Secondary Logo

Journal Logo

Outcome Measurement Tools

Predictive Outcome Measures Versus Functional Outcome Measures in the Lower Limb Amputee

Gailey, Robert S. PhD, PT

Author Information
JPO Journal of Prosthetics and Orthotics: January 2006 - Volume 18 - Issue 6 - p P51-P60
  • Free


John Ware1 wrote, “Life has two dimensions: quantity and quality.” The distinction between the two entities is well illustrated in the common greeting “may you have a long and healthy life” (p. 473). Length of life is expressed in terms of average life expectancy, mortality rates, death due to specific causes, and numerous other indicators. When defining the second dimension, quality of life encompasses standard of living, the quality of housing and the neighborhood in which one lives, job satisfaction, and many other factors.1 With regard to the amputee, it has been well established that for a number of vascular amputees who have undergone an amputation due to the magnitude of vascular compromise, life expectancy is relatively short. For the amputee who has lost a limb to tumor, trauma, or congenital condition, amputation has little or no bearing on life expectancy. Medical interventions have improved care to the point where the impact of amputation on longevity has decreased tremendously.

So, as quantity of life improves, what about quality of life? The World Health Organization (WHO) defined health as a “state of complete physical, mental, and social well-being and not merely the absence of disease or infirmity.”2 Health connotes “completeness” – nothing is missing from the person; it connotes “proper functions” – all is working efficiently. That would suggest that if a prosthesis has the ability to improve the quality of life for the amputee, then it should be provided. This is not to say that the “best” prosthesis belongs on all patients, primarily because what is considered the “best” for one person may be a hindrance to another. For example, a prosthetic foot that provides a mechanical high energy return at terminal stance, propelling the limb into the swing phase of gait will be a tremendous asset for a strong person with a fast cadence. Conversely, the same foot might only throw a frail, elderly person off balance, resulting in the fear of falling and a tentative gait pattern. Unfortunately, we are not sure what the best prosthesis is for the different amputee populations.

The dilemma becomes apparent when trying to determine what are the best prosthetic components for each individual amputee. Common indictors such as materials, time of fabrication, complexity of design, and cost do not dictate the “best” components for each individual. Matching functional ability with the proper components is the solution for optimizing physical performance. This is not a novel concept. In fact, if questioned, most clinicians would agree that marriage of the correct prosthetic components for the appropriate level of function is one of the primary goals of the rehabilitation team. However, this goal is apparently not so easy to achieve.

At the very root of the problem is the inability to define a “successful prosthetic ambulator.” There are many interpretations for this common goal, from simply using the prosthesis “about an hour per day”3 to “prosthetic use without external support on a daily basis.”4 There is no agreement anywhere with regard to the threshold of “successful prosthetic use.” The question is whether there is one threshold or many as severe authors have advocated by virtue of offering multilevel functional scales and indices.

There is a wide variety of indices published within the literature.5–10 In the United States today, it appears that only one index is of significant consequence to Medicare and managed care providers, the Durable Medial Equipment Regional Carrier (DMERC) K codes or Medicare's Functional Classification Level (MFCL) index.11 Its exceptional importance is solely because assignment to a particular level of function within this index determines the level of prosthetic care, and, consequently, the financial reimbursement a prosthetist will receive for the prosthetic services rendered. After an extensive review of all available indices, the descriptors and selection of number of levels chosen for the MFCL index appears to be well constructed. This assumption is predicated on the basis that the MFCL index has a range from a fully dependent bed-bound amputee to a recreational athlete capable of higher level activities. Definitions clearly describe observable differences between each of the five levels using seemingly commonly assessed functional skills. It must be kept in mind, however, that as attractive as the MFCL may appear to be after initial examination, it was created based on content validity of a Medicare subcommittee and was not based on any formal scientific research.

The use of a functional index seems to be reasonable, but how is the amputee classified? Currently, a physician or prosthetist classifies the amputee according to functional level. “Functional level” defined by Medicare is a measurement of the capacity and potential of the patient to accomplish his or her expected postrehabilitation daily function.11 Functional level is determined by three determinants: 1) the patient's past medical history, 2) the patient's current condition, including the status of the residual limb and nature of other medical problems, and 3) the patient's desire to ambulate. In reality, the physician or prosthetist looks at health status or the number of comorbidities and asks the patient if he or she wants to walk again. Helm et al.12 and Gailey et al.13 did not find comorbidity to be one of the strongest predictors regarding functional outcome, although many others had.14–17 Unfortunately, Medicare offers no standardized assessment index or formula for use by the clinician when determining the severity of the comorbidities associated with the amputee. Regarding an amputee's motivation to ambulate, it is not difficult to imagine that most amputees interested in prosthetic fitting will say they are very motivated to learn to walk and return to an active life style, especially when their response will influence the level of prosthetic care they will receive.

