Establishing the Reliability and Validity of Measurements of Walking Time Using the Emory Functional Ambulation Profile.

A combination of impairments can lead to decreased ability to ambulate and an increased risk of falls.[1] Stroke, for example, is a cause of impairment often associated with a decline in ambulation. Assessments are often undertaken to predict the ability of individuals to use a variety of skills when performing tasks necessary for daily living, leisure, vocational pursuits, and other required behaviors.[2] The ability to ambulate successfully should be assessed because this behavior contributes to meaningful activity.[3] Some authors[4-6] contend that an assessment of ambulation should include environmental variables encountered in daily living such as different terrains, obstacles, and stairs. For the purpose of this study, "functional ambulation" is defined as the ability of a person to walk with maximal independence and in the least time under various environmental circumstances. In this regard, changes in quantification of walking time and use of assistive devices may be useful in predicting the impact of specific interventions.

Inevitably, cognition,[7] balance,[8,9] vision and joint position sense,[10] strength,[11] speed[12] endurance[13] and adaptability to environmental demands[4-6] contribute to successful ambulation. Changes in these variables often are evident in individuals poststroke,[14,15] and these changes may be manifested as slow gait speed and altered stance phases[16,17] or as compromised ability to regain balance, control movement, or adjust energy expenditure.[18] Individuals poststroke may have difficulty adapting to environmental demands, such as rising from a chair, stepping over an obstacle, or ascending stairs.

Measures of gait have been used in both laboratory and clinical settings.[19] Laboratory gait analyses may be useful because quantitative measures included in these analyses, such as decreased walking speed and decreased stride length, have been associated with an increased risk of falling.[20] Laboratory tests often require electronic equipment, such as electrogoniometers, electrodes, footswitches, computers,[21] and video cameras.[22] These technical gait analyses may be costly and time-consuming, require extensive training to administer,[3,23] and often do not assess walking across commonly encountered terrains.[24] Some more clinically applicable tests of ambulation are easy to administer and require only a stopwatch.[25,26] Both laboratory gait analyses and simple tests of gait speed typically do not test the ability to move around obstacles and over different surfaces.

The Functional Ambulation Profile was first described by Nelson,[26] who marked the plantar surface of the foot or shoe and recorded foot contact aspects of gait, such as stride length, cadence, and so on. The conceptualization of a functional ambulation profile presented here (ie, the Emory Functional Ambulation Profile [E-FAP]) is quite different and is an inexpensive and easily administered assessment of ambulation. The E-FAP was designed to provide quantitative information about ambulation by measuring time to walk over a standardized array of surfaces and obstacles and accounts for the use of an assistive device. The 5 subtasks in the E-FAP (5-m walks on a floor and on a carpet, an "up & go" task, negotiating an obstacle course, and stair climbing) are included to represent environmental challenges commonly encountered in everyday life.

For the E-FAP to be useful to clinicians, patients, health care decision makers, and third-party payers, reliability and validity of the E-FAP measurements must be established. Subjects who have had strokes are likely to demonstrate altered ambulation[14,18] and should exhibit different scores on an ambulation test than subjects without impairment. Therefore, construct validity would be supported if scores on the E-FAP separate subjects who have had strokes from subjects without impairment.

Because no gold standard or criterion measure of ambulation exists, the concurrent validity of the E-FAP will be evaluated by comparing the E-FAP with tests of gait speed and balance. The Timed 10-Meter Walk Test is reported to yield reliable and concurrent, valid measurements of gait speed in patients who have had strokes.[17] An increase in gait speed is positively correlated with an improved level of mobility in elderly people.[23] Criterion tests of balance include the Berg Balance Test[8] and the Functional Reach Test,[27] both shown to yield valid measurements of balance in elderly people (Berg Balance Test: subjects' mean age=83 years [SD=6.9]; Functional Reach Test: age range = 70-87 years). Good scores on balance scales are positively correlated with high levels of independent mobility in patients poststroke.[17] If the E-FAP scores, in turn, are correlated with scores on these tests of gait speed and balance, concurrent validity of the E-FAP measurements would be supported. Furthermore, unlike the other 2 tests, the E-FAP provides information to health care decision makers about how clients ambulate in a variety of environments. These data can be obtained repeatedly during interventions to detect the effectiveness of treatment and to help design home programs. This knowledge is also important in making decisions about the need for caregivers or altering a home environment or for assessing employment opportunities.

The reliability and validity of the E-FAP component measures should be assessed before the tool can be applied to assess the benefit of interventions to improve walking. Therefore, the research questions under consideration were: (1) Does the E-FAP differentiate between subjects who have had strokes and subjects without impairment? and (2) Is the E-FAP correlated with previously validated measures of walking speed and balance?

Method

Subjects

The 56 volunteer subjects tested were obtained by convenience sampling. Twenty-eight subjects without impairment (8 women, 20 men) were matched by age ([+ or -] 5 years), height ([+ or -] 15.24 cm), and gender with 28 subjects with a history of strokes. Subject characteristics are presented in Table 1. Eighteen subjects who had strokes required assistive devices: 6 subjects used canes, 2 subjects used ankle-foot orthoses (AFOs), 8 subjects used canes and AFOs, and 2 subjects used quad canes and AFOs. Ten subjects who had strokes did not use an assistive device. Twenty-seven subjects without impairment reported no falls; 1 subject without impairment reported a fall during the year preceding the study. Fifty-five subjects demonstrated intact joint position sense; joint position sense for 1 subject with a stroke secondary to aphasia was not tested.

