psychiatrists, nurse practitioners, and research raters use it to track treatment-resistant depression, esketamine, TMS, and trial outcomes. It measures severity and change; it never diagnoses. This page covers how to document and interpret MADRS results, with a fictional sample note.
Trained clinician raters: psychiatrists, psychiatric nurse practitioners and nurses, psychologists, and research raters; no license or publisher qualification level applies, and reliability depends on training and a structured interview; the MADRS-S is completed by the patient
Interventional psychiatry, TMS, esketamine, and ketamine programs, prescribers managing treatment-resistant depression, payer utilization reviewers, clinical trial sponsors and monitors, referring clinicians, and quality abstractors, who will still need a PHQ-9
4 to 8 chart lines (version and rater, interval, item profile, total with the convention named, baseline and change with the rule named, separate risk line, decision) · interview about 15 to 20 minutes with a structured guide (clinical team estimate)
Clinician-rated depression severity scale (10 items scored 0 to 6 on defined and intermediate steps, summed to 0 to 60, higher is worse; the MADRS-S is a separate nine-item self-report with its own total)
Baseline and serial ratings in treatment-resistant depression, esketamine and TMS courses, ketamine and ECT programs, antidepressant trials, payer authorization and continuation packets, and measurement-based care where somatic symptoms confound a self-report
Montgomery and Åsberg 1979 (British Journal of Psychiatry); copyright claimed by Stuart Montgomery and the Royal College of Psychiatrists; severity bands and remission cutoffs are later conventions; accepted by many payers, mandated by none, not accepted by the CMS and HEDIS depression-remission measures; described here for documentation, no items reproduced
The Montgomery-Åsberg Depression Rating Scale (often written Montgomery-Asberg, and pronounced as the word MADRS) is a clinician-rated measure of depression severity published by Stuart A. Montgomery and Marie Åsberg in the British Journal of Psychiatry in April 1979 under the title "A new depression scale designed to be sensitive to change." The authors started from the 65-item Comprehensive Psychopathological Rating Scale, identified the 17 depressive symptoms most common in 54 English and 52 Swedish patients, and kept the ten items that changed most with antidepressant treatment and correlated best with overall change in 64 patients from four drug trials. Each item is rated 0 to 6, with descriptions at the even steps and intermediate points between them, and the ten ratings sum to a total of 0 to 60; there are no weights and no reverse-scored items. Described at the topic level, the items cover sadness as observed by the rater and as reported by the patient, inner tension, reduced sleep, reduced appetite, concentration difficulty, lassitude, reduced ability to feel, pessimistic thoughts, and suicidal thoughts. The original instructions let the rater choose the interval between ratings, weekly or otherwise, provided it is recorded, and the 1979 reliability work paired psychiatrists with general practitioners and with psychiatric nurses, so nothing in the instrument limits rating to physicians. Two derivatives are charted under their own names: the Structured Interview Guide for the MADRS, or SIGMA (Williams and Kobak, 2008), which supplies scripted probes, rates the past week, and reported a total-score intraclass correlation of 0.93 across 162 interviews by 81 rater pairs, and the MADRS-S (Svanborg and Åsberg, 1994), a separately published nine-item self-report with a recall window of about three days and its own scoring.
Three facts carry the documentation. First, the MADRS measures severity and change in a depression that has already been diagnosed by interview; it is not a diagnostic instrument, and no total confirms or excludes major depressive disorder. Second, the numbers that get attached to it are conventions with dates and authors, not parts of the scale: the 1979 paper has no severity bands, the bands in general use (essentially absent up to 6, mild from 7 to 19, moderate from 20 to 34, severe from 35) are attributed to a short 1986 report by Snaith and colleagues, Müller and colleagues (2003) put the moderate-to-severe boundary at 31, response is conventionally a reduction of at least 50 percent from baseline, and published remission cutoffs run from 4 to 12, so a defensible entry names the version, the rater, the interval, the convention, the baseline, and the rule applied. Third, the suicidal thoughts item is a single screening item, so any rating above zero triggers a separate suicide risk assessment or C-SSRS, and an improving total never stands in for one. The PHQ-9 is the neighbor most often confused with it: a patient-reported screening and monitoring scale that United States quality measures count and the MADRS cannot replace, where the MADRS earns its place as the clinician-rated outcome in treatment-resistant depression, esketamine and TMS programs, and trials.
