Prostheticsand OrthoticsInternational,1998, 22,178-185Three measures of functional outcome forlower limb amputees: a retrospective reviewS. P. TREWEEK and M. E. CONDIEProstheticsNational Centre for Training and Education inand Orthotics, University of Strathclyde, Glasgow, Scotland,AbstractOutcome measures are becoming increasinglyimportant in health care. Functional outcomemeasures are of particular importance for lowerlimb amputees since much of the rehabilitationprocess is concerned with increasing mobilityand personal independence.The Scottish PhysiotherapyAmputeeResearch Group (SPARG) has used threemeasures of functional outcome: the BarthelIndex, Russek's classification and theLocomotor Index. The review reported hereinvolves 938 patients having a primaryamputation at the transtibial or transfemorallevel between October 1992 and July 1997.Differences in function due to age and level ofamputation are well known clinically and themeasures were compared by looking at theirability to detect these differences.The Barthel Index lacked sensitivity becauseof ceiling effects and should not be consideredas a suitable functional outcome measure foramputee patients. Russek's classification doesdetect significant differences but requires a largenumber of patients making it unsuitable forsingle hospital investigations. The LocomotorIndex demonstrates significant differences dueto age and amputation level despite fewerpatients being assessed by this measure duringthe period covered by this paper. The range ofthe Locomotor Index can be extended to covermore active amputees by considering its'advanced activities' subscale separately.The Locomotor Index is a promising measureUK.and should be considered by rehabilitation teamslooking for a valid, reliable and sensitivefunctional outcome measure for use with lowerlimb amputees.IntroductionClinicians involved in the rehabilitation oflower limb amputees increasingly need to useoutcome measures to demonstrate that they areproviding a clinically effective service.Functional assessment measures are of particularimportance for this group of patients since muchof the rehabilitation process is associated withimproving mobility and personal independence.Recognising that physiotherapy is a centralcomponent of all amputee rehabilitationprogrammes, the Scottish PhysiotherapyAmputee Research Group (SPARG) wasestablished in 1991 to evaluate t of amputees and to disseminate theresults (see Physiotherapy 79, p.649). The groupcomprises every senior physiotherapist inScotland (population approximately 5.5 million)with a clinical responsibility for amputeepatients; at present, 26 physiotherapists fall intothis category. In addition, SPARG has membersrepresenting the British Association ofChartered Physiotherapists in AmputeeRehabilitation, the British Association ofProsthetists and Orthotists and the David MurrayFoundation (a Scottish charitable organisationworking with amputees). SPARG also worksclosely with the Scottish Vascular Audit Groupwhose membership comprises all consultantvascular surgeons in Scotland and with theInformation and Statistics Division at theScottish Office Department of Health.One of SPARG's core activities is to conductAllcorrespondencetobeaddressedtoM r s M . E. Condie, National Centre for Training andEducationinProstheticsandOrthotics,University of S t r a t h c l y d e , 131 St. J a m e s Road,Glasgow G 4 O L S , Scotland, U K .178
Functionaloutcome fora regular, nationwide audit of the rehabilitationcare received by lower limb amputees inScotland (Condie et al., 1996). The audit isbased around a document known as a DischargeSummary Form (DSF) and custom software thatallows data collected on the DSFs to be storedand subsequently analysed. The DSF was firstused in October 1992 and a DSF is nowcompleted for virtually all lower limb amputeesin Scotland. Physiotherapists at ten amputatinghospitals use the custom software to enter andanalyse their own data. Other hospitals returntheir DSFs to the SPARG Coordinator wherethey are dealt with centrally. Analysis for thewhole of Scotland is done by merging eachhospital's data into a single database.As part of this audit work, SPARG hasattempted to measure the functional abilities oflower limb amputees at the time of dischargefrom hospital by including a functionalassessment section on the