| | The Stroke Upper-Limb Activity Monitor: Its Sensitivity to Measure Hemiplegic Upper-Limb Activity During Daily LifeAbstract de Niet M, Bussmann JB, Ribbers GM, Stam HJ. The Stroke Upper-Limb Activity Monitor: its sensitivity to measure hemiplegic upper-limb activity during daily life. ObjectiveTo test the Stroke Upper-Limb Activity Monitor (Stroke-ULAM), which uses electrogoniometry and accelerometry to measure the amount of upper-limb usage in stroke patients in daily life conditions, for its sensitivity to discriminate between moderately recovered and well-recovered stroke patients and control subjects. DesignCross-sectional study. SettingAt home or a rehabilitation center. ParticipantsSeventeen patients with stroke and 5 control subjects. InterventionsNot applicable. Main Outcome MeasureLevel of usage of upper limb and the percentage of affected upper-limb activity compared with unaffected upper-limb activity (proportion). ResultsThe level of usage of the affected upper limb of stroke patients was lower than that of the nondominant upper limb of control subjects (electrogoniometry, 97.8°±92.3°/min vs 286.2°±46.5°/min, P<.01; accelerometry 1.0±0.5g/min vs 2.4±0.8g/min, P<.01). Stroke patients had lower proportions than control subjects in both electrogoniometry (22.6%±18.0% vs 84.6%±9.8%, P<.01) and accelerometry (39.2%±21.4% vs 93.3%±5.0%, P<.01). Well-recovered stroke patients had significantly higher proportions compared with moderately recovered patients on both electrogoniometry and accelerometry. ConclusionsThe Stroke-ULAM sensitively measures actual performance, and therefore can be a valuable addition to the mostly capacity-oriented tools currently used to evaluate upper-limb function. Proportion is preferred to the level of usage. STROKE IS THE MAIN CAUSE of long-term disability in Western society.1 The majority of stroke cases involve infarctions in the middle cerebral artery region and about 40% of stroke survivors have partial or total loss of function of the hemiplegic upper limb.2 The International Classification of Functioning, Disability and Health3 distinguishes 3 levels of human functioning: body functions and structures, activity, and participation. The activity level is divided into capacity (a person’s ability to execute a task or an action) and performance (what a person actually does in his/her current environment). Although rehabilitation interventions can focus on capacity and/or performance, the eventual aim is to improve performance. Nevertheless, studies related to upper-limb functioning after stroke focus more on measurement of capacity than on measurement of performance. Although related to each other, performance and capacity have to be regarded as different constructs.3, 4, 5, 6 The concept of learned nonuse is a good illustration of the difference between capacity and performance: people with learned nonuse do not involve their hemiplegic upper limb in daily activities to the full extent of their capacity to do so.7, 8 In evaluating the efficacy of subacute and chronic stroke rehabilitation, upper-limb function should be assessed at the level of performance as well as capacity. Many objective evaluation tools have been developed to test upper-limb capacity after stroke.9, 10, 11 In contrast, upper-limb performance in daily life has not received much attention due to the lack of validated measurement tools. Observations in daily life could be used but they are time consuming and may interfere with the patient’s natural behavior.12, 13 Other possible methods include self-reports and questionnaires, but these kinds of instruments are inherently subjective, focus on other aspects of daily functioning (eg, experienced problems), and not all relevant aspects of upper-limb functioning in daily life can be measured this way. From our previous experience, we know that performance measured with an objective monitoring device is generally different from performance measurements obtained with interviews or self-reports.14 To evaluate the actual level of performance, there is a need to measure performance objectively, unobtrusively and in detail, especially with the concept of learned non-use in mind. Several monitoring devices have been developed for the objective measurement of upper-limb activity. In general, these portable devices are based on sensors that are sensitive to movement of body segments, and their data are stored—possibly after some pre-processing—in that device. Uswatte et al15, 16 presented an upper-limb activity-monitoring device based on accelerometers (sensors that measure acceleration in one plane) that detect movement and nonmovement of the upper limb. Vega-Gonzales and Granat17 also developed a device to measure upper-limb activities using pressure transducers on wrist and shoulder. The Upper Limb Activity Monitor (ULAM), as developed by Schasfoort et al,18 is a similar instrument, but it adds the capability to detect body postures and body motions,19 as well as upper-limb movements. Using a combination of sensors