The reality concerning the evolution of the prosthetic prescription process is that with the exception of some additional paperwork, the overall system has not changed at all. The prosthetist is still required to secure a physician's prescription prior to the fabrication of the prosthesis. If the physicians are to classify the amputees, they will continue to prescribe as they always have since there are no formal assessment guidelines. If they choose to “underprescribe” for fear that the patient will not reach rehabilitation potential, or in an attempt to contain medical costs, they can continue to do so. In many cases, the physicians have continued to leave the classification decision to the prosthetist as they have historically done by cosigning the necessary documentation. In most cases, this is very appropriate because the prosthetist is more familiar with all the prosthetic options available. As a result, the prosthetist decides the functional level of the amputee based on clinical experience without a formal assessment tool. If the prosthetists want to continue to “underfit” amputees, they can always use “lesser” prosthetic components. Conversely, if they choose to raise the level of prosthetic components beyond the amputees’ functional capability, for whatever reason, they can misclassify the amputee, or petition the physician or Medicare by providing the required documentation necessary to elevate the level of prosthetic care.

The product of the MFCL system is the same subjective assignment of prosthetic components to the amputee with the minor addition of some documentation. There is no genuine system of checks and balances as originally intended. However, the most significant contribution of the MFCL process may be the formal definition of “prosthetic success” in the form of the MFCL index that describes a hierarchy of “prosthetic successes.” However, without a measurement tool that objectively and effectively assigns the amputee to each of the appropriate predetermined levels of “prosthetic success” by determining the amputee's “readiness to ambulate,” the same system that has prevailed throughout history continues to exist.

The most common argument opposing the use of functional assessment tools is that professionals have the ability to accurately determine the needs of the amputee based on their years of clinical experience. After reviewing the findings of Stephen and Aitken,18 where significant differences were found in assessing mobility and self-care of the amputee who were evaluated by clinicians on the same rehabilitation team, the validity of such arguments must be questioned. If members of the same rehabilitation team find it difficult to agree where everyone theoretically has met in team meetings, treated common patients, exchanged professional information, and followed similar treatment policies, what is the likelihood that uniform agreement can be reached across multiple professions and amputee rehabilitation facilities? To further add potential confusion to the classification of amputees, Medicare offered very little in the form of guidelines or definitions concerning the five MFCL. Clearly, an objective assessment tool was needed to provide standardization to the process of assigning a functional level to the amputee.


The signs of a patient getting better are often clear not only to the clinician but also to family members and the patient. The concept of functional assessment is to “measure change.” To measure change, the WHO has identified three areas of measure: 1) With impairment at the organ system level, has the underlying condition improved? 2) With disability as the functional consequence of impairment, has physical function improved? 3) With handicap as the social and societal consequence of disability, is the patient better off overall for the treatment?12

Functional outcome measures should consistently measure what they say they are going to measure.19 Consequently, we have the basis of reliability and validity. Reliability is a fundamental way to reflect the amount of error, both random and systematic, inherent in any measurement.20 Reliability is best measured by comparing the variance between observations on an individual with the variance across the subject range using the interclass correlation coefficient (ICC).21 Validity is the degree to which an instrument measures what it is intended to measure.22 Validity reflects the ability of the instrument to measure the “true” factor of interest. In rehabilitation, there is no gold standard of function, so validation is a roundabout process. We can compare one measure with others that seem to reflect the same phenomena (concurrent validity).2

The measurement tools identified from the literature can be divided into two categories: subjective or self-report assessment method, and the more objective performance-based assessment.


Frequently, the preferred choice of assessment instrument is a self-report because of the ease of use and the ability to reflect the patient's perspective. In addition, there may be variances or inconsistencies with other methods of evaluation.23 Professional report permits a rigid use of defined test item characteristics and can be performed by the clinician rather quickly, but the opportunity for subjectivity or bias is apparent if the evaluator is also the treating clinician. With these forms of assessment tools, the patient, family member, or clinician is asked to complete a questionnaire or similar type survey and a test score or descriptive profile is generated.

