Selecting, Administering, and Interpreting Outcome Measures among Adults with Lower-Limb Loss: An Update for Clinicians

Abstract

Purpose of review:

To summarize outcome measurement research among adults with lower-limb loss (LLL) for clinicians.

Recent findings:

Houghton Scale, Prosthetic Evaluation Questionnaire-mobility subscale (PEQ-m), Prosthetic Limb Users Survey of Mobility (PLUS-M™), Activities-Specific Balance Confidence Scale (ABC), Amputee Mobility Predictor (AMP), Comprehensive High-Level Activity Mobility Predictor, Four Square Step Test (FSST), Narrowing Beam Walking Test (NBWT), L Test, 10 Meter Walk Test (10MWT), and 6 Minute Walk Test (6MWT) are appropriate for evaluating individual patient changes post-LLL. Post-LLL, Socket Comfort Score, Patient-Specific Functional Scale, Patient-Reported Outcomes Measurement Information System 29-Item Profile, Timed Up and Go, and 2 Minute Walk Test may be more appropriate for evaluating groups. Minimal detectable change is available for 15/20 reviewed measures. Many measures differ between mobility levels.

Summary:

Quick, reliable measures for evaluating patient functional change include ABC, PEQ-m, PLUS-M™, FSST, and L Test; when resources allow, NBWT, 10MWT, 6WMT and AMP may be considered.

INTRODUCTION

During examinations of adults with lower-limb loss (LLL), clinicians collect various types of data, including subjective information (e.g., amputation date, etiology), clinical metrics (e.g., weight, range-of-motion), and self-report and performance-based outcome measures. Outcome measure data provide useful information that motivates patients, guides clinical decisions, facilitates communication among healthcare providers, improves patient outcomes, and justifies care to payers.[1] Major barriers to outcome measure use in clinical practice include difficulties with selection, administration, and interpretation.[1]

Outcome measure selection takes into consideration patient-specific factors (e.g., time since amputation, amputation level, comorbidities), as well as the goals of the evaluation (e.g., to determine mobility level classification, to predict adverse outcomes) or re-evaluation (e.g., to justify need for prosthetic componentry or future services). Administration and re-administration of outcome measures relies on standardized procedures, so differences can be interpreted as patient change void of procedural variations. Data interpretation challenges clinicians to consider measurement error, as well as other factors impacting self-report or performance (e.g., socket fit, pain, fatigue, depression).

Comprehensive clinical examinations include subjective information, impairment-based metrics (e.g., range-of-motion), and self-report and performance-based outcome measures evaluating various domains. Evaluating various domains (e.g., balance, functional mobility) provides a more complete picture of the patient. Domains can be population-specific (e.g., socket fit) or non-specific (e.g., gait). Performing both prosthesis-specific measures and outcome measures used across patient populations (where reference data from healthy individuals is available) allows for comparison of the patient’s status to both peers with LLL and similarly-aged adults without LLL. Comparison to peers with LLL reassures patients they are on a typical trajectory post-amputation. That said, post-amputation outcomes are largely mediocre. For example, patients post-LLL report low levels of community participation[2] and most do not return to pre-amputation employment[3]. Consequently, comparison to peers with LLL may result in overestimating recovery. Conversely, comparison to similarly-aged adults without LLL can identify persistently-impaired domains for targeted intervention, so patients can reach their maximal potential post-amputation, effectively enhancing community and work reintegration.

Clinicians are encouraged to include both self-report and performance-based measures in LLL examinations. While self-report measures assess patient perception, performance-based measures assess capacity under a set of conditions (e.g., a given socket fit, pain presentation, fatigue-level). Discordance between self-report and performance-based measures evaluating the same domain is common, with patients often overestimating their status.[4–6] While questionnaires are often more easily administered in busy practice settings, supplementing self-report measures with performance-based measures not only improves the patient’s self-assessment (as patients become more accurate in reporting after task performance[7]) but data discrepancies can help fine-tune clinical interventions. For example, a patient with LLL with poor balance-confidence but moderate balance performance may benefit from an intervention supplemented with cognitive-behavioral therapy strategies.[8]

Included outcome measures in this review are prosthesis-specific or have been used in LLL research. For each measure, psychometric properties (e.g., reliability, validity) for adults with LLL are discussed. Specific attention is given to between-days, test-retest reliability, which allows calculation of minimal detectable change (MDC; i.e., change surpassing measurement error considered to be ‘true change’).[9] MDC can be calculated at the 90% (MDC₉₀) or 95% (MDC₉₅) confidence level. Changes meeting or exceeding MDC₉₀ may be considered acceptable when evaluating clinical interventions (e.g., exercise-based, prosthetic modification). Conversely, changes exceeding MDC₉₅ may be more appropriate when evaluating surgical interventions, as severe adverse events are more likely; thus, the potential negative consequences of a wrong clinical decision are greater.[10] For group comparisons (e.g., evaluating results for groups of patients with LLL to determine quality-of-services), test-retest reliability, as indicated by an intraclass correlation coefficient (ICC) of ≥.70 is desired.[11, 12] For evaluating change over time within a single patient, an ICC of ≥.90 is desirable.[13] For measures with ordinal scoring by the clinician (i.e., Amputee Mobility Predictor, Berg Balance Scale), high inter-rater reliability enables re-evaluations to be conducted by a different practitioner than the baseline evaluator. Minimally clinical important difference (MCID; i.e., the smallest change reflecting a meaningful change as perceived by the patient, family, clinician, or other entity,[14]) is provided when available. Floor (e.g., patient is unable to complete test) and ceiling (e.g., demonstration of further improvement, e.g., MDC, is unlikely) effects are also provided, as clinicians should select another measure for evaluating change in these cases.

SELF-REPORT OUTCOME MEASURES

Self-report measures, i.e. questionnaires, may be administered in paper or electronic format. Administration mode should be determined based on clinic workflow. Having interpretation sheets readily available allows support personnel to assist with scoring and eliminates the need for memorization of MDCs, MCIDs, reference values, and other criteria necessary for interpreting data.

POPULATION-SPECIFIC

The following self-report measures were developed specifically for adults with LLL who are using a prosthesis, and thus, are not appropriate for use during the pre-prosthetic rehabilitation phase.

Socket Fit

Socket Comfort Score (SCS).