DSF. This wouldprovide an additional, standardised outcome bywhich to compare the rehabilitation programmesin use throughout Scotland. Since 1992, SPARGhas used three outcome measures: the BarthelIndex (Kullman, 1987; Mahoney and Barthel,1965), Russek's classification (Kullman, 1987;Russek, 1961) and the 'Locomotor Index' partof the Prosthetic Profile for Amputees(Gauthier-Gagnon and Grisé, 1994; Grisé et al.,1993). None of these measures was used for thewhole period covered by this paper (1/10/92 31/7/97). The Barthel Index was used between1/10/92 and 30/9/95, Russek's classificationbetween 1/10/92 and 30/4/97 and the LocomotorIndex from 1/10/96 onwards. As is clear fromthese dates, the functional assessment section ofthe DSF generally contained two measures.Completing the functional assessment part of theDSF took the physiotherapist less than fiveminutes.The Barthel Index was originally developed asa means of assessing the level of eletal disorders. It consists of tenitems, each of which is rated in terms of whetherthe patient is able to perform a particular taskindependently (see Appendix, Table A1). Scoresfor the ten items are summed to give an overallscore out of 100. The validity of the BarthelIndex is well documented (Shah and Cooper,1993) and the Index has also been found to bereliable (Collin et al., 1988). Although notamputees179developed for amputees, some authors(Kullman, 1987; Goldberg, 1984) have used theBarthel Index with this group of patients andfound it to be a useful indicator of functionalabilities and rehabilitation outcome. Further, theBarthel Index is widely used and the RoyalCollege of Physicians (1992) and others (Wadeand Collin, 1988; Shah and Cooper, 1993)recommend its use as a 'gold standard' formeasuring rehabilitation outcomes. For thesereasons, and because some SPARG membershad used Barthel with elderly (non-amputee)patients, SPARG chose the Barthel Index as oneof its functional outcome measures in 1992.The Russek's classification, unlike the BarthelIndex, was developed for use with lower limbamputees. It is a six-point scale (see Appendix,Table A2) used to assess a patient's functionalabilities when using his/her prosthesis. A scoreof six is awarded when the prosthesis providesfull restoration of function and a score of onemeans that the prosthesis offers no advantage tothe patient. In addition to the basic six-pointscale, Kullman (1987) used the positive andnegative factors concerning the patient, thestump and the prosthesis listed by Russek (1961)in his original publication. The number ofpositive and negative factors was used tocorrelate walking ability (as measured by thesix-point scale) with prognosis prior to receivingthe prosthesis. Russek found, for example, thatthe presence of one negative factor usuallydecreased walking ability by one point on thescale. SPARG, however, was not concerned perse with prognosis at admission, but in assessingin a simple way functional abilities at dischargeand so used only the six-point scale. Altner et al.(1980),forexample,usedRussek'sclassification in this way to assess the pre- andpost-amputation functional abilities of blindlower limb amputees.To the best of the authors' knowledge, there isno published work concerning the validity andreliability of Russek's classification. ion, to be a useful method ofevaluating an amputee's walking abilities and,because of this, SPARG considered itworthwhile to try this measure with amputeepatients in Scotland.The Locomotor Index is part of a moredetailed assessment measure known as theProsthetic Profile for Amputees (PPA)
180S. P. Treweekdeveloped by Gauthier-Gagnon and colleaguesat the University of Montreal (Gauthier-Gagnonand Grisé, 1994; Grisé et al., 1993). TheLocomotor Index is a self-standing 14 pointmeasure with each item on the scale scored fromzero to three according to the patient's degree ofindependence in performing a given activity (seeAppendix, Table A3). This scoring system givesa minimum score of zero and a maximum scoreof 42. The index can be divided into two sevenpoint subscales covering basic and advancedactivities. Gauthier-Gagnon et al. (1993) foundthese subscales to be clinically useful with theadvanced subscale discriminating between theleast and most able amputees (Gauthier-Gagnon,1995, personal communication). The authors ofthe PPA have demonstrated