Functional level assessment of individuals with transtibial limb loss: Evaluation in the clinical setting versus objective community ambulatory activity

Abstract

The functional level (K level) of prosthetic users is used to choose appropriate prosthetic components, but ratings may highly subjective. A more objective and robust method to determine K level may be appealing. The aim of this study was to determine the relationship between K level determined in the clinic to K level based on real world ambulatory activity data collected by StepWatch. Twelve individuals with transtibial limb loss gave informed consent to participate. K level assessments performed in the clinic by a single treating prosthetist were compared with a calculated estimate based on seven days of real world ambulatory activity patterns using linear regression. There was good agreement between the two methods of determining K level with R²= 0.775 (p < 0.001). The calculated estimate of K level based on actual ambulatory activity in real world settings appears to be similar to the treating prosthetist’s assessment of K level based on gait observation and patient responses in the clinic. Clinic-based ambulatory capacity in transtibial prosthetic users appears to correlate with real world ambulatory behavior in this small cohort. Determining functional level based on real world ambulatory activity may supplement clinic-based tests of functional capacity.

Keywords

Amputee ambulation gait K-level prosthesis

Introduction

The United States’ Medicare Functional Classification Level (MFCL) K level is a 0 to 4 point scale of functional level for those with limb loss.¹ The MFCL levels are described in Table 1. Higher K level rated individuals generally receive more technologically advanced and expensive components such as microprocessor-controlled knees and ankles, and carbon fiber “energy storage and return” prosthetic feet. In the United States, current Centers for Medicare/Medicaid Services (CMS) guidelines allow reimbursement to providers of prosthetic components if there is evidence that the individual’s K level matches the prosthetic components provided. This evidence must be in the patient’s medical record and noted by the evaluating physician; without this evidence of appropriate functional level, CMS may demand repayment of the reimbursement paid to the prosthetic provider for the components. The Inspector General of Health and Human Services² has noted a 27% increase in the cost of prosthetic care for beneficiaries of CMS while the number of individuals with limb loss has dropped by 2% (2005–2009). Although the Inspector General’s observation is legitimate, it was not peer-reviewed, and it does not take into account the possible mobility, quality of life, health care cost improvements for the individuals with limb loss who received more expensive prosthetic components.

Table 1.

Descriptions for the Medicare Functional Classification Level.¹

K level	Description
K0	Does not have the ability or potential to ambulate or transfer safely with or without assistance and a prosthesis does not enhance quality of life or mobility
K1	Has the ability or potential to use a prosthesis for transfers or ambulation on level surfaces at fixed cadence. Typical of the limited and unlimited household ambulator
K2	Has the 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	Has the 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 that demands prosthetic utilization beyond simple locomotion
K4	Has the ability or potential for prosthetic ambulation that exceeds the basic ambulation skills, exhibiting high impact, stress, or energy levels, typical of the prosthetic demands of the child, active adult, or athlete

Prosthetists routinely evaluate patients’ functional level, but there is no standard methodology for this assessment. The K level language is somewhat vague and each prosthetist may have a slightly different interpretation of the functional level characteristics. This brings up an important issue regarding an individual’s capacity to perform at a high functional level during clinical tests versus high functional level behavior in the community. For example, Ashe et al. have shown that more than half of older community dwelling adults classified by clinic-based tests as high functional capacity exhibit low functional level behavior in the community.³ Mudge et al. have shown that clinic-based assessments of functional walking capacity show improvement after interventions for post stroke individuals, but these individuals do not demonstrate increased functional walking behavior in the community.⁴

Several clinic-based functional task test batteries have been created for those with limb loss.^5,6 There is some published evidence that some capacity tests correlate with clinical evaluation of K level for prosthetists,⁵ but very little evidence exists that high functional capacity individuals with limb loss choose to demonstrate highly functional activity levels during typical community behavior. The opposite might also be true; that individuals who demonstrate low functional capacity on clinic-based tests might choose to have higher levels of activity in the community. There are efforts underway to develop a new mobility instrument (questionnaire) for individuals with limb loss,⁷ but published evidence of reliability, validity and sensitivity have not yet appeared in the literature. The problem with any questionnaire assessing functional level in prosthetic users is that it will be susceptible to bias because a high score means more advanced and expensive prosthetic componentry for the prosthetic user. There is an inherent incentive to over-rate one’s functional performance.