Interventional psychiatry is the MADRS's home: esketamine clinics, TMS programs, ketamine and ECT services, and treatment-resistant depression consults rate it at baseline and at intervals through a course, because it was the primary outcome of the landmark TMS trial (O'Reardon and colleagues, 2007, week-4 change in 301 patients) and of the esketamine development program, and because US payers list it among the scales they accept for authorization and continuation. Psychiatrists, psychiatric nurse practitioners, and trained nurses rate it in those programs; research raters rate it, usually with the SIGMA and after calibration, in antidepressant trials, where the total is the endpoint. General outpatient psychiatry reaches for it when a patient's pain, sleep disorder, or medical illness inflates the somatic items of a self-report, or when minimization is suspected and a clinician's judgment should carry the rating. In Canada and Australia it lives in rTMS services and research rather than in any payer rule. The neighbors are fixed. The PHQ-9 remains the screening and quality-measure instrument and is what many TMS clinics track with day to day; the BDI-II is the psychology self-report; a rating above zero on the suicidal thoughts item is documented in a suicide risk assessment or with the C-SSRS; serial totals follow the conventions of the outcome measure note; the score supports the decisions recorded in the medication management note, the psychiatric consultation note, and the treatment plan review; and the baseline-and-current-score paragraph feeds the prior authorization request, which owns the letter mechanics while this page owns the scale-specific content those letters must carry.
No law, payer, or professional body publishes a MADRS note format; what exists is a 1979 paper with rating instructions, a 2008 structured guide, trial protocols, and payer criteria that ask for a baseline and a recent score on an accepted scale. What survives review is an entry that names the instrument, the rater, and the interview method; states the interval rated; describes the item profile at the topic level; gives the total with its severity convention cited; reports the baseline, the change, and the response or remission rule applied; documents a separate risk assessment whenever the suicidal thoughts item is rated above zero; and records the treatment context and the decision. Each element below carries the pitfall that most often undermines it.
Instrument, rater, and interview method. Write which instrument produced the number: the clinician-rated MADRS, a SIGMA-guided MADRS, or the MADRS-S self-report with its form and total range. Name the rater, their role, and their training or calibration, and the language or translation used. Clinician and self-report versions have different items, windows, and totals, and a structured interview changes reproducibility, so a reviewer, a payer, or a trial monitor needs this line before the total means anything. Pitfall: "MADRS 27" with no version, rater, or method, or a MADRS-S total submitted where the payer or protocol asked for the clinician scale.
Interval rated. State the period the rating covers and the days since the previous rating. The 1979 instructions let the rater choose the interval, weekly or otherwise, provided it is recorded; the SIGMA and most trial protocols rate the past week; the MADRS-S asks about roughly the past three days. Serial totals are comparable only when the interval is the same, so say what it was rather than assuming the reader knows the convention. Pitfall: A bare total with no interval, or weekly and monthly ratings trended as if they covered the same period.
Item profile at the topic level. Say which domains carry the total and which have resolved, in your own words: sadness as observed and as reported, inner tension, sleep, appetite, concentration, initiative and energy, emotional responsiveness, pessimism, and suicidal thoughts. The ten item scores live on the outcome flowsheet; the narrative describes the pattern, because a 24 carried by mood and pessimism and a 24 carried by sleep, appetite, and tension call for different decisions. Never paste anchor descriptions into the note. Pitfall: A total with no profile, or the scale's anchor text copied into the narrative as if it were the patient's account.
Total and the named severity convention. Chart the total over 60 and, if a severity label is used, the convention and its source: the grade scores of Snaith and colleagues (1986) are the usual reference, Müller and colleagues (2003) place the moderate-to-severe boundary lower, and some programs adopt their own. Write the label as severity and change, never as a diagnosis, and keep the diagnosis on the line where the evaluation that made it lives. Pitfall: "MADRS 33, moderate depression" as a diagnosis, or a band presented as the scale's official interpretation.
Baseline, change, and the rule applied. Give the baseline date and total on the same instrument, method, and rater type, the current total, the absolute change, the percent change, and the response or remission rule you are applying with its source (the plan's policy, the protocol, or CANMAT). State the result against that rule and keep the payer's continuation criterion separate from the remission convention; describe change from the previous visit rather than labeling it. Pitfall: "Responded" or "in remission" with no baseline, no percentage, and no rule, or a percentage computed from the wrong visit.
