Hucknall, and Arnold Health Centre. We are particularly grate-
ful to the patients who took part in the study.
Contributors: See bmj.com
Funding: Department of Health.
Competing interests: None declared.
1 Felson DT, Naimark A, Anderson JJ, Kazis L, Castelli W, Meenan RF. The
prevalence of knee osteoarthritis in the elderly: the Framingham
osteoarthritis study. Arthritis Rheum 1987;30:914-8.
Badley EM, Tennant A. Disablement associated with rheumatic disorders
in a British population:problems with activities of daily living and level of
support.Br J Rheumatol 1993;32:601-8.
FisherNM, Prendergast DR, Gresham
rehabilitation: its effect on muscular and functional performance of
patients with knee osteoarthritis. Arch Phys Med Rehabil 1991;72:367-74.
Deyle GD, Henderson NE, Matekel MP, Ryder MG, Garber MB, Allison
osteoarthritis of the knee. A randomized, controlled trial. Ann Intern Med
O’Reilly SC, Muir KR, Doherty M. Screening for knee pain in
osteoarthritis: which question? Ann Rheum Dis 1996;55:931-3.
Bellamy N, Buchanan WW, Goldsmith CH, Campbell J. Validation study
of WOMAC: a health status instrument for measuring clinically-
important patient-relevant outcomes following total hip or knee arthro-
plasty in osteoarthritis. J Ortho Rheumatol 1988;1:95-108.
3GE, Calkins E. Muscle
7Brazier JE, Harper R, Jones NM, O’Cathain A, Thomas KJ, Usherwood T,
et al. Validating the SF-36 health survey questionnaire: new outcome
measure for primary care. BMJ 1992;305:160-5.
Zigmond AS, Snaith RP. The hospital anxiety and depression scale. Acta
Psychiatr Scand 1983;67:361-70.
Tornvall G. Assessment of physical capabilities with special reference to
the evaluation of maximum voluntary isometric muscle strength. Acta
Physiol Scand 1963;58(suppl 201):1-102.
10 Altman RD, Hochberg MC, Murphy WA, Wolfe F. Atlas of individual
radiographic features in osteoarthritis. Osteoarthritis Cart 1995;3(suppl
11 Altman DG. Confidence intervals for the number needed to treat. BMJ
12 Cook RJ, Sackett DL. The number needed to treat: a clinically useful
measure of treatment effect. BMJ 1995;310:452-4.
13 Van Barr ME, Dekker J, Oostendorp RA, Bijl D, Voorn TB, Lemmens JA.
The effectiveness of exercise therapy in patients with osteoarthritis of hip
or knee: a randomised clinical trial. J Rheumatol 1998;25:2432-9.
14 Van Baar ME, Dekker J, Oostendorp RAB, Bijl D, Voom TB, Bijlsma JW,
et al. Effectiveness of exercise in patients with osteoarthritis of hip or
knee: nine months’ follow up. Ann Rheum Dis 2001;60:1123-30.
15 Ettinger WH, Burns R, Messier SP, Applegate W, Rejeski WR, Morgan T,
et al. A randomized trial comparing aerobic exercise and resistance exer-
cise with a health education program in older adults with knee
osteoarthritis. JAMA 1997;277:25-31.
(Accepted 20 June 2002)
The SCOFF questionnaire and clinical interview for eating
disorders in general practice: comparative study
Amy J Luck, John F Morgan, Fiona Reid, Aileen O’Brien, Joan Brunton, Clare Price, Lin Perry,
J Hubert Lacey
Standards 2 and 3 of the national service framework for
mental health outline the need to improve health care
for patients with anorexia nervosa and bulimia nervosa.1
Healthcare workers in primary care are at the forefront
of screening and managing these disorders. Assessment
tools available to primary healthcare professionals can
take a long time to administer and may need to be inter-
preted by specialists2; this may limit improvements in
care. A screening tool was developed, but only to
facilitate epidemiological research.3
The SCOFF questionnaire is a brief and memora-
ble tool designed to detect eating disorders and aid
treatment (see figure). It showed excellent validity in a
clinical population and reliability in a student popula-
tion.4 5We assessed the SCOFF questionnaire in
Participants, methods, and results
We invited sequential women attenders (aged 18-50) at
two generalpractices in
participate. We gave participants information sheets
thatdescribed the study. Women
consented to participate were asked the SCOFF ques-
tions in a separate room; this took about two minutes.