The aim of this study was to evaluate the relationship between the MFCL K level of 12 transtibial amputees determined by their treating clinical prosthetist and K level determined by a custom algorithm based on real world functional activity level from wearable sensor (StepWatch⁸) data in their typical community environment.

Methods

Twelve individuals with limb loss participated in this Institutional Review Board approved study (Table 2). The protocol was explained to each participant, and each participant signed an informed consent form. Participant recruitment was open to adults with dysvascular or traumatic causes of transtibial limb loss, including those with diabetes provided their disease was controlled and limb health was adequate on both limbs. The inclusion criteria were adults with transtibial limb loss, at least one year of successful ambulation on their prosthesis, a stable gait pattern, and fluent in English. Exclusion criteria for the study were underlying conditions that could affect gait such as chronic obstructive pulmonary disease or symptomatic cardiovascular disease.

Table 2.

Demographic data of study participants with K level determined by the prosthetist and using the calculated K level based on seven days of real world ambulation.

Participant	Age (years old)	Body mass (kg)	Body height (m)	Prosthetist K level	Calculated K level	K level difference	Gender	Limb loss side
1	68	70	1.68	2	3	1	F	R
2	69	57	1.57	2	2.9	0.9	M	L
3	55	73	1.7	2	3.1	1.1	M	R
4	41	110	1.96	3	3.7	0.7	M	L
5	74	75	1.78	3	3.3	0.3	M	R
6	65	91	1.83	3	3.3	0.3	M	L
7	58	86	1.78	3	3.5	0.5	M	L
8	53	99	1.88	3	4	1	M	L
9	66	78	1.73	3	3.4	0.4	M	L
10	49	75	1.75	4	4.2	0.2	M	R
11	39	111	1.91	4	3.9	−0.1	M	R
12	47	61	1.7	4	4.6	0.6	M	L
Mean	57	82	1.77	3	3.6	0.6
SD	12	18	0.11	0.7	0.5	0.4
Range	41–74	57–111	1.57–1.96	2–4	2.9–4.6	−0.1–1.1

F: female; M: male; R: right; L: left

Each participant’s K level was determined by a single treating prosthetist in the clinic setting by the typical method of gait observation and questioning the participants about habitual walking behavior, environmental barriers and community ambulation challenges. After collecting these K level rating data in the clinic, each participant was fitted with a StepWatch Activity Monitor on their prosthesis (proximal to the prosthetic foot) to record their ambulatory activity over a seven day period in the community. The StepWatch records step counts of amputees with 99.6% accuracy,⁸ and is especially accurate at slow walking speeds when other monitors under-count steps.⁹ StepWatch has demonstrated high accuracy in step counts for a wide range of gait pathologies including post-stroke,^10,11 multiple sclerosis and Parkinson’s disease,¹² cerebral palsy,^13–17 Duchene muscular dystrophy,^18–20 intermittent claudication,^21–25 and many others.^26–37

The data from the StepWatch was downloaded and a custom algorithm was utilized to determine each individual’s K level. This algorithm had four total components that used both community activity data (StepWatch data – three components) and clinically determined K level (clinician’s judgment – one component) for calculation. These four components were averaged to determine the final calculated K level estimate; therefore the StepWatch data contributed 75% of the calculated score and the clinician’s judgment contributed 25% to the final calculated K level estimate. The calculated K level ranged from 0.0 to 4.9 point scale similar to the K level score. Participants were considered within a K level until they achieved the next whole number rating: a 2.9 is still a K2 ambulator (no rounding for the calculated value).

