Volume 88, Issue 7 , Pages 896-900, July 2007
Activity Levels Among Lower-Limb Amputees: Self-Report Versus Step Activity Monitor
Article Outline
Abstract
Stepien JM, Cavenett S, Taylor L, Crotty M. Activity levels among lower-limb amputees: self-report versus step activity monitor.
Objective
To determine the accuracy of self-reported activity by community-dwelling, lower-limb amputees.
Design
Descriptive study.
Setting
A regional prosthetics outpatient service.
Participants
Seventy-seven unilateral lower-limb amputees at least 6 months after prosthetic rehabilitation.
Interventions
Not applicable.
Main Outcome Measures
Measured activity counts (in steps/min) and self-reported activity (rest, low, medium, high) in 15-minute intervals over 1 week were recorded for each participant.
Results
Participants averaged 3063±1893 steps per day. Strong agreement (γ≥0.7) between self-reported and measured activity was found for only 34% of participants between the hours of 9:00 am to 9:00 pm. The measured and self-reported proportion of time spent in various states of activity also showed poor agreement (rest, r=.41; low level activity, r=.39; medium level, r=.26; high level, r=.40). There was no bias toward either over- or under-reporting.
Conclusions
The majority of participants were unable to accurately self-report their activity levels (sleep excluded) as compared with measured activity levels. This may have important implications for prescribing appropriate prosthetics and for clinicians who provide patients with advice on promoting health.
Key Words: Amputees, Physical effort, Rehabilitation, Validation studies
PROMOTING ACTIVITY AND FITNESS is an important component of clinical encounters with amputees inasmuch as they often have significant comorbidities and face challenges in conventional exercise approaches. Inappropriate prescription of a prosthesis significantly affects an amputee’s comfort and mobility,1 and also has financial implications for funding agencies. Current practice is to use scales such as Medicare k-levels and Otto Bock Mobis,2 both of which take into consideration daily ambulatory activity levels and a patient’s weight to guide the prescription of an appropriate prosthesis. Clinically, activity levels are commonly determined from patients’ self-reports and evidence on the reliability of these reports in community-dwelling amputees is lacking. The prescription of inappropriate prostheses may have an impact on activity and therefore make it difficult for clinicians to compel amputees to participate in interventions that promote increased activity.
Previous research involving other adult populations has shown that validating measured ambulatory activity against self-reported ambulatory activity leads to conflicting results, including strong positive correlations,3, 4 poor-to-moderate positive correlations,5, 6, 7 or overestimation of activity frequency and intensity.8, 9 Recall limitations are also associated with self-report measures. Measured ambulatory function in these studies has been quantified by several instruments, including pedometers,3, 4 accelerometers,5, 8 and heart rate monitors.9
To determine daily activity in the amputee population, and particularly concerning prosthetic use, a device that can be attached to the prosthesis should yield high quality data. A step activity monitor (SAM) can be attached to a prosthetic limb to record the activity of that limb only; it is a combination of an accelerometer and step counter and is therefore better able than pedometers to detect movement by people with different gait patterns.10 The StepWatch3a Activity Monitor has been shown to accurately record the number of steps taken in normal walking and in climbing or descending stairs.10 StepWatch3 has also been shown to produce significantly less absolute errors in steps taken than do pedometers11 and can record steps per minute, therefore determining intensity of activity at any given time. Previous studies10 have shown that StepWatch3 has an overall accuracy of 99.7% when used in the lower-limb amputee population. As a self-report measure, an activity diary is most likely the best method with which to compare self-reported levels of activity to SAM.
Therefore, our goals in this study were to quantify the number of steps lower-limb amputees take per day, and to determine whether they accurately self-reported their daily activity levels using an activity diary when those levels were compared with their levels as measured by the SAM.
Methods
Participants
Participants were recruited from a regional prosthetic service between May and October 2005 through a letter of invitation sent to eligible participants. Telephone calls were then made to confirm that they received the letter, that they still fulfilled the study criteria, and that they wanted to participate.