Physical performance can be the most objective measure, providing more accurate information of the patient's physical abilities. Weaknesses do apply such as performance variation due to “good days” and “bad days,” artificially created environments, and performance tasks that may illustrate only the patient's ability at a given point in their rehabilitation sequence, not at the base or final outcome. Few instruments in this category of testing have been designed specifically for the amputee. However, a fair number have been applied to the amputee, with varying success. The most notable self-report, professional report, and performance-based measures used with lower limb amputees are:


  • 1) Amputee Activity Survey24
  • 2) Sickness Impact Profile25
  • 3) Reintegration to Normal Living26
  • 4) Prosthetic Profile of the Amputee27
  • 5) SF-36 Health Status Profile28
  • 6) Prosthetic Evaluation Questionnaire29
  • 7) Orthotic Prosthetic User's Survey30


  • 1) Barthel Index31
  • 2) Functional Independence Measure32


  • 1) Six-minute Walk33
  • 2) Functional Ambulation Profile34
  • 3) Two-minute Walk35
  • 4) Performance-Oriented Assessment of Mobility Problems36
  • 5) Timed Get-Up and Go37
  • 6) Berg Balance Measurement38
  • 7) Duke Mobility Skills Profile Test39
  • 8) Functional Reach Test40
  • 9) Amputee Mobility Predictor13


The first item of focus is to determine if an index appropriately defines and differentiates the levels of ambulatory ability of the amputee. Although several indices used throughout the literature were identified, only one can be used for this investigation, the MFCL. Because it has been adopted by the United States Medicare system, by default it must be used. The second issue concerns the existence of a method of measurement that adequately assesses and describes the current functional ambulatory ability of the amputee to establish construct validity. As discussed, there are three alternatives, self-report, professional report, or performance-based functional assessment measures. The third task is to construct an assessment instrument to predict the amputee's future ability or readiness to ambulate. Clearly, the optimal situation would be to have a single test that could appraise the current ambulatory status of the amputee, differentiate among the five functional levels of the MFCL, and predict the amputee's readiness to ambulate.

The difficulty in designing a tool for assessing “readiness to ambulate” arises by virtue of the term “readiness,” which implies “prepared for” or “likely to do something.” Therefore, the instrument must be a predictor tool for how well an amputee will ambulate without actually observing the amputee ambulating. The Amputee Mobility Predictor (AMP) is the only functional assessment instrument that has demonstrated the ability to determine functional level or predict functional ability for amputees.12


The reliability study was performed with a small group of subjects (n = 24) with a mean age of 68 years. Inter-rater reliability of the AMP was very good, as both raters’ observations were extremely comparable during both AMP tests with ICC scores of 0.992 for the AMPnoPRO and 0.996 with the AMPPRO. The AMP intra-rater reliability — AMPnoPRO and AMPPRO were 0.966, and 0.964, respectively — also demonstrated good instrument stability with little variation from the initial testing period to follow-up testing 3 weeks later. Because of the high intra-rater reliability, it can be ensured that not only will the test results be consistent regardless of who tests the amputee, but also that the test outcomes will remain constant across repeated tests, exclusive of any dramatic changes in functional status. Therefore, a variety of appropriate allied health professionals within a clinic, once trained, may administer the AMP with reasonable confidence. This would be valuable in a multidisciplinary rehabilitation team situation where many disciplines are responsible for amputee follow-up care.


Although the ability of the AMP to measure change in functional status throughout rehabilitation was not specifically examined, the AMP has sufficiently high reliability between raters and trials. This implies that if change were to occur with an amputee, the AMP would have the ability to detect the change in functional status. The likely ability to measure change in functional status would be an additional clinical benefit not only for monitoring progress or decline but also in the ability to clearly document any variations in functional status. Finally, the effects of wearing or not wearing a prosthesis during AMP testing did not influence results, as all scores were very similar regardless of rater or time of rating.


Results indicate that there was a significant relationship between the prescribed prosthetic components and the assigned MFCL groups. However, there was only 46% agreement with the prescribed foot/ankle components and 69% agreement in knee prescription with respect to MFCL assignment. Interestingly, it is a general belief by some third-party payers and other skeptics within the field of prosthetics that amputees are frequently “overprescribed” or inappropriately receive prosthetic componentry that was intended for higher functioning amputees. A closer look at the MFCL assignments that were based on the amputee's current functional level and not “functional potential” reveals that 30% of amputees were “underprescribed” or received prosthetic foot/ankle assemblies designated for amputees functioning at a lower level. Knee units were appropriately prescribed, especially at the K3 and 4 level. The foot/ankle assemblies include all the participants in the study, because all should have received a foot from the K1 level and higher. The knee figures exclude transtibial amputees because they would not receive a prosthetic knee unit. Tables 1 and 2 provide a summary of MFCL assignments and the foot and knee units that were prescribed for the subjects.

Table 1
Table 1:
MFCL assignments and prosthetic foot unit prescription
Table 2
Table 2:
MFCL assignments and prosthetic knee unit prescription


To determine the AMP's ability to test “readiness to ambulate,” the construct validity had to be evaluated. The MFCL's descriptions defined and distinguished the amputee's “readiness to ambulate,” establishing known groups, and allowing the AMPnoPRO and AMPPRO to measure the construct “readiness to ambulate.” As a result, the average AMP score for subjects in each of the MFCLs should be clearly different. When construct validity known groups method was tested with the use of analysis of variance to see if the AMPnoPRO and AMPPRO were capable of discriminating between the MFCLs, the results were extremely encouraging. Not only were the differences among the four groups significant, but the AMPnoPRO, or testing without the use of the prosthesis, proved to be a very competent instrument as well. The ultimate result was that the test, which could be administered with or without the use of a prosthesis and would clearly discriminate between the functional levels of an index, clearly represents the range of functional levels observed with lower limb amputees.