Three parameters, which contributed 75% to the calculated K level, were extracted from the StepWatch community activity data that were mathematical analogs of the descriptions of the K level attributes in the original HCFA1 document: 1) the average step rate from 60 of each participant’s most active non-contiguous minutes of activity was calculated for each of the seven days. Essentially this estimate is the average of the fastest, densest (most steps per minute) short term walking the individual can perform during community ambulation. This portion of the algorithm is intended to emulate the “potential to ambulate” section of the MFCL description (for example, “MFLC-2—Has the 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.”)¹ 2) The ratio of Low:Medium:High step rates was calculated (Low 0–15 steps/min; Medium 16–40 steps/min; High >40 steps/min) for each of the seven days. Primarily this evaluates the ability of the individual to walk at different cadences during community ambulation. This portion of the algorithm is intended to emulate the “cadence variability” section of the MFCL description (for example, “MFLC-3—Has the 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 that demands prosthetic utilization beyond simple locomotion.”)¹ 3) The total number of daily steps was used to estimate energy expenditure for each of the seven days based on an equation (EEtotal (kcal) = 2.033 kcal*kg⁻¹*weight (kg) + 0.368 kcal*steps – 86.1 kcal) developed by Foster et al.³⁶ This estimates the metabolic cost of walking in the community. This portion of the algorithm is intended to emulate the “energy level” section of the MFCL description (for example, “MFLC-4—Has the ability or potential for prosthetic ambulation that exceeds the basic ambulation skills, exhibiting high impact, stress, or energy levels, typical of the prosthetic demands of the child, active adult, or athlete.”)¹

Each of these StepWatch components (1) rating of average step rate for 60 non-contiguous minutes; 2) rating of Low:Medium:High step rate ratios; 3) rating of energy expenditure based on total daily steps) uses steps per minute data in different ways and the calculated values are mapped to a 0.0 to 4.9 point scale for each. This results in individual K levels for “potential to ambulate”, “cadence variability” and “energy level”, all on 0.0 to 4.9 point scales. The fourth component comes from the K level judged by the clinical prosthetist. Together these four components make up the calculated K level score by averaging them. To evaluate the concurrent validity, simple linear regression was utilized to determine the correlation between the prosthetist’s clinical judgment of K level and the K level determined by the algorithm.

Results

All participants completed the study and were rated by the prosthetists after seven days of community ambulation. The prosthetist rated three participants as K level 2, six participants as K level 3, and three participants as K level 4. The linear regression revealed a significant relationship between the prosthetist’s K level rating and the calculated K level (p < 0.001) with an R² of 0.775. The calculated K level was consistently slightly higher than the prosthetist’s rating for all but one individual, with the mean difference 0.6 K levels, the standard deviation 0.4 K levels and a range of −0.1 to +1.1 difference between the prosthetist’s K level and the calculated K level. The relationship between the two K level methods is shown in Figure 1.

Figure 1.

K level rating of 12 transtibial amputees by a single treating prosthetist using usual clinical methods and calculated by the algorithm from StepWatch data in their community setting.

Discussion

This study aimed to evaluate the correlation between the K level for 12 adults with transtibial limb loss determined by their clinical prosthetist and the K level determined by the calculated estimate from their daily ambulatory behavior. There was good agreement between the two methods of determining K level with R²= 0.775 (p < 0.001) and a very strong positive correlation (R = 0.88). The calculation tended to rate the participant’s K level slightly higher than the prosthetist, but by definition the decimal portion of the calculated K level is truncated. The calculated method also has greater resolution than the single integer value for K levels given by the prosthetist. It could be that some individuals were given K2 rating for example when they were actually a very high K2, but not quite a K3 in the opinion of the prosthetist. The lack of an additional significant figure in the prosthetist K level adds a certain level of inaccuracy that may have altered the observed linear relationship between functional level measures.

