The SCOFF is a five-question screen for anorexia nervosa and bulimia nervosa. Its name is a mnemonic for the five yes/no questions - making yourself Sick, loss of Control over eating, loss of more than One stone in three months, believing oneself Fat when others disagree, and food Dominating life. Each "yes" scores one point, giving a total of 0 to 5.
It screens for the likelihood of an eating disorder; it does not diagnose one, grade severity, or distinguish between diagnoses. Two or more positive answers indicate that a fuller eating-disorder history and examination are warranted.
02 - Origin & purpose
Where it comes from.
The SCOFF was developed by John Morgan, Fiona Reid and Hubert Lacey at St George's Hospital Medical School, London, and published in the BMJ in 1999. The questions were generated with input from eating-disorder patients and specialists, deliberately modelled on the CAGE: a memorable verbal screen a clinician can use without a form.
The intention was to give non-specialists - GPs, general psychiatrists, practice nurses - a tool quick enough for routine consultations in populations where eating disorders are easily missed. In the original case-control study it detected all cases of anorexia and bulimia at a threshold of two, and a subsequent primary-care validation supported its use in unselected patients.
03 - Scoring & cutoffs
How scoring works.
One point per "yes"; total 0-5. The validated threshold is 2 or more, which in the original study gave sensitivity 100% and specificity 87.5% against DSM-IV anorexia and bulimia. In UK use the third question refers to losing more than one stone (about 6.35 kg) in three months; when administering it elsewhere, use a locally meaningful equivalent and note the adaptation. The SCOFF score carries no severity gradient - a 5 is not "worse" than a 2; both are positive screens requiring the same follow-up questions about weight history, compensatory behaviours, binge episodes and physical risk.
Score
Severity
Interpretation
0–1
Below threshold
Below the screening threshold for an eating disorder.
2–5
At or above threshold
Two or more positive answers indicate a likely eating disorder; further assessment is recommended.
04 - Validation evidence
How well it performs.
The original 1999 study (116 eating-disorder cases vs 96 controls, young women) reported sensitivity 100% and specificity 87.5% at ≥2. Luck and colleagues' 2002 primary-care validation found sensitivity 84.6% and specificity 89.6%, detecting every case of anorexia and bulimia and seven of nine EDNOS cases.
The most comprehensive synthesis, Kutz et al. 2020 (25 studies), pooled sensitivity at 0.86 (95% CI 0.78-0.91) and specificity at 0.83 (95% CI 0.77-0.88), but found accuracy was highest in case-control studies of young women with anorexia or bulimia and notably lower in samples with more men, binge eating disorder, or community recruitment - and concluded the evidence does not yet support the SCOFF for screening the full DSM-5 range of eating disorders in primary care. An earlier meta-analysis (Botella et al. 2013, 15 studies) had pooled sensitivity 0.80 and specificity 0.93.
0.86
SENSITIVITY (POOLED, KUTZ 2020)
0.83
SPECIFICITY (POOLED, KUTZ 2020)
100%
SENSITIVITY (ORIGINAL 1999, ≥2)
87.5%
SPECIFICITY (ORIGINAL 1999, ≥2)
05 - How it compares
How it compares to the alternatives.
Instrument
Items
Time
When to reach for it
SCOFF
5
~1 min
Verbal-friendly first-line screen for anorexia and bulimia in general settings.
EAT-26
26
~5 min
Fuller attitudes-and-behaviours screen with a ≥20 referral threshold; better documentation trail, still not diagnostic.
EDE-Q
28
~10 min
The standard self-report severity measure with global and subscale scores; use when you need to quantify and track symptoms, not just screen.
ESP (Eating disorder Screen for Primary care)
5
~1 min
Alternative brief screen; unlike the SCOFF it asks directly about dissatisfaction with eating patterns and past eating disorders.
BEDS-7
7
~2 min
Purpose-built brief screen for binge eating disorder - the diagnosis the SCOFF is weakest on.
06 - When to use it
Right tool, wrong tool.
The SCOFF is intended as a first-line screen in general settings - primary care, general psychiatry, practice nursing - where eating disorders are easily missed. It suits opportunistic use when weight, food or body-image concerns surface in a consultation.
A positive screen (2 or more) should lead to a full eating-disorder history, examination and medical-risk assessment, not to a diagnosis or a severity rating.
Reach for it when
-Opportunistic screening in primary care or general psychiatry when weight, food or body-image concerns surface.
-A structured verbal prompt in consultations where a written questionnaire would be intrusive.
-Young adult populations where anorexia and bulimia prevalence is highest.
Reach for something else when
-Diagnosing or grading the severity of an eating disorder.
-Monitoring change during treatment - use the EDE-Q.
-Screening specifically for binge eating disorder or ARFID.
-Medical-risk assessment of an established eating disorder - the SCOFF asks nothing about physical state.
-Relying on a negative screen when observation suggests otherwise, particularly in men and older patients.
07 - Confidence & precision
Reading the score with care.
No SEM or meaningful-change threshold exists, and none is meaningful for a five-item binary screen: the SCOFF yields a category (positive/negative at ≥2), not a measurement. With pooled specificity around 0.83, expect roughly one in six people without an eating disorder to screen positive in typical settings - the follow-up conversation, not the score, carries the clinical weight. For quantifiable, trackable symptom measurement use the EDE-Q.
08 - Limitations
What it cannot tell you.
Validated primarily in young women with anorexia and bulimia; sensitivity drops in men, binge eating disorder and community samples (Kutz 2020). It predates DSM-5 and does not cover ARFID or reflect current binge-eating criteria.
The "one stone" item is UK-specific and needs local adaptation. Self-report and verbal administration can perform differently, and denial or minimisation - common in eating disorders - directly lowers sensitivity. A positive screen carries no information about medical risk, which must be assessed separately (e.g. per MEED guidance).
See how Aisel removes friction where it costs most. A 20-minute walkthrough tailored to your clinic.
We value your privacy
We use cookies to analyse site usage and improve your experience. Analytics and embedded media (e.g. YouTube) only load if you accept. Read our cookie policy.