With SCS, prosthetic socket comfort is rated 0=‘most uncomfortable fit’ to 10=‘most comfortable socket fit imaginable’.[15] As patients are familiar with rating pain on the Visual Analogue Scale (0=‘no pain’ to 10=‘worst imaginable pain’[16]), clinicians are encouraged to carefully review SCS anchors with the patient to ensure accurate reporting. SCS between-days, test-retest reliability has been reported (ICC=.77 paper; ICC=.79 electronic format) and indicates appropriateness for group comparisons, although use of the same format for SCS administration is recommended as mixed-mode administration (e.g., paper at baseline, electronic at follow-up) has poorer reliability (ICC=.63).[17] Low SCS ratings correlate with evidence of poor fit (e.g., persistent redness), including prosthetist judgement of poor fit.[15] The MDC is 2.73 points,[17] thus, a 3-point change (e.g., 5/10 at baseline, 8/10 at follow-up) indicates an intervention (e.g., prosthetic modification) has resulted in a ‘true change’ (e.g., improved socket fit).

Prosthesis Use & Mobility

Houghton Scale.

Houghton Scale is a 4-item questionnaire evaluating prosthesis wear time and use, assistive device use, and perceived stability walking on flat surfaces, slopes, and rough ground.[18] Scores range from 0–12.[18] Higher scores indicate greater prosthesis use and prosthesis-enabled mobility. Test-retest reliability (ICC=.96)[19] suggests Houghton Scale is appropriate for evaluating individual patient change, although MDC is not reported. Scores ≤5 may indicate household ambulators, scores of 6–8, limited-community/household ambulators, and scores ≥9, community-ambulators.[20] Prior research suggests lack of floor and ceiling effects,[20, 19] but scores of 12/12 require use of another measure with items posing greater mobility-challenges to evaluate subsequent patient improvement.

Prosthetic Evaluation Questionnaire-mobility subscale (PEQ-m).

PEQ-m consists of 12 items assessing prosthesis-enabled ambulation and transfers.[21] Each item is scored 0=‘unable’ to 4=‘no problems’, resulting in a total score of 0–48 points or an average score of 0–4; for both, higher scores indicate better prosthesis-enabled mobility.[14] Test-retest reliability is high (ICC=.90–.95) suggesting appropriateness for assessing individual changes pre- and post-intervention. MDC₉₅ is .65/4 points or 5.5/48 points (16.3% and 11.5% of the maximal scores, respectively), and the MCID for patients undergoing rehabilitation is 8/48 points (i.e., 16.7% of the maximal possible score).[14, 17] No floor or ceiling effects are reported.[17] When individuals are classified with the Houghton Scale, PEQ-m scores (maximum: 48) are significantly different between ambulatory groups, i.e., household ambulators (14.8±9.4points), limited-community/household ambulators (26.4±8.8points), and community-ambulators (36.6±9.9points).[20]

Prosthetic Limb Users Survey of Mobility (PLUS-M™).

PLUS-M™ evaluates perceived ability to carry out household and outdoor ambulation activities for individuals aged ≥18 years with prosthesis experience.[22] Each item is scored 1=‘unable to do’ to 5=able to do ‘without any difficulty’; items are totaled.[22] Totals are associated with T-scores and percentiles, where higher values indicate better mobility.[22] When a patient’s T-score is above or below 50, he/she is above or below the mean with respect to the reference sample of adults with LLL, respectively. Each T-score corresponds to a percentile, and the percentile can be interpreted as “the individual reports better mobility than [insert percentile] of adults with unilateral amputation from the reference sample”.[22] A reference sample of adults with bilateral lower-limb loss (ankle-level or higher, but hip level or lower) is also available.[22] Test-retest reliability (using T-score) has been established for paper (12-item and 7-item: ICCs=.97), electronic (12-item: ICC=.95; 7-item: ICC=.94), and computer-adapted (ICC=.92) administration methods.[17] MDC₉₀ is 4.50, 4.69, and 6.42 for the 12-item, 7-item, and computer-adapted version, respectively.[17] Clinicians should be sure to use the correct (a) table for converting raw scores during paper form administrations, (b) MDC when comparing baseline and follow-up data based on the test version, and (c) reference sample data based on unilateral or bilateral limb involvement. To evaluate mobility over time, the 12-item PLUS-M™ is recommended.[17] Known-groups construct validity has been established among adults with LLL, as increasing PLUS-M™ T-scores are associated with increasing functional mobility level classification (i.e., K-level).[22] PLUS-M™ floor and ceiling effects have not been found.[17]

NON-SPECIFIC

The following self-report measures were developed for other patient populations, but have undergone psychometric testing among adults who are using a lower-limb prosthesis.

Balance

Activities-Specific Balance Confidence Scale (ABC).

The ABC asks individuals to rate their confidence that they will not lose their balance or become unsteady with 16 activities (e.g., reaching, walking on icy sidewalk).[23] Each item is scored 0%=‘no confidence’ to 100%=‘complete confidence’, and items are averaged. Test-retest reliability (ICC=.90) and MDC₉₅ of 14.36% suggest appropriateness for evaluating change in individual patient status.[23] Among adults with a unilateral transfemoral or transtibial amputation, greater balance-confidence has been associated with better mobility and greater community participation.[24] A shortened version, the 6-item ABC,[25] while reliable (ICC=.92), requires a greater MDC₉₅ (i.e., 21.45%) to detect change, has poor agreement with the 16-item ABC, and therefore, is not recommended following LLL.[23]

A revised ABC, using an ordinal score of 0=‘no confidence’ to 4=‘complete confidence’ (where items are averaged) has been advocated for adults with LLL.[26] Test-retest reliability (ICC=.95) is similar to the original, 16-item ABC with a MDC₉₅ of .58 (14.5% of maximal score)[17] regardless of administration mode (i.e., paper, electronic). Floor and ceiling effects are not found,[17] and individuals with higher ABC scores are more likely to be classified into higher mobility classifications, suggesting known-groups construct validity.[27] Further, when individuals with LLL are classified with the Houghton Scale, ABC is significantly different between ambulatory groups, i.e., household ambulators (1.0±0.8), limited-community/household ambulators (1.7±0.7), and community-ambulators (2.9±0.8).[20]

Multiple Domains

Patient-Specific Functional Scale (PSFS).

An advantage to administering the PSFS is the patient self-identifies and rates 5 items, scored 0=‘can’t perform the activity at all’ to 10=‘can perform the activity fully’.[28] Patients may list work-specific tasks or recreational and sports activities that may not be found on other questionnaires. As with SCS, clinicians are encouraged to review anchors with patients to ensure accurate reporting. Higher total scores indicate better function. Test-retest reliability (ICC=.83)[28] suggests PSFS appropriateness for evaluating groups of patients with LLL over time, but not necessarily individual patients, negating the advantage of item individualization. MDC₉₅ for adults with unilateral LLL is 11.2 points,[28] i.e., 22.4% of the maximal possible score, which may be difficult to meet, reducing the measure’s clinical utility.

Patient-Reported Outcomes Measurement Information System 29-item Profile (PROMIS-29).