It should be noted that the MFCL K levels have no published evidence of reliability or construct validity, and have only been accepted by clinicians and researchers because of reimbursement policies in the United States. For many prosthetists, the key feature that distinguishes a K3 ambulator from a K2 ambulator is cadence variability. It is implicit in the description that individuals who have multiple walking speeds will exhibit cadence variability, be more functional community ambulators and deserve more expensive and high-performance prosthetic components such as carbon fiber “energy storage and return” feet or microprocessor-controlled knees.

These data also show that there is some variability between clinic-based estimate of capacity and real world behavior, but there is still a fairly high relationship between the two for this small cohort of participants. One confounding effect is that individuals may have been limited in their community behavior by the very prosthetic components provided based on their clinical K level assessment, creating a self-fulfilling cycle of component, capacity and behavior. Previous work has demonstrated that some K2 individuals can become K3 individuals if given a microprocessor-controlled knee.³⁸ Additional carefully controlled studies with larger numbers of subjects are needed to determine if specific prosthetic components can increase or decrease performance on clinic-based tests of capacity or alter real world functional behavior.

This initial study on the concurrent validity of the calculated K level algorithm compared with typically derived clinician impression of K level focused only on transtibial prosthesis users, and additional work is needed for individuals with other levels of limb loss. Also, quantifying steps may not be as informative as other metrics that relate to mobility and health of prosthetic users. Objectively quantifying the loads applied to the prosthetic and contralateral limb during ambulation may provide a more direct link between prosthetic component performance, individual functional performance and limb loss comorbidities such as osteoarthritis, obesity, diabetes, cardiovascular disease, and stroke. Systems that remotely monitor forces and loads during ambulation in the community may provide early intervention opportunities to halt medical problems before they progress to the point that only expensive interventions are effective. Work in this area quantifying the moment impulse during the gait of prosthetic users appears to hold some promise for understanding limb and joint loading in these individuals.³⁹ By quantifying the loading dose perhaps more effective interventions could be developed and monitored for lower limb prosthetic users.

Although this study was conducted in North America and the K levels were created by the United States CMS, K levels have become the de-facto functional level in several international locations, including the Netherlands,^6,40,41 Canada,⁴² and Germany.⁴³ There seems to be a general consensus that expensive, higher performance prosthetic components should be provided to those who will utilize the added performance of these more expensive components.

Conclusion

This study sought to relate clinic-based assessments of K level with a calculated functional level estimate based on ambulation in real world settings for transtibial amputees. There was good agreement between the two methods of determining K level with R²= 0.775 (p < 0.001). Determining functional level based on real world ambulatory activity may supplement clinic-based tests of functional capacity and increase objectiveness in K level assessment.

Footnotes

Acknowledgement

AOPA had no input into the decision to publish this manuscript, no influence upon the design, conduct, findings or presentation of these data.

Declaration of Conflicting Interests

The author(s) declared the following potential conflicts of interest with respect to the research, authorship, and/or publication of this article: MO, TK, and DB work or worked for Orthocare Innovations, the manufacturers of the sensor worn by the participants in this study, and the developers of the algorithm used to determine K level from community activity data.

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This work was supported by the American Orthotic and Prosthetic Association (AOPA). AOPA has had no influence upon the design, data collection, analysis, or interpretation of this research study, no involvement in the decision to publish these results, and no involvement in the writing of this manuscript.

References

HCFA. Common Procedure Coding System HCPCS 2001, Washington, DC: US Government Printing Office, 2001.

Levinson

. Questionable billing by suppliers of lower limb prostheses. Office of the Inspector General (Oot IG). Health and Human Services, Washington, DC: HHS, 2011.

Ashe

Eng

Miller

. Disparity between physical capacity and participation in seniors with chronic disease. Med Sci Sports Exerc 2007; 39: 1139–1146.