Table 1. Frequency, Mean, Standard Deviation, and Range for Characteristics of Subjects With Strokes and Subjects Without Impairment

Characteristic/Group    n     [bar] X    SD         Range

Age (y)
  Without impairment     28    56.43     13.82       34-89
  With strokes           28    56.04     12.80       38-88
Height (cm)
  Without impairment     28   173.54      7.49   162.56-190.50
  With strokes           28   175.19      8.53   154.94-190.50
Months since onset of
  stroke                 28    13.59     12.30        1-51
Side of lesion
  Left                   13
  Right                  15

To meet inclusion criteria, subjects were required to perform the following tasks without another individual's assistance: (1) follow simple spoken Commands, (2) ascend and descend 5 stairs, (3) get in and out of a chair, (4) walk 10 m on a hard-surfaced floor, and (5) walk 10 m on a carpeted floor. Subjects with strokes could participate if they used an AFO, a cane, a quad cane, a hemiwalker, or any combination of these devices. No other type of assistive device, however, was allowed. The subjects without impairment had to have: (1) no more than 2 falls during the year preceding the study, (2) an absence of a serious medical diagnosis, including pulmonary, cardiac, neurologic, systemic, or musculo-skeletal problems or a history of stroke, (3) the ability to ambulate without an assistive device, and (4) an absence of assistance with activities of daily living.

All subjects participating in the study provided written informed consent. The personal physician of each subject with a stroke approved his or her participation in the study.

Measurements and Instrumentation

Emory Functional Ambulation Profile. Five subtasks comprise the E-FAP: (1) a 5-m walk on the hard-surfaced floor, (2) a 5-m walk on the carpeted floor, (3) performance of an "up & go" task, (4) negotiation of an obstacle course, and (5) ascent and descent of 4 stairs.[28] Subtasks were completed in the above sequence according to a standardized protocol (Appendix). The number of seconds taken to complete each subtask was recorded. The time to complete each subtask was multiplied by a factor corresponding to the level of assistive device used (Figure). Inclusion of the assistance factor as part of the E-FAP serves several purposes. First, this inclusion allows the score to reflect differences in the amount of assistance required by an individual. The assistance factor increases relative to the amount of support offered by the assistance device. Second, use of an assistance factor permits differentiation of individuals who are walking at the same speed but with different assistive devices. Moreover, the assistance factor can reflect changes in gait speed as patients progress from one device to another. Third, use of the assistance factor offers a clinically descriptive picture for other health care professionals to interpret the E-FAP. The 5 subtask scores were summed to yield an E-FAP total score. For each subject with a stroke, an E-FAP total score also was computed without the multiplication factor for assistive device. Thus, calculations were made for all subjects with strokes using both an assistance factor and no assistance factor.

[Figure ILLUSTRATION OMITTED]

Timed 10-Meter Walk Test. Each subject was instructed to walk at a comfortable, normal pace for 10 m. Only the middle 6 m, however, was timed to eliminate the effects of acceleration and deceleration.[29] Start and stop of performance time coincided with the toes of the leading foot crossing the 2-m mark and the 8-m mark, respectively. From these data, the speed was calculated by dividing the middle 6 m by the time (in seconds) required to walk the 6 m.[15]

Berg Balance Test. The Berg Balance Test was administered to each subject according to the standard protocol.[30] The Berg Balance Test consists of 14 tasks performed in the following sequence: (1) sit to stand, (2) stand unsupported, (3) sit with back unsupported, (4) stand to sit, (5) transfer, (6) stand with eyes closed, (7) stand with feet together, (8) reach forward with an outstretched arm, (9) retrieve object from floor, (10) turn to look behind, (11) turn 360 degrees, (12) place alternate foot on stool, (13) stand with one foot in front of the other foot, (14) stand on one foot. Performance of each task was scored on a scale from 0 to 4, with 0 representing minimal completion of the task and 4 representing full completion of the task according to test criteria. A total score of 56 represents perfect performance.

Functional Reach Test. The score on the Functional Reach Test was the distance (in centimeters) of the subject's reach as determined by the total excursion of the subject's third metacarpal in the nonhemiparetic arm of the subjects with strokes or the dominant arm of the subjects without impairment.[27] Each subject was instructed to flex the test arm forward to 90 degrees and then reach forward as far as possible without taking a step. One practice trial was performed followed by 3 separate test measurements. The average of the 3 measurements was the subject's final score. With the exception of an AFO, use of assistive devices for subjects with strokes was not allowed during this test.

Leg length was measured bilaterally (in centimeters) for each subject.[31] Each subject assumed a supine position on the treatment table with anterior superior iliac spines (ASISs) and medial malleoli exposed. Following palpation of the superior aspect of the ASIS, the tip of the tape measure was placed on this landmark. Next, the most prominent point on the medial malleoli was palpated, and the tape measure was extended to this landmark and leg length was then measured.

Each subject self-reported age and gender information, and this information was verified by inspection of legal identification (eg, driver's license). Information about time since onset of stroke and side of lesion was obtained by self-report or medical chart review. The height of each subject was measured by having each subject remove his or her shoes and stand upright against a wall marked in inches. Measurement was recorded in inches and later converted to centimeters.

Joint position sense was assessed bilaterally for each subject at the following joints: shoulder, hip, knee, and ankle.[31] Each subject was instructed to keep his or her eyes open while the extreme positions of "bent" and "straight" were demonstrated for each joint. Next, the subject was asked to close his or her eyes while the investigator grasped the limb and moved the joint passively into one of the extreme positions. The subject was asked to identify the position of the joint as bent or straight, and the investigator recorded joint position sense as "intact" if the subject correctly identified the position of the joint or "not intact" if the subject's response was incorrect.

Prior to test administration, blood pressure was assessed using a sphygmomanometer and stethoscope on the brachial artery.[31] Any subject with a resting blood pressure greater than 170/100 mm Hg was excluded from the study.[32] Pulse rate was measured (in beats per minute [bpm]) by palpation of the radial artery and timed with digital watches.[31] Any subject with a resting heart rate greater than 100 bpm was excluded from the study.[32] One subject was excluded from the study because of high blood pressure. A physician examined that subject. All other subjects met the blood pressure and heart rate criteria before or after the testing session.