The 6-minute walk and the AAS were selected as concurrent tests to be used as the gold standards. The 6-minute walk had an obvious advantage for testing ambulation as it is a performance-based assessment that is a walking test. Another advantage taken into consideration was that it was time tested, especially important with elderly and cardiovascular patients.41,42 The AAS was the self-report subjective instrument that appeared most relevant to the amputee.

Tables 3–6 provide the means, standard deviations, and range comparison for each of the MFCLs to the 6-minute walk, AAS, AMPPRO, and AMPnoPRO, respectively. As stated previously, all means were statistically significant among groups, with the one exception of the AAS scores between level 3 and level 4. Further examination of the standard deviation and range prompts further discussion of the clinical relevance of each of these instruments.

Table 3
Table 3:
Six-minute walk distance means, standard deviations, and range
Table 4
Table 4:
Amputee activity survey means, standard deviations, and range
Table 5
Table 5:
AMPnoPRO means, standard deviations, and range
Table 6
Table 6:
AMPPRO means, standard deviations, and range

The 6-minute distance mean values given in Table 3 are distinctly separated and without overlap with the standard deviations. The range of the minimum and maximum distances recorded shows some overlap as isolated subjects may record values outside of the group means. For example, the range of level 2 is 16 to 480 meters, whereas level 3 is 48 to 475 meters, demonstrating that some level 2 amputees walked just as far as level 3 amputees. Although the 6-minute walk has demonstrated the ability to differentiate among the various MFCLs, individual 6-minute walk distance results in some cases may not be absolutely reflective of a particular amputee's MFCL.

The AAS mean scores and standard deviations in Table 4 were significantly separated in the lower functional levels but failed to significantly distinguish differences between levels 3 and 4. The range of scores was found to be 62, 131, 107, and 59 points for levels 0 and 1, 2, 3, and 4, respectively. Regardless of the wide range in test scores, the lower scores across all levels were found to be somewhat close together, whereas the high score for levels 2, 3, and 4 were very close, with only a 5-point difference among all three.


The MFCL is a widely used index and the AMP was able to discriminate among the different levels, but one of the most important questions still remained to be answered: Can it predict the amputee's “readiness to ambulate” using the distance walked 6 minutes as the dependent variable? The AMP, and the selected variables of age, comorbidity, and time after amputation, were used to determine predictive validity, with the AMP instruments proving to have the strongest contribution. The fact that the AMP scores are not totally explained by age, comorbidity, and time after amputation in its ability to predict functional ambulation further illustrates that the AMP contributes more than any other variable.

The 6-minute walk has been determined to be one of the best indicators for ambulation ability, exercise tolerance, and ability to differentiate levels of cardiac status and is safe for patients with cardiac compromise and vascular disease.42 If a clinician wants to know a patient's “readiness to ambulate,” they are essentially seeking to know how far they can walk within a given time. The ability of the AMPnoPRO to successfully predict the distance that amputees will walk in 6 minutes clearly enhances the overall clinical value of this measurement tool.

The addition of the three variables age, comorbidity, and time after amputation not only offers the ability to explain a greater percentage of the variance but also allows clinicians to incorporate factors that are traditionally perceived as indicators for functional success. A multiple regression model was examined to determine whether the AMP could predict function when other known predictors of function (age, comorbidity and time since amputation) are introduced. The predictive validity of the AMP held with the addition of age, comorbidity, and time since amputation along with the AMPnoPRO and then again with the AMPPRO as examined against the 6-minute walk distance.

Multiple regression results revealed strong support for the ability of the AMP to predict distance walked in 6 minutes — 69% and 62% for the AMPnoPRO and AMPPRO, respectively. Furthermore, the squared partials indicated that 39% and 26% of the contribution comes from the AMPnoPRO and AMPPRO, respectively, with these two predictors providing the greatest individual contributions to the overall prediction. With the majority of the contribution coming from the AMP instruments, the power of administering an objective form of measurement strengthens the argument that formal testing has benefits over simply looking at factors such as the amputee's age, time after amputation, and the number of comorbidities when determining ambulation potential. In other words, when determining the level of function or ambulation readiness, the results of the AMPnoPRO or AMPPRO contribute far more information when predicting the outcome than any other variable.