A researcher blind to the woman’s score on the SCOFF
questionnaire conducted a clinical diagnostic interview
lasting 10-15 minutes, based on criteria from the Diag-
nostic and Statistical Manual of Mental Disorders (fourth
edition). Women identified by the interview as having
an eating disorder were invited to discuss this and were
offered the contact number for the Eating Disorders
Association. The local research and ethics committee
approved the study.
Of the 341 women who agreed to take participate,
one (who had a body mass index of 17 (weight (kg)/
height (m)2)) had anorexia nervosa, three had bulimia
nervosa, and nine had an “eating disorder not
otherwise specified.” A receiver operating curve set the
optimal threshold for the questionnaire at two or more
positive answers to the five questions. With this cut off,
the SCOFF questionnaire detected all four cases of
anorexia nervosa and bulimia nervosa and seven of
nine cases of eating disorders not otherwise specified
(figure). The questionnaire had a sensitivity of 84.6%
(95% confidence interval 54.6% to 98.1%). In the 328
women confirmed not to have an eating disorder, the
questionnaire indicated 34 false positives. It had a
specificity of 89.6% (86.3% to 92.9%), positive predic-
tive value of 24.4% (12.9% to 39.5%), and negative
predictive value of 99.3% (97.6% to 99.9%).
The SCOFF questionnaire detected all cases of
anorexia and bulimia nervosa. It is an efficient screen-
ing tool for eating disorders.
Two missed cases of eating disorders not otherwise
specified reflect the reality of clinical situations, in
which denial or non-disclosure by patients may occur.
One of the patients in whom the diagnosis was missed
later disclosed disordered eating behaviour. It may be
more difficult and perhaps less pertinent to detect
patients who do not meet full criteria for anorexia ner-
vosa or bulimia nervosa.
The positive predictive value of the questionnaire is
low because of the low prevalence of eating disorders
in this sample, which was consistent with the Western
of London, London
Amy J Luck
John F Morgan
senior lecturer in
lecturer in community
lecturer in psychiatry
clinical research fellow
J Hubert Lacey
lecturer in medical
Faculty of Health
and Social Care
J F Morgan
BMJ VOLUME 325 5 OCTOBER 2002bmj.com
population as a whole. Overinclusion is acceptable for
screening instruments designed for disorders with
high mortality rates,particularly as the questionnaire is
short and easy to administer. Positive results should be
followed by further questioning rather than by
The SCOFF questionnaire is efficient at detecting
cases of eating disorders in adult women in primary
care. We recommend its use by healthcare profession-
als in primary care. Future work will assess the validity
of the questionnaire in other populations, such as ado-
lescents, in whom the prevalence may be higher.
Study to be attributed to the Department of Psychiatry at St
George’s Hospital Medical School, University of London,
London. We thank K Kennett for her help with data collection.
We also thank Wandle Valley Surgery and Brocklebank Group
Practice, particularly T Coffey, who provided study patients. We
thank the volunteers for their kind participation.
Contributors: All authors contributed to the conception and
design of the study. AL, AO’B, JB, and CP were responsible for
the collection and management of data. AL, FR, and JM
analysed and interpreted data. AL wrote the paper and all other
authors oversaw the writing and edited the paper. JM and HL
will act as guarantors.
Competing interests: None declared.
1Department of Health. A national service framework for mental health: mod-
ern standards and service models. London: Stationery Office, 1999.
Garner DM, Olmstead MA, Polivy J. Development and validation of a
multidimensional eating disorder inventory for anorexia nervosa and
bulimia. Int J Eat Disord 1983;2:15-34.
3Beglin SJ, Fairburn CG. Evaluation of a new instrument for the detection
of eating disorders in community samples. Psychiatry Res 1992;44:191-
Morgan JF, Reid F, Lacey JH. The SCOFF questionnaire: assessment of a
new screening tool for eating disorders.BMJ 1999;319:1467-8.