that the measure hasface and construct validity and strong test-retestagreements show the measure also to be reliable(Gauthier-Gagnon and Grisé, 1994; Grisé et al.,1993).SPARG now uses only the Locomotor Index.The aim of this paper is to review its experiencesof these three measures and to explain whySPARG would now recommend the LocomotorIndex as the only one of these measures thatshould be considered an appropriate measure offunctional outcome for lower limb amputees.MethodologyData collected on amputees who had anamputation between 1/10/92 and 30/7/97 werereviewed. The patient group considered in thispaper comprises a subgroup of 938 unilateralamputees who had an amputation at either thetranstibial (74%) or transfemoral (26%) level,were fitted with a prosthesis and had theirfunctional abilities assessed using at least one ofthe Barthel Index, Russek's classification or theLocomotor Index. There were 573 men and 346women. The sex of the remaining 19 patientswas not recorded but these patients are includedand M. E.Condiein the analysis since patients were notsubdivided by sex. The mean age was 67 with astandard deviation of 15 years; 78% of patientswere 60 or over. The dominant aetiology wasperipheral vascular disease which accounted for87% of patients, increasing to 92% for patientsover 40. A third of patients with peripheralvascular disease were also recorded as beingdiabetic. The remaining amputations resultedfrom trauma (5%), tumours (2%), congenitaldeformities (1%) and various other causesincluding infection (5%).A total of 546 patients were assessed using theBarthel Index, 772 using Russek's classificationand 195 using the Locomotor Index. Thedifferent numbers of patients assessed using thethree measures simply reflects the different timeperiods for which the measures were in use.Basic details of patients assessed using each ofthe three measures are given in Table 1. Manypatients appear twice in Table 1 becauseRussek's classification was used, at differenttimes, together with the Barthel Index and theLocomotor Index. Of the 546 Barthel Indexpatients, 514 were also assessed using Russek'sclassification while 61 of the Locomotor Indexpatients were also assessed using Russek'sclassification.The three measures are compared here usingtwo a priori predictions. Firstly, youngeramputees should score significantly higher thanolder amputees and, secondly, transtibialamputees should score significantly higher thantransfemoral amputees. The consensus view ofSPARG is that a measure demonstrating both ofthese results is a better measure than one thatdoes not.ResultsSummaries of median scores on the BarthelIndex, Russek's classification and LocomotorIndex are given in Tables 2-4. The Mann-Table 1. Basic details of patients assessed using the Barthel Index, Russek's classification and the Locomotor Index.PVD peripheral vascular disease.
Functionaloutcome foramputees181Table 2. Median Barthel Index scores achieved by transtibial (TT) and transfemoral (TF) amputees. Amputees have beencompared by level (TT v TF) and by age (e.g. TT 40 v TT 40). CI confidence interval.Table 3. Median Russek's classification scores achieved by transtibial (TT) and transfemoral (TF) amputees. Amputeeshave been compared by level (TT v TF) and by age (eg. TT 40 v TT 40). CI confidence interval.Table 4. Median Locomotor Index scores achieved by transtibial (TT) and transfemoral (TF) amputees. Amputees havebeen compared by level (TT v TF) and by age (TT 40 v TT 40). There were too few young transfemoral amputees for ameaningful comparison of median scores for the two age groups. CI confidence interval.Whitney test has been used for significancetesting as all three measures are ordinal makingparametric tests inappropriate. Only the BarthelIndex does not show a significant difference inmedian score between transtibialandtransfemoral amputees. Younger (40 years oldwas chosen as the cutoff) transtibial amputeesscore significantly higher (p 0.001) than olderamputees for all three measures. Russek'sclassification also demonstrates a significantdifference (p 0.001) in median score due to agefor transfemoral amputees. Unfortunately, therewere too few (four) transfemoral amputees belowthe age of 40 assessed using the LocomotorIndex and a meaningful comparison of medianscores for the two age groups was not possible.