PROMIS-29 is a generic health-related quality-of-life survey for individuals aged ≥18 years that includes the following eight subscales: physical function, anxiety, depression, fatigue, sleep disturbance, satisfaction with participation (v1.0)/ability to participate in social roles and activities (≥v2.0), pain interference, and pain intensity.[29] Items are scored using a 7-day recall window, except for physical function, which emphasizes current capabilities.[30] Higher scores indicate more of the construct being assessed, so high scores for physical function and participation are preferred, while lower scores on all other subscales are preferred. Scores are totaled for each of the first 7 subscales and then converted to a T-score, where 50 represents the mean of the reference population, i.e., the United States general population, and 40 and 60 represent 1 standard deviation below or above the mean, respectively.[29] Test-retest reliability among adults with LLL (ICC=.79–.88) indicates clinical utility for group comparisons using paper or electronic forms for PROMIS-29 v1.0, with the following MDC₉₀: physical function=6.13, anxiety=7.81, depression=6.71, fatigue=7.74, sleep disturbance=7.61, satisfaction with participation=9.53, and pain interference=8.51.[17] The 8^th subscale, pain intensity, is rated on a 0–10 scale, where 0=‘no pain’ and 10=‘worst imaginable pain’. Among adults with LLL, between-days, test-retest reliability for pain intensity is reported (ICCs=.85–.89; MDC₉₀=1.97).[17] Unfortunately, among adults with LLL, floor effects are possible for pain intensity rating, as well as PROMIS-29 v1.0 anxiety, depression, and pain interference subscales, while ceiling effects are possible for physical function and participation subscales.[17] Psychometric evaluation of the updated PROMIS-29 v2.1 among adults with LLL may determine if the revised measure is appropriate for evaluating individual patient changes.

PERFORMANCE-BASED OUTCOME MEASURES

The following performance-based outcomes were developed to evaluate lower-limb prosthesis-enabled mobility or have been used in LLL research. For population-specific measures, multiple domains are assessed, including transitions, balance, functional mobility, and gait. Non-specific measures generally focus on a single-domain. Figure 1 provides course layouts. Table 1 provides details regarding test-retest reliability, MDC, MCID, equipment and administration time requirements, patient instructions, and scoring. Tables 1 and 2 provide data to assist with results interpretation.

Figure. 1.

Course layouts are provided for the (A) Four Square Step Test, (B) 10 Meter Walk Test, (C) Figure-of-8 Walk Test, (D) Timed Up and Go, (E) L Test of Functional Mobility, and (F) 6 or 2 Minute Walk Tests, where either a linear or looped course may be used. Each ‘X’ represents a cone.

Table 1.

Outcome Measure Details For Selection, Administration, and Interpretation among Patients with Lower-Limb Loss