Eligible participants were identified from a patient register, using the following criteria: unilateral lower-limb amputation; residing in unsupported care and using their prosthesis for primary ambulation (with or without the use of gait aids), a minimum of 6 months post-prosthetic rehabilitation, older than 18 years, and cognitively capable of completing questionnaires and maintaining a self-report diary of activity.
Study Procedure
During their first visit, participants were administered the Mini-Mental State Examination (MMSE)12 to determine their cognitive status (a score of ≥25 met the inclusion criteria). Participant characteristics, including age, sex, stump length, reason for amputation, and years since amputation, were recorded. Functional ability while wearing the prosthesis was determined with the Locomotor Capabilities Index (LCI).13
Participants were instructed in the use of the activity diary, which consisted of a table for each day of the week, with rows corresponding to time in 15-minute increments and the columns corresponding to 4 defined levels of activity. Those 4 predetermined levels were defined per leg as follows: resting (no steps taken), low (1–15 steps per minute), medium (16–40 steps per minute), and high (40+ steps per minute). The manufacturer of the SAM defined these activity levels from a normative data collection.
The StepWatch3 Activity Monitor was fitted to the participant’s prosthesis. The SAM was programmed to record 8 days of activity to ensure that 6 days of complete data were collected for each participant.
Participants returned the SAM and diary on their second visit.
Statistical Analysis
We analyzed all data for the group as a whole and then separately for the subjects with transtibial and transfemoral amputations. All demographic data were reported as mean ± standard deviation (SD). The characteristics of patients who declined to participate versus those of the study participants were compared using t tests for continuous data and chi-square tests of association for categorical data.
Activity levels for each 15-minute period recorded in both the diary and by the SAM were analyzed as one 6-day period and compared for agreement using the γ statistic.14 The γ statistic is a measure of concordance of observations for ordinal data. It takes on a positive value if the number of concordant pairs of observations is greater than the number of discordant pairs, a negative value if the number of discordant pairs is greater than the number of concordant pairs, and zero if the number of concordant pairs equals the number of discordant pairs. In this study, we accepted γ statistics greater than 0.7 as representing a strong concordance, between 0.5 and 0.7 as indicative of moderate concordance, and less than 0.5 as representative of poor agreement.
Comparisons between self-report and measured activity were made in 3 ways. First, γ statistics were calculated using all available data for each participant. Second, because it was likely that it would be easier to estimate activity during sleep rather than when a subject was awake, and this would therefore affect the rate of agreement between self-report and measured activity, we considered only the period between 9:00 am and 9:00 pm on each day of data collection, when the majority of participants were likely to be active. Histograms summarized the distribution of γ statistics for the study sample. Finally, to remove the effect of any time calibration discrepancies between the SAM and a participant’s self-report, the number of measured and self-reported periods in each activity state was also calculated for 9:00 am to 9:00 pm on the 6 full days of data collection. The Pearson correlation coefficient was calculated for self-report and measured activity for each activity level. The self-reported activity was plotted against measured activity levels and visually compared with a line of perfect agreement. Missing diary data points were imputed as rest.
Approval for this study was obtained from the Repatriation General Hospital Research and Ethics Committee.
Results
Seventy-seven unilateral lower-limb amputees completed the study. The 59 subjects who declined to participate were similar to participants in age, sex, amputation length, cause of amputation, and time since amputation. Among all eligible subjects, the rates of consent to participate differed for each of the causes for amputation: trauma, 61%; vascular, 49%; cancer, 85%; congenital, 43%; and infection, 20%. All participants wore the SAM and all completed the diaries for the 6 complete days of the study period.
Participant characteristics are shown in table 1. Men were in the majority of the sample (78%), the most common amputation length was transtibial, and approximately one half of the participants had their amputations because of a traumatic incident. Transtibial amputees had been amputees for significantly less time than transfemoral amputees and took significantly more steps per day, but there were no other differences between the 2 groups. The sample had good locomotors capabilities, as measured by the LCI, for both the basic (20.6±1.7) and advanced (19.3±3.6) scores.
