Phantom pain: a sensitivity analysis.

Abstract

To analyse how decisions to dichotomise the frequency and impediment of phantom pain into absent and present influence the outcome of studies by performing a sensitivity analysis on an existing database. Five hundred and thirty-six subjects were recruited from the database of an orthopaedic workshop and filled out a questionnaire in which the following items were assessed: demographics, side, date, level and reason of amputation, presence and frequency of phantom sensations, phantom pain and stump pain, and impediment due to phantom pain. The prevalence of phantom pain ranged from 7-72% when different cut off points for the frequency of phantom pain were applied. The significance of the various risk factors for the prevalence of phantom pain changed when different cut off points were applied. Only stump pain and phantom sensations were significant risk factors for all cut off points. Risk factors for the impediment of phantom pain changed when different cut off points were applied and these risk factors were different from those for the prevalence of phantom pain. The choice of cut off points influences the outcome of phantom pain studies considerably. This study provides some insight into the differences in prevalence and risk factors found in literature.