The suicidal thoughts item and the separate assessment. Record the item rating and treat any rating above zero, any rise, or any clinical concern as the trigger for a separate, dated risk assessment (the suicide risk assessment structure or the C-SSRS), with the safety plan status and where the formulation is filed. A falling total is not evidence that risk has resolved; the item is a screen, not the assessment. Pitfall: Risk inferred from an improved total, or the single item charted as the risk assessment.
Treatment context, decision, and companion records. Chart the treatment status at the time of rating (TMS session number, esketamine induction or maintenance week, medication changes since the last rating, adherence), the decision the score supports (continue, intensify, change, taper, request continuation), and what is filed separately: the REMS session record and Patient Monitoring Form for esketamine, the PHQ-9 kept for any quality measure that counts only the PHQ-9, and the authorization packet built from the baseline and current totals. Pitfall: A MADRS entered for a CMS or HEDIS remission measure that counts only the PHQ-9, or the REMS safety record folded into the outcome entry.
MADRS DOCUMENTATION BLOCK Date: [ ] Setting: [ ] Clinician: [ ] Visit: [baseline / serial rating / continuation request / trial visit] Instrument: [clinician MADRS / SIGMA-guided MADRS / MADRS-S (form and range)] Rater: [name, role, training or calibration] Language: [ ] Interval rated: [past 7 days / other: state it] Days since last rating: [ ] Item profile (topic level): [domains carrying the total; domains resolved; item scores on the flowsheet, no anchor text] Total: [ ] of 60 Convention: [label and source, e.g. Snaith and colleagues 1986; severity and change, not a diagnosis] Baseline: [date] [ ] of 60, same method Change: [ ] points, [ ] percent Rule applied: [response or remission definition and source: payer policy, protocol, guideline] Result: [meets / does not meet] Suicidal thoughts item: [rating; any rating above zero: separate dated risk assessment, tool used, safety plan status, filed at ...] Treatment context: [TMS session / esketamine phase, REMS record separate / medication changes since last rating / adherence] Interpretation: [congruence with interview and collateral; minimizing or over-reporting noted; confounders] Decision: [continue / intensify / change / authorization packet: baseline, current, percent change, instrument, rater, interval] Other measures: [PHQ-9 for quality measures, charted separately] 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 scale items or anchor descriptions.
Scenario: a US interventional psychiatry clinic's esketamine maintenance visit at week 12, where the psychiatric nurse practitioner rates a SIGMA-guided MADRS before the dose, compares it with the induction baseline under the plan's continuation criterion and the clinic's remission convention, documents the separate risk assessment the suicidal thoughts item triggered earlier in the course, and keeps the REMS session record and the PHQ-9 as their own documents. All details are fictional.
Patient: J.M., 46 · Setting: Interventional psychiatry clinic, esketamine maintenance visit (week 12) · Clinician: L. Okonkwo, PMHNP-BC · Note date: 09/16/2026
Measure: Clinician-rated MADRS administered with the Structured Interview Guide (SIGMA) by me before today's dose, in English; I completed the clinic's rater training and calibration exercise before the induction phase. Rating covers the past seven days; 7 days since the 09/09/2026 rating. All ten items rated; item scores are on the outcome flowsheet. Total 14 of 60. Induction baseline 08/04/2026, same method and rater: 34 of 60. End of induction 08/28/2026: 19 of 60. Prior rating 09/09/2026: 17 of 60.
Item profile: The total is carried by reduced emotional responsiveness, low initiative, and concentration difficulty. Observed and reported sadness, inner tension, pessimism, sleep, and appetite have improved to low ratings since induction. The suicidal thoughts item is rated 0 this week; it was rated above zero at induction and at the 09/09/2026 visit.
Total and convention: 14 of 60. Under the grade scores of Snaith and colleagues (1986), the convention this clinic cites, 14 falls in the mild range; at induction, 34 fell at the top of the moderate range under the same convention and above the moderate-to-severe boundary that Müller and colleagues (2003) proposed, which is why the source is named. The MADRS measures severity and change and does not establish a diagnosis; the diagnosis of recurrent major depressive disorder, treatment-resistant, rests on the intake psychiatric evaluation and is unchanged.
