Background: Few epidemiological studies assessing the prevalence of chronic dermatosis are available in France and most of these studies have used biased samples drawn from specific subpopulations. As several recent studies have mentioned that self-reported diagnosis either underestimates or overestimates disease prevalence, the validity of such data is questionable.
Objectives: To evaluate the agreement between self-reported and dermatologists' diagnoses for five chronic dermatoses: acne, eczema, fungal infection, psoriasis and seborrhoeic dermatitis, and to analyse the factors associated with patients' diagnosis knowledge.
Patients/methods: A cross-sectional study was conducted on the national day of skin tumour screening in April 2003. A self-administered questionnaire was completed by patients and a standard form was completed by dermatologists after clinical examination. A total of 4,622 adults were examined by a dermatologist in 134 screening centres across France. The agreement between self-reported and dermatologists' diagnoses was analysed using the kappa index. Characteristics of patients who were aware of their diagnosis were compared with those of patients who were not, by using multivariate logistic regression models.
Results: The demographic characteristics of the sample differed from those of the French population. The self-reported prevalence was significantly lower than the actual prevalence for all diseases except eczema. Underestimations ranging from 23% to 35% were observed in nearly all subgroups of our population. The agreement between self-reported and dermatologists' diagnoses was low to moderate. Many cases of chronic skin diseases were diagnosed in patients who did not report them. Diagnosis knowledge was poorer in those above 54 years of age and better for patients treated for the condition and those with impairment of social life.
Conclusions: We observed a low agreement between self-reported and dermatologists' diagnoses for five chronic diseases. Self-reports underestimated the actual prevalence of four of five common diseases.