182S. P. TreweekTable 5 compares median scores achievedusing the full Locomotor Index with scoresachieved using only the advanced activitiessubscale. The advanced activity subscaleappears to provide the same level of sensitivityas the full Index but with median values thatrepresent a smaller proportion of the maximumscore, i.e. ceiling effects appear to be reduced.The basic activities subscale (data not shown)also shows significant differences due to age andlevel when considered separately but the medianvalues obtained from this subscale represent ahigher proportion of the maximum score. Forexample, the median basic activities score forthe transtibial amputees included in Table 4 is 19or 90% of the maximum score.DiscussionFunctional outcome measures should be valid,reliable and unidimensional (Tennant andYoung, 1997). In addition to these psychometricand measurement properties, a clinically usefulfunctional outcome measure should reflectclinical experience. A functional outcomemeasure for use with lower limb amputeesshould, at the very least, demonstrate astatistically significant difference in the scoresobtained by patients who have been subdividedby age and level of amputation. A measure thatdoes not show that transtibial amputees havemore functional capacity on average thantransfemoral amputees must be considereddubious since clinical observation shows thatthere is a very real difference. The same is trueof young and old amputees. Ideally, the measureshould also have a good range such that floorand ceiling effects do not adversely affect theresponsiveness of the measure. Only after thesebasic criteria have been met can the measure beand M. E.Condieused to investigate less obvious, and perhapsspeculative, causes of variation in, functionaloutcome.The Barthel Index is standardised, valid andreliable but lacks sensitivity when used withamputee patients. The data presented in Table 2do not show a significant difference in medianBarthel score between transtibialandtransfemoral amputees. Further, the very highmedian scores are an indication that ceilingeffects (i.e. a predominance of maximum scoresin some items) are a significant problem. Theability of the Barthel to respond to clinicallyimportant change has been questioned by otherauthors (Ashburn et al., 1993; Simpson andForster, 1993; Smith, 1993) with the mobilityand transfer sections receiving particularcriticism. These two sections are arguably themost important for amputees as these patientsgenerally have few problems with feeding,grooming and toileting. The feeding andgrooming sections of the Index highlight themajor shortcoming of the Barthel when usedwith amputees: it asks the wrong questions.Amputees uniformly score very high on thesequestions which greatly reduces the measure'ssensitivity. These problems make the BarthelIndex completely inappropriate for use as afunctional outcome measure with lower limbamputees and SPARG stopped using it in 1995.Russek's classification, unlike the Barthel,does show significant differences betweenpatients of different age and level of amputation(Table 3). While these results would seem to beencouraging, the Russek's six-point scale meansthat large numbers of patients are required toshow these differences. Annual studiesconducted by SPARG (Condie et al., 1996;Treweek and Condie, 1996) have generallyTable 5. Comparison of median scores achieved using the full Locomotor Index with scores achieved using only theadvanced activities subscale. Amputees have been compared by level (TT v TF) and by age (TT 40 v TT 40).CI confidence interval.
Functionaloutcome forfailed to replicate the results seen in Table 3. Thenumber of patients assessed via the Russek'sclassification in these annual studies was around330 which does not appear to be enough fordifferences due to level of amputation to reachsignificance. The situation for a single hospitalconducting a local study of functional outcomeis even worse since the number of primary lowerlimb amputees per year is often less than 30. Theeffect of age is, however, clear and has beenseen in previous work (Treweek and Condie,1996). The poor sensitivity of Russek'sclassification led SPARG to stop using it in1997.The Locomotor Index (Table 4) demonstratessignificant differences due to age (for transtibialamputees) and level of amputation despite themuch smaller number of patients that havecurrently been assessed using this measure. It isalso likely that the