Change and rule applied: Absolute change from the induction baseline: 20 points; percent change: 58.8 percent reduction (34 to 14) on the same instrument, method, and rater. This meets the plan's continuation criterion, a reduction of at least 50 percent from baseline on a standardized depression rating scale with a score obtained within the past month. Remission under the convention this clinic has adopted from CANMAT, a MADRS total of 10 or less, has not been reached, so J.M. is charted as a responder and not in remission. Change from 09/09/2026: 3 points, described rather than labeled.
Risk: The suicidal thoughts item is a single screening item, not a risk assessment. Because it was rated above zero at induction and on 09/09/2026, a separate risk assessment was completed at each of those visits and was repeated today: C-SSRS screener negative for the past month, no ideation reported since 09/09/2026, no suicidal behavior during the course, protective factors unchanged (spouse, work, treatment engagement), safety plan from 08/04/2026 reviewed with J.M. and unchanged. Formulation and plan are filed in the risk section of the record.
Treatment context: Esketamine nasal spray, maintenance phase, week 12, dose and interval unchanged per the treatment plan; concurrent oral antidepressant unchanged since before induction; no other medication changes; adherence confirmed; alcohol and sleep reviewed and unchanged. Today's session safety documentation (observation start and end times, blood pressure at the intervals the program requires, sedation and dissociation checks, readiness for discharge) is on the REMS session record, and the Patient Monitoring Form will be submitted within the program's window; that record is separate from this outcome entry and does not depend on the MADRS.
Decision and record: Continue maintenance esketamine at the current interval. Continuation authorization request submitted today with the instrument and version, the rater, the interval rated, the induction baseline (34 on 08/04/2026), the current total (14 on 09/16/2026), and the percent change. Next MADRS at the 09/30/2026 visit, same method. The clinic's PHQ-9 was also collected today for quality-measure reporting and is charted separately; it is not used to compute the MADRS change. J.M. reviewed the scores and agrees 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. No scale items or anchor text are reproduced.
Writing these after every session? BastionGPT drafts complete notes from bullets, dictation, or a transcript.
Generate a note from bulletsUnited States: the number is a convention; the processes around it are real. No federal or state LAW names the MADRS or prescribes a scoring note. The esketamine REMS is LAW in effect: under the FDA-required program, esketamine is dispensed and administered only in healthcare settings enrolled in the program, outpatients are enrolled, every patient is monitored for at least two hours after each dose with blood pressure reassessed at about 40 minutes, and the program's Patient Monitoring Form is completed after every outpatient session and submitted within seven days; the current prescribing information (2026) covers treatment-resistant depression as monotherapy or with an oral antidepressant, and none of it requires a MADRS, which is why the safety record and the outcome entry are separate documents. PAYER POLICY is where the MADRS appears, as one accepted option and with rules that differ by plan. UnitedHealthcare's commercial esketamine criteria (effective July 1, 2026) require a documented baseline on at least one of the BDI, HAM-D, MADRS, PHQ-9, or QIDS and, for reauthorization, a baseline and a score from within the last month showing remission or a positive clinical response, with no MADRS number set in the policy; Carelon's clinical criteria for esketamine (published April 1, 2025) list seven standard rating scales (BDI, GDS, HAM-D, IDS-SR, MADRS, PHQ-9, QIDS), define the inadequate prior response as less than a 50 percent reduction on one of them, and require at least a 50 percent reduction from baseline for continuation; Aetna's Massachusetts TMS request form (version 12-25) accepts the GDS, PHQ-9, BDI, HAM-D, MADRS, QIDS, or IDS-SR before and after treatment, asks for greater than 50 percent improvement for retreatment, and offers a MADRS below 10 as a typical remission definition, a payer-form convention rather than a property of the scale. Medicare coverage of TMS runs through Medicare Administrative Contractor local coverage determinations (Novitas L34998, for example), each with its own documentation expectations for severity and response; check the current LCD and billing article for your contractor rather than assuming a MADRS is required or sufficient. QUALITY PROGRAMS are the trap: CMS Quality ID 370, Depression Remission at Twelve Months (2026 specification), is built on an initial PHQ-9 or PHQ-9M score greater than nine and remission as a score below five at twelve months, and the NCQA HEDIS measure Depression Remission or Response for Adolescents and Adults (DRR-E) defines response as a PHQ-9 reduction of 50 percent or more and remission as a PHQ-9 below five within four to eight months of the elevated score, so a clinic can satisfy an esketamine or TMS authorization with a MADRS and still owe a PHQ-9 for the measure. CONVENTION supplies the rest: trial statistical analysis plans define response as a reduction of at least 50 percent from baseline and remission at a protocol-specific total, and the same protocols specify rater training, the structured guide, and how missing items are handled, none of which binds a clinic unless it adopts them and says so in the chart.
