© 2000 by Oxford University Press
A new classification of acute protein-energy malnutrition
A Regional Institute of Community Mental Health Care, Leonard Springerlaan 39, 7941 GX Meppel, The Netherlands B Department of General Practice, Erasmus University Rotterdam, Rotterdam, The Netherlands Z Corresponding author Fax: 31522 51983 E-mail: gernaat@worldonline.n1
Current schemes to classify protein-energy malnutrition (PEM) in hospitalized children below 5 years of age in developing countries are of limited usefulness, because age is often not known and wasted children are insufficiently discriminated. In this study, a newly proposed scheme, based on the presence of wasting (weight-for-height 
2z scores) and oedema, was compared with clinical and the original Wellcome criteria to classify PEM in 538 under-fives hospitalized for PEM in a rural hospital in Nchelenge, Zambia. Wasting in the absence of oedema was clinically false-positively diagnosed in 34.0 per cent, whereas the Wellcome classification gave rise to 28.6 per cent false-positive and 39.9 per cent false-negative diagnosis. In the presence of oedema, clinical diagnosis of wasting was 44.7 per cent false-positive and 41.4 per cent false-negative, while the Wellcome scheme had a false-negative rate of 54.3 per cent. We conclude that the Wellcome classification did not add to the clinical diagnostic accuracy of PEM as opposed to the proposed scheme, which merits further study. In an Appendix the use and calculation of z scores of weight-for-height is further explained.