Clinically, the most valuable result of this finding would be the predictive formula that could be applied to all prosthetic candidates:

Υ (6-min distance) = − -12.239 − -1.226 (age) + 0.129 (amp. mo) + 7.956 (AMPnoPRO) − 6.235 (comorbidity score)

To estimate the distance walked in 6 minutes, substitute the amputee's age, number of months since amputation, AMPnoPRO score, and comorbidities from the index score into the formula. Multiply by the base predictor scores that were generated with parameter estimate scores for age, time since amputation, AMPnoPRO score, and comorbidity index score, and the values from the basic K1 foot and knee values are produced. The base value is the distance the amputee would walk in 6 minutes with a basic K1 foot and knee unit. Figures 1 through 3 provide the values that represent the predicted additional distance an amputee would walk if provided with the corresponding combination of components. The clinical advantage is that the rehabilitation team member could obtain the necessary information from the amputee, determine the base distance they would be able to walk with a basic prosthesis, and then predict the functional value of upgrading the prosthesis. Although it remains to be completely determined just how well the predictive formula works with all levels of amputees, the formula presents some promising possibilities and is worth further examination.

Figure 1.
Figure 1.:
Predicted 6-minute walk distance for a 25-year-old, healthy TFA, AMPnoPRO score 38. Base distance with K1 components 233 m; add cadence knee and multiaxial foot 330 m; cadence knee and DR foot 349 m; if a TTA with DR foot 408 m. DR = dynamic response.
Figure 2.
Figure 2.:
Predicted 6-minute walk distance for a 75-year-old, unhealthy TFA, AMPnoPRO score 25. Base distance with K1 components 76 m; cadence knee and multiaxial foot 173 m; cadence knee and DR foot 192 m; if a TTA with DR foot 250 m.
Figure 3.
Figure 3.:
Predicted 6-minute walk distance for a TFA with basic K1 components would be 233 m; if a TFA had a cadence control knee and DR foot 349 m, the same distance as a TTA with a basic K1 foot; a TTA with DR foot 408 m.

Figure 1 illustrates the difference in the distance walked in 6 minutes that a 25-year-old transfemoral amputee (TFA) with no comorbidity and an AMPnoPRO score of 38 is predicted to achieve. The base distance with basic prosthetic components would be 233 meters. Add a cadence control knee and a multiaxial foot, and the distance would increase to 330 meters. Add a cadence knee and a dynamic response foot, and the distance would be 349 meters, and if he were a transtibial amputee (TTA) with his own anatomical knee and a dynamic response foot, he could travel 408 meters in 6 minutes.

Figure 2 uses a 75-year-old TFA with three comorbidities and an AMPnoPRO score of 25. The base distance would be 76 meters. Add a cadence control knee and multiaxial foot and the distance would increase to 173 meters. Add a cadence knee and dynamic response foot the distance would be 192 meters, and if he were a TTA with his anatomical knee and a dynamic response foot, he could travel 250 meters in 6 minutes.

What has yet to be determined is the actual functional benefit of each specific component. For example, in the last illustration, would a 75-year-old unhealthy man actually benefit from a dynamic response foot, or would the “spring” from the foot throw him off balance and therefore decrease walking speed as a result of his fear of falling? Conversely, there could be a population of amputees who would benefit significantly in the distance covered during ambulation if provided higher-level prosthetic components and not restricted by their functional classification level. The actual functional benefit of specific prosthetic components needs to be determined before a sophisticated system of matching amputee function and prosthetic componentry can be instituted.

One interesting note, Figure 3 demonstrated the value of an anatomical knee over a prosthetic knee. If a young healthy TFA were fitted with a basic prosthesis, a predicted distance of 233 meters would be walked with a maximum predicted distance of 349 meters if fitted with a cadence-controlled knee and dynamic response foot. However, if a TTA were fitted with a basic prosthesis, the base distance would equal the maximum distance for TFA of 349 meters and if upgraded to a dynamic response foot, the TTA's distance over 6 minutes would increase to 408 meters.

The value of these equations is that with use of either the AMPnoPRO or the AMPPRO and the amputee's age, the number of months since amputation, and the comorbidity score from the index (in the case of the AMPPRO), a relatively accurate prediction of how far the amputee will walk in 6 minutes can be made. In turn, this would provide clinicians with the objective documentation necessary to support a decision for prosthetically fitting an amputee.


The author also recognized when creating the AMP that several other criteria had to be met for the AMP to be clinically accepted. The AMP had to require a minimal amount of equipment, a brief administration time with a somewhat simplistic scoring system. The AMP's administration had to be easily understood by a variety of clinicians with diverse educational qualifications such as physicians, prosthetists, physical therapists and possibly nurses. Finally, the AMP had to test what it was designed to test and therefore predict the amputee's readiness to ambulate, which under the current Medicare system meant that it could differentiate between the five MFCLs.