Canada and Australia set process, not a number, and the evidence explains why the chart must name its conventions. In Canada no federal or provincial LAW names the MADRS. Ontario Health's provincial rTMS program (PAYER POLICY) funds rTMS for adults 18 and older with treatment-resistant unipolar depression, takes referrals from a physician, nurse practitioner, or psychiatrist, determines eligibility by clinical assessment, and describes performance measures and data collection without naming a rating scale or a MADRS threshold; public drug plans largely do not cover esketamine, after the Canadian Drug Expert Committee recommended in December 2020 that it not be reimbursed for major depressive disorder and British Columbia listed it as a non-benefit effective August 3, 2021, so US-style continuation criteria do not transfer. CANMAT's 2023 guideline update (CONVENTION) recommends measurement-based care, lists the MADRS among clinician-rated scales and the PHQ-9 among patient-rated ones, defines early improvement as a reduction of at least 20 percent within two to four weeks of starting treatment and response as at least 50 percent from baseline, gives a MADRS total of 10 or less as its example of symptom remission (with 7 or less on the HAM-D and 4 or less on the PHQ-9), and treats a reduction under 20 percent at two to four weeks as a trigger to consider dose optimization or a switch. In Australia, rTMS has been on the Medicare Benefits Schedule since 1 November 2021 (LAW, through regulation; bare item numbers): 14216 and 14217 cover initial prescription and mapping and a course of up to 35 sessions for adults with major depressive disorder who have not improved on at least two classes of antidepressant and have had no previous rTMS, prescription and mapping are performed personally by a psychiatrist trained in rTMS, and the retreatment items 14219 and 14220 (up to 15 sessions) require a satisfactory response to the initial course, assessed with a validated major depressive disorder tool at least four months after the initial service, followed by relapse; no item names the MADRS. The RANZCP's Professional Practice Guideline 16 on rTMS administration and its position statement 79 (both November 2018; CONVENTION) call for regular monitoring of response and side effects without prescribing a scale. On the evidence: the 1979 paper reported inter-rater correlations of 0.89 to 0.97 between psychiatrists and general practitioners or psychiatric nurses and better separation of responders from non-responders than the comparison measure, and it contains no severity bands; the bands in general use are attributed to a three-page report by Snaith, Harrop, Newby, and Teale (1986); Müller and colleagues (2003) found that a total of 31 best separated moderate from severe depression against the Hamilton scale and the CGI in 85 patients (sensitivity 93.5 percent, specificity 83.3 percent); Riedel and colleagues (2010) found in 846 inpatients that a total of 7 or less best matched CGI-defined remission and a reduction of about 46 percent best matched CGI-defined response; Hawley and colleagues (2002) proposed remission below 9 or 10, Zimmerman and colleagues (2004) proposed 4 or less as a narrow and 9 or less as a broad definition, and the SUSTAIN-1 esketamine trial defined stable remission as 12 or less in at least three of the last four weeks of its optimization phase. Williams and Kobak (2008) reported a total-score intraclass correlation of 0.93 across 162 interviews by 81 rater pairs with the SIGMA; Carmody and colleagues (2006) found the scale unifactorial in 233 and 985 outpatients, with alphas of 0.90 and 0.92 and about twice the measurement precision of the 17-item Hamilton scale at average severity; Quilty and colleagues (2013) fit four symptom factors in 821 outpatients; Turkoz and colleagues (2021) linked a 6-point change to one CGI-S category and a 12-point change to two in 565 patients with treatment-resistant depression; and Leucht and colleagues (2017) linked a total below 8 to a CGI-S rating of borderline ill and a 48 to 57 percent reduction to a rating of much improved. The MADRS-S (Svanborg and Åsberg, 1994) has nine items, asks about the past three days, correlated 0.87 with the BDI in a 2001 comparison, and is published with 0 to 3 half-step item scoring and a 0 to 27 total (Fantino and Moore, 2009) while versions with 0 to 6 items and a 0 to 54 total also circulate, so the exact form belongs in the chart, and dozens of translations are catalogued by ePROVIDE, none of which proves equivalence, so the language and version are charted too.
