SCOFF eating disorder questionnaire

What this assessment is

SCOFF is a five-question screen for warning signs associated with eating disorders.

John Morgan, Fiona Reid, and J. Hubert Lacey introduced it in a 1999 BMJ study. The name is an acronym drawn from the original questions.

SCOFF is designed to raise suspicion of a possible eating disorder. It does not diagnose one.

That is its purpose.

Who it was designed for

The original validation involved women aged 18 to 40 in a specialist setting. A later general practice study tested women aged 18 to 50.

The evidence base has grown. A later evidence review for the U.S. Preventive Services Task Force pooled results across 10 adult studies, while finding much less evidence for adolescents.

The questionnaire is now used more widely, including with adolescents and people of different genders. Performance from the original studies should not be assumed to transfer unchanged to every group.

Use in a particular age group, language, culture, or clinical setting needs suitable evidence and a clear follow-up process.

How it is administered

SCOFF uses five yes-or-no responses and takes only a few minutes.

It is short.

The questions do not share one recent recall period. They ask about warning signs and experiences that may occur over different periods.

Use the complete standard form. Rewording a question, changing a response choice, or adding a new timeframe can change how people answer.

What it measures

SCOFF covers broad warning signs related to eating control, weight and body image, and the effects of disordered eating.

It does not provide a full account of food intake, compensatory behaviors, weight history, physical symptoms, or daily impairment.

The EAT-26 is a longer eating-attitudes screen. It uses a different response format and includes referral criteria that are not part of the SCOFF score.

How scoring works

Each endorsed response scores 1. Each non-endorsed response scores 0. Add the five values for a total from 0 to 5.

The standard positive-screen rule is a total of 2 or higher.

This is one decision point, not a set of severity bands. A score of 4 is not a diagnosis of a more severe eating disorder than a score of 2.

Do not create bands.

Keep it as a count.

Complete responses are needed for the standard total. Do not treat a missing response as a no.

How to interpret a SCOFF score

A total of 2 or higher raises suspicion of an eating disorder and supports further assessment.

In the original specialist study, this rule identified all cases in the small validation sample and correctly screened out most controls. In the later general practice study, sensitivity was lower. It identified about 85 in 100 participants with an eating disorder and correctly screened out about 90 in 100 without one.

Those figures describe study samples, not a guarantee for an individual.

Accuracy varies.

A result below 2 does not rule out an eating disorder. Continue assessment when symptoms, behavior, physical changes, or the person's concerns still suggest a problem.

What the score cannot tell you

SCOFF cannot identify the type of eating disorder, determine its severity, or show whether someone is medically stable.

It cannot assess all eating-disorder presentations. People may also underreport symptoms because of shame, fear, limited insight, or uncertainty about what a question means.

Urgency depends on the full picture, including current intake, compensatory behavior, rapid change, fainting, hydration, vital signs, mental state, and other health concerns. The score alone cannot make that decision.

Evidence and limitations

SCOFF was deliberately designed for brevity and high sensitivity. That makes it useful as an initial screen, but false positives and false negatives still occur.

The early studies were small and focused on adult women in the United Kingdom. Later research has examined broader groups, with different estimates of accuracy.

The limits remain.

Use the questionnaire to decide whether more assessment is needed. Do not use it to confirm or dismiss a diagnosis.