Identical digital, manual stopwatches with readings to 1/100 of a second were used to measure performance time on all timed tests. Digital stopwatches were calibrated manually to 1/10 of a second prior to and throughout data collection. Interrater agreement was determined for leg length, joint position sense, and subject characteristics (ie, gender, age, height) by repeated observations of 2 investigators. To establish reliability between investigators prior to data collection, agreement between the investigators was obtained over 4 trials. Agreement was defined as exact for subject characteristics and less than or equal to one-half standard deviation of data reported in previous studies, as follows: Timed 10-Meter Walk Test= [+ or -] 2.9 seconds,[12] Berg Balance Test total score= [+ or -] 2.9,[30] Functional Reach Test= [+ or -] 1.9 in(*),[33] E-FAP total score= [+ or -] 2.9 seconds (based on Timed 10-Meter Walk Test data), leg length= [+ or -] 1.0 cm,[31] and joint position sense=exact. Two investigators scored each of the 4 tests concurrently throughout the study in order to assess interrater reliability at the conclusion of the study.

Procedure

Two of 4 investigators were randomly assigned to each subject for a data collection session. The investigators initially contacted and interviewed each subject by telephone or in person to assess qualification for the study. Each subject was asked to wear comfortable walking shoes to the 1-hour data collection session. At the start of the data collection session, each subject answered a questionnaire to ensure qualification. One investigator took preliminary measurements, including blood pressure, pulse rate, height, joint position sense, and leg length. The 4 tests (E-FAP, Timed 10-Meter Walk Test, Berg Balance Test, and Functional Reach Test) were administered in random order for each subject. For example, the first subject received the following sequence: the Berg Balance Test (subject completed all 14 items in sequence), the Timed 10-Meter Walk Test, the E-FAP (subject completed all 5 subtasks in sequence), and the Functional Reach Test. Each subject was offered a 2-minute rest period between tests, if needed. Otherwise, each subject progressed immediately to the next test. For each test, one investigator demonstrated the test and gave specific instructions to the subject. A second investigator guarded the subject during each task. Each investigator independently recorded performance data on a separate data collection form.

On completion of the data collection session, one investigator remeasured the subject's blood pressure and pulse rate. If blood pressure exceeded 170/100 mm Hg or the pulse rate exceeded 100 bpm, the subject was asked to remain seated for 5 minutes, and blood pressure and pulse rate were reassessed. If blood pressure or pulse rate did not descend, a physician was contacted. The subject was dismissed at the conclusion of the data collection session when blood pressure and pulse rate values were below 170/100 mm Hg and 100 bpm, respectively.

Data Analysis

The mean, standard deviation, and minimum and maximum values were determined for each test score, age, height, and leg length per group and for time since onset of stroke in the subjects with strokes. Gender in both groups and side of lesion in the subjects with strokes were summarized as frequency of occurrence. Measurements obtained by the primary researcher were used in all analyses except determination of interrater reliability.

All test scores and leg-length measurements were interval data, except scores from the Berg Balance Test, which were ordinal data. Normality of distribution and homogeneity of variance for each test score and leg-length measurement were tested using the Walled test and the Sphericity test, respectively. In the subjects without impairment, data meeting these assumptions included the E-FAP total score and scores for the floor, up & go, obstacles, and stairs subtasks. Their scores on the Functional Reach Test, the Timed 10-Meter Walk Test, and the leg-length measure also met these assumptions. Scores on the Functional Reach Test, the Timed 10-Meter Walk Test, and the leg-length measure met these assumptions in the subjects with strokes. If data did not meet an assumption, parametric test results are reported only when the nonparametric tests yielded identical results. Otherwise, nonparametric test results are identified.

The agreement between raters' scores on the E-FAP, the Timed 10-Meter Walk Test, the Berg Balance Test, the Functional Reach Test, each subtask of the E-FAP, and the leg-length measure were analyzed using a nested random-effects analysis of variance (ANOVA) to calculate the intraclass correlation coefficient (ICC[2,1]).[34] The E-FAP total score and scores on the Timed 10-Meter Walk Test, the Berg Balance Test, and the Functional Reach Test were compared between groups and tested using an unpaired t test per variable. The differences between times on subtasks of the E-FAP (floor, carpet, up & go, obstacles, stairs) were compared within each group using a Friedman ANOVA by ranks and a Tukey Honestly Significant Difference post hoc test. The difference in scores on each subtask of the E-FAP between groups was tested using a Wilcoxon rank sum test with a Bonferroni adjustment. The relationship between the E-FAP total score and scores on the Timed 10-Meter Walk Test, the Berg Balance Test, and the Functional Reach Test was tested using a Pearson product-moment correlation coefficient or a Spearman rank correlation. The correlation between average leg length and scores on the Timed 10-Meter Walk Test for each group and for both groups combined was determined using the Pearson product-moment correlation coefficient.

For all statistical tests, the alpha level was [is less than] .05 and the power was .90. Power was based on effect size of a 10% difference between subtasks and between groups. Effect size was based on previously collected E-FAP data from the Program in Restorative Neurology at the Emory Clinic, Atlanta, Ga, on subjects with strokes. To obtain a power of .90, data were collected on 28 subjects without impairment and 28 subjects with strokes.

Results

The E-FAP total scores demonstrated high interrater reliability in the subjects without impairment (ICC[2,1] =.997) and in the subjects with strokes (ICC[2,1] =.999) (Tab. 2). The distribution of ICC values for other tests among all subjects was high (ICC[2.1] =.980-1.000; Tab. 3).

