Objective - To perform a pilot study on the value of the Goldberg Depression Scale as an instrument for monitoring improvement in depressed patients treated with anti-depressants in general practice.

Design - A comparative study using simultaneous ratings on the observer-based 17-item Hamilton Depression Scale and the patientrated Goldberg Depression Scale.

Setting - General practice.

Patients - Twenty-one patients meeting the ICD-10 criteria of a moderate depressive episode were assessed at the time of inclusion and through three follow-up visits.

Main outcome measures - Scores on the Goldberg Depression Scale compared to the Hamilton Depression Scale.

Results - An acceptable internal and external validity of the Gold berg Depression Scale was demonstrated. The Loevinger coefficient varied from 0.25 at the time of diagnosis to 0.57, 0.65 and 0.69 by visits two, three and four. Factor analysis identified only one general factor explaining 50% or more of the variants, except at visit 1. When the Goldberg Depression Scale was correlated to the Hamilton Depression Scales, a coefficient of 0.74 was obtained (p < 0.001).

Conclusion - This pilot study indicates that the Goldberg Depression Scale is suitable for monitoring improvement in depressed patients treated in general practice. Further studies are recommended.