Rating scales

The CGI (Clinical Global Impression) scale

April 12, 2026 8 min read

The Clinical Global Impression scale (CGI) was developed by William Guy at the National Institute of Mental Health and published in 1976 in the ECDEU Assessment Manual for Psychopharmacology. It has become the most widely used outcome measure in psychiatry, not because it is sophisticated but because it is short, intuitive, and clinically meaningful. A trained clinician can complete it in less than a minute. The FDA accepts it as a co-primary or secondary endpoint in nearly every psychiatric trial.

In one sentence

The CGI is a one- to three-item clinician rating of overall illness severity (CGI-S) and treatment-related change (CGI-I), each on a 1–7 scale.

The three components

The first two — CGI-S and CGI-I — are the workhorses.

Why such a simple scale survives

The CGI does something no other instrument does: it asks an experienced clinician for a single, holistic judgement. This is exactly the kind of judgement a senior psychiatrist makes every day in the clinic. By asking it explicitly and recording it numerically, the CGI captures clinical wisdom in a form that can be aggregated across trials and across centres. The trade-off is that it depends entirely on rater experience — a green clinician's CGI-S of 4 may not be the same as a 30-year veteran's.

Anchoring CGI to other scales

One of the most cited papers in modern schizophrenia methodology is by Stefan Leucht and colleagues (2005, Neuropsychopharmacology), which mapped CGI-S to PANSS and BPRS thresholds. Their estimates, often called the "Leucht conversions":

This mapping is what allows researchers to translate continuous PANSS change into clinically intuitive labels.

How it is administered

The CGI is completed by the treating clinician (or a trained rater) at every assessment point — typically baseline and follow-up visits. There is no patient interview specific to the CGI; it is a synthesis of everything the rater knows about the patient. For trials, it is usually rated immediately after a longer assessment such as the PANSS, drawing on that interview content.

What "very much improved" means in practice

The CGI-I anchor descriptions tie loosely but consistently to other measures. In schizophrenia trials, "much improved" (CGI-I 2) typically corresponds to a 25–30% reduction in PANSS total; "very much improved" (CGI-I 1) corresponds to 50% or more. This is one reason both scales are usually reported together — the CGI-I gives a global feel, the PANSS gives the numbers.

Strengths

Limitations

Variants and adaptations

Several disorder-specific CGI versions exist, including the CGI-BP (bipolar disorder) and the CGI-SCH (schizophrenia, with separate ratings for positive, negative, depressive, and cognitive symptom domains). The CGI-SCH is useful when a rater wants the brevity of the CGI but a domain-level breakdown.

What this means for patients and families

The CGI is one of the rating scales most likely to actually appear in your clinical chart, because it is so quick. If your psychiatrist's note says "CGI-S = 4, CGI-I = 3 (minimally improved)," they are recording a clinical impression in a standardised way. You can ask what would constitute a "much improved" CGI-I rating at the next visit, which is a useful structuring conversation for shared decision making.

How regulators think about it

The FDA, the European Medicines Agency, and most national regulators accept CGI-I as a secondary endpoint in psychiatric trials. Some early trials and accelerated-approval contexts have used CGI as a co-primary endpoint. The FDA Drugs portal includes briefing documents that frequently cite CGI changes alongside scale-specific outcomes like PANSS.


This article is for educational purposes only and is not medical advice, diagnosis, or treatment. Always consult a qualified mental health professional. If you or someone you know is in crisis, call or text 988 in the US, or your local emergency number.

Frequently asked questions

Is the CGI a self-report scale?
No. It is a clinician judgement, drawing on interview, observation, collateral information, and the rater's overall clinical experience.
What does CGI-I 'minimally improved' mean?
A small but real improvement that is clinically detectable but not yet clearly meaningful for the patient's life. CGI-I scores of 1 ('very much improved') and 2 ('much improved') are usually treated as response in trials.
Why is the CGI still used if newer scales exist?
Because nothing else captures a clinician's overall judgement as efficiently. Rich scales like the PANSS quantify symptoms; the CGI quantifies the gestalt that comes from clinical training.

Try Frida — your calm companion

Frida helps people living with schizophrenia track moods, manage medication, and build stability. 7-day free trial.

Get the app →