When the MFCL was first designed, it was for the expressed purpose of classifying amputees into functional classes so that prosthetic componentry could be prescribed according to the needs of the amputee and would in turn control the cost of prosthetic care. Therefore, it would be beneficial if the AMP could also predict the functional value of the different classes of prosthetic foot and knee components. Because the prosthetic components are classified according to function, then each higher class of components should have some measurable improvement on the function of the amputee. If ambulation performance was the measurable criteria, then the amputee should be able to walk a greater distance when provided with a higher level of prosthetic componentry.

Many researchers have attempted to demonstrate functional differences in prosthetic foot components. The most common dependent variable was oxygen consumption, although significant difference between dynamic and nondynamic response feet has never been established.43,44 Muscle power outputs and mechanical energy characteristics have also been examined with differences found between prosthetic feet, but once again functional effects were not observed.45,46 To date, the literature has yet to yield a report that offers either a method of testing or a performance-based result that offers any physiological or functional differences among prosthetic components. However, the very implication of the Medicare K-modifiers (MFCL index for prosthetic components Table 7) is that a functional difference does exist among prosthetic components. On examination of the hierarchy of components as classified by HCFA (Medicare), it appears that cost is a principle justification for the placement of specific components.

Table 7
Table 7:
DMERC prosthetic policy definition of foot/ankle assemblies and knee units by Medicare functional classification levels

Two supplemental questions regarding componentry can be asked in an attempt to further understand what contribution specific variables have on the 6-minute walk. First, does componentry affect the 6-minute walk after controlling for age, time after amputation, comorbidity, and the AMPnoPRO? Second, what are the specific contributions of the different classifications of foot/ankle and knee components?


The most common argument against prediction of prosthetic ability would be the variable of compliance. In fact, Mueller and Delitto47 found that only compliance and medical problems after prescription showed a significant difference between successful and nonsuccessful long-term TFA prosthetic users. Surely these are strong predictors and would provide very valuable information except that both compliance and medical problems after prescription happen after the fact and the therapist would be unable to foresee either problem.47 Prediction attempts to forecast the future; there are no absolutes, only the best guess possible.

Clearly the advantage to such a prediction tool is that the advantage of improved prosthetics could be proposed before fabrication of the prosthesis. The determination of the potential level of function would rest with the prescription decision-makers, who would have much more objective data on which to base their decision. This would enable the value of each classification of prosthetic knee and foot component to be more easily depicted and the ramifications to the amputee of the ultimate decision more easily understood in familiar terms.


  • 1) The number of subjects was adequate for the purposes of this study. Where an increase in sample size would have been more beneficial is in the distribution of subjects across all the MFCL levels. Although this was not a problem for this validity study because neither group was considered functional prosthetic users and therefore could not perform the AMPPRO, the 6-minute walk test or the AAS, future studies examining the characteristic and cut scores would benefit from larger numbers.
  • 2) Because different diagnoses have varied impact on the amputee's overall health, the Melchiorre comorbidity index for amputees appeared to offer the best account of health status. The Melchiorre index did offer a good number of diagnoses common to the amputee, however, there was a sense that the list of medical conditions could be further defined and possibly a more accurate system developed. The areas that may have benefited from additional items in particular dealt with the area of diabetes, vascular disease, and cardiopulmonary compromise. That is not to say that the current index is not appropriate and in fact proved to be a very adequate tool.
  • 3) The prediction equations require cross-validation. Are the prediction equations capable of predicting the distance walked in 6-minutes with all MFCL of amputees and with all classifications of prosthetic components? This would require a greater number of subjects or the ability to vary the prosthetic componentry on the same group of amputees.
  • 4) The ability to measure change over time must be determined.


Indices of functional disability are being used increasingly to rate the status of patients studied in clinical research or treated in clinical practice.1 As the healthcare industry demands new mechanisms for the classification of patients to ease the complexities of reimbursement, the necessity for reliable and valid measures will increase. In the case of the amputee, Medicare and subsequently managed care companies have imposed such a system where the amputee must be classified; consequently, the level of prosthetic care is imposed. The Medicare Functional Classification Levels (MFCL) were created to describe five levels of functional ability and in turn designate the appropriate prosthetic componentry for each class level.

The Amputee Mobility Predictor (AMP) was designed to determine the amputee's readiness for ambulation. The AMP has the ability to be administered both without and with the use of a prosthesis. Therefore, it can be administered before the initial fitting of the first prosthesis or it can assess an amputee who has used a prosthesis for years.

The AMP is also relatively easy to administer in 15 minutes or less, with a simple scoring system, requiring very little equipment or space. Testing revealed that the AMP has a high inter-rater and intra-rater reliability and appears to be a very practical clinical tool. The high reliability of AMP suggests that multiple disciplines could administer the test with training and that results will be consistent over time.