Locomotor Index would haveshown significant differences due to age withtransfemoral amputees had more of thesepatients been available for assessment. Theseresults suggest that the Locomotor Index is moresensitive than both the Barthel and Russek'sclassification. There is, however, a tendency forthe median values to be at the higher end of theLocomotor Index's scale, the most strikingexample of this being the median score foryoung transtibial amputees. The original authorsof the Locomotor Index calculated mean scoresand obtained similar values to the medianspresented here: mean Locomotor Index score of30.7 (out of a possible 42) for 396 amputees ofmixed amputation level (Gauthier-Gagnon,1995, personal communication). Their meanscore for the advanced activity subscale was13.0 out of a possible 21.By considering the advanced activity subscaleseparately, it is possible to reduce the medianvalue as a proportion of the maximum scorewithout losing sensitivity. This extendstheuse ofthe Locomotor Index to more active, but elderly,amputees although no improvement is seen foryounger active amputees. Perhaps it is too muchto expect the same functional outcome measureto be suitable for a fit, 25 year old traumaticamputee and a 75 year old amputee withperipheral vascular disease and diabetes.Although the seven-point advanced activitiessubscale gives similar results to the full 14-pointLocomotor Index, the temptation to drop theseven items of the basic activities subscaleamputees183should be resisted since only the full measure hasbeen validated. The use of the complete measureis the approach recommended by the originalauthors and should be used until the subscales arefound to be valid and reliable when used alone.ConclusionSPARG has gained a great deal of experiencewith functional outcome measures during thefive year period covered by this paper. HadSPARG the benefit of this experience in 1992,the Barthel Index and Russek's classificationwould not have been chosen as functionaloutcome measures. The Barthel Index has verypoor sensitivity and although Russek'sclassification does demonstrate significantdifferences due to age and level of amputation,this six-point scale requires a large number ofpatients to achieve this. Differences infunctional outcome having more subtleexplanations than age and level of amputationare likely to require even more patients.Conversely, the Locomotor Index givessignificant results for smaller numbers ofpatients and the advanced activities subscaleallows the range of the measure to be increasedto include some of the more active amputees.The Locomotor Index is a promising measureof functional outcome for lower limb amputeesand this is the only measure SPARG currentlyuses. A system of post-discharge functionalassessment based around the Prosthetic Profilefor Amputees and including the LocomotorIndex is now being developed. This will allowmonitoring of long-term functional ability andraises the prospect of being able to link elementsof acute rehabilitation care to long-termfunctional outcome. This will provide somemuch needed information about long-termclinical effectiveness and give a more evidencebased foundation to some aspects of amputeerehabilitation.REFERENCESALTNER P C , RUSIN J J , D E B O E R A ( 1 9 8 0 ) . Rehabilitationof blind patients with lower extremity amputations.Arch Phys Med Rehabil 61, 8 2 - 8 5ASHBURNJ,PARTRIDGEC,DESOUZAL(1993).Physiotherapy in the rehabilitation of stroke: a reviewClin Rehabil 7 , 3 3 7 - 2 4 5 .COLLIN C , W A D E D T , D a V I E S S, HORNE V ( 1 9 8 8 ) . TheBarthel ADL Index: a reliability study. Int Disabil Stud10, 6 1 - 6 3 .
184CondieS. P. TreweekME, JONES D, SCOTT H,TreweekSP (1996). Aone-year national survey of patients having a lowerlimb amputation. Physiotherapy 82, 14-20.and M. E. CondieRUSSEK AS ( 1 9 6 1 ) . Management of lower extremityamputees. Arch Phys Med Rehabil 42, 6 8 7 - 7 0 3 .SHAH S, COOPER B ( 1 9 9 3 ) . Commentary on 'a criticalGaUTHIER-GaGNON C, GRISÉ M-CL (1994). Prostheticprofile of the amputee questionnaire: validity andreliability: Arch Phys Med Rehabil 75, 1309-1314.evaluation of the Barthel Index'. Br J Occup Ther 56,70-72.SIMPSONGOLDBERG RT (1984). New trends in the rehabilitation oflower extremity amputees. Rehabil Lit 45, 2-11.GRISÉ M-CL,GAUTHIER-GAGNON C,MARTINEAUGG(1993), Prosthetic profile of people with lowerextremity amputation: conception and design of afollow-up questionnaire. Arch Phys Med