Outcome Measure	Psychometric Data in LLL	Required Equipment	Time (min)	Instructions	Scoring	Available Data in Adults without LLL for Comparison
POPULATION-SPECIFIC
Amputee Mobility Predictor (AMP)	Between-days, test-retest reliability: ICC=.88[28]-.98[31] Inter-rater Reliability: ICC=.99[31] MDC90: 3.4 points[28] MCID: NR	2 standard chairs with armrests; 4”- high shoebox; sets of 3 stairs; ruler; pencil; tape; stopwatch 3.7m linear space[31]	<15[31]	Use with unilateral amputations and bilateral transtibial LLL (use bilateral AMP if higher level(s) LLL) [31, 32] Follow form instructions[31] 1 trial/item[31] Assistive device allowed[31]	Add scores from 21 items[31] Score: 0–47 points[31] Higher scores indicate better functional mobility[31]	NR
Comprehensive High-Level Activity Mobility Predictor (CHAMP)	Between-days, test-retest reliability: ICC=.97[35] MDC95: 3.7 points[35] MCID: NR No ceiling effect[36]	2 chairs/ walker; 8–18, 6- inch cones; tape; counter; stopwatch 13 × 10m floor space[35]	<15[35]	Pre-requisites: (a) AMP ≥37 points[35, 36] AND (b) 6MWT ≥250m[35] OR 2MWT >97m[36] Demonstrate and read instructions[35] 2–3 trials/item[35] 30–60s between trials[35]	Each item scored 0–10[35] Sum all 4 items (score: 0–40)[35] Higher scores indicate better balance, mobility, and agility	Male service members, aged 18–40 years (n=97): 35.4±1.2 points[33]
NON-SPECIFIC
5-Times Sit-to-Stand Test (5XSTS)	Between-days, test-retest reliability: NR MDC: NR MCID: NR	Standard chair (with or without armrests); stopwatch 1.5m × 1.5m floor space	<5	Pre-requisite: Ability to transfer sit-to-stand without upper extremities Clinician demonstrates[41] “With arms folded across the chest, stand up straight and sit down as quickly as possible 5 times.”[41] Start timer on “Begin” and end timer at end of 5th stand[41] or when the buttock returns to the chair after the 5th stand[82] Allow rest breaks between trials	Inability to complete 5 stands = invalid test Record faster time of 2 trials[41] Faster times indicate better transfers and functional leg strength	Adults, aged 41±11 years (n=32): 8.2±1.7s[83] Adults:[84] Aged 60–69 years (n=4184): 11.4s (95%CI: 11.4–11.4) Aged 70–79 years (n=8450): 12.6s (95%CI: 12.6–12.6) Aged 80–89 years (n=344): 12.7s (95%CI: 10.7–14.8)
Four Square Step Test (FSST)	Between-days, test-retest reliability: ICC=.97[42] MDC90: 2.0s[42] MCID: NR Floor effect[43]	4 canes; taped course (Figure 1); stopwatch 3×3m floor space	<5[85]	Pre-requisite: Not appropriate if using walker[85, 86] Clinician demonstrates[42, 85] “Complete the sequence as fast as possible without touching the canes. Both feet must make contact with each square. If possible, face forward during the entire sequence.”[42, 82, 85] Start timer when 1st foot contacts box 2, and stop when both feet return to box 1[82, 85] 1 practice trial[42, 43, 82, 85] Do not call out numbers during test as this may pace the patient, affecting the time Trial is invalid if incorrect sequence, loss of balance, or cane is touched[42, 85] May repeat trial if invalid[42]	Record faster time of 2 trials[42, 85] Faster times indicate better dynamic balance If necessary, patient may turn to clear canes, increasing the test time	Adults, aged 74±6 years (n=27): median 8.70s (IQR: 7.36–10.01)[85] Adults, aged 84±6 years (n=45): 15.8±8.3s[86]
Functional Reach (FXR)	Between-days test-retest reliability: NR MDC: NR MCID: NR No floor or ceiling effect[46]	Yardstick secured to wall, adjustable to height of patient’s acromion[44] 1.2m of wall space	<2	Align acromion with yardstick[44] Use patient’s right arm if bilateral LLL or ipsilateral arm if unilateral LLL[44] “Make a fist. Reach forward as far as possible without losing your balance or taking a step” [44, 45] Do not specify reaching strategy[44, 45] 5 trials[44] Trial considered invalid if wall touched or balance lost (e.g., step occurs); repeat trial[44, 45]	Record distance (in cm) from starting position of 3rd metacarpal[44, 45] Average trials #3–5[44] Greater distance from starting position to ending position indicates better balance	Adults:[45] Aged 20–40 years M (n=16): 42.5±4.9cm F (n=28): 37.2±5.5cm Aged 41–69 years M (n=22): 38.0±5.6cm F (n=28): 35.1±5.6cm Aged 70–87 years M (n=20): 33.4±3.9cm F (n=14): 26.6±9.0cm
Narrowing Beam Walking Test (NBWT)	Between-days test-retest reliability: ICC=.90[42] MDC90: 0.16[42] MCID: NR No floor or ceiling effect with higher-functioning adults with LLL[43, 47] Possible floor effect among LLL with lower mobility levels (e.g., cane use)[47]	Narrowing beam with 4 segments (each 1.83m with 3.8cm height): 18.6cm, 8.6cm, 4.0cm, and 2.0cm, with marks placed every 15.24cm on beam[47] 7.32m linear floor space[47]	<5	Pre-requisite: Able to walk 10 feet without assistive device (e.g., cane, walker)[47] Begin with foot of choice on beam and other foot on ground[42] “On ‘Go’, walk along the beam as far as you can. Keep your arms crossed over your chest as you walk. Once you move your arms away from your body or step off the beam, I will ask you to stop.”[42, 43] Measure distance when patient uncrosses arms or steps off the beam[42] No practice trial[42] Perform 5 trials[43, 47]	Use average of trials #3–5 and divide by 22. [42, 43, 47] Score is normalized distance[43, 47] Score: 0 (no distance covered)-1.0 (full beam) [47] Must walk past .61m mark to receive a score >0[42]	NR
Berg Balance Scale (BERG)	Between-days test-retest reliability: NR Inter-rater Reliability: ICC=.95[51]-.99[87] MDC: NR MCID: NR Ceiling effect[50, 51, 43, 46]	Standard chairs: 1 with and 1 without armrests; 7–9” step stool; ruler; slipper; stopwatch 2×2m floor space	<20[51]	14 items[88] Read instructions for each item Each item scored 0 (unable to perform task) – 4 (able to perform task independently) [88]	Score: 0–56 points[88] Higher scores indicate better balance in daily life[55]	Community-dwelling older adults: Aged 73±8 years (n=35): 52.7±4.0 points[89] Aged 80±5 years (n=182): 50.7±7.7 points[90]