Construct validity known groups method was used to determine the variation among the five levels of the MFCL and revealed that the AMP could differentiate among each of the classes both without and with the use of a prosthesis. Concurrently, the 6-minute walk and Amputee Activity Survey (AAS) demonstrated an equal ability for discrimination among the MFCL, suggesting that the five levels do represent differences in functional ability that can be measured and that the AMP has the ability to measure the differences.

Finally, the predictive criterion validity demonstrated that not only does the AMP continue to be the strongest contributor when explaining the relationship between overall predictability and the 6-minute walk but that it continues to hold its value when other variables such as age, comorbidity, and time after amputation are included. Moreover, when prosthetic components are added to the regression model, the contribution of foot and knee units can be identified. Not only does this finding offer a predictive formula that with further research may produce a means to determine the clinical value of providing specific K-modifier prosthetic components to an individual amputee, but this is the first time that influence of prosthetic componentry to function has been clinically measured.

The AMP is the first clinically applicable measurement tool for amputees with the efficiency, reliability, validity and predictive potential necessary to possibly differentiate among functional classes as per the MFCL. The AMP can be used either without or with the use of a prosthesis, providing objective information that may furnish the clinician with greater insight into the amputee's readiness to ambulate and the selection of prosthetic components to permit them to achieve an optimal gait. Without doubt, the AMP offers clinicians a valuable tool for the justification and prescription of prosthetic care for the amputee.


Tables 1 through 7 were reprinted with permission from R.S. Gailey (1999): The Amputee Mobility Predictor: A Functional Assessment Instrument for the Prediction of the Lower Limb Amputee's Readiness to Ambulate. University of Strathclyde.