Mudge

Barber

Stott

. Circuit-based rehabilitation improves gait endurance but not usual walking activity in chronic stroke: A randomized controlled trial. Arch Phys Med Rehabil 2009; 90: 1989–1996.

Gailey

Roach

Applegate

. The amputee mobility predictor: An instrument to assess determinants of the lower-limb amputee’s ability to ambulate. Arch Phys Med Rehabil 2002; 83: 613–627.

Theeven

Hemmen

Stevens

. Feasibility of a new concept for measuring actual functional performance in daily life of transfemoral amputees. J Rehabil Med 2010; 42: 744–751.

Morgan

Amtmann

Abrahamson

. Use of cognitive interviews in the development of the PLUS-M item bank. Qual Life Res 2014; 23: 1767–1775.

Coleman

Smith

Boone

. Step activity monitor: Long-term, continuous recording of ambulatory function. J Rehabil Res Dev 1999; 36: 8–18.

Karabulut

Crouter

Bassett

Jr . Comparison of two waist-mounted and two ankle-mounted electronic pedometers. Eur J Appl Physiol 2005; 95: 335–343.

10.

Mudge

Stott

. Test–retest reliability of the StepWatch Activity Monitor outputs in individuals with chronic stroke. Clin Rehabil 2008; 22: 871–877.

11.

Mudge

Stott

Walt

. Criterion validity of the StepWatch Activity Monitor as a measure of walking activity in patients after stroke. Arch Phys Med Rehabil 2007; 88: 1710–1715.

12.

Schmidt

Pennypacker

Thrush

. Validity of the StepWatch Step Activity Monitor: Preliminary findings for use in persons with Parkinson disease and multiple sclerosis. J Geriatr Phys Ther 2011; 34: 41–45.

13.

Van Wely

Dallmeijer

Balemans

. Walking activity of children with cerebral palsy and children developing typically: A comparison between the Netherlands and the United States. Disabil Rehabil 2014; 36: 2136–2142.

14.

Bjornson

Zhou

Stevenson

. Relation of stride activity and participation in mobility-based life habits among children with cerebral palsy. Arch Phys Med Rehabil 2014; 95: 360–368.

15.

Bjornson

Zhou

Stevenson

. Walking activity patterns in youth with cerebral palsy and youth developing typically. Disabil Rehabil 2014; 36: 1279–1284.

16.

Kang

Bjornson

Barreira

. The minimum number of days required to establish reliable physical activity estimates in children aged 2–15 years. Physiol Meas 2014; 35: 2229–2237.

17.

Ishikawa

Kang

Bjornson

. Reliably measuring ambulatory activity levels of children and adolescents with cerebral palsy. Arch Phys Med Rehabil 2013; 94: 132–137.

18.

McDonald

Coleman

Henricson

. Assessment of StepWatchTM activity monitoring in phase 2b study of ataluren (PTC124TM) in nmDMD/BMD. Neuromusc Disord 2009.

19.

McDonald

Widman

Walsh

. Use of step activity monitoring for continuous physical activity assessment in boys with Duchenne muscular dystrophy. Arch Phys Med Rehabil 2005; 86: 802–808.

20.

McDonald

Widman

Abresch

. Utility of a step activity monitor for the measurement of daily ambulatory activity in children. Arch Phys Med Rehabil 2005; 86: 793–801.

21.

Gardner

Parker

Montgomery

. Efficacy of quantified home-based exercise and supervised exercise in patients with intermittent claudication: A randomized controlled trial. Circulation 2011; 123: 491–498.

22.

Gardner

Montgomery

Parker

. Physical activity is a predictor of all-cause mortality in patients with intermittent claudication. J Vasc Surg 2008; 47: 117–122.

23.

Gardner

Montgomery

Scott

. Association between daily ambulatory activity patterns and exercise performance in patients with intermittent claudication. J Vasc Surg 2008; 48: 1238–1244.

