Outcome measures of physical function in adult unilateral lower limb amputees during prosthetic rehabilitation: use in clinical practice and psychometric properties

The aim of this thesis is to inform clinicians and researchers of the reliability and responsiveness of the most commonly used outcome measures in prosthetic rehabilitation in the UK. In addition, this thesis supports the call for more studies of high methodological quality to provide evidence of the psychometric properties of outcome measures of physical function in lower limb amputees. A survey (study I) of Allied Health Professionals established that the outcome measures used most often during prosthetic rehabilitation in the UK were: the Timed Up and Go (TUG), a timed walk test, the Locomotor Capability Index (LCI) and its modified version (LCI-5), the Socket Comfort Score (SCS) and the Special Interest Group in Amputee Medicine (SIGAM) Mobility Grades. A standardised quality checklist (COSMIN) was used in a systematic review (study II) to measure the methodological quality and strength of evidence of the published literature that reported on the psychometric properties of outcome measures used to measure physical function during prosthetic rehabilitation. The review found mixed methodological quality ratings and many studies with small sample sizes rendering the strength of the evidence indeterminate. A limited number of studies commented on limits of agreement and measurement error when reporting on reliability. Even fewer studies reported on responsiveness with only one reporting minimally clinically important difference (MCID) values. Values for consistency, agreement and measurement error, were calculated for the top five commonly used outcome measures as identified from the survey, using a test-retest study design with a period of 7 days between tests (study III). Minimum detectable change (MDC) values were calculated for the SIGAM, LCI-5, TUG and 2MWT. The EQ-5D-5LTM, a measure of the global health of the respondent, was also included as knowledge of its psychometric properties in a population of pwLLA is unknown. However, reliability could not be confirmed for the EQ-5D-5L or the SCS in this population. A longitudinal study (study IV), based during the early rehabilitation period (mean 84 days) following provision of a primary prosthesis, gathered data to calculate indices of responsiveness for the same six outcome measures. Effect sizes were presented for five measures: SIGAM, LCI-5, TUG, 2MWT, SCS and EQ-5D-5L. Minimal clinically important difference values were also presented for the first time for all the outcome measures in this population. A patient reported change questionnaire was used as the anchor in a Receiver Operator Characteristic (ROC) curve analysis to establish the MCID values.sub_phyunpub2604_ethesesunpu