This study was designed to determine a cost-effective use of psychologist resources in multimodal cognitive-behavioural treatments (MMCBT) for chronic neck/shoulder pain. A randomised controlled trial was conducted with 66 patients divided in two groups. The first group (A) was treated following the approach of MMCBT with the clinical psychologist only functioning as a "coach" to the other health professionals. In this group, the psychologist had on average 5 hours of input per patients. The second group (B) was treated with the same inpatient MMCBT but with the behavioural component administered by the clinical psychologist directly to the patients. In this second group the psychologist had on average 17 hours of input per patient in the entire intervention. The outcome variables included physical, emotional and social factors, and sick-leave. Both groups showed significant improvements over time. The improvements were evident only in sub-groups, specifically in women. The only significant difference between the groups was in "perceived helplessness" favouring the "psychologist contact" setting. It is concluded that in terms of input of clinical psychology, the treatment setting with the "coaching" technique proved to be the most cost-effective use of the psychologist in the two treatment settings investigated.