on the wrist, trunk, and legs, the ULAM relates upper-limb usage to different body postures (eg, walking, sitting).14 The ULAM has been validated and applied in patients with complex regional pain syndrome type I (CRPS I)14 and is able to discriminate upper-limb usage of patients from that of healthy subjects. Furthermore, the ULAM also measures outcomes other than those based on questionnaires.20 The ULAM has recently been adapted for use in stroke patients, because the original ULAM was developed specifically for the CRPS I population. At that time, we assumed that CRPS I has consequences mainly affecting the amount of movement, and that accelerometers seemed to be adequate for purposes of measurement. In contrast to the CRPS I group, we assumed that—in stroke subjects—the intensity of upper-limb movement would be too low (ie, due to very slow movement of the arm) for valid measurement of upper-limb usage. Additionally, the range and quality of elbow movement is (part of) the focus of many clinical tests and movement analyses and could be measured with electrogoniometers.21, 22, 23 In personal discussions, experts have confirmed that elbow movement during daily life could be an indicator for upper-limb usage; these experts were rehabilitation specialists, physical therapists, and occupational therapists of our university medical center and of the rehabilitation center that we collaborate with. All are experienced people in the field of stroke treatment. Based on these findings, we extended the ULAM by 1 electrogoniometer on each elbow, which produced the Stroke Upper Limb Activity Monitor (Stroke-ULAM). The Stroke-ULAM registers both electrogoniometric data and accelerometric data. The aim of the current study was to examine the sensitivity of the Stroke-ULAM, that is, the ability of the Stroke-ULAM to discriminate between groups. Measurements were performed in a stroke population and in a healthy control group. It was assumed that, on the group level, subjects with more capacity will have more upper-limb activity in daily life. Therefore, the Stroke-ULAM should be able to detect differences in upper-limb usage between moderately recovered stroke patients, well-recovered stroke patients, and healthy subjects. Furthermore, the differences in sensitivity between measurements based on electrogoniometry versus accelerometry was examined, primarily to investigate the possibility of reducing the number of sensors used. Methods  Participants A total of 22 subjects volunteered to participate in the study; 17 stroke patients (of whom 5 patients were hospitalized at the Rijndam Rehabilitation Center and had a partly structured activity pattern during the day as a consequence of therapy) and 5 healthy control subjects (table 1). We divided the stroke population into moderately recovered and well-recovered subgroups, based on the score on the upper-extremity part of the Fugl-Meyer Assessment (FMA).24 Subjects with scores of 45 or higher were defined as well-recovered; scores below 45 defined those subjects as moderately recovered. We set the cutoff point at 45 because, with higher scores, patients have full wrist functioning, whereas with lower scores, patients generally lack full wrist functioning. The inclusion criteria were: knowledge of the Dutch language and the ability to walk inside the house or rehabilitation center. Control subjects were excluded if they reported any limitation in daily life activities. Patients were excluded if they suffered from any other medical condition besides stroke that might interfere with upper-limb function. Patients in the chronic phase had all been inpatients of the Rehabilitation Centre Rijndam. Control subjects were from the authors’ social circles. The study was approved by the Medical Ethics Committee of the Erasmus University Medical Centre Rotterdam and all participants gave informed consent. Device and Apparatus The Stroke-ULAM is the ULAM as described previously,25 with an additional electrogoniometera for each elbow. The ULAM consists of 5 piezoresistive acceleration sensorsb (4 uniaxial, 1 biaxial; size, 1.0×1.0×0.5cm), placed on the lateral side of the left and right thigh (the sensitive axis in sagittal plane while standing), on the sternum (sensitive axes in sagittal and longitudinal plane while standing), and on the upper limbs, just proximal to the wrist joint (sensitive axis perpendicular to the upper limb in sagittal direction in anatomic position). Two electrogoniometers were added to the ULAM configuration on both the left and right elbow to measure elbow angles (fig 1). The electrogoniometer consists of 2 parts: a distal end block was attached to the forearm and a proximal end block to the upper arm, both with the center axis of the end block parallel to the segment axis. All sensors were fixed on Rolian Kushionflex using double-sided tape and subsequently attached to the skin. Raw signals of both accelerometers and electrogoniometers were stored on a Vitaport II digital recorderc that was carried in a bag around