The Hospital Anxiety and Depression Scale (HAD) have been used in Mexico in drug abusers, burned patients, older people, with renal insufficiency and high-risk pregnant women. The aim of this study was to determine reproducibility and accuracy of the questionnaire in a sample of obese subjects. A group of 75 obese patients (BMI > 27) without diabetes mellitus were invited to participated in the study. Diagnosis of anxiety or depression was made by an structured interview based on the DSM-IV criteria, and they were requested to complete the HAD. All subjects were randomized for the manoeuvre sequence. Sensibility specificity, positive predictive value and negative value, and unweighted kappa coefficient (for concordance) were calculated for the two procedures. The questionnaire reproducibility was assessed buy test-retest with other 25 independent subjects. Internal validity was estimated by alpha Cronbach, Guttman and intraclass correlation coefficients. Mean age was 39.7 +/- 11.5 years and BMI 39.1 +/- 9.6. The best cut off point for anxiety was 8 points (Kappa 0.68) and for depression 7 points (Kappa 0.73). Mean age for test-retest was 39.2 +/- 14.5 years and BMI 45.3 +/- 14.6. The alpha-Cronbach was 0.84 for the first tes. and 0.86 for the second. Intraclass coefficient correlation was 0.946. The HAD is applicable for obese subjects, it is reproducible and concordant with a structured interview.