MADRS: Scoring, Cutoffs & Interpretation

01 - What it measures

What this scale measures.

The MADRS measures the severity of depressive symptoms over the previous week. A clinician rates ten items - apparent sadness, reported sadness, inner tension, reduced sleep, reduced appetite, concentration difficulties, lassitude, inability to feel, pessimistic thoughts and suicidal thoughts - each on a 0 to 6 scale, and sums them to a total between 0 and 60. Ratings are made on the basis of a semi-structured interview rather than a fixed questionnaire, so the clinician's judgement of the patient's account and presentation both contribute to the score.

The item set is deliberately weighted towards the psychological core of depression. Somatic and psychomotor features that dominate older instruments are largely absent, which is what allows the total to move cleanly when mood improves rather than being held up by sleep, appetite or physical complaints that resolve on a different timescale. The MADRS therefore reads as a measure of episode severity and its trajectory, not as a diagnostic test and not as a full description of the depressive syndrome.

02 - Origin & purpose

Where it comes from.

Stuart Montgomery and Marie Asberg published the scale in the British Journal of Psychiatry in 1979 under the title "A new depression scale designed to be sensitive to change". The ten items were not written from scratch. They were selected empirically from the 65-item Comprehensive Psychopathological Rating Scale as the items that moved most during antidepressant treatment, using combined English (n = 54) and Swedish (n = 52) patient samples - a bilingual derivation intended to reduce cultural bias in the resulting item set.

The purpose was explicitly comparative. The Hamilton Depression Rating Scale was already dominant, but a substantial share of its items covered somatic, anxiety and insomnia features that dilute the treatment-change signal. Montgomery and Asberg optimised for responsiveness instead of syndrome coverage. Snaith and colleagues added the now-standard severity grade scores in 1986, Svanborg and Asberg published a self-rated version derived from the same CPRS lineage in 1994, and the MADRS went on to become a primary endpoint in a large share of regulatory antidepressant trials.

03 - Scoring & cutoffs

How scoring works.

Each of the ten items is rated 0 to 6 for the past week. Descriptive anchors are defined at 0, 2, 4 and 6; the odd values of 1, 3 and 5 are intermediate points the rater uses when the presentation sits between two anchors. No item is reverse-scored, and the total is the simple sum, giving a range of 0 to 60. Because ratings depend on a semi-structured interview, some rater familiarity with the anchor descriptions is assumed.

The severity bands below follow Snaith and colleagues (1986). Two further conventions matter more in practice than the bands themselves: response is usually defined as a reduction of at least 50% from baseline, and remission as a total of 10 or below. That remission threshold is a convention rather than a settled finding - criterion scores from 6 to 12 appear across trials, Zimmerman and colleagues derived an optimal value of 9 or below for a broad definition and 4 or below for a strict one, and Leucht and colleagues' equipercentile linking to the Clinical Global Impression put "almost complete absence of symptoms" below 8. Read any band as a description of symptom load, never as a diagnosis.

04 - Validation evidence

How well it performs.

Internal consistency is strong and stable across samples: Cronbach's alpha was 0.89 in the original work and 0.90 to 0.92 in larger outpatient samples (n = 233 and n = 985), with median item-total correlations of 0.75 to 0.78. Inter-rater reliability is the scale's particular strength - correlations of 0.89 to 0.97 have been reported for total and difference scores across rater pairs, including a psychiatrist paired with a general practitioner or with a nurse, which is what makes the MADRS usable across mixed clinical teams rather than only by specialist raters.

Convergent validity against the Hamilton scale is high (r = 0.88 to 0.92). Muller and colleagues (2003) found a MADRS total of 31 or above separated severe from moderate depression with 93.5% sensitivity and 83.3% specificity. The self-rated MADRS-S tracks the clinician version closely (r = 0.80 to 0.94), which supports its use for between-visit monitoring. One caveat on structure: Rasch analysis has found that neither the ten-item MADRS nor the 17-item HAM-D meets strict criteria for unidimensionality, and factor analyses typically recover about four factors - sadness, neurovegetative, detachment and negative thoughts. The total is best read as a severity index, not as a pure single construct.

