Background. Good communication is a crucial clinical skill. Previous research demonstrated better clinical outcomes when practitioners and patients agree about the nature of patients' core presenting complaints. We investigated the nature of this agreement and its impact on outcome among depressed primary care patients.

Method. We compared presenting problem formulations completed by patients, GPs and therapists in a primary care randomized controlled trial of cognitive-behavioural therapy and non-directive counselling for depression. Participants compiled formulations from a list of 13 potential problems of self-completed questionnaires. Subjects scored at least 14 on the Beck Depression Inventory (BDI) at baseline. Outcome measure for this study included BDI at 4 and 12 months, failure to attend for therapy when referred, dropout from therapy and patient satisfaction.

Results. Among 464 trial patients, 395 received therapy. Patient baseline problem formulations included significantly more items than GPs, who identified significantly more items than therapists. Agreement levels varied according to a range of patient and professional variables. While patients in complete agreement with their therapists about their main problem after assessment had lower average BDI scores at 12 months (9·7 v. 12·8, P=0·03); we found no other significant associations between the extent of agreement and clinical outcome. There were significant (but relatively weak) associations between agreement and aspects of patient satisfaction.

Conclusion. Our results suggest that detailed mutual understanding of the presenting complaints may be less important than agreement that the core problem is psychological, and that referral for psychological therapy is appropriate.