A novel prognostic approach to defining chronic pain was developed in a US primary care low back pain population, using a combination of information about pain history, current status and likely prognosis. We tested whether this method was generalizable to a UK population. A prospective cohort of 426 patients who consulted with back pain at one of five UK general practices, and who returned follow-up information 1-year later were included. A baseline risk score was calculated based on pain severity and prognostic measures (depression, diffuse pain, pain duration), and cut-points from the US study applied for the risk of future clinically significant back pain, as defined by Chronic Pain Grades 2-4. New cut-points were also derived for the UK population using identical methods. The cut-points for probable and possible chronic pain developed in the US population (80% and 50% probability of future clinically significant back pain, respectively) were appropriate for the UK population, but the cut-point for classifying people at low risk (20% probability) was not replicated in the UK sample. The newly derived cut-points in the UK sample were similar; they remained the same for probable chronic pain, were slightly increased for possible chronic pain, and slightly reduced for those at intermediate or low risk. This method for defining chronic pain prospectively, using risk thresholds for future clinically significant pain, appears to be generalizable to a UK back pain population, particularly for identifying probable chronic pain, and may be generalizable to other primary care low back pain populations.