Timed Up and Go (TUG)	Between-days test-retest reliability: ICC=.88 (line/cone)[28, 42] MDC90: 3.6s[28] (line); 2.2s (cone) [42] MCID: NR Ceiling effect[46, 57]	Standard chair with armrests; taped course (Figure 1); stopwatch[56] 4m linear space	<2	“On the word ‘Go’, get up from a chair and walk to the line on the floor [or around the cone], turn around, return to the chair and sit down again. Please walk at a normal comfortable pace.” [28, 42, 56] Start stopwatch on ‘Go’ and end when buttock 1st touches chair on return[42, 56] Assistive device allowed[46] 1 practice trial[28] 1–2 recorded trials[28, 42]	Record time to complete test (if 1 trial) or fastest trial[28, 42] Faster times indicate better functional mobility	Adults:[91] Aged 60–69 years (n=176): 8.1s (95%CI: 7.1–9.0) Aged 70–79 years (n=798): 9.2s (95%CI: 8.2–10.2) Aged 80–89 years (n=1102): 11.3s (95%CI: 10.0–12.7)
L Test of Functional Mobility (L Test)	Between-days test-retest reliability: .95-.99 (dual-task: .93-.98) [59] MDC95: 2.2–3.2s (dual-task: 2.9–3.8)[59] MCID for patients undergoing intervention (e.g., new socket, additional therapy): 4.5s[58]	Standard chair without armrests; [58] taped course (Figure 1); stopwatch See Figure 1 for space requirements	<2[51]	Clinician demonstrates “On the word ‘Go,’ stand up from the chair, walk to the 1st line, turn 90 degrees, and walk to the 2nd line, then turn 180°, and return to sit in the chair. Please walk at a normal comfortable pace.” [58, 92] Assistive device allowed[57, 59, 92]	Record time to complete the test[59] Faster times indicate better functional mobility	Adults:[93] 60–69 years (n=34): 20.1±3.1s 70–79 years (n=31): 27.3±8.9s 80+ years (n=40): 35.9±16.6s
10 Meter Walk Test (10MWT)	Between-days test-retest reliability: ICC=.96[42] for time for 10m MDC90: 0.91s for time for 10m[42] MCID: NR	Taped course (Figure 1); stopwatch 14m linear space[42, 60]	<5	“On ‘Go’ walk at your preferred comfortable walking speed [or as fast as safely possible without running] to the finish line.”[34, 42, 94] Assistive device allowed[60] 1 practice trial[42] Start timer when patient crosses 2m mark and stop timer when patient crosses 12m mark[42, 60] 2 trials per condition[42] Calculation of gait speed: distance/time = speed (m/s)[60]	Record best of 2 trials (per condition)[42] Faster times (s) or greater speeds (m/s) indicate better gait	Adults: Self-selected [fast] gait speed:[94] 20–29 years (n=37) M: 1.39±.15m/s [2.53±.29m/s] F: 1.41±.18m/s [2.47±.25m/s] 30–39 years (n=36) M: 1.46±.09m/s [2.46±.32m/s] F: 1.42±.13m/s [2.34±.34m/s] 40–49 years (n=43) M: 1.46±.16m/s [2.46±.36m/s] F: 1.39±.16m/s [2.12±.28m/s] 50–59 years (n=43) M: 1.39±.23m/s [2.07±.45m/s] F: 1.40±.15m/s [2.01±.26m/s] 60–69 years (n=36) M: 1.36±.21m/s [1.93±.36m/s] F: 1.30±.21m/s [1.77±.25m/s] 70–79 years (n=42) M: 1.33±.20m/s [2.08±.36m/s] F: 1.27±.21m/s [1.74±.28m/s]
Figure-of-8 Walk Test (F8WT)	Between-days test-retest reliability: NR Inter-rater Reliability: NR MDC: NR MCID: NR	2 cones; taped course (Figure 1); stopwatch	<2	Demonstrate “Walk around the 2 cones as quickly and accurately as possible.”[41] 2 trials[41] Assistive device allowed[41]	Record best time[41] and number of steps[67, 68] Faster times and fewer steps indicate better gait[67]	NR
6 Minute Walk Test (6MWT)	Between-days test-retest reliability: ICC=.97[28] MDC90: 45m[28] MCID: NR	2–4 cones to mark course (Figure 1); rolling measurement wheel; stopwatch; blood pressure equipment, pulse oximeter, chair[70, 95] Long corridor (≥30m) [28, 95] or loop (Figure 1)	<10	Pre-requisite: Assess vital signs to establish patient safety for administration:[95] Relative contraindications: systolic BP ≥200mmHg OR diastolic ≥120mHg) Absolute contraindication: Sp02 ≤ 85% on room air as assessed with pulse oximetry “Cover as much ground as possible in [2/6] minutes. Walk as fast as possible, but safely. You may rest at any point, but the clock will not stop, so start walking again when able.” Allow to sit (or lean against wall) and rest if needed[28, 70, 95] Refrain from conversation during testing to maximize effort Standardized encouragement at each minute, e.g., “Keep up the good work, you have [X] minutes left.” [70, 95] Monitor pulse oximetry during test Halt test if SpO2<80%[95] Participants may not run or jog[95] Assistive device allowed[28]	Record distance walked (in m) by trailing patient with rolling measurement wheel; walk slightly behind to avoid pacing the patient[95] Greater distances indicate better walking endurance[95] With retesting, be sure to use the same course layout as the number of turns can impact the distance walked[95, 96]	Adults:[97] 60–64 years (n=861) M: 616±84m F: 551±77m 65–79 years (n=1566) M: 577±94m F: 519±92m 70–74 years (n=1813) M: 560±93m F: 501±90m 75–79 years (n=1451) M: 507±115m F: 465±104m 80–84 years (n=784) M: 479±110m F: 422±107m 85–89 years (n=470) M: 436±130m F: 390±118m 90–94 years (n=234) M: 369±135m F: 326±115m
2 Minute Walk Test (2MWT)	Between-days test-retest reliability: ICC=.83[28] MDC90: 34m[28] MCID: NR No floor or ceiling effect[46]	Same at 6MWT	<5[51]	Same as 6MWT	Same as 6MWT	Adults:[75] 18–54 years (n=799) M: 201m (95%CI: 197–205) F: 183m (95%CI: 181–185) 55–59 years (n=53) M: 191m (95%CI: 177–205) F: 176m (95%CI: 168–185) 60–64 years (n=77) M: 179m (95%CI: 165–193) F: 166m (95%CI: 158–175) 65–69 years (n=44) M: 184m (95%CI: 171–198) F: 155m (95%CI: 141–170) 70–74 years (n=65) M: 172m (95%CI: 164–181) F: 146m (95%CI: 137–155) 75–79 years (n=33) M: 158m (95%CI:140–175) F: 141m (95%CI: 122–160) 80–85 years (n=66) M: 144m (95%CI: 133–156) F: 134m (95%CI: 126–143)

