The false negative rate for the clinical diagnosis (FNR) for malignant mela
noma is reported to be of the order of 20-50%. The aim of this study was to
investigate possible predictor variables for FNR, with particular referenc
e to the features and score of the seven-point check-list within the total
population (778) of histologically proved malignant melanomas presenting in
Leicestershire between 1982 and 1996. The FNR was 18.5%. The check-list wo
uld have failed as a referral indication in only 0.8-1.1% of the lesions. T
he major check-list features occurred more commonly than the minor features
, excepting size greater than or equal to 7 mm, confirming the diagnostic i
mportance of the major criteria. The FNR was unaffected by age or sex. More
rarely involved sites had higher rates (31-42%), and the face was a partic
ularly difficult diagnostic site. Clinical features of lesions associated w
ith a higher FNR were lack of irregular pigmentation and shape, altered sen
sation, the presence of inflammation and size < 7 mm. The FNR was inversely
related to the total score and major feature score, but directly related t
o the minor score. The minor features, in addition to the major features, a
re potentially valuable in avoiding false negative diagnoses and we suggest
their retention as part of the check-list, There was only one patient, in
whom the diagnosis of melanoma was initially missed and who was not biopsie
d on presentation to hospital, who re-presented after 1 year. However, the
study illustrates the importance of avoiding a false negative diagnosis as
there was marked delay in the excision of such lesions.