24.

Hodges

Sandercock

Das

. Randomized controlled trial of supervised exercise to evaluate changes in cardiac function in patients with peripheral atherosclerotic disease. Clin Physiol Funct Imaging 2008; 28: 32–37.

25.

Gardner

Montgomery

Scott

. Patterns of ambulatory activity in subjects with and without intermittent claudication. J Vasc Surg 2007; 46: 1208–1214.

26.

Braun

Dillon

Sheiko

. Reliably estimating ambulatory activity in youth with arthrogryposis. Disabil Rehabil 2015.

27.

Segal

Shofer

Hahn

. Functional limitations associated with end-stage ankle arthritis. J Bone Joint Surg Am 2012; 94: 777–783.

28.

Sheiko

Bjornson

Lisle

. Physical activity assessment in adolescents with limb salvage. J Pediatr 2012; 161: 1138–1141.

29.

Nguyen

Burr

Gill

. Validation of the StepWatch device for measurement of free-living ambulatory activity in patients with chronic obstructive pulmonary disease. J Nurs Meas 2011; 19: 76–90.

30.

Ishikawa

Stevens

Kang

. Reliability of daily step activity monitoring in adults with incomplete spinal cord injury. J Rehabil Res Dev 2011; 48: 1187–1194.

31.

Downs

Leonard

Hill

. Initial assessment of the StepWatch Activity Monitor to measure walking activity in Rett syndrome. Disabil Rehabil 2011.

32.

Schulte

Schubert

Winter

. Step activity monitoring in lumbar stenosis patients undergoing decompressive surgery. Eur Spine J 2010; 19: 1855–1864.

33.

Orendurff

Walker

Jovanovic

. Intensity and duration of intermittent exercise and recovery during a soccer match. J Strength Cond Res 2010; 24: 2683–2692.

34.

Winter

Muller

Brandes

. Level of activity in children undergoing cancer treatment. Pediatr Blood Cancer 2009; 53: 438–443.

35.

Orendurff

Schoen

Bernatz

. How humans walk: Bout duration, steps per bout, and rest duration. J Rehabil Res Dev 2008; 45: 1077–1089.

36.

Foster

Lanningham-Foster

Manohar

. Precision and accuracy of an ankle-worn accelerometer-based pedometer in step counting and energy expenditure. Prev Med 2005; 41: 778–783.

37.

Algase

Beattie

Leitsch

. Biomechanical activity devices to index wandering behavior in dementia. Am J Alzheimers Dis Other Demen 2003; 18: 85–92.

38.

Kahle

Highsmith

Hubbard

. Comparison of nonmicroprocessor knee mechanism versus C-Leg on Prosthesis Evaluation Questionnaire, stumbles, falls, walking tests, stair descent, and knee preference. J Rehabil Res Dev 2008; 45: 1–14.

39.

Kobayashi

Orendurff

Arabian

. Effect of prosthetic alignment changes on socket reaction moment impulse during walking in transtibial amputees. J Biomech 2014; 47: 1315–1323.

40.

Theeven

Hemmen

Rings

. Functional added value of microprocessor-controlled knee joints in daily life performance of Medicare Functional Classification Level-2 amputees. J Rehabil Med 2011; 43: 906–915.

41.

Theeven

Hemmen

Geers

. Influence of advanced prosthetic knee joints on perceived performance and everyday life activity level of low-functional persons with a transfemoral amputation or knee disarticulation. J Rehabil Med 2012; 44: 454–461.

42.

Dudek

Khan

Lemaire

. Ambulation monitoring of transtibial amputation subjects with patient activity monitor versus pedometer. J Rehabil Res Dev 2008; 45: 577–585.

43.

Thiele

Westebbe

Bellmann

. Designs and performance of microprocessor-controlled knee joints. Biomed Tech (Berl) 2014; 59: 65–77.