Data presented as mean±standard deviation, unless otherwise noted.

Abbreviations: LLL=lower-limb loss; NR=not reported; ICC=intraclass correlation coefficient; MDC=minimal detectable change; MCID=minimal clinically-important difference; CI=confidence interval; IQR=interquartile range; M=male; F=female

Table 2.

Available Data for Interpreting Results

Outcome Measure	Sample Characteristics and Size (n)	Score
POPULATION-SPECIFIC
Amputee Mobility Predictor (AMP), points	Male service members, aged 18–40 years, 3.2±1.9 years post-LLL[35]
	- Unilateral transtibial (n=60)	45.7±1.1
	- Unilateral transfemoral (n=32)	43.4±1.2
	- Bilateral LLL (n=26)	41.3±2.8
	Unilateral LLL ≥2 years prior, aged 66±13 years (n=44)[28]	40±4
	Unilateral LLL 10±12 years prior, aged 55±19 years[31]†
	- K0–1 (n=18)	25.0±7.4
	- K2 (n=43)	34.6±6.5
	- K3 (n=67)	40.5±3.9
	- K4 (n=39)	44.7±1.8
	Unilateral transtibial or transfemoral ≥1 years prior[34]
	- K3, aged 60±12 years (n=35)	40.4±0.4
	- K4, aged 46±12 years (n=20)	44.9±0.6
Comprehensive High-Level Activity Mobility Predictor (CHAMP), points	Male service members, aged 18–40 years, 3.2±1.9 years post-LLL[35]
	- Unilateral transtibial (n=60)	26.9±5.4
	- Unilateral transfemoral (n=32)	19.7±3.3
	- Bilateral LLL (n=26)	18.8±4.5
	Unilateral LLL, aged 18–60 years[36]
	- Male (n=33)	19.1±6.8
	- Female (n=12)	14.5±7.4
	- Transtibial (n=35)	19.5±6.9
	- Transfemoral (n=10)	12.4±5.4
	- K3-level (n=25)	14.2±6.7
	- K4-level (n=20)	22.6±4.6
NON-SPECIFIC
5-Times Sit-to-Stand (5XSTS), s	Unilateral LLL ≥1 year prior[41]
	- Transtibial
	○ K3, aged 50 (95%CI: 46–55) years (n=28)	8.85 (95%CI: 8.07–9.63)
	○ K4, aged 40 (95%CI: 35–46) years (n=22)	8.06 (95%CI: 7.17–8.94)
	- Transfemoral
	○ K3, aged 50 (95%CI: 40–61) years (n=10)	13.21 (95%CI: 11.90–14.51)
	○ K4, aged 38 (95%CI: 26–49) years (n=7)	10.15 (95%CI: 8.58–11.72)
Four Square Step Test (FSST), s	Bilateral transfemoral, aged 27.7±8.4 years[98]
	- Full-length prostheses (n=17)	22.0±10.2
	- Stubby-length prostheses (n=5)	21.4±5.3
	Unilateral transtibial or transfemoral, aged 55±21 years, 16±15 years prosthesis use (n=60)[42]	10.4±5.3
	Unilateral LLL ≥1 year prior, modified (without canes)[41]
	- Transtibial
	○ K3, aged 50 (95%CI: 46–55) years (n=28)	8.82 (95%CI: 8.30–9.34)
	○ K4, aged 40 (95%CI: 35–46) years (n=22)	7.80 (95%CI: 7.21–8.38)
	- Transfemoral
	○ K3, aged 50 (95%CI: 40–61) years (n=10)	10.34 (95%CI: 9.47–11.21)
	○ K4, aged 38 (95%CI: 26–49) years (n=7)	8.67 (95%CI: 7.63–9.72)
Functional Reach (FXR), cm	Transmetatarsal, 2.6±3.1 years prior, aged 63.2±9.2 years (n=15)[44]	19.1±8.6
Narrowing Beam Walking Test (NBWT), 0–1	Transtibial or transfemoral, 14±13 years prior, aged 49±15 years (n=40)[47]	0.41±0.06
	Unilateral transtibial or transfemoral, aged 55±21 years, 16±15 years prosthesis use (n=60)[42]	0.42±0.21
	Unilateral transtibial or transfemoral 14±13 years prior, aged 49±15 years (n=40)[43]
	- Transtibial (n=25)	0.46 (95%CI: 0.39–0.54)
	- Transfemoral (n=15)	0.30 (95%CI: 0.21–0.40)
	- Non-fallers (n=16)	0.51 (95%CI: 0.43–0.58)
	- Fallers (n=24)	0.33 (95%CI: 0.25–0.42)
	- K1/K2 (n=17)	0.26 (95%CI: 0.18–0.34)
	- K3/K4 (n=23)	0.51 (95%CI: 0.45–0.57)
Berg Balance Scale (BERG), points	Unilateral transtibial >1 year prior, aged 18–35 years (n=24)[99]	54±1
	Unilateral transtibial[53]
	- Active, aged 26.7±7.8 years, 12±7 years of prosthesis use (n=12)	54.2±1.0
	- Sedentary, aged 33.0±6.7 years, 14±8 years prosthesis use (n=12)	52.1±3.9
	Unilateral or bilateral LLL, 21±15 years prior, aged 54±12 years (n=30)[51]	51±5
	Unilateral transfemoral[52]
	- Aged 44±18 years (n=18)	52±3
	- Aged 70±5 years (n=12)	41±5
Timed Up and Go (TUG), s	Unilateral transtibial or transfemoral ≥1 years prior[34]
	- K3, aged 60±12 years (n=35)	12.8±0.5
	- K4, aged 46±12 years (n=20)	9.5±0.8
	Unilateral transtibial or transfemoral, aged 55±21 years, 16±15 years prosthesis use (n=60)[42]	10.4±2.8
	Unilateral LLL ≥2 years prior, aged 66±13 years (n=44)[28]	12.3±4.5
	Unilateral LLL 4 (range: 0.3–17) years prior[56]
	- Transtibial, aged 74 (range: 61–86) years (n=27)	23.8±23.0
	- Transfemoral, aged 73 (range:68–81) years (n=5)	28.3±12.2
L Test of Functional Mobility (L Test), s	Unilateral transtibial[53]
	- Active, aged 27±8 years, 12±7 years prosthesis use (n=12)	11.3±1.8
	- Sedentary, aged 33±7 years, 14±8 years prosthesis use (n=12)	16.4±9.7
	Unilateral LLL, aged 61±7 years (n=19)[92]
	- At discharge from inpatient rehabilitation	72.4±44.6
	- 4-months post-discharge from rehabilitation	48.8±30.0
	Unilateral LLL 12±15 years prior, aged 56±14 years (n=93)[57]	32.6±14.9
	Unilateral transtibial[59]
	- Dysvascular etiology, 0.3±0.3 years prior, aged 60±8 years (n=20)	31.3±7.3
	○ Dual-task	36.8±10.5
	- Non-dysvascular etiology, 2±2 years prior, aged 56±14 years (n=20)	22.5±3.6
	○ Dual-task	29.0±5.4
	Unilateral transfemoral/bilateral LLL 1±1 years prior, aged 58±15 years (n=20) [59]	36.2±19.1
	- Dual-task	41.2±23.0
10 Meter Walk Test (10MWT), m/s	Self-selected Gait Speed (using 6m timed from 10m course)
	Unilateral transtibial or transfemoral ≥1 years prior[34]
	- K3, aged 60±12 years (n=35)	0.88±0.04
	- K4, aged 46±12 years (n=20)	1.21±0.05
	Unilateral LLL, aged 63±13 years, at discharge from rehabilitation (using 10m timed from 12m course)[100]‡
	- K1 (n=6)	0.17 (IQR: 0.15–0.19)
	- K2 (n=43)	0.38 (IQR: 0.25–0.54)
	- K3 (n=54)	0.63 (IQR: 0.50–0.71)
	- K4 (n=7)	1.06 (IQR: 0.95–1.18)
	Fast Gait Speed (using 6m timed from 10m course)
	Unilateral transtibial or transfemoral ≥1 years prior[34]
	- K3, aged 60±12 years (n=35)	1.12±0.05
	- K4, aged 46±12 years (n=20)	1.56±0.07
Figure-of-8 Walk Test (F8WT), s	Fast Speed
	Unilateral LLL ≥1 year prior[41]
	- Transtibial
	○ K3, aged 50 (95%CI: 46–55) years (n=28)	6.39 (95%CI: 5.94–6.83)
	○ K4, aged 40 (95%CI: 35–46) years (n=22)	5.80 (95%CI: 5.29–6.30)
	- Transfemoral
	○ K3 aged 50 (95%CI: 40–61) years (n=10)	8.34 (95%CI: 7.59–9.09)
	○ K4 aged 38 (95%CI: 26–49) years (n=7)	6.84 (95%CI: 5.94–7.73)
6 Minute Walk Test (6MWT), m	Male service members, aged 18–40 years, 3±2 years post-LLL[35]
	- Unilateral transtibial (n=60)	661±87
	- Unilateral transfemoral (n=32)	542±67
	- Bilateral LLL (n=26)	523±110
	Unilateral transtibial, aged 46±14 years, 8±9 years of prosthesis use (n=13)[70]	545±65
	Unilateral transtibial or transfemoral ≥1 years prior[34]
	- K3, aged 60±12 years (n=35)	311±19
	- K4, aged 46±12 years (n=20)	427±27
	Unilateral LLL ≥2 years prior, aged 66±15 years (n=44)[28]	332±115
	Unilateral/bilateral LLL, aged 60±15 years[72]
	- Aged <50 years (n=23)	514±113
	- Aged ≥50 years (n=63)	351±116
	- Trauma/Congenital etiology (n=45)	457±120
	- Cancer-related etiology (n=5)	445±90
	- Dysvascular etiology (n=21)	345±104
	- Infection-related etiology (n=6)	261±76
	- Diabetes-related etiology (n=9)	256±122
2 Minute Walk Test (2MWT), m	Unilateral/bilateral LLL, aged 66±3 years (n=290)[101]
	Start of inpatient rehabilitation (Male/Female)	30±20 / 23±12
	- At rehabilitation discharge (Male/Female)	46±32 / 29±14
	- 3-months post-discharge (Male/Female)	81±47 / 50±27
	Unilateral LLL ≥2 years prior, aged 66±13 years (n=44)[28]	114±36
	Unilateral/Bilateral LLL 13±11 years prior, aged 55±15 years (n=52)[73]	111±49

^†K-level Classification Definitions: K0=no ability or potential to transfer safely with or without assistance and the prosthesis does not enhance quality-of-life or mobility. K1=ability or potential to use a prosthesis for transfers or ambulation on level surfaces at fixed cadence; typical of a limited and unlimited household ambulator. K2=ability or potential for ambulation with the ability to traverse low level environmental barriers such as curbs, stairs, or uneven surfaces; typical of the limited community ambulator; K3=ability or potential for ambulation with variable cadence; typical of the community ambulator who has the ability to traverse most environmental barriers and may have vocational, therapeutic, or exercise activity demanding prosthetic utilization beyond simple locomotion. K4=ability or potential for prosthetic ambulation exceeding basic ambulation skills, exhibiting high impact, stress, or energy levels; typical of the prosthetic demands of the child, active adult, or athlete.