05 - How it compares

How it compares to the alternatives.

06 - When to use it

Right tool, wrong tool.

The MADRS is built for measuring change in an identified depressive episode, and that is where it earns its keep: baseline severity at the start of a treatment episode, repeat ratings every two to four weeks while a medication or therapy is being titrated, and a defensible response or remission judgement at the end. It is the standard severity endpoint in antidepressant trials, so a MADRS trajectory is directly comparable to published evidence in a way that most alternatives are not.

It is a poor choice as a first-pass screen. It requires a clinician, an interview and roughly fifteen minutes, which is a heavy instrument to point at an unselected population - a brief self-report screener does that job better and cheaper. It also cannot make a diagnosis. A high score describes severity; whether the presentation meets criteria for a major depressive episode, and whether it is unipolar or bipolar, remains a matter for clinical assessment. Item 10 is a single suicidality rating and is not a substitute for a structured risk assessment.

07 - Confidence & precision

Reading the score with care.

A MADRS total is an interval, not a point. With inter-rater correlations of 0.89 to 0.97, agreement between competent raters is good but not perfect, and shifts of two or three points between visits should not be over-interpreted - particularly when the rater has changed. There is no widely replicated published standard error of measurement for the MADRS, so the honest position is to anchor change on the established conventions rather than on a precise error band.

Those conventions are: a reduction of at least 50% from baseline for response, and 10 or below for remission. Leucht and colleagues' equipercentile linking to the Clinical Global Impression provides the most useful calibration - a percentage reduction of roughly 48% to 57% corresponded to being rated "much improved", and 80% to 84% to "very much improved". Where percentage change is impractical, an improvement of around 6 points is commonly treated as the smallest change a clinician would recognise as meaningful. Percentage change is preferable to absolute change on this scale, because the same six points mean something different starting from 38 than from 14.

08 - Limitations

What it cannot tell you.

The MADRS is a severity measure, not a diagnostic instrument. It cannot distinguish a unipolar from a bipolar depressive episode, and it will not tell you whether diagnostic criteria are met. Its narrow focus is also a coverage gap: psychomotor agitation and retardation, guilt as a discrete feature, and most somatic symptoms are absent or only indirectly represented, so an agitated or heavily somatic presentation can score lower than clinical impression suggests.

It requires clinician time and a semi-structured interview, and rating quality depends on familiarity with the anchor descriptions. Ratings cover only the past week, so it is insensitive to longer-term course and to chronic low-grade presentations. Item 10 rates suicidal thinking on a single 0 to 6 dimension and must never be used as a risk assessment in its own right. Finally, the standard Snaith bands were derived in adult psychiatric populations; performance differs in bipolar depression, older adults, adolescents and medically ill groups, where population-specific cutoffs have been proposed and no single threshold transfers cleanly.

09 - Licensing, explained

How licensing works.

Clinician-rated scale, free to use for educational purposes and routine clinical assessment. Please reproduce it in full and keep the original attribution.

Common questions from clinicians

FAQ

References

Source literature.

1. [1]Montgomery SA, Asberg M A new depression scale designed to be sensitive to change (1979) ↩
2. [2]Snaith RP, Harrop FM, Newby DA, Teale C Grade scores of the Montgomery-Asberg Depression and the Clinical Anxiety Scales (1986) ↩
3. [3]Muller MJ, Himmerich H, Kienzle B, Szegedi A Differentiating moderate and severe depression using the Montgomery-Asberg Depression Rating Scale (MADRS) (2003) ↩
4. [4]Hawley CJ, Gale TM, Sivakumaran T Defining remission by cut off score on the MADRS: selecting the optimal value (2002) ↩
5. [5]Leucht S, Fennema H, Engel RR, Kaspers-Janssen M, Szegedi A What does the MADRS mean? Equipercentile linking with the CGI using a company database of mirtazapine studies (2017) ↩
6. [6]Svanborg P, Asberg M A new self-rating scale for depression and anxiety states based on the Comprehensive Psychopathological Rating Scale (1994) ↩