031). Three cases of delay to prosthesis included: wound (2) and orthopaedic (1) complications. Figures 3–5 (available in the eAddenda) illustrate the percentages

of true to false positives for the clinical prediction rules time frames. This shows the clinical utility of using the clinical prediction rules for any one individual and the risk of appropriate classification. There were no significant associations between selleck compound library having a number of clinical prediction rules variables for the time frames and cessation of prosthetic use due to death, based on 29 deceased participants from the retrospective cohort (p = 0.164) and eight deceased participants from the prospective cohort (p = 0.170). Few studies have examined factors at the time of discharge in order to determine prosthetic use into the future. This is the first study to propose and validate clinical prediction rules for timelines of 4, 8 and 12 months post-discharge that use statistical optimisation modelling to select a parsimonious set of variables from the rehabilitation model of care, which predict increased likelihood of prosthetic non-use. Previous research has examined univariate associations with poor outcomes.5 In the present study, a much wider range of perioperative and demographic factors were examined and confirmed that a large number of factors are significantly associated with prosthetic non-use. These were grouped into

intrinsic, amputation and functional domains. The major point of difference DAPT from surgical studies12, 21 and 35 was that causative factors for amputation were not associated with non-use. The key point of this research, however, was that multivariate predictive models were used to determine a predictive model of outcome at out four time points. Three clinical prediction rules were derived and validated, as the results for the 4-month and 6-month outcomes were identical. These results validate that a subgroup of early prosthetic non-users exist and can be targeted. The high level of concordance between retrospective and prospective prosthetic non-use survival curves demonstrates that

there was no substantial change in clinical practice (contamination) during the validation study. These findings call for development of a model of care that optimises outcome for these individuals. Rehabilitation may focus on optimising transfers, wheelchair mobility, physical fitness and mental wellbeing rather than prosthetic gait. The present study found that having a very high number of comorbidities was significantly predictive of prosthetic non-use at 4 months, but not at later time periods. This was an interesting finding, as depending on how effectively comorbidities are managed they may become worse with age.32 However, this finding suggests that if prosthetic use can be sustained for the first 4 months post-discharge in the presence of this disease burden, then such systemic conditions may not be highly related to non-use at a later time.