Rehabil 74,862-870.M S , FoRSTERA(1993).Assessingelderlypeople: should we all be using the same scales?Physiotherapy 79, 8 3 6 - 8 4 1 .SMITH A ( 1 9 9 3 ) . Beware of the Barthel.Physiotherapy79, 8 4 3 - 8 4 4 .TENNANT A, YOUNG C ( 1 9 9 7 ) . Coma tocommunity:continuity in measurement. Phys Med Rehabil:StateArt Rev 11, 3 7 5 - 3 8 4 .KULLMANN L (1987). Evaluation of disability and ofresults of rehabilitation with the use of the BarthelIndex and Russek's classification. Int Disabil Stud 9,68-71.MAHONEY FI, BARTHEL D W (1965). Functionalevaluation: the Barthel Index. Md State Med J 14, 6 1 65.TREWEEK SP, CONDIE M E ( 1 9 9 6 ) . A survey of the lowerlimb amputee population in Scotland between 1 / 1 0 / 9 3and 3 0 / 9 / 9 4 ; annual audit report of the ScottishPhysiotherapy Amputee Research Group (SPARG)(available through SPARG.)W A D E D T , COLLIN C ( 1 9 8 8 ) . The Barthel A D L Index: aRoyal College of Physicians of London and BritishGeriatric Society (1982). Standardised assessmentscales for elderly people.-London: RCP and BGS.standard measure of physical disability? Int DisabilStud 10, 6 4 - 6 7 .AppendixTable A 1 . The Barthel Index. Patients unable to do a particular activity score zero for that activity.Table A2. Russek's classification. Note that some authors choose to reverse the scoring,i.e. 1 'Full restoration', 6 'Not feasible'.
FunctionalTable A3. The Locomotor Indexoutcome for amputees185
functional outcome for lower limb amputees. Methodology Data collected on amputees who had an amputation between 1/10/92 and 30/7/97 were reviewed. The patient group considered in this . in the analysis since patients were not subdivided by sex. The mean age was 67 with a standard deviation of 15 years; 78% of patients were 60 or over. The .
Patient Reported Outcome Measures References. This document provides information about patient-reported outcome measures (PROMs) . These measures have special considerations outside the more common structure , process , and outcome measures . This information supplements the information found in the Blueprint, Chapter 5,
Functional outcome assessment documented, care plan not documented, documentation the patient is not eligible for a care plan at the time of the encounter . OR . Functional Outcome Assessment not Documented, Reason Not Given . Performance Not Met: G8541: Functional outcome assessment using a standardized tool not documented, reason not given . OR
Country: Viet Nam Initiation Plan Project Title: For preparation and implementation of Project on Leveraging Viet Nam's Social Impact Business Ecosystem in Response to COVID-19 (ISEE-COVID) Expected UNDAF/CP Outcome(s): Outcome 1.1; Outcome 2.1; Outcome 4.1 Expected CPD Output(s): Outcome 1; Outcome 2; Outcome 3 Initiation Plan Start/End Dates: 20 May 2021 - 20 August 2021 (3 months)
graphical comparison of an agency's actual outcome (1) with its expected outcome (using a "national" comparison group) and (2) with its risk-adjusted outcome for the prior year. An enumeration of the outcome measures with a brief rationale for why the current group of 41 outcomes was selected is provided in Section 5 of this document.
levels of the organization. Dan Petersen suggests that only process-oriented metrics be used at the lower managerial or unit levels and activity measures (with some outcome measures) primarily used for the middle-upper management levels. Pure outcome measures should be reserved for the executive leve
Measures outcome rather than care coordination actions Cost Measures Select utilization measures in alignment with model of care Measures may need to vary by practice and patient population Measures outcome rather than care coordination actions What accountability mechanisms exist to track the effectiveness of care coordination? 1.
Numeric Functional Programming Functional Data Structures Outline 1 Stuff We Covered Last Time Data Types Multi-precision Veriﬁcation Array Operations Automatic Differentiation Functional Metaprogramming with Templates 2 Numeric Functional Programming Advanced Functional Programming with Templates Functional Data Structures Sparse Data Structures
each measure for at least 50% of the EP's patients Of these measures, EP would report on at least 2 outcome measures OR If 2 outcome measures are not available, report on at least 1 outcome measure and at least 1 resource use, patient experience of care, efficiency/appropriate use, or patient safety measure