1. Ware J. Standards for validating health measures: definition and content. J Chron Dis 1987;40:473–480.
2. Studenski S, Duncan P. Measuring rehabilitation outcomes. Geriatric Rehabil 1993;9:823–830.
3. Anderson A, Cummings V, Levine S, Kraus A. The use of lower extremity prosthetic limbs by elderly patients. Arch Phys Med Rehabil 1967;48:533–538.
4. Moore T, Barron J, Hutchinson F, et al. Prosthetic usage following major lower extremity amputation. Clin Orthop 1989;238:219–224.
5. Hanspal R, Fisher K, Chakrabarty B, et al. Mobility grades in amputee rehabilitation. Clin Rehab 1991;5:344.
6. Kerstein M, Zimmer H, Dugdale F, Lerner E. Rehabilitation after bilateral lower-extremity amputation. Arch Phys Med Rehabil 1975;56:309–311.
7. Kullman L. Evaluation of disability and of results of rehabilitation with the use of the Barthel index and Russek's classification. Int Disabil Stud 1987;9:68–71.
8. Narang I, Mathur B, Singh P, Jape V. Functional capabilities of lower limb amputees. Prosthet Orthot Int 1984;8:43–51.
9. Russek A. Management of lower extremity amputees. Arch Phys Med 1961;42:687–703.
10. Volpicelli L, Chambers R, Wagner F. Ambulation levels of bilateral lower-extremity amputees. J Bone Joint Surg 1983;65A:599–605.
11. HCFA Common Procedure Coding System HCPCS 2001. Washington DC: US Government Printing Office; 2001. ch 5.3.
12. Helm P, Engel T, Holm A, et al. Function after lower limb amputation. Acta Orthop Scand 1986;57:154–157.
13. Gailey R, Roach K, Applegate B, Nash M. The Amputee Mobility Predictor (AMP): an instrument to assess determinants of the lower limb amputee's ability to ambulate. Arch Phys Med Rehab 2002;83:613–627.
14. Greive A, Lankhorst G. Functional outcome of lower-limb amputees: a prospective descriptive study in a general hospital. Prosthet Orthotics Int 1996;20:79–87.
15. Nissen S, Newman W. Factors influencing reintegration to normal living after amputation. Arch Phys Med Rehabil 1992;73:548–551.
16. Seeman T, Charpentier P, Berkman L, et al. Predicting changes in physical performance in a high-functioning elderly cohort: McArthur studies of successful aging. J Gerontol 1994;49:M97–M108.
17. Steinberg F, Sunwoo I, Roettger R. Prosthetic rehabilitation of geriatric amputee patients: a follow-up study. Arch Phys Med Rehabil 1985;66:742–745.
18. Stephen P, Aitken R. Morbidity survey of lower limb amputees. Clin Rehabil 1987;1:181–186.
19. Applegate W, Blass J, Williams T. Instruments for the functional assessment of older patients. N Engl J Med 1990;17:1207–1214.
20. Streiner D, Norman G. Health measurement scales: a practical guide to their development and use. Oxford: Oxford University Press, 1995.
21. Feinstein A, Josephy B, Wells C. Scientific and clinical problems in indexes of functional disability. Ann Intern Med 1986;105:413–420.
22. Portney L, Watkins M. Foundations of clinical research: Application to practice. Norwalk: Appleton and Lange, 1993.
23. Siu A, Reuben D. Hierarchical measures of physical function in ambulatory geriatrics. J Am Geratr Soc 199038:1113–1119.
24. Day H. Amputee rehabilitation: finding a niche. Prosthet Orthot Int 1998;22:92–101.
25. Bergner M, Bobbitt R, Carter W, Gilson B. The Sickness Impact Profile: development and final revision of a health status measure. Med Care 1981;19:787–805.
26. Wood-Dauphinee S, Opzoomer A, Williams J, et al. Assessment of global function: the reintegration to normal living index. Arch Phys Med Rehabil 1988;69:583–590.
27. Gauthier-Gagnon C, Grise M. Prosthetic profile of the amputee: Handbook of documents developed within the framework of a prosthetic follow-up study. Montreal, Quebec, Canada: Ecole de Readaptation, Faculte de Medecine, Universite de Montreal, 1992.
28. McHorney C, Ware J, Raczek A. The MOS 36-item short-form health survey (SF-36), II: psychometric and clinical tests of validity in measuring physical and mental health constructs. Med Care 1993;31:247–263.
29. Legro M, Reiber G, Smith D, et al. Prosthetic evaluation questionnaire for persons with lower limb amputations: assessing prosthesis-related quality of life. Arch Phys Med Rehabil 1998;79:931–938.
30. Heinemann AW, Bode RK, O’Reilly C. Development and measurement properties of the Orthotics and Prosthetics Users’ Survey (OPUS): a comprehensive set of clinical outcome measures. Prosthet Orthot Int 2003;27:191–206.
31. Mahoney F, Barthel D. Functional evaluation: the Barthel index. Maryland State Med J 1965;14:61–65.
32. Davidoff G, Roth E, Haughton J, Ardner M. Cognitive dysfunction in spinal cord injury patients: sensitivity of functional independence measure subscales vs neuropsychologic assessment. Arch Phys Med Rehabil 1990;71:326–329.
33. Cooper K. A means of assessing maximal oxygen uptake. JAMA 1968;203:201–204.
34. Nelson R, Dm G, Everhart J, et al. Lower-extremity amputation in NIDDM: 12 year follow-up study in Pima Indians. Diabetes Care 1988;11:8–16.
35. Butland R. Two-, six-, 12-minute walking tests in respiratory disease. Br Med J 1982;284:283–284.
36. Tinetti M. Performance-oriented assessment of mobility problems in elderly patients. J Am Geriatr Soc 1986;34:119–126.
37. Mathias S, Nayak USL, Isaacs B. Balance in elderly patients: the “get-up and go” test. Arch Phys Med Rehabil 1986;67:387–389.
38. Berg KO, Maki BE, Williams JI, et al. Clinical and laboratory measures of postural balance in an elderly population. Arch Phys Med Rehabil 1992;73:1073–1080.
39. Hogue C, Studenski S, Duncan P. Key aspects of recovery: Improving nutrition, rest, and mobility. Funk S, Chanpagne M, eds. New York: Springer, Inc., 1990.
40. Duncan P, Weiner D, Chandler J, Studenski S. Functional reach: a new clinical measure of balance. J Gerontol 1990;45:M192–M197.
41. Gujral J, McNally P, O’Malley B, Burden A. Ethnic differences in the incidence of lower extremity amputation secondary to diabetes mellitus. Diabet Med 1993;10:271–274.
42. Peeters P, Mets T. The 6-minute walk as an appropriate exercise test in elderly patients with chronic heart failure. J Gerontology 1996;51A:M147–M151.
43. Gailey R, Wenger M, Raya M, et al. Energy expenditure of transtibial amputees during ambulation at self-selected pace. Prosthet Orthot Int 1994;18:84–91.
44. Torburn L, Perry J, Ayyappa E, Shanfield S. Below-knee amputee gait with dynamic elastic response prosthetic feet: a pilot study. J Rehab Res Dev 1990;27:369–384.
45. Gitter A, Czerniecki J, DeGroot D. Biomechanical analysis of the influence of prosthetic feet on below knee amputee walking. Am J Phys Med Rehab 1991;70:142–148.
46. Winter D, Sienko S. Biomechanics of the below knee amputee gait. J Biomech 1988;21:361–367.
47. Mueller M, Delitto A. Selective criteria for successful long term prosthetic use. Phys Ther 1985;65:1037–1040.
© 2006 American Academy of Orthotists & Prosthetists