Table 1. Patient Characteristics
| Characteristics | Transtibial (n=54) | Transfemoral (n=23) | Total (N=77) |
|---|---|---|---|
| Age (y) | 62±16 | 58±13 | 60±15 |
| Sex (male/female) | 44/10 | 16/7 | 60/17 |
| Reason for amputation, n (%) | |||
 Trauma | 28(51.9) | 11(47.8) | 39(50.6) |
| Vascular | 20(37.0) | 3(13.0) | 23(29.9) |
| Other | 6(11.1) | 9(39.1) | 15(19.5) |
| Years since amputation⁎ | 21±20 | 30±16 | 24±19 |
| MMSE score | 28.3±1.6 | 28.2±1.6 | 28.3±1.6 |
| LCI | |||
| Basic score | 20.4±2.0 | 21.0±0.0 | 20.6±1.7 |
| Advanced score | 19.0±3.9 | 20.0±2.8 | 19.3±3.6 |
| Total score | 39.5±5.4 | 41.0±2.8 | 40.0±4.8 |
| No. of steps per day⁎ | 3395±1965 | 2284±1472 | 3063±1893 |
⁎P<.05 (transtibial vs transfemoral). |
Figure 1 shows the frequency of activity measured by the diary and SAM at each specified activity level. Figure 2A shows the distribution of the γ statistic for the 24-hour period for the entire group. The γ statistic for 87% of the respondents was 0.7 or higher, suggesting strong agreement between the self-reported and measured activity for each 15-minute period. Only 3 participants reported a γ statistic of less than 0.5, which is indicative of poor agreement. When analyzed separately, both transtibial and transfemoral amputees showed strong agreement (85% vs 91%).
Fig 1.
Frequency of activity level measured by self-report diary and SAM for activity measured over the (A) 24-hour period and (B) between 9:00 am and 9:00 pm.
Fig 2.
Distribution of γ statistics for the (A) entire 24-hour period and (B) activity measured between 9:00 am and 9:00 pm.
When the γ statistics were recalculated for the period when the majority of participants would be awake (9:00 am to 9:00 pm), there was less agreement between self-reported and measured activity. The γ statistic for 34% of the participants was 0.7 or higher, and for 34% of the participants the γ statistic was less than 0.5 (fig 2B). Transtibial amputees had low agreement, with only 24% of participants having a γ statistic 0.7 or higher while 43% had less than 0.5; transfemoral amputees showed much better agreement, with 87% showing strong or moderate concordance.
Examination of agreement of level of activity between self-report and measured levels for all participants was greater for the 24-hour period (73%) than between 9:00 am and 9:00 pm (60%). As shown in table 2, a larger proportion of the sample under-reported activity in both the 24-hour period (16%) and the 9:00 am to 9:00 pm period (23%), compared with over-reports of activity (11% in 24-h period, 18% in the 9:00 am to 9:00 pm period). There was a similar measure of level of agreement when transtibial and transfemoral amputees were examined separately (71% vs 79% for the 24-h period, 57% vs 67% between 9:00 am and 9:00 pm).
Table 2. Agreement of Level of Activity for 15-Minute Periods Reported in Diary When Compared With SAM
| Reporting Period | Diary | |||
|---|---|---|---|---|
| Rest | Low | Medium | High | |
| 24-Hour Period | ||||
| StepWatch | ||||
| Rest | 25529 | 3578 | 285 | 1 |
| Low | 7476 | 11456 | 1507 | 60 |
| Medium | 139 | 446 | 344 | 57 |
| High | 9 | 11 | 38 | 62 |
| 9:00 am to 9:00 pm | ||||
| StepWatch | ||||
| Rest | 5322 | 2783 | 252 | 1 |
| Low | 5174 | 9486 | 1361 | 51 |
| Medium | 109 | 395 | 295 | 40 |
| High | 7 | 7 | 16 | 44 |
When time of activity was discounted, there was still little agreement between self-reported and measured activity when compared for each participant for the days on which complete data were recorded between 9:00 am and 9:00 pm (fig 3). The correlations between self-reported and measured activity were generally low overall for each level of activity (rest, r=.41; low, r=.39; medium, r=.26; high, r=.40). Transtibial amputees had results similar to those of the overall group (rest, r=.35; low, r=.30; medium, r=.26; high, r=.55), while transfemoral amputees showed higher correlations for rest (r=.51) and low (r=.60), but poor correlations for medium (r=−.03) and high (r=.34) levels.