Table 2. Frequency, Mean, Standard Deviation, Range, and Intraclass Correlation Coefficient (ICC[2,1]) Among Multiple Raters for Emory Functional Ambulation Profile Scores With Assistance Factor for Subjects With Strokes and Subjects Without Impairment

Test/Group             n    [bar] X     SD

Total
  Without impairment   28     33.35     4.93
  With strokes         28    313.71   369.02

Floor
  Without impairment   28      3.88     0.52
  With strokes         28     34.22    44.95

Carpet
  Without impairment   28      3.87     0.61
  With strokes         28     33.65    42.66

Up & go
  Without impairment   28      8.74     1.39
  With strokes         28     76.24    70.25

Obstacles
  Without impairment   28     10.64     1.79
  With strokes         28    123.31   181.42

Stairs
  Without impairment   28      6.22     1.02
  With strokes         28     46.12    62.90

Test/Group                Range         ICC

Total
  Without impairment   26.16-44.61     .997
  With strokes         35.96-1795.84   .999

Floor
  Without impairment    3.15-4.94      .977
  With strokes          4.16-231.40    .999

Carpet
  Without impairment    3.13-5.44      .964
  With strokes          3.90-198.25    .999

Up & go
  Without impairment    6.63-12.07     .990
  With strokes          8.50-311.90    .982

Obstacles
  Without impairment    8.09-14.44     .998
  With strokes          7.50-961.85    .999

Stairs
  Without impairment    4.81-8.35      .880
  With strokes          7.06-311.22    .999

Table 3. Mean, Standard Deviation, Range, Intraclass Correlation Coefficient Among Multiple Raters (ICC[2,1]), and Results of Independent t Tests Comparing Each Functional Test Score Between Subjects With Strokes (n=28) and Subjects Without Impairment (n=28)

Test/Group                                 [bar] X     SD

Emory Functional Ambulation Profile
  total score
  Without impairment                         33.35     4.93
  With strokes with AF(a)                   313.71   369.02
  With strokes without calculating AF(b)     97.48    70.46

Timed 10-Meter Walk Test (m/s)
  Without impairment(c)                       1.43     0.24
  With strokes                                0.77     0.35

Berg Balance Test
  Without impairment(c)                      55.36     1.45
  With strokes                               45.16     8.54

Functional Reach Test (cm)
  Without impairment(c)                      32.11     5.88
  With strokes                               21.92     6.57

Test/Group                                     Range         ICC

Emory Functional Ambulation Profile
  total score
  Without impairment                       26.16-44.61      .997
  With strokes with AF(a)                  35.96-1795.84    .999
  With strokes without calculating AF(b)   35.96-386.90

Timed 10-Meter Walk Test (m/s)
  Without impairment(c)                     1.00-1.86       .980
  With strokes                              0.13-1.48       .998

Berg Balance Test
  Without impairment(c)                    49.00-56.00     1.000
  With strokes                             19.00-56.00      .995

Functional Reach Test (cm)
  Without impairment(c)                    19.00-49.30      .992
  With strokes                              9.70-34.00      .995

(a) Independent t test, P < .0002; without impairment < with strokes. AF=assistance factor.

(b) Independent t test, P < .0000; without impairment < with strokes.

(c) Independent t test, P < .0000; without impairment < with strokes.

The E-FAP total scores for subjects with strokes were higher (ie, slower times), both with (P [is less than] .0002) and without (P [is less than] .0000) an assistance factor included in the scores, than E-FAP total scores for subjects without impairment (Tab. 3). The scores of the subjects with strokes were higher on each E-FAP subtask (P [is less than] .0001, Wilcoxon rank sum test with Bonferroni adjustment), both with and without an assistance factor, than the scores of the subjects without impairment. The E-FAP subtask scores varied within each group (P [is less than] .0001, Friedman 2-way ANOVA by ranks). Among the subjects without impairment, floor and carpet subtask scores were not different; however, floor and carpet subtask scores were lower than stairs subtask scores. Up & go subtask scores were lower than obstacles subtask scores for the subjects without impairment (Tabs. 2 and 4). The pattern of differences among subtasks was the same in the subjects with strokes as in the subjects without impairment, with the following exceptions: stairs subtask scores equaled floor subtask scores for the subjects with strokes with an assistance factor, and stairs subtask scores equaled up & go subtask scores for the subjects with strokes without an assistance factor.

Table 4. Significant Results of Tukey Honestly Significant Difference Post Hoc Test for Within-Group Differences on Subtasks for Subjects With Strokes and Subjects Without Impairment(a)

Subtask/Group                              Up & Go

Floor
  Without impairment                    Up & go > Floor
  With strokes with AF                  Up & go > Floor
  With strokes without calculating AF   Up & go > Floor

Carpet
  Without impairment                    Up & go > Carpet
  With strokes with AF                  Up & go > Carpet
  With strokes without calculating AF   Up & go > Carpet

Up & go
  Without impairment
  With strokes with AF
  With strokes without calculating AF

Obstacles
  Without impairment
  With strokes with AF
  With strokes without calculating AF

Subtask/Group                               Obstacles

Floor
  Without impairment                    Obstacles > floor
  With strokes with AF                  Obstacles > floor
  With strokes without calculating AF   Obstacles > floor

Carpet
  Without impairment                    Obstacles > carpet
  With strokes with AF                  Obstacles > carpet
  With strokes without calculating AF   Obstacles > carpet

Up & go
  Without impairment                    Obstacles > up & go
  With strokes with AF                  Obstacles > up & go
  With strokes without calculating AF   Obstacles > up & go

Obstacles
  Without impairment
  With strokes with AF
  With strokes without calculating AF

Subtask/Group                               Stairs

Floor
  Without impairment                    Stairs > floor
  With strokes with AF
  With strokes without calculating AF   Stairs > floor

Carpet
  Without impairment                    Stairs > carpet
  With strokes with AF                  Stairs > carpet
  With strokes without calculating AF   Stairs > carpet

Up & go
  Without impairment                    Stairs < up & go
  With strokes with AF                  Stairs < up & go
  With strokes without calculating AF

Obstacles
  Without impairment                    Stairs < obstacles
  With strokes with AF                  Stairs < obstacles
  With strokes without calculating AF   Stairs < obstacles

(a) Minimum within-group difference was .27 for subjects without impairment, .65 for subjects with strokes with assistance factor (AF), and .42 for subjects with strokes without calculating AF.

