Abstract

OBJECTIVE: This study aimed to construct a decision aid to estimate the likelihood of independence with a prosthesis following rehabilitation for limb loss secondary to advanced ischaemia (acute or chronic limb-threatening ischaemia) or diabetic foot disease (DFD). A secondary aim was to determine whether prosthetic independence is a surrogate marker of long term survival. METHODS: A retrospective cohort study of a prospectively maintained database of unilateral amputations due to ischaemia or DFD entering rehabilitation between 2007 and 2020 was performed. Predictors of independent prosthetic mobility (IPM) were utilised in construction of the IPM prediction model, which underwent bootstrap internal and criterion validation through correlation with predictors of other measures of function: Timed Up and Go (TUG) and 2 minute walk test. Kaplan-Meier and Cox regression analyses were performed to address the secondary aim. RESULTS: Of the 771 patients included, only 49.9% of amputees achieved IPM. Independent negative predictors of IPM were age > 75 years, female sex, higher amputation level, active malignancy, cerebrovascular disease, end-stage renal disease, and cognitive impairment. The model yielded high discrimination (C statistic 0.778), and internal validation was demonstrated with bootstrapping (C statistic 0.778), confirming no over-optimism. There was a strong correlation between IPM, TUG, and 2 minute distance and their predictors, confirming strong criterion validity. The IPM group had a median survival of 93.7 (80.7, 105) months, whereas the non-IPM group fared worse with a median survival of 56.6 (48.5, 66.7) months (p < .001). CONCLUSION: An internally validated decision aid for estimating the likelihood of independence with a prosthesis after major amputation was constructed. A strong association between female sex and poorer prosthetic mobility was observed. Prosthetic function was shown to be a surrogate marker of long term survival. Future research will involve external validation studies to confirm the generalisability of the decision aid in clinical practice.