Fig 3.
Correlation of comparisons between self-report and measured activity levels for (A) rest, (B) low level, (C) medium level, and (D) high level. Points on the graphs represent the number of observations made for each activity level for each participant with the StepWatch and diary. Line on graphs represents the line of perfect agreement (r=1.0).
Discussion
In this study we have shown that amputees do not reliably self-report their daily activity. We did not find any systematic bias toward either over- or under-reporting in the entire group, or when considering transtibial and transfemoral amputees separately. It does appear that transfemoral amputees are better at determining their activity levels than are transtibial amputees and this may be attributed to the fact that transfemoral amputees took significantly less steps than did transtibial amputees and these levels of activity are comparable to previous work.15
Little information is available on the walking patterns of amputees, which makes it difficult for clinicians to identify subtle declines that may require intervention. In this study we found that on average, amputees took about 6000 steps per day (SAM counts × 2 for comparison to pedometry), although there was a wide range measured and transtibial amputees took significantly more steps than did transfemoral amputees. These results are comparable with those of other studies that used the SAM in lower-limb amputee15 and stroke16, 17 populations. These levels fall well below the 10,000 steps per day recommended to the healthy adult population to maintain a healthy lifestyle.18 It is not known if this level of activity is required of the lower-limb amputee population to maintain good health. It has been shown that at least one fifth of chronic stroke victims have a decline in mobility from between 1 to 3 years after they leave the rehabilitation setting, with inactivity being shown to be a risk factor.19 It seems likely that amputees may also experience a decline in mobility because of inactivity.
The StepWatch Activity Monitor is a step activity monitor designed specifically for the amputee population; it has the advantage of being able to record steps per minute and therefore determine activity intensity and not just the number of steps taken per day, as recorded by a pedometer. The ability to determine an amputee’s activity intensity and not just the number of steps taken per day, allowed us to measure the level of perception our subjects had about their activity intensity, something a pedometer cannot measure. This feature also gives clinicians greater insight into the prosthetic needs of the patients. Participants’ compliance in using the SAM was high, with all them wearing it for the entire study period. Although there is an initial expense in purchasing SAM units and docking stations, their day-to-day use is inexpensive. Only a little time is needed to fix the monitor to the patient. Various self-report measures, such as diaries,3, 4 quality of life (QOL) surveys,7 or specific self-report measures of activity5, 8 have been used to estimate the amount of ambulatory activity of a variety of populations on a daily basis. Self-report of ambulatory activity may be influenced by a social desirability bias that can lead to over-reporting of daily activity.20
The implications of this can affect individual amputees, as well as funding agencies such as insurance companies. If daily activity is under-reported, patients may not be prescribed the prostheses that will let them ambulate most effectively; this could potentially affect their QOL. If, however, people over-report daily activity, then funding bodies or the people themselves are potentially paying for prostheses that are unnecessary and do not provide any improvement in ambulation or QOL over less complicated and potentially less expensive prostheses. This ability to accurately determine activity, along with several other factors—including vocational and recreational needs and exposure to hilly terrain—is important to consider in determining the appropriate prostheses.
Study Limitations
This study must be interpreted with several caveats in mind. We recruited a convenience sample from a list of existing clients rather than stratifying our recruiting to randomly select a representative proportion of amputation lengths, causes, or ages. While we were able to show that subjects who declined to join the study were similar to those who did with regard to age, sex, and years since their amputation, there were differences in consent rates according to the reason for the amputation; the overall consent to participate rate of 57% was low. Also, the calculation of number of steps and the inability to determine the terrain covered while taking those steps meant that we were unable to directly compare the amount of activity by 1 subject with that of another, because people of different heights or health status may have differing stride lengths. Consequently, they take a different number of steps in traveling the same distance. The keeping of a diary to determine self-reported activity is not a method used clinically, therefore our results cannot be directly compared with the information given to clinicians by amputees when a decision about a suitable prosthesis is being made.