Perry L,Morgan J,Reid F,O’Brien A,Brunton J,Luck A,et al.Oral versus
written administration of the SCOFF. Int J Eating Disorders (in press).
(Accepted 28 February 2002)
Resource implications and health benefits of primary prevention
strategies for cardiovascular disease in people aged 30 to 74:
mathematical modelling study
The authors of this primary care paper, Tom Marshall and
Andrew Rouse, have alerted us to some errors in the
costings given for the follow up of patients (27 July, pp 197-
9). They have confirmed that despite this the conclusions of
the article remain the same.
Firstly, in table 2 in the web version (bmj.com), the costs
should read, downwards: £3567; £18 290; £32 628; £3567;
£6758; £24 489; £36 233; £14 983; £25 975; £46 270; and
£34 950. The heading for that column should read: “Cost
per event prevented.’’
Secondly, the first three paragraphs of the results
section should read as below.
Technical efficiency:maximising benefits within total resources
For any given allocation of resources to primary prevention
of cardiovascular disease more cardiovascular events can be
prevented under RM strategies than the equivalent JBR
strategies. A primary care team can prevent 5.7 events for
£40 934 under strategy RM-2 or 5.7 events for £28 090
under RM-3. The most efficient strategy for a primary care
team with a budget of £40 934 is therefore RM-3.A primary
care team can prevent 7.6 events for £116 233 under
strategy RM-1 or 7.6 events for £86 696 under RM-2. The
most efficient strategy for a primary care team with a budget
of £116 233 is therefore RM-2.For a primary care team with
a budget of over £116 233 the most efficient strategy is
Maximising efficiency within available clinical staff time
[After second sentence] At one clinic a month there is not
sufficient clinical time to assess all eligible adults. JBR
strategies therefore cannot be implemented. Strategy RM-3
can prevent 4.0 cardiovascular events at a cost of £3567 per
event prevented. RM-2 can prevent 1.1 more events at an
incremental cost of £18 290 per event prevented. RM-1 can
prevent 2.5 more events than RM-2 at an incremental cost
of £32 628 per event prevented.
Compared with one clinic a month, allocating two
clinics a month to RM-3 can prevent 1.6 more events at a
cost of £6758 per event prevented. Allocating two clinics a
month to RM-2 prevents a further 1.7 events at an
incremental cost of £24 489 per event prevented. Two clin-
ics a month following strategy RM-1 prevents a further
3.5 cardiovascular events at an incremental cost of £36 233
per event prevented.
0 0.25 0.500.75
One point is given for every "yes" answer.
• Do you ever make yourself Sick because
you feel uncomfortably full?
• Do you worry you have lost Control over
how much you eat?
• Have you recently lost more than One
stone in a 3 month period?
• Do you believe yourself to be Fat when
others say you are too thin?
• Would you say that Food dominates your
A score of 2 indicates possible anorexia
nervosa or bulimia nervosa
Receiver operating curve showing the effect of cut-off points (1 to 5) for the SCOFF
questionnaire to detect cases and non-cases of eating disorders. 1 to 5=minimum number of
positive responses to questionnaire
Submitting articles to the BMJ
We are now inviting all authors who want to submit a paper to
the BMJ to do so via the web (http://submit.bmj.com).
Benchpress is a website where authors deposit their
manuscripts and editors go to read them and record their
decisions. Reviewers’ details are also held on the system, and
when asked to review a paper reviewers will be invited to access
the site to see the relevant paper. The system is secure, protected
by passwords, so that authors see only their own papers and
reviewers see only those they are meant to.
Anyone with an internet connection and a web browser can use
The system provides all our guidance and forms and allows
authors to suggest reviewers for their paper. Authors get an
immediate acknowledgement that their submission has been
received, and they can watch the progress of their manuscript.
The record of their submission, including editors’ and reviewers’
reports, remains on the system for future reference.
The system itself offers extensive help, and the BMJ’s editorial
office will help authors and reviewers if they get stuck.
Benchpress is accessed via http://submit.bmj.com or via a link
BMJ VOLUME 3255 OCTOBER 2002bmj.com