Abstract
Background and Purpose. Body weight support (BWS) treadmill training has recently been shown to be effective for gait training following stroke, but few researchers have measured the usefulness of this intervention in enhancing function, and there are no reports in which BWS overground ambulation was studied. The purposes of this case report were (1) to report the feasibility and patient tolerance for using a BWS system for overground ambulation, (2) to measure the function of patients with chronic stroke (2 years post-stroke) prior to and following BWS treadmill and overground ambulation training, and (3) to describe a protocol used for patient treatment progression using BWS treadmill training. Case Descriptions. The participants were 2 women, aged 87 and 93 years, who had strokes more than 2 years before data collection. A 10-m timed walk test, the Berg Balance Scale, the gait portion of the Tinetti Gait and Balance Assessment, and a measure of step length were administered. Intervention consisted of BWS ambulation training 3 times a week for 6 to 7 weeks. Each day there was BWS treadmill and overground training. Outcomes. Participant A improved most in 10-m walking time and Berg Balance Scale score. Participant B exhibited improvements in step length and 10-m walking time. Discussion. The outcomes suggest that very old patients with chronic functional deficits secondary to cerebrovascular accident tolerated BWS treadmill and overground ambulation training and made improvements following this intervention.
- Ambulation training
- Body weight support
- Partial weight bearing
- Stroke
Following a cerebrovascular accident (CVA), many people are unable to ambulate in the community1; therefore, regaining optimal mobility is the primary goal for many patients. Conventional efforts to improve the mobility of patients following a CVA have focused on preambulation mat programs with progression to standing and finally to ambulation.1 This approach often includes techniques that are supposed to facilitate movement patterns by use of a variety of sensory inputs.2–4 The effectiveness of these techniques, however, is in question.1
One recently described task-oriented approach to gait rehabilitation incorporates a body weight support (BWS) system.5–8 Through controlled reduction of weight bearing during ambulation, the BWS system attempts to provide postural support and promote coordination of the lower extremities. The decrease in weight bearing is intended to minimize the demands on the muscles, thus, in theory, allowing the patient to develop more effective and efficient movement strategies. The controlled environment also may increase patient confidence by providing a safe way to practice walking.9 Using the system also frees the therapist from physically supporting patients during gait training and provides the therapist with the opportunity to step back and evaluate the patient's gait. As body support is slowly decreased, demands on the patient for postural control and balance are increased. Therefore, BWS ambulation is thought to address posture, balance, and coordination while allowing training in a safe, efficient, and task-oriented manner.7,8,10
Researchers have used BWS systems for patients with spinal cord injury11–13 and stroke.9,14–,18 Improvements in temporal/distance gait variables (stride length, cadence, gait speed, and swing and stance symmetry) and electromyographic patterns of patients receiving BWS training following stroke have been noted in the literature.14–18 Results offer promise for increasing the ability of patients to walk after a CVA; however, there are limitations in these studies. Results have primarily focused on improvements in gait using a BWS system with treadmill ambulation.14–18 Body weight support training with overground ambulation training has rarely been discussed in the literature. Proponents of BWS training contend that the task-oriented nature of BWS (ie, practicing gait in its entirety rather than focusing on preambulation skill training) is one of the keys to its presumed effectiveness.19 Body weight support overground ambulation may be more task specific and more effective than BWS treadmill training for improving ambulation of patients following stroke because the walking surface is the same as a typical walking surface. In only 2 studies with patients following CVA have functional measures been used rather than temporal/distance variables or electromyography to demonstrate the effectiveness of BWS intervention.17,18 Therefore, the purposes of this case report were (1) to report the feasibility and patient tolerance for using a BWS system with overground ambulation, (2) to investigate whether the intervention could affect overall functional mobility of patients with prior stroke as measured by functional assessment tools commonly used by clinicians, and (3) to describe and discuss the protocol we used for progression of BWS ambulation training.
Case Description
Assessment Tools
Prior to BWS training, the participants were evaluated to determine whether they had difficulties with gait and activities of daily living that would be amenable to intervention. Several tests and measurement scales commonly used by physical therapists were administered to measure each participant's functional level. The same therapist administered each test to maintain consistency. All tests were selected based on recommendations from the post-stroke clinical practice guidelines published by the US Department of Health and Human Services.20
Balance was assessed using the Berg Balance Scale. This scale has high degrees of interrater and intrarater reliability with elderly individuals21 and is able to detect change in status of patients with acute stroke.22 We also used a 10-m timed walk test to determine gait speed, using a stopwatch to time each participant. This test has been found to relate to falls, balance, and functional mobility in the elderly population.23 Each participant was also asked to walk at a comfortable speed over a sheet of ShuTrack carbon paper* that provided a template of each participant's footprints. Average step length was determined using measurements acquired from the template to allow us to assess gait symmetry.
The gait section of the Tinetti Test of Balance and Mobility and observational gait analysis were used to provide information about the quality and symmetry of gait.24 Finally, participants, nursing staff, and available family were asked about changes in functional status following intervention. For example, we asked whether there were noticeable differences in willingness to walk or interest in walking, distances walked, or assistance required for walking or transfers.
Participant A
Participant A was an 87-year-old woman who had a CVA 10 years previously. She had been a resident at an extended care facility for 3.5 years. She last received physical therapy 3 years prior to the collection of data for this case report. Relevant past medical history included diagnoses of breast cancer with mastectomy 13 years previously, hypertension, osteoarthritis, and ataxia. Her medications included aspirin, acetaminophen, Tenormin,† Darvocet,‡ Xanax,§ and potassium. She was independent with all activities of daily living except showering, which required use of a shower chair and assistance of one person.
During our observation of gait, lateral movements beyond the weight shift expected during single-leg stance were noted bilaterally. We also believe there was decreased hip extension bilaterally at the end of the stance phase, and shorter step length on the left. She had moderate kyphosis and genu valgum bilaterally. She was independent during walking on smooth surfaces using a rolling walker. Her 10-m walking time (using a rolling walker) was 28.7 seconds, and she was able to ambulate to the cafeteria (75 m) but required 2 sitting rest stops. Her average step length on the was 33.8 cm on the right and 27.2 cm on the left. Her step length ratio (short step:long step) was 0.80, and she scored 6/12 on the gait portion of the Tinetti Gait and Balance Assessment.
When asked about difficulties with ambulation, she reported that her primary difficulties were balance and swaying from side to side during walking. She complained of slowness when walking and stated that she avoided stairs because she feared falling; however, she reported being able to negotiate 1 to 2 steps with use of a handrail.
Her difficulty with balance and sway while ambulating was demonstrated by her inability to perform activities while standing without physical assistance or support of her walker. Balance difficulties were reflected by her score of 26/56 on the Berg Balance Scale. Table 1 gives a summary of her pre-intervention performance