the waist, with a sample frequency of 32Hz. After the measurement the raw data were downloaded onto a personal computer. Protocol The monitoring period was 24 hours except for the patients who received inpatient rehabilitation. In this latter group, the monitoring period was 12 hours for practical reasons, such as nursing care. The 12-hour period for inpatients started at 9:00 am and lasted until 9:00 pm. The FMA was performed at the earliest 1 day in advance of the monitoring period. Subjects were instructed to continue their ordinary life, although swimming and taking a bath or shower were prohibited during the monitoring period. The research questions were not revealed to the subjects prior to the monitoring period to prevent adaptations of daily activity patterns. Data Analysis and Transformation Based on feature signals derived from the measured accelerometer signals of the legs and trunk, and by using activity-specific settings in the analysis software and a minimal distance-based detection method, each second of the measurement 1 body posture or motion (eg, sitting, standing, lying, walking, cycling) from a set of body postures and motions was automatically detected. Additionally, from each accelerometer signal of the upper limb the intensity was calculated by calculating the root mean square of the signal after band-pass filtering (finite impulse response, 0.3−16Hz) and downscaling the sample frequency to 1Hz. Details on the data analysis can be found in Bussmann19 and Schasfoort18 and colleagues. The elbow angle signals derived from the electrogoniometers were filtered using a recursive fourth-order Butterworth low-pass filter with a cutoff frequency of 2Hz. After filtering these data, the derivative of these signals was calculated. Subsequently, this derivative was summed over periods of 1 second, and finally the data were rectified. To eliminate measurement bias and to use only relevant elbow angular movement, a threshold was added at the end of the analysis of 6.4° per second. This threshold was based on standardized measurements in which the electrogoniometers were moved over known angles and several threshold values were tested. The actual outcome measure for the electrogoniometers was total elbow movement per minute (flexion, extension). The measurements with a 24-hour duration were trimmed to the period from 9:00 am to 9:00 pm to allow comparison with the data of the hospitalized subjects (also measured from 9:00 am to 9:00 pm). Outcome Measures The Stroke-ULAM has 2 main outcome measures: (1) an absolute measure for each upper limb (level of usage) and (2) a relative measure indicating the level of usage of the affected upper limb compared with the unaffected upper limb (proportion). These outcome measures were determined by both electrogoniometric and accelerometric data. The electrogoniometry level of usage (for both affected and unaffected upper limb) is the elbow joint movement of the upper limb per minute (in degrees per minute), whereas the proportion is the level of usage of the affected upper limb divided by the level of usage of the unaffected upper limb. The level of usage for accelerometry is expressed as the intensity per minute (in g/min). The intensity depends on the variability of the raw acceleration signal around the mean value, that is, the higher the variability, the higher the intensity. The accelerometric proportion is calculated in the same way as for electrogoniometry. For the control subject the proportion is the level of usage of the nondominant upper limb divided by the level of usage of the dominant upper limb. The outcome measures were calculated over the time periods during the 12-hour measurement period that subjects were sitting or standing. For descriptive purposes, also the percentage was calculated that subjects performed body motions (eg, walking, cycling, general movement, climbing stairs), and the percentage that subjects were sitting or standing (see table 1). Statistical Analysis We used the Mann-Whitney U test to compare the proportions and levels of usage between the control and patient group, and between the patient subgroups. The level of usage of the affected upper limb was compared with the unaffected upper limb with the Wilcoxon test. Nonparametric tests were preferred to parametric tests because of the relatively small numbers in each group. The Pearson correlation coefficient was used to compare the outcome measures of electrogoniometry and accelerometry. The level of significance was set at .05 in all tests. Results Because 5 patients were measured in the rehabilitation center instead of at their home, the data analysis for these subjects was performed twice: once over the total measurement period and once excluding the therapy periods. Because no differences were found between these two analyses, the total measurement period of these 5 patients was included in our analysis. Sensitivity of Accelerometry The results for the accelerometry were comparable with those of the electrogoniometry (see table 