The gait speed of the subjects with strokes was slower (P [is less than] .0000, Tab. 3) than that of the subjects without impairment on the Timed 10-Meter Walk Test. Berg Balance Test scores were lower (P [is less than] .0000, Tab. 3) in subjects with strokes than in subjects without impairment, indicating poorer balance in subjects with strokes. The distance reached by the subjects with strokes was less (P [is less than] .0000, Tab. 3) than the distance reached by subjects without impairment on the Functional Reach Test.

The relationships among test scores are presented in Table 5. In the subjects with strokes, E-FAP scores with and without an assistance factor were negatively related to the Timed 10-Meter Walk Test scores (P [is less than] .0001) and the Berg Balance Test scores (P [is less than] .0009); that is, slow times on the E-FAP correlated with slow gait speeds and poor balance. The E-FAP scores were not related (P [is greater than] .05) to Functional Reach Test scores in the subjects with strokes. The Berg Balance Test scores were positively related to the Timed 10-Meter Walk Test scores (P [is less than] .0004, Tab. 5) and to the Functional Reach Test scores (P [is less than] .0004, Tab. 5) in the subjects with strokes. In the subjects without impairment, E-FAP scores were negatively related (P [is less than] .0001) to Timed 10-Meter Walk Test scores but were not related to Berg Balance Test scores or Functional Reach Test scores (P [is greater than] .05).

Table 5. Pearson Product-Moment Correlation Coefficient (r) Among All Four Functional Tests (Emory Functional Ambulation Profile [E-FAP], Timed 10-Meter Walk Test, Berg Balance Test, and Functional Reach Test) for Subjects With Strokes and Subjects Without Impairment

Relationship                                r        P

E-FAP total score with:
  Timed 10-Meter Walk Test
    Without impairment                    -.759   .0001(a)
    With strokes with AF(b)               -.708   .0001(a)
    With strokes without calculating AF   -.783   .0001(a)
  Berg Balance Test
    Without impairment(c)                 -.206   .2929
    With strokes with AF                  -.602   .0007(a)
    With strokes without calculating AF   -.592   .0009(a)
  Functional Reach Test
    Without impairment                    -.230   .2384
    With strokes with AF                  -.301   .1192
    With strokes without calculating AF   -.358   .0613
Timed 10-Meter Walk Test with:
  Berg Balance Test
    Without impairment(c)                  .052    .7946
    With strokes                           .627    .0004(a)
  Functional Reach Test
    Without impairment                     .307    .1123
    With strokes                           .349    .0690
Berg Balance Test with:
  Functional Reach Test
    Without impairment(c)                  .352    .0664
    With strokes                           .619    .0004(a)

(a) P [is less than or equal to] .05.

(b) AF = assistance factor.

(c) Spearman correlation coefficient.

Discussion

The E-FAP is a clinical tool designed to measure walking time in 5 environmental circumstances and to account for the use of assistive devices. The E-FAP can be quickly administered in less than 20 minutes and is inexpensive and quantitative. High interrater reliability results (ICC=.997) indicate that the E-FAP may be used by multiple raters with consistency in scores between raters. The ability of the E-FAP to differentiate subjects with strokes from subjects without impairment supports the construct validity of the E-FAP for subjects with strokes. The E-FAP scores' correlation with scores on the Timed 10-Meter Walk Test and the Berg Balance Test in subjects with strokes provides evidence for the concurrent validity of the E-FAP in these subjects.

Limitations of this study were the relatively small number of subjects and the inclusion of only subjects with strokes and subjects without impairment. In addition, all levels of mobility were not represented, because subjects were required to ambulate without the assistance of another person. In the subjects with strokes, however, nearly equal distribution of right- and left-sided lesions and the wide range of times since onset of stroke increased the representation of people with strokes.

The E-FAP may be used to differentiate between subjects based on differences in functional ambulation. The scores of the subjects with strokes on the other 3 tests were poorer than were the scores of the subjects without impairment, indicating function is decreased in people who have had a stroke. Gait deficiencies in the subjects with strokes may be indicated by dependency on assistive devices. Eighteen subjects with strokes required assistive devices for ambulation. The E-FAP scores were different between groups both with and without an assistance factor, suggesting the difference between scores was not obtained artificially by including an assistance factor in the subjects with strokes.

For purposes of clinical application, ordering of subtasks according to increasing time to complete each subtask may be desirable. As a result, subtasks that potentially place high demands on endurance are not encountered at the beginning of the test, and the maximum number of subtasks can be completed before fatigue limits further participation. In contrast, subtasks requiring the most time to complete may be ordered first, before fatigue is an issue. In either case, the order of subtasks according to increasing time may be of interest in future revision and development of the E-FAP. Subtask order difference between groups according to increasing time may suggest that subjects with strokes responded differently to subtask demands in comparison with subjects without impairment.