The suicidal thoughts item deserves its own rule. It is one screening item on a severity scale, rated on the same 0 to 6 steps as the others, and it is not a suicide risk assessment: a rating above zero, a rise from the previous rating, or any clinical concern prompts a separate, dated assessment of ideation, intent, plans, access to means, prior behavior, protective factors, and disposition, documented on the suicide risk assessment or C-SSRS structure and filed in the risk section, and a falling total never implies that risk has resolved. The esketamine prescribing information states that the drug's effectiveness in preventing suicide or in reducing suicidal ideation or behavior has not been demonstrated, one more reason to keep the risk record separate from the outcome entry. If you or a client needs immediate support: call or text 988 (US), 9-8-8 (Canada), or Lifeline 13 11 14 (Australia).
The MADRS is copyrighted by Stuart Montgomery and the Royal College of Psychiatrists, as recorded on ePROVIDE, and is made available for an individual clinician's or researcher's own use, copied in full with the full citation, with distribution to others or republication requiring the College's permission. The 1979 article is copyright Royal College of Psychiatrists and is published in the British Journal of Psychiatry by Cambridge University Press, the copy hosted by the American Psychological Association carries a 1978 Stuart Montgomery copyright line, and the SIGMA (Williams and Kobak, 2008) is copyright Royal College of Psychiatrists. The permission statement printed with the scale allows an individual clinician or researcher to copy it for their own use without asking, in full and acknowledging the source, and requires permission from the Royal College of Psychiatrists for copying and distribution to others or for republication in print, online, or by any other means; the College routes permission requests, including rating-scale requests, through the Request Permission link on the article's Cambridge Core page and may charge a fee. An EHR build that reproduces the item and anchor text, a printed pack handed to other clinicians, or a public web calculator that displays the scale is a distribution or republication question for the rights holders, and no blanket license for those uses had been published as of September 2026; recording independently written item scores, a total, and a narrative, as the template on this page does, is a different matter. No trademark is claimed for the name. 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. The 1979 paper contains no severity bands; the bands in general use trace to a three-page 1986 report by Snaith and colleagues, and the moderate-to-severe boundary moved from 35 to 31 when Müller and colleagues tested it against the Hamilton scale and the CGI in 85 patients. In 846 inpatients the values that best matched clinician-judged remission and response were 7 or less and a reduction of about 46 percent (Riedel and colleagues, 2010), not 10 and 50; published remission cutoffs run from 4 (Zimmerman and colleagues, 2004) to 12 (the SUSTAIN-1 esketamine trial). With the structured guide the total-score intraclass correlation reached 0.93 across 81 rater pairs (Williams and Kobak, 2008); a 6-point change corresponded to one CGI-S category in 565 patients with treatment-resistant depression (Turkoz and colleagues, 2021); and the CMS and HEDIS remission measures count only the PHQ-9. No published audit counts how often the version, rater, interval, or baseline is missing from a MADRS chart entry; the payer criteria described under compliance considerations are what turn those omissions into a denial. The BastionGPT Clinical Advisory Board sees the same errors most often in MADRS documentation reviews:
BastionGPT is specifically trained, tuned, and clinically tested on psychiatric progress notes and outcome-measure 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.
A MADRS total is the sum of ten clinician ratings, each on a 0 to 6 scale with descriptions at the even steps and intermediate points between them, so the total runs from 0 to 60 and higher means more severe. The original 1979 paper contains no severity bands. The bands in general use, essentially absent or recovered up to 6, mild from 7 to 19, moderate from 20 to 34, and severe from 35 to 60, are attributed to a short 1986 report by Snaith and colleagues, and later work disagrees at the boundaries: Müller and colleagues (2003) found that a total of 31 best separated moderate from severe depression against the Hamilton scale and the CGI, and some tools splice several papers into five bands. So a chart entry reads "total 24 of 60; moderate under the Snaith and colleagues (1986) grade scores" rather than "MADRS 24 = moderate," names the source whenever a label is used, and never turns the label into a diagnosis: the MADRS measures severity and change in someone whose depression has already been diagnosed by interview. The item profile matters as much as the total, because the same 24 can be carried by mood and pessimism or by sleep, appetite, and tension.