Conclusions
This study has shown that self-reported activity by lower-limb amputees in an experimental setting is not accurate and is therefore unlikely to be accurate in a clinical setting. This suggests that an objective measure of daily activity is more appropriate when prosthetic devices are being prescribed for individual amputees; funding agencies also benefit from such measures. The StepWatch Activity Monitor, which has been specifically designed for use with lower-limb amputees, may be an appropriate objective measure of daily activity in this population.
Supplier
Acknowledgments
We thank Jackie O’Connor, BPO, for assistance in conceptualizing the study, and Lynne Giles, MPH, for her statistical assistance.
References
- . Mechanical properties of prosthetic limbs: adapting to the patient. J Rehabil Res Dev. 2001;38:299–307
- . Indications for Otto Bock microprocessor knees. 2004;Available at: http://www.ottobockus.com/assets/documents/C-leg_Compact_Indications.pdf. Accessed March 8, 2007
- . Measuring physical activity during pregnancy. West J Nurs Res. 2005;27:722–734
- . Self-reported physical activity validated by a pedometer: a pilot study. Public Health Nurs. 2006;23:88–94
- . Development and validation of a pregnancy physical activity questionnaire. Med Sci Sports Exerc. 2004;36:1750–1760
- . Failure to generalize determinants of self-reported physical activity to a motion sensor. Med Sci Sports Exerc. 1992;24:904–910
- . Ambulatory activity in men with diabetes: relationships between self-reported and real-world performance-based measures. J Rehabil Res Dev. 2004;41:571–579
- . Validity and reliability of a physical activity recall instrument among overweight and non-overweight men and women. J Sci Med Sport. 2003;6:477–491
- . Seven-day activity and self-report compared to a direct measure of physical activity. Am J Epidemiol. 1984;120:818–824
- . Step activity monitor: long-term, continuous recording of ambulatory function. J Rehabil Res Dev. 1999;36:8–18
- . Step activity monitor: increased accuracy in quantifying ambulatory activity. J Orthop Res. 1999;17:703–708
- . Mini-mental state (A practical method for grading the cognitive state of patients for the clinician). J Psychiatr Res. 1975;12:189–198
- . Prosthetic profile of people with lower extremity amputation: conception and design of a follow-up questionnaire. Arch Phys Med Rehabil. 1993;74:862–870
- . Statistical methods for rates and proportions. 2nd ed. New York: Wiley-Interscience; 1981;
- . Prosthetic intervention effects on activity of lower-extremity amputees. Arch Phys Med Rehabil. 2006;87:717–722
- . Accelerometer monitoring of home- and community based–ambulatory activity after stroke. Arch Phys Med Rehabil. 2004;85:1997–2001
- . Steps after stroke (Capturing ambulatory recovery). Stroke. 2005;36:1305–1307
- . How many steps/day are enough? (Preliminary pedometer indices for public health). Sports Med. 2004;34(1):1–8
- . Susceptibility to deterioration of mobility long-term after stroke. Stroke. 2006;37:167–171
- Improving question wording in surveys of culturally diverse populations. Ann Epidemiol. 1997;7:334–342
- a Cyma Corp, 6405 218th St SW, Ste 100, Mountlake Terrace, WA 98043-2180.
No commercial party having a direct financial interest in the results of the research supporting this article has or will confer a benefit upon the author(s) or upon any organization with which the author(s) is/are associated.Reprints are not available from the author.
PII: S0003-9993(07)00225-0
doi:10.1016/j.apmr.2007.03.016
© 2007 American Congress of Rehabilitation Medicine and the American Academy of Physical Medicine and Rehabilitation. Published by Elsevier Inc. All rights reserved.
Volume 88, Issue 7 , Pages 896-900, July 2007