Concurrent validity of the E-FAP is supported by correlation of the E-FAP scores with scores on the Timed 10-Meter Walk Test and the Berg Balance Test. The Timed 10-Meter Walk Test scores and the E-FAP scores were correlated in both groups possibly because each test measures walking speed. However, the E-FAP measures gait speed in various environmental conditions versus solely across a standard, hard-surfaced floor. Gait speed has been correlated with other clinical tests of function, including the Fugl-Meyer Test, Barthel Index, and Berg Balance Test.[17] The Berg Balance Test and the E-FAP may correlate because both involve multiple tasks requiring balance, strength, and endurance. The Berg Balance Test has previously been correlated with level of independent mobility.[8] Therefore, because the E-FAP correlates with the Timed 10-Meter Walk Test and the Berg Balance Test, E-FAP performance also may reflect function.

The E-FAP and the Functional Reach Test did not correlate in either group because the Functional Reach Test measures a person's stability during performance of a single task, whereas the E-FAP is designed to incorporate multiple tasks that an individual might encounter during everyday functioning. Weiner et al[33] indicate the Functional Reach Test seems less influenced by strength and endurance and, instead, represents a more "pure" balance measure. This measure may be influenced by limitations in shoulder or intervertebral mobility. The E-FAP and the Berg Balance Test, however, each incorporate measures of balance, strength, and endurance during multiple functional activities.

In the subjects with strokes, the Functional Reach Test may have correlated with the Berg Balance Test because both tests measure an element of balance, which the E-FAP is not sensitive to in this specific group. The Functional Reach Test is one of the 14 tasks included in the Berg Balance Test; thus, a correlation between the 2 tests could be expected. The E-FAP, the Berg Balance Test, and the Functional Reach Test seem to be related based on relative dynamic challenges to balance. Although the Functional Reach Test involves a relatively less dynamic task, the Berg Balance Test incorporates a continuum of balance challenges ranging from less dynamic to more dynamic.

The E-FAP score correlation with scores of the other 3 tests in the group without impairment is of interest because these analyses may reveal information about the E-FAP's ability to detect differences in subjects with high levels of function, a consideration as patients' function improves. In individuals without impairment, the E-FAP may not correlate with the Berg Balance Test because scores on the Berg Balance Test are ordinal data, whereas E-FAP scores are interval data and, thus, are more sensitive. Richards et al[17] found that the Berg Balance Test is not as sensitive in subjects with faster walking speeds in comparison with slower walkers. Furthermore, Berg et al[30] suggest the utility of the Berg Balance Test is limited in active people with minimal balance deficits.

Clinical Implications

In the current health care environment, decisions regarding reimbursement for patient care often are made by a variety of individuals, many of whom have not had clinical training. Consequently, meaningful tests should be concise, efficient, simple, and easy to comprehend by a number of health care decision makers. They must be also interpreted easily by health care professionals and individuals authorized to decide on reimbursement for services. The E-FAP may meet these criteria. The E-FAP requires minimal equipment and can be administered fairly quickly. The test can be administered in most settings by clinicians or personnel under their supervision. If further research supports the predictive validity of the E-FAP, health care professionals can use the subtasks from the E-FAP to determine whether a patient is ready to return to home-based ambulatory activities by assessing data from the individual subtasks. In addition, improvements can be seen over different conditions when administered more than once.

Future Studies

The E-FAP measures may be clinically useful because reliability, construct validity, and concurrent validity were supported in this study for people with strokes. Additional studies, however, are necessary to establish the validity of the E-FAP. Currently, the E-FAP cannot be used to assess a patient requiring any manual assistance from another person. Therefore, future development of the E-FAP will consider methods to account for patients requiring manual assistance. The sensitivity of the E-FAP to patient progress in rehabilitation over time or over other surfaces or inclines also should be investigated. Finally, additional research studies are needed to establish the reliability and validity of E-FAP measures in other subject populations.

Conclusion

The E-FAP scores differentiated between subjects with strokes and subjects without impairment and correlated with previously validated measures of walking speed and balance (Timed 10-Meter Walk Test and Berg Balance Test) in the subjects with strokes. Therefore, the E-FAP may be provide reliable and valid clinical measurements of functional ambulation in individuals with strokes.

(*) 1 in=2.54 cm.

References

[1] Tinetti ME, Ginter SF. Identifying mobility dysfunctions in elderly patients: standard neuromuscular examination or direct assessment? JAMA. 1988;259:1190-1193.

[2] Eisenberg MG. Dictionary of Rehabilitation. New York, NY: Springer Publishing Co; 1995:103.

[3] Fox KM, Felsenthal G, Hebel JR, et al. A portable neuromuscular function assessment for studying recovery from hip fracture. Arch Phys Med Rehabil. 1996;77:171-176.

[4] Cohen JJ, Sween JD, Walker JM, Smith KB. Establishing criteria for community ambulation. Topics in Geriatric Rehabilitation. 1987;3:71-77.

[5] Robinett CS, Vondran MA. Functional ambulation velocity and distance requirements in rural and urban communities: a clinical report. Phys Ther. 1988;68:1371-1373.

[6] Means KM. The obstacle course: a tool for the assessment of functional balance and mobility in the elderly. J Rehabil Res Dev. 1996;33:413-429.

[7] Friedman PJ, Baskett JJ, Richmond DE. Cognitive impairment and its relationship to gait rehabilitation in the elderly. N Z Med J. 1989;102: 603-606.

[8] 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.

[9] Harada N, Chiu V, Damron-Rodriguez, et al. Screening for balance and mobility impairment in elderly individuals living in residential care facilities. Phys Ther. 1995;75:462-469.

[10] Tinetti ME, Speechley M. Prevention of falls among the elderly. N Engl J Med. 1989;320:1055-1059.

[11] Koch M, Gottschalk M, Baker, DI, et al. An impairment and disability assessment and treatment protocol for community-living elderly persons. Phys Ther. 1994;74:286-298.

