Sixteen observers measured eight anatomical parameters on digitalised images of six acute distal radial fractures using the Patient Archiving Communication System software and repeated the measurements at least 2 weeks later. Inter- and intraobserver reliability was calculated using intraclass correlation coefficients and tolerance limits. The highest interobserver agreement was demonstrated in the dorsal tilt (intraclass correlation coefficient 0.858; tolerance limit 14.2 degrees ). When compared with the results of a study looking at observer reliability in measurement of healed distal radial fractures, the reliability of computerised measurements is not significantly different from those achieved by manual techniques (dorsal tilt interobserver tolerance limits on computer system 16 degrees , compared to 15 degrees using ruler and protractor). These results suggest that the recommended radiological reduction limits for distal radius fractures of <10 degrees change in palmar tilt, <2 mm radial shortening, <5 degrees change in radial angle and a <1 to 2 mm articular step cannot be reliably measured.