Whatever rule you name, and you must name one, because neither is built into the scale. Response is conventionally a reduction of at least 50 percent from the baseline total; trial statistical analysis plans state it that way, CANMAT's 2023 update uses it, and several payers copy it. Remission has no consensus number: a total of 10 or less is the most common operational definition and CANMAT's example, Hawley and colleagues (2002) proposed below 9 or 10, Zimmerman and colleagues (2004) proposed 4 or less as a narrow and 9 or less as a broad definition, the SUSTAIN-1 esketamine trial required 12 or less in at least three of the last four weeks of its optimization phase, and Riedel and colleagues (2010) found in 846 inpatients that 7 or less best matched clinician-judged remission and that a reduction of about 46 percent best matched clinician-judged response. Chart the baseline date and total on the same instrument and method, the current total, the absolute and percent change, and the rule with its source, and keep the payer's criterion and your clinic's remission convention apart: a patient can meet a plan's continuation criterion and still not be in remission under the convention you cite. Change from the previous rating, as opposed to baseline, is described rather than labeled.
No US payer reviewed for this page requires the MADRS by name; policies list it as one accepted instrument, and their rules differ (PAYER POLICY, not law). UnitedHealthcare's commercial esketamine criteria (effective July 1, 2026) require a documented baseline on at least one of the BDI, HAM-D, MADRS, PHQ-9, or QIDS and, for reauthorization, a baseline and a score from within the last month showing remission or a positive clinical response, with no numerical MADRS threshold in the policy. Carelon's esketamine criteria (published April 1, 2025) list seven scales, the MADRS among them, and require at least a 50 percent reduction from baseline for continuation. Aetna's Massachusetts TMS request form (version 12-25) accepts the GDS, PHQ-9, BDI, HAM-D, MADRS, QIDS, or IDS-SR before and after treatment and asks for greater than 50 percent improvement for retreatment, while offering a MADRS below 10 as a typical remission reference. Medicare TMS coverage runs through local coverage determinations with their own documentation expectations; check your contractor's current LCD. So the defensible sentence is "meets the plan's stated continuation criterion," with the baseline, the current total, and the arithmetic in the note, not "MADRS response is 50 percent." One trap: the CMS Depression Remission at Twelve Months measure (Quality ID 370) and the HEDIS DRR-E measure count only the PHQ-9 and PHQ-9M, so a MADRS that satisfies the authorization earns nothing there. The letter mechanics belong to the prior authorization request page; this page owns the scale-specific payload.
Choose the MADRS when a clinician-rated severity and change measure is wanted: interventional psychiatry, treatment-resistant depression, trials, a payer packet that names it, or a patient whose pain, sleep disorder, or medical illness inflates the somatic items of a self-report, because the MADRS weights mood, anhedonia, and cognitive symptoms more heavily, and Carmody and colleagues (2006) found it unifactorial with about twice the measurement precision of the 17-item Hamilton scale at average severity. Choose the PHQ-9 for screening, primary care and collaborative care monitoring, and anything reported to CMS Quality ID 370 or HEDIS DRR-E, which count only the PHQ-9; it is faster, patient-reported, and mapped to DSM symptom criteria, and many TMS clinics track with it day to day. The HAM-D is the historical comparator with broader anxiety and somatic content, still named in many payer lists and older protocols. Whichever you choose, keep the same instrument and method across a course of treatment and, when a payer or trial names an instrument, use that one; a series that switches scales midway cannot show change. No validated conversion between the MADRS and the PHQ-9 is established for North American practice, so do not convert one into the other in the chart; the BDI-II page covers the other common self-report.
