The aim of this study was to establish whether a splint programme is effective in preventing the deterioration of already evident swan neck and boutonniere deformities in rheumatoid arthritis (RA). A randomised clinical trial was done on 34 RA patients with swan neck and 34 RA patients with boutonniere deformities. The literature revealed that several factors such as age, gender, socio-economic status, time after onset of the illness and lifestyle could influence results. Care was taken to allocate, as far as possible equal numbers of patients with these attributes to the experimental and control groups. It became clear from the literature that swan neck and boutonniere deformities could manifest themselves in different forms and stages or grades of deterioration. Various splints to halt the downward spiral of the deformity were recommended, without proper scientific verification, by the authors. No specifications as to which splint was recommended for which form or grade of deformity or instructions for wearing of the splints were included. For the study the PIP hyperextension splint and the three-point-PIP extension splint was chosen for the swan neck and boutonniere deformities respectively. Patients were followed up for one year. Results were marginally positive for the prevent ion of swan neck deformities by the hyperextension splint programme, but results for the three-point PIP extension splint programmes were negative for grade I boutonniere deformities. Loss of flexor muscle strength was evident in almost all the groups (experimental and control) but more so for grade I swan neck and grade I boutonniere deformities. The variability of measurements were found to be large. Many possible sources of variation were identified, which included biological differences between people, different courses the illness could take and a weak test-retest reliability of some goniometer measurements. This fact and the relatively small sample subgroups caused some results to be not significant on the 5% level. From the significant findings, and other not significant tendencies that were too persistent to ignore, linked to the different manifestations and grades of swan neck and boutonniere deformities, recommendations were made. These suggestions will have to be tested by experimentation.