[12] Potter JM, Evans AL, Duncan G. Gait speed and activities of daily living function in geriatric patients. Arch Phys Med Rehabil. 1995;76: 997-999.

[13] Peel C. Age-related changes in the cardiopulmonary system. In: Irwin S, Tecklin JS, eds. Cardiopulmonary Physical Therapy. St Louis, Mo: Mosby; 1995:292-307.

[14] Perry J, Garrett M, Gronley JK, Mulroy SJ. Classification of walking handicap in the stroke population. Stroke. 1995;26:982-989.

[15] Wade DT, Wood VA, Heller A, et al. Walking after stroke: measurement and recovery over the first 3 months. Scand J Rehabil Med. 1987;19:25-30.

[16] Salive ME, Guralnik J, Glynn RJ, et al. Association of visual impairment with mobility and physical function. J Am Geriatr Soc. 1994;42: 287-292.

[17] Richards CL, Malouin F, Wood-Dauphinee S, et al. Gait velocity as an outcome measure of locomotor recovery after stroke. In: Craik RA, Oatis C, eds. Gait Analysis: Theory and Applications. St Louis, Mo: Mosby; 1995:355-364.

[18] Bohannon RW. Gait performance of hemiparetic stroke patients: selected variables. Arch Phys Med Rehabil. 1987;68:777-781.

[19] Mathias S, Nayak USL, Isaacs B. Balance in elderly patients: the "get-up and go" test. Arch Phys Med Rehabil. 1986;67:387-389.

[20] Thorbahn LD, Newton RA. Use of the Berg Balance Test to predict falls in elderly persons. Phys Ther. 1996;76:576-585.

[21] Shumway-Cook A, Baldwin M, Polissar NL, Gruber W. Predicting the probability for falls in community-dwelling older adults. Phys Ther. 1997;77:812-819.

[22] Wolfson L, Whipple R, Amerman P, Tobin JN. Gait assessment in the elderly: a gait abnormality rating scale and its relation to falls. J Gerontol. 1990;45:M12-M19.

[23] Friedman PJ, Richmond DE, Baskett JJ. A prospective trial of serial gait speed as a measure of rehabilitation in the elderly. Age Ageing. 1988;17:227-235.

[24] Tinetti ME. Performance-oriented assessment of mobility problems in elderly patients. J Am Geriatr Soc. 1986;34:119-126.

[25] Robinson JL, Smidt GL. Quantitative gait evaluation in the clinic. Phys Ther. 1981;61:351-353.

[26] Nelson AJ. Functional Ambulation Profile. Phys Ther. 1974;54: 1059-1065.

[27] Duncan PW, Weiner DK, Chandler JM, Studenski SA. Functional reach: a new clinical measure of balance. J Gerontol. 1990;45: M192-M197.

[28] Wolf SL. A method for quantifying ambulatory activities. Phys Ther. 1979;59:767-768.

[29] Bohannon RW, Andrews AW, Thomas MW. Walking speed: reference values and correlates for older adults. J Orthop Sports Phys Ther. 1996;24:86-90.

[30] Berg KO, Wood-Dauphinee S, Williams JI, Gayton D. Measuring balance in the elderly: preliminary development of an instrument. Physiotherapy Canada. 1989;41:304-311.

[31] Catlin P, Barrett S, Binder-Macleod S, et al. Competencies and Criteria for the Physical Therapist. Atlanta, Ga: Division of Physical Therapy, Department of Rehabilitation Medicine, Emory University School of Medicine; 1983.

[32] Irwin S, Blessey RL. Patient evaluation. In: Irwin S, Tecklin JS, eds. Cardiopulmonary Physical Therapy. St Louis, Mo: Mosby; 1985:64-102.

[33] Weiner DK, Duncan PW, Chandler J, Studenski SA. Functional reach: a marker of physical frailty. JAm Geriatr Soc. 1992;40:203-207.

[34] Portney LG, Watkins MP. Foundations of Clinical Research: Applications to Practice. East Norwalk, Conn: Appleton & Lange; 1993.

Appendix.

Emory Functional Ambulation Profile (E-FAP) Protocol

The E-FAP is composed of 5 subtasks: (1) Floor, (2)Carpet, (3) Up & Go, (4) Obstacles, and (5) Stairs. Each subject is given a rest period between performances of the subtasks long enough for the researcher to explain and demonstrate the next component. Each subject poststroke is instructed to use an assistive device as needed and to wear a gait belt during performance of all subtasks. The researcher designated as primary researcher demonstrates, provides instructions, and answers the subject's questions. Primary researcher and secondary researcher each record performance times for all 5 subtasks on separate data collection forms. Upon completion of the entire data collection session, each researcher calculates a total E-FAP score.

Introduction

Primary researcher provides an explanatory overview of the 5 subtasks comprising the E-FAP. Prior to performance of each subtask, primary researcher explains and demonstrates the subtask. Subject is informed that performance of each subtask is timed and is instructed to ask clarification questions at any time.

Floor

Setup: A 1-m strip of masking tape is placed on the hard-surfaced floor at the starting point. Five meters ahead of the starting point, a 2-cm piece of masking tape marks the end point. A small piece of tape is used to mark the end point so that subjects do not decelerate in anticipation of the finish line.

1. Primary researcher explains while demonstrating the Floor subtask: "When I say `go,' walk at your normal, comfortable pace until I say `stop.'"

2. Primary researcher assists subject as needed in placing toes on starting line tape.

3. Primary researcher says "go," and primary and secondary researchers simultaneously press stopwatches to begin timing.

4. Subject walks toward primary researcher, who is standing 1 m beyond the end point of the 5-m distance. Secondary researcher walks alongside the subject as the subject traverses the 5-m distance.

5. Primary and secondary researchers simultaneously press stopwatches to stop as subject's lead foot crosses the end point. Primary researcher tells subject to stop when he or she is beyond the end point.