2). The accelerometric level of usage of the affected upper limb was lower for stroke patients compared with the control subjects (P<.01). Moderately recovered patients also had lower levels of usage compared with the well-recovered patients (P<.01). The level of usage of the affected upper limb differed (ie, was lower) from that of the unaffected upper limb in all patient groups, but not for the control group (see table 2). The accelerometers were also discriminative for the proportion between stroke patients and control subjects, and between moderately recovered and well-recovered stroke patients (P<.01, P<.01, respectively). Thus, accelerometry is sensitive to differentiate between the level of usage of the affected upper limb and unaffected upper limb as well as between the level of usage of the patients and control subjects. Comparison of Electrogoniometry and Accelerometry The proportions of electrogoniometry and accelerometry were discriminative between the proportions for patients and for control subjects, and all proportions of electrogoniometry and accelerometry correlated very well (r=.938). The proportions for electrogoniometry were on average lower than for accelerometry (36.73%±31.14% vs 51.47%±29.77%, P<.01). Additionally, the range of individual proportion values was larger for electrogoniometry than for accelerometry (2.23%−97.09% vs 12.48%−100.01%). Figure 2 shows the individual levels of usage and proportions for both the affected and unaffected upper limb as well as the proportion values (upper row shows data derived from electrogoniometry). The results of this study show that the Stroke-ULAM outcome measures, based on electrogoniometry or accelerometry, are sensitive to differences between patients and controls, and to differences between well-recovered and moderately recovered stroke patients. The proportions based on electrogoniometry of moderately recovered patients are lower than when based on accelerometry (see fig 2). Discussion The results of the healthy subjects show proportions close to 100%; although we measured only 5 healthy subjects, these data are very similar to the results of 10 healthy subjects that participated in a study by Schasfoort et al14 who also found proportions close to 100%. Because moderately recovered patients have little to no control over the affected upper limb, it is unlikely that the affected upper limb is actually involved in daily life activities. As a consequence, tasks are performed with the unaffected upper limb, resulting in lower proportions (and lower levels of upper-limb usage for the affected upper limb). The low proportions are best seen in the proportions of the electrogoniometry; the unaffected elbow joint movement is used up to 45 times as much as the affected elbow joint movement. A possible explanation for the difference in proportions between both methods is the effect of rigid upper-limb movements. Electrogoniometers will not detect activity during rigid upper-limb movements because the elbow joint is not moved. Accelerometers do detect rigid upper-limb movements because upper-limb usage will cause acceleration and deceleration profiles. Rigid upper-limb movements are assumed to represent involuntary movements, for example related to trunk activity rather than isolated actions of the affected upper limb. Rigid upper-limb movements are reflected to a lesser extent by electrogoniometry than by accelerometry; therefore, electrogoniometry probably reflects a more realistic proportion of intentional upper-limb usage. In this study, 2 outcome measures were used to indicate upper-limb usage: (1) an absolute measure for each upper limb (level of usage), and (2) a relative measure between the level of usage of the affected and unaffected upper limb (proportion). The proportion is the result of the level of usage of both upper limbs and shows the relative use of the affected upper limb in comparison with the unaffected upper limb. The proportion is thus dictated by the relative use of both upper limbs and not by the factual or absolute use (these are represented by the level of usage). This has several advantages over the level of usage outcome measures. First, proportion depends less on the actual amount of activity of a subject, which is not only influenced by a patient’s capacity but also by external factors such as the social environment and aid tools. It measures the contribution of the affected upper limb to activities of daily life of the upper limbs compared with the unaffected side, that is, compensation strategies of the unaffected upper limb will result in a lower contribution of the affected side and therefore in lower proportions. One of the main targets of rehabilitation therapies (eg, the constraint-induced movement therapy) is that the affected upper limb will be involved more during activities of daily life. The Stroke-ULAM can measure the increase in this involvement. Also, it is an easy-to-use outcome measure because it allows one to compare upper-limb usage of several subjects