Any clinician trained in the interview and the scale: one aim of the 1979 paper was an instrument usable by raters without specialist psychiatric training, its reliability work paired psychiatrists with general practitioners and with psychiatric nurses (inter-rater correlations of 0.89 to 0.97), and nothing in the instrument limits rating to physicians, so a psychiatric nurse practitioner or nurse in a TMS or esketamine program can rate it; whether a payer, employer, or protocol requires a particular credential is a separate question. The scale was published without standard interview questions, a known source of rater variability; the SIGMA (Williams and Kobak, 2008) supplies scripted probes and a past-week frame and reported a total-score intraclass correlation of 0.93 across 162 interviews by 81 rater pairs. Formal rater training, calibration, and independent raters are trial-protocol requirements, not properties of the scale; in a clinic the defensible minimum is a rater who knows the procedure, uses the same method every time, and records it. Chart the rater's name and role, whether the SIGMA was used, and any training or calibration, so a reviewer can tell a structured rating from an unstructured one and a clinician rating from a self-report.
No and no. The MADRS-S (Svanborg and Åsberg, 1994) is a separately published patient-completed scale with nine items, a recall window of about three days, and its own scoring: the form validated by Fantino and Moore (2009) scores each item 0 to 3 with half-steps for a 0 to 27 total, while versions scoring 0 to 6 per item for a 0 to 54 total also circulate, so the form and its total range belong in the chart. Self-ratings and clinician ratings agree well in group data, but the instruments differ in rater, items, window, and total, and none of the payer policies reviewed for this page names the MADRS-S; they name the MADRS and, separately, self-report scales such as the PHQ-9, BDI, or QIDS. Submit the instrument the payer or protocol named, label a self-report MADRS-S every time, never trend MADRS-S totals against clinician MADRS totals, and, when minimization is suspected, prefer the clinician scale, whose instructions allow the rater to draw on observed behavior and other clinical information when a definite answer cannot be obtained.
A separate, dated risk assessment, every time. The item is one screening question on a severity scale and is not a risk assessment: document the item rating, then the assessment of ideation, intent, plans, access to means, prior behavior, protective factors, and disposition on the suicide risk assessment structure or with the C-SSRS, the safety plan status, and where the formulation is filed. Repeat it whenever the rating rises, and do not read a falling total as resolved risk; the esketamine prescribing information states that the drug's effectiveness in preventing suicide or in reducing suicidal ideation or behavior has not been demonstrated. Keep the risk record in the risk section rather than inside the outcome entry, so a payer packet built from the MADRS series is never the only place the assessment lives. If you or a client needs immediate support: call or text 988 (US), 9-8-8 (Canada), or Lifeline 13 11 14 (Australia).
At the topic level the ten items cover sadness as observed by the rater and as reported by the patient, inner tension, reduced sleep, reduced appetite, concentration difficulty, lassitude, reduced ability to feel, pessimistic thoughts, and suicidal thoughts, each on a 0 to 6 scale; this page does not print or paraphrase the anchor descriptions, and the scale itself is the source for rating. Rights: the ePROVIDE record lists the copyright as claimed by Stuart Montgomery and the Royal College of Psychiatrists, the 1979 article is the College's copyright, and the permission statement printed with the scale allows an individual clinician or researcher to copy it in full, with the full citation, for their own use, while copying for distribution to others or republication in print, online, or by any other means requires the College's permission, which it handles through the Request Permission link on the article's Cambridge Core page and may charge for. An EHR template that reproduces the item and anchor text, a printed pack for other clinicians, and a public web calculator that displays the scale are distribution or republication questions for the rights holders; no blanket license for them had been published as of September 2026, and the free copies online are not evidence of one. Recording your own item scores, the total, and a narrative in the chart, as this page's template does, reproduces nothing.
Yes. Give it the facts (instrument and guide, rater and training, interval, item scores or the profile, total, baseline date and total, the rule your payer or protocol applies, the suicidal-thoughts item rating with the separate assessment, treatment status, and the decision) and it drafts the full entry: version and interval, the profile at the topic level, the total with the convention named, the change with the rule applied, the separate risk line, the decision, and the baseline-and-current-score paragraph for a continuation request, ready for your review. It can also check a finished note for a total with no version or interval, a band charted as a diagnosis, a response claim with no baseline or rule, a MADRS-S mixed into a clinician series, a suicidal-thoughts rating with no separate assessment, or a MADRS offered where the measure counts only the PHQ-9. 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.