6. Primary and secondary researchers record times on separate data collection forms.

Carpet

Setup: A piece of short pile carpet, no less than 7 m long and 2 m wide, is taped securely to the floor. Starting point is marked with a 1-m strip of masking tape. End point is marked exactly 5 m from the starting point with a 2-cm piece of masking tape. Both starting point and end point are at least 1 m from the edge of the carpet.

1. Primary researcher explains while demonstrating the Carpet subtask: "When I say `go,' walk at your normal, comfortable pace until I say `stop.'"

2. Primary researcher assists subject as needed in placing toes on starting line tape.

3. Primary researcher says "go," and primary and secondary researchers simultaneously press stopwatches to begin timing.

4. Subject walks toward primary researcher who is standing 1 m beyond the end point of the 5-m distance. Secondary researcher walks alongside the subject as the subject traverses the 5-m distance.

5. Primary and secondary researchers simultaneously press stopwatches to stop timing as subject's lead foot crosses the end point. Primary researcher tells the subject to stop when he or she is beyond the end point.

6. Primary and secondary researchers record times on separate data collection forms.

Up & Go

Setup: Standard armchair with a 46..cm seat height is placed on the hard-surfaced floor. Three meters away, a 1-m strip of black tape is placed on the floor.

1. Primary researcher explains while demonstrating the Up & Go subtask: "Next, you will sit in this chair with your back against the back of the chair and your arms resting on the armrests. When I say `go,' you will stand up from the chair, walk at your normal comfortable pace past this line where I will be standing, turn around, walk back to the chair, and sit down, making sure your back is against the back of the chair."

2. Subject assumes sitting position in the chair. Primary researcher stands at the 3-m point marked with masking tape. Secondary researcher stands beside the chair and prepares to walk with the subject.

3. Primary researcher says "go," and primary and secondary researchers simultaneously press stopwatches to begin timing.

4. Primary researcher monitors line to ensure both of subject's feet cross the line before turning around.

5. Primary and secondary researchers stop timing when subject is fully seated with back against the chair.

6. Primary and secondary researchers record times on separate data collection forms.

Obstacles

Setup: A 1-m piece of masking tape is placed on a hard-surfaced floor to mark the starting point. A brick is placed on the floor at the 1 1/2-m mark and the 3-m mark. A 40-gal rubber trash can is placed at the 5-m mark.

1. Primary researcher explains while demonstrating the Obstacles subtask: "When I say `go,' walk forward at your normal, comfortable pace and step over each brick. Then, walk around the trash can from either the left or right. Then walk back stepping over the bricks again. Continue walking until I say `stop.'"

2. Primary researcher assists subject as needed in placing toes on starting line.

3. Primary researcher says "go," and primary and secondary researchers simultaneously press stopwatches to begin timing.

4. When subject begins walking, primary researcher steps back 1 m beyond the end line while secondary researcher walks with subject.

5. When subject's foot crosses the end line, primary and secondary researchers simultaneously press stopwatches to stop timing. Primary researcher tells the subject to "stop" when he or she is beyond the end line.

6. Primary and secondary researchers record times on separate data collection forms.

Stairs

Setup: Stairs with 4 steps, hand railings, and the following measurements are utilized: 26.04-cm stair depth, 75.57-cm stair width, 15.24-cm stair height, 76.20-cm platform depth, and 75.57-cm platform width. A 1-m piece of masking tape is placed 25 cm from the base of the first step.

1. Primary researcher explains while demonstrating the Stairs subtask: "When I say `go,' walk up the stairs at your normal, comfortable pace to the top of the stairs, turn around, and come back down. You may use the handrails if needed. I will follow behind you for safety."

2. Primary researcher assists subject as needed in placing toes on starting tape.

3. Primary researcher says "go," and primary and secondary researchers simultaneously press stopwatches to begin timing.

4. Primary researcher follows subject up stairs to guard.

5. Primary and secondary researchers press stopwatches to stop timing when subject's nonlead foot comes in firm contact with the floor.

6. Primary and secondary researchers record times on separate data collection forms.

Scoring the Emory Functional Ambulation Profile

1. Researchers multiply time recorded for each subtask by appropriate factor according to level of assistive device used during that subtask.

2. Researchers record the product in the cell corresponding to appropriate subtask and level of assistive device.

3. Researchers repeat this procedure for each column/subtask.

4. Researchers sum the 5 subtask scores to obtain the E-FAP total score.

5. All total scores also are computed without the factor for assistance device for purposes of statistical testing.

SL Wolf, PT, PhD, FAPTA, is Director, Division of Research, Professor, Department of Rehabilitation Medicine, and Associate Professor, Department of Anatomy and Cell Biology, Emory School of Medicine, Atlanta, Ga. Address all correspondence to Dr Wolf at Center for Rehabilitation Medicine, 1441 Clifton Rd NE, Atlanta, GA 30322 (USA) ([e-mail]).

PA Catlin, PT, EdD, is Professor and Director, Division of Physical Therapy, Department of Rehabilitation Medicine, Emory University School of Medicine.

K Gage, K Gurucharri, R Robertson, and K Stephen were students in the Master of Physical Therapy Program, Division of Physical Therapy, Department of Rehabilitation Medicine, Emory University School of Medicine, at the time of the study. This study was completed in partial fulfillment of their degree requirements.

The Emory Program in Restorative Neurology (PROREN) provided subjects and resources. The Emory University Center for Rehabilitation Medicine physicians and physical therapy staff, Heather Baer, MD, and George Cotsonis, MA, also provided assistance.

This study was approved by the Human Investigations Committee at the Emory University School of Medicine.

This article was submitted November 9, 1998, and was accepted July 20, 1999.