^‡Data presented as median (IQR) rather than mean±standard deviation.

Abbreviations: LLL=lower-limb loss; CI=confidence interval; IQR = interquartile range

POPULATION-SPECIFIC

Amputee Mobility Predictor (AMP).

The AMP is a 21-item, reliable (Table 1) and valid tool for assessing static and dynamic balance, transfers, and gait of patients with unilateral amputations and bilateral transtibial amputations.[31, 32] For patients with bilateral amputations at the transfemoral level or transfemoral/transtibial level amputations, the modified AMP may be more appropriate,[32] but test-retest reliability is unknown. Scores differ on the AMP with respect to amputation level and functional classification (Table 2), suggesting known-groups validity.[31, 27, 33, 34]

Comprehensive High-Level Activity Mobility Predictor (CHAMP).

The CHAMP, a 4-item battery, was created to assess high-level mobility in service members, aged 18–40 years, with traumatic LLL.[33, 35] Developers recommend administering the CHAMP to those with (a) AMP ≥37points AND (b) 6 Minute Walk Test (6MWT) ≥250m;[35] data is provided to assist with results interpretation (Tables 1 and 2). A subsequent study including a more heterogeneous sample (e.g., both males and females, aged 18–60 years, with various amputation etiologies) suggests appropriateness for patients with (a) AMP ≥37points and (b) 2MWT >97m and provides data to assist with results interpretation based on sex, age, amputation level, and functional mobility level (Table 2).[36] Perhaps the greatest hurdle to widespread use, however, is space requirements (Table 1).

NON-SPECIFIC

Transfers and Functional Strength

5 Times Sit-to-Stand (5XSTS).

After LLL, adults may experience difficulty with rising from a chair.[37, 38] 5XSTS assesses sit-to-stand transfer ability and functional lower-extremity strength.[39] While between-days, test-retest reliability (ICC=.97– 1.00) has been established among various populations,[39, 40] it is unknown for adults with LLL. 5XSTS time increases with age (Table 1) and may help differentiate between adults with unilateral transfemoral amputation with differing mobility levels (Table 2).[41]

Balance

Four Square Step Test (FSST).

FSST assesses dynamic balance through multi-directional stepping. Among adults with unilateral LLL, between-days, test-retest reliability (ICC=.97) suggests appropriateness for evaluating individual patient change.[42] MDC₉₀ is 2.0s (i.e., ~20% of the mean).[42] FSST can be modified by substituting lines on the floor for canes,[41] which may increase valid trials and reduce FSST’s known floor effect.[43] FSST without canes may help differentiate between functional mobility levels among adults with unilateral LLL.[41]

Functional Reach (FXR).

FXR assesses anteroposterior trunk stability, which is impaired among adults with amputation (Tables 1 and 2).[44] While test-retest reliability (ICC=.92) indicates FXR appropriateness for evaluations of patients without LLL,[45] reliability is not reported among patients with LLL. Nevertheless, of all performance-based measures reviewed, FXR may be the most appropriate for assessing balance among heterogeneous patients with LLL since it lacks both floor and ceiling effects.[46] Further, comparison data for adults without LLL across the lifespan is available allowing benchmarking of patients post-LLL (Table 1).[ 45]

Narrowing Beam Walking Test (NBWT).

While NBWT was developed to overcome ceiling effects of other balance measures, floor effects are still a concern, particularly for patients with LLL using assistive devices.[47] Between-days, test-retest reliability (i.e., ICC=.90) indicates appropriateness for assessing changes in individuals over time.[42] Barriers to clinical use include the need to (a) acquire the beam with the specified dimensions, (b) have ≥7.32m of linear space to conduct the test, and (c) store the beam. For middle-aged adults with unilateral amputations who walk without assistive devices, known-groups validity is supported by differences between fallers and non-fallers, amputation-levels, and lower (i.e., K1, K2) versus higher (i.e., K3, K4) mobility levels (Table 2).[43]

Berg Balance Scale (BERG).

While BERG has established between-days, test-retest reliability among older adults (ICC=.95–.97),[48, 49] reliability for patients with LLL remains unknown (Table 1). Items on the BERG found to be most difficult for patients with LLL include tandem stance, stool taps, and turning 360°.[50] Although BERG scores may not vary by amputation level,[51] scores are worse with greater mobility deficits,[52] assistive device use,[52] and reduced physical activity (Table 2)[53]. Due to conflicting evidence as to whether BERG assists with falls-risk assessment among patients with LLL[54, 51] and older adults in the general population[55], clinicians are cautioned with using BERG to evaluate fall risk.

Functional Mobility

Timed Up and Go (TUG).

TUG includes two transfers, walking 6m, and a turn. Between-days, test-retest reliability indicates appropriateness for group-level evaluations, but MDC₉₀, i.e., 2.2–3.6s, may be difficult to achieve at 22–29% of the mean.[28, 42] TUG times vary with age, amputation level, time since amputation,[28, 56] and mobility level, [34] complicating comparisons to peers with LLL (Table 2). Reference data is available for older adults without LLL (Table 1).

L Test of Functional Mobility (L Test).

L Test is a modified TUG, which includes two transfers, walking 20m, and two turns, intended to overcome TUG’s ceiling effect.[57, 58] Among adults with LLL, L Test has been evaluated during a dual-task condition challenging cognitive load, i.e., serial subtractions by 3’s from a randomly selected number between 100 and 150.[59] Reliability suggests appropriateness of standard and dual-task L Tests for evaluating individual patient changes. With L Test MDCs at ~7–14% of means,[59] ‘true changes’ may be detectable, but times may vary by age, amputation level or etiology, time since amputation, and physical activity level (Tables 1 and 2).