or different periods of time without concerns about different activity patterns. Overall, the results suggest that, within a clinical context, the proportion is the most appropriate outcome measure of upper-limb usage in daily living conditions. As mentioned above, the electrogoniometer proportion is theoretically preferred to the accelerometric proportion as the outcome measure, but accelerometers are smaller, cheaper, less vulnerable, and are already frequently used in activity monitoring studies. The technical vulnerability of the electrogoniometers was the major cause of the incomplete measurements and is a major source of concern for future measurements. A decision as to which sensor is to be preferred could not be made on the basis of this study and should be explored in future research. In this study, we investigated the sensitivity of the Stroke-ULAM to measure the upper-limb usage during sitting and standing, but not during the whole measurement period. We assumed that an upper-limb activity monitor that is able to discriminate between body motions and postures would increase the sensitivity of upper-limb usage outcome measures.14 Uswatte et al16 also studied the use of accelerometers to measure upper-limb usage and also used proportion based on accelerometry as the main outcome measure. They used accelerometers on the wrist and were not able to distinguish between body movements and postures and that has some potential limitations. The most important limitation is that movement of the whole human body (eg, during walking) will be reflected in the signals from the upper limbs. As a consequence, such activities will be regarded as being upper-limb usage, whereas this is not the case. To explore the effect of walking and other body movements, we analyzed our proportion data over the whole measurement period for the stroke patients. The proportion for accelerometry increased from 39.2%±21.4% to 44.3%±21.5% (P<.001) and for electrogoniometry from 22.3%±18.0% to 24.7%±18.5% (P=.015). This increase was expected, although the difference in increase between accelerometry and electrogoniometry was expected to be larger, because movements of the upper limbs caused by walking (and most other body movements) would be reflected more in accelerometers than in electrogoniometry. This analysis nevertheless shows that body movements do increase the proportion, and without measuring body postures and motions, an increase in proportion (eg, after an intervention) could be the effect either of increased upper-limb usage or of increased body motion. These data support our assumption that it is important to measure upper-limb usage only during sitting and standing. Study Limitations The relatively small number of patients who participated in this study could have influenced the results, although the results do reflect the expectations. A larger number of measurements would probably strengthen the results rather than weaken them. Furthermore, the proportion may be partly determined by the dominance of the affected upper limb (be it the dominant or nondominant upper limb prior to stroke). Although this is an interesting topic, we do not analyze this issue. In most cases, the unaffected upper limb becomes the dominant side after stroke because ultimately it becomes the most capable upper limb. Conclusions The Stroke-ULAM is sensitive enough to detect differences between upper-limb usage of moderately recovered stroke patients, well-recovered stroke patients, and control subjects. The Stroke-ULAM can be a valuable addition to the mostly capacity-oriented tools currently used to evaluate upper-limb function during the rehabilitation process. The outcome measure can give the physicians or therapists valuable information about the improvement (or lack thereof) of the performance of the upper limbs. Suppliers References 1. 1Ward NS, Cohen LG. Mechanisms underlying recovery of motor function after stroke. Arch Neurol. 2004;61:1844–1848. MEDLINE |
CrossRef
2. 2Jorgensen HS, Nakayama H, Raaschou HO, Vive-Larsen J, Stoier M, Olsen TS. Outcome and time course of recovery in stroke (Part I: Outcome. The Copenhagen Stroke study). Arch Phys Med Rehabil. 1995;76:399–405. Abstract |
Full-Text PDF (817 KB)
|
CrossRef
3. 3World Health Organization. International classification of functioning, disability and health (ICF). Geneva: WHO; 2001;. 4. 4Roth EJ, Heinemann AW, Lovell LL, Harvey RL, McGuire JR, Diaz S. Impairment and disability: their relation during stroke rehabilitation. Arch Phys Med Rehabil. 1998;79:329–335. Abstract |
Full-Text PDF (913 KB)
|
CrossRef
5. 5Wade DT. Epidemiology of disabling neurological disease: how and why does disability occur?. J Neurol Neurosurg Psychiatry. 1997;63(Suppl 1):S11–S18. 6. 6De Haan R, Horn J, Limburg M, Van Der Meulen J, Bossuyt P. A comparison of five stroke scales with measures of disability, handicap, and quality of life. Stroke. 1993;24:1178–1181. MEDLINE 7. 7Taub E, Crago JE, Burgio LD, et al. An operant approach to rehabilitation medicine: overcoming learned nonuse by shaping. J Exp Anal Behav. 1994;61:281–293. MEDLINE |
