Can women at risk of cervical abnormality be identified?

Objective To determine if use of a detailed risk factor profile accurately predicts the presence of cytological abnormality of the cervix or improves the appropriateness of referral for colposcopic assessment when women are found to have these abnormalities.
Design Cross-sectional survey.
Setting Family planning clinic.
Population 1219 consecutive women, aged between 15 and 19 years, attending for contraceptive advice. Variables included age, social class, educational status, hormonal and obstetric history, smoking and alcohol habits, history of sexually transmitted diseases, the age of first intercourse, number of sexual partners, duration of each relationship, frequency of intercourse, contraception used and the age of each partner.
Main outcome measures Presence or absence of cytological abnormality and the presence or absence of histological abnormality in those with cytological abnormality referred for colposcopic assessment.
Results Univariate analysis confirmed many of the known associations of cervical abnormality. Discriminant analysis identified five independent significant predictors of cytological abnormality and four independent predictors of dyskaryotic cytology. At best models, derived from identified variables correctly predicted 10.1 % of individuals with cytological abnormality and 13.5% of those with dyskaryotic cytology. Of those referred for colposcopic assessment because of abnormal cytology, models were able to predict 23.5% of those with histological evidence of cervical intraepithelial neoplasia.
Conclusions Despite the availability of detailed information regarding the known correlates of cervical neoplasia in this age group, it was not possible to identify the majority of women with cervical abnormality. It is concluded that the strength of these associations is not sufficient to allow useful prediction of membership of a high risk group.