Gait

10 Meter Walk Test (10MWT).

10MWT performed at self-selected and fast gait speeds provides insight regarding typical walking speed and patient capacity for negotiating community environments, e.g., crosswalks. Two meters for acceleration and deceleration is provided at either course end with the goal of steady-state walking during the timed 10 central meters.[60, 42] Between-days, test-retest reliability (ICC=.96) using a 14m course indicates appropriateness for evaluating individual changes in self-selected gait speed in patients with LLL.[42] For space-limited clinicians considering shorter courses, gait speed assessment over <5m is not recommended,[61] but total course lengths of 10m (i.e., 6m timed) have been shown to have acceptable between-days, test-retest reliability in other populations (e.g., Parkinson’s: ICC=.96–.97; MDC₉₅=.18–.25m/s).[62] Gait speed may help differentiate between functional mobility levels (Table 2).[34]. As multiple factors negatively impact gait speed post-LLL,[63–66] comparison to age-matched adults without LLL reduces interpretation complexity (Table 1).

Figure-of-8 Walk Test (F8WT).

F8WT requires planning a path and executing two turns, assessing gait during cognitive load.[67] Reliability has been established for self-selected F8WT time (ICC=.84), number of steps for course completion (ICC=.82), and smoothness of movement (ICC=.61) among older adults,[67] but not specifically post-LLL. Given smoothness ICC,[67] reliability of evaluating changes in F8WT quality-of-movement between-sessions is questioned. Among adults ≥6 months post-unilateral LLL, the F8WT has recently been modified to include two novel conditions (5 trials/condition) performed at the patient’s self-selected speed: (1) carrying a tray with two cups of water while being asked not to spill the water to further increase cognitive load and (2) walking on soft, uneven terrain created by securing additional foam slates underneath foam mats.[68] F8WT time is better among adults with LLL participating in organized physical activity (4.5±0.9s) when compared to sedentary peers (7.2±2.3s), although number of steps are not significantly different (10±2 versus 12±3 steps).[53] F8WT performed at fast speed differentiates between mobility levels among adults with a unilateral transfemoral amputation (Table 2), informing administration recommendations in LLL (Table 1).[41]

Walking Endurance

6 Minute Walk Test (6MWT).

6MWT is a clinical exercise test that assesses aerobic capacity through prolonged walking.[69] As adults with LLL reach 72–87% of age-predicted maximal heart rate during the 6MWT,[70] vital sign assessment prior to 6MWT administration (to ensure patient safety) is best practice.[71] Continuous pulse oximetry is recommended during testing, with test cessation if SpO₂ drops <80%.[71] Between-days, test-retest reliability (ICC=.97; MDC₉₀ ~13% of mean)[28] suggests clinical utility for evaluating individual patient changes. Post-LLL, 6MWT may help differentiate between higher mobility levels,[34] but performance also varies by age and amputation etiology (Table 2).[72] Normative data for adults without LLL is available (Table 1).

2 Minute Walk Test (2MWT).

Test-retest reliability (ICC=.83; MDC₉₀ ~28–30% of mean)[73] suggests using 6WMT over 2MWT whenever possible (Table 1). Whereas 6MWT helps differentiate between higher mobility levels, 2MWT assists with discriminating between lower (e.g., K2) and higher (i.e., K3) mobility levels using a ≥113m cut-point.[72] 2MWT is highly predictive of 6MWT using the equation: 6MWT(m)=3.1(2MWT)-54.5,[72] which allows clinicians to compare baseline 2MWT to follow-up 6MWT, when necessary. Among adults with LLL, 2MWT distance is associated (r=.78) with community-walking, specifically peak activity data (i.e., highest 30 minutes of step activity/day).[74] When interpreting 2MWT data after LLL, consideration of age, sex, and time since amputation appears important (Tables 1 and 2). Reference data for adults without LLL is available (Table 1).[75]

EVALUATIONS

Functional Classification.

In addition to showing change in patient status post-LLL, use of outcome measures can assist clinicians with objectively classifying functional mobility (Table 2) and predicting adverse health outcomes. One functional mobility classification, i.e., K-level, is provided by the United States’ Centers for Medicare and Medicaid Services.[76] Physical and occupational therapists, as well as prosthetists, may be asked to assist physicians in patient K-level classification.[76]

Predicting Adverse Outcomes

Failure to Achieve Community-Walking Level.

Wong and colleagues conducted a derivation study to generate a clinical rule predicting adults with LLL (unilateral or bilateral, any level and etiology) who would not be community-ambulators (defined as Houghton Scale <9/12) at 1-year follow-up.[77] Individuals meeting 1 of the following increase their probability for not being a community ambulatory from 59.3% to 81.4%: (i) initial Houghton Scale ≤7/12, (ii) ABC ≤65%, (iii) BERG item ‘retrieve object from floor’ ≤3/4, and/or (iv) BERG item ‘look over shoulders’ ≤3/4, while individuals meeting ≥2/4 criteria, increase their probability to ≥92.1%.[77] Validation of this prediction rule is necessary before widespread clinical adoption, particularly given the study’s sample size (n=40).[77]

Prosthesis Nonuse.

Prosthesis nonuse may be defined as prosthesis wear for cosmesis only or permanent prosthesis abandonment for weekday locomotor activities.[78] Reasons for prosthesis abandonment include residual limb wounds, pain, and comorbidities.[60] Following transtibial-level or higher amputation with prosthetic rehabilitation, at 1-year follow-up, the following (administered during prosthetic rehabilitation) are predictors of prosthesis nonuse: (i) FSST ≥36.6s, (ii) TUG ≥21.4s, (iii) 10MWT self-selected speed ≤.44m/s, and (iv) 6MWT ≤191m. Relative risk of prosthesis nonuse is (i) 2.76, (ii) 3.17, (iii) 2.76, and (iv) 2.84X greater for patients with each predictor, respectively.[60] From a screening perspective, the 6MWT has the highest sensitivity of the four tests, resulting in correct classification of 80.6% of patients who abandon their prosthesis at follow-up.[60]

Falls.

More than 50% of adults with LLL fall, which is 2X the rate of the older adult general population.[79] Among adults with LLL, nearly 30% report multiple falls and 26% report injury.[80] With transtibial amputation, the following outcomes have been associated with an increased risk of multiple falls within the 6 months following discharge from prosthetic rehabilitation: (i) FSST ≥24s and (ii) TUG ≥19s.[81] Cut-points, however, should only be used acutely post-amputation among adults with unilateral transtibial amputation (not those with transfemoral or bilateral LLL).

CONCLUSIONS

When selecting outcome measures for patients post-LLL, considerations include the rationale for collecting data (i.e., will data be used to inform individual patient-care or care for a group of patients), measure test-retest reliability, whether MDC is achievable, known-groups validity (which may assist with classification), predictive validity (which may assist with risk assessment), known floor/ceiling effects, and resources available. Administration, scoring, and interpretation suggestions are provided for 20 measures, which may be used following LLL. Areas for further scientific inquiry are highlighted to enhance future assessments of patients following lower-limb amputation.