CrossRef
8. 8Sterr A, Szameitat A, Shen S, Freivogel S. Application of the CIT concept in the clinical environment: hurdles, practicalities, and clinical benefits. Cogn Behav Neurol. 2006;19:48–54. MEDLINE |
CrossRef
9. 9van der Lee JH, Roorda LD, Beckerman H, Lankhorst GJ, Bouter LM. Improving the Action Research Arm test: a unidimensional hierarchical scale. Clin Rehabil. 2002;16:646–653. MEDLINE |
CrossRef
10. 10van der Lee JH, de Groot V, Beckerman H, Wagenaar RC, Lankhorst GJ, Bouter LM. The intra- and interrater reliability of the action research arm test: a practical test of upper extremity function in patients with stroke. Arch Phys Med Rehabil. 2001;82:14–19. Abstract | Full Text |
Full-Text PDF (59 KB)
|
CrossRef
11. 11Kopp B, Kunkel A, Flor H, et al. The Arm Motor Ability Test: reliability, validity, and sensitivity to change of an instrument for assessing disabilities in activities of daily living. Arch Phys Med Rehabil. 1997;78:615–620. Abstract |
Full-Text PDF (843 KB)
|
CrossRef
12. 12Melanson EL, Freedson PS. Physical activity assessment: a review of methods. Crit Rev Food Sci Nutr. 1996;36:385–396. MEDLINE |
CrossRef
13. 13Montoye HJ, Taylor HL. Measurement of physical activity in population studies: a review. Hum Biol. 1984;56:195–216. MEDLINE 14. 14Schasfoort FC, Bussmann JB, Zandbergen AM, Stam HJ. Impact of upper limb complex regional pain syndrome type 1 on everyday life measured with a novel upper limb-activity monitor. Pain. 2003;101:79–88. Abstract | Full Text |
Full-Text PDF (386 KB)
|
CrossRef
15. 15Uswatte G, Miltner WH, Foo B, Varma M, Moran S, Taub E. Objective measurement of functional upper-extremity movement using accelerometer recordings transformed with a threshold filter. Stroke. 2000;31:662–667. MEDLINE 16. 16Uswatte G, Foo WL, Olmstead H, Lopez K, Holand A, Simms LB. Ambulatory monitoring of arm movement using accelerometry: an objective measure of upper-extremity rehabilitation in persons with chronic stroke. Arch Phys Med Rehabil. 2005;86:1498–1501. Abstract | Full Text |
Full-Text PDF (161 KB)
|
CrossRef
17. 17Vega-Gonzalez A, Granat MH. Continuous monitoring of upper-limb activity in a free-living environment. Arch Phys Med Rehabil. 2005;86:541–548. Abstract | Full Text |
Full-Text PDF (202 KB)
|
CrossRef
18. 18Schasfoort FC, Bussmann JB, Stam HJ. Ambulatory measurement of upper limb usage and mobility-related activities during normal daily life with an upper limb-activity monitor: a feasibility study. Med Biol Eng Comput. 2002;40:173–182. MEDLINE |
CrossRef
19. 19Bussmann JB, Martens WL, Tulen JH, Schasfoort FC, van den Berg-Emons HJ, Stam HJ. Measuring daily behavior using ambulatory accelerometry: the Activity Monitor. Behav Res Methods Instrum Comput. 2001;33:349–356. MEDLINE 20. 20Schasfoort FC, Bussmann JB, Stam HJ. Correlation between a novel upper limb activity monitor and four other instruments to determine functioning in upper limb complex regional pain syndrome type I. J Rehabil Med. 2005;37:108–114. MEDLINE |
CrossRef
21. 21Buckley MA, Yardley A, Johnson GR, Carus DA. Dynamics of the upper limb during performance of the tasks of everyday living—a review of the current knowledge base. Proc Inst Mech Eng [H]. 1996;210:241–247. MEDLINE |
CrossRef
22. 22Steenbergen B, Van Thiel E, Hulstijn W, Meulenbroek RG. The coordination of reaching and grasping in spastic hemiparesis. Hum Mov Sci. 2000;19:75–105.
CrossRef
23. 23Roby-Brami A, Jacobs S, Bennis N, Levin MF. Hand orientation for grasping and arm joint rotation patterns in healthy subjects and hemiparetic stroke patients. Brain Res. 2003;969:217–229. MEDLINE |
CrossRef
24. 24Fugl-Meyer AR. Post-stroke hemiplegia assessment of physical properties. Scand J Rehabil Med Suppl. 1980;7:85–93. MEDLINE 25. 25Schasfoort FC, Bussmann JB, Martens WL, Stam HJ. Objective measurement of upper limb activity and mobility during everyday behavior using ambulatory accelerometry: the upper limb activity monitor. Behav Res Methods. 2006;38:439–446. a Department of Rehabilitation Medicine, Erasmus Medical Center, Rotterdam, The Netherlands b Rehabilitation Center Rijndam, Rotterdam, The Netherlands.  Reprint requests to Johannes B. Bussmann, PhD, Dept of Rehabilitation Medicine, Erasmus Medical Center, Rm H-022, ’s Gravendijkwal 230, 3015 CE Rotterdam, The Netherlands
Supported by Kinderfonds Adriaanstichting Rotterdam. 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. PII: S0003-9993(07)00407-8 doi:10.1016/j.apmr.2007.06.005 © 2007 American Congress of Rehabilitation Medicine and the American Academy of Physical Medicine and Rehabilitation. Published by Elsevier Inc. All rights reserved. | |
|
FULL TEXT
00407-8/journalimage?img=PIIS0003999307004078.gr1.lrg.gif&fig=fig1&kwhquery=&issn=0003-9993&ishighres=false&allhighres=false&free=yes','journalimage');)
00407-8/journalimage?img=PIIS0003999307004078.gr2.lrg.gif&fig=fig2&kwhquery=&issn=0003-9993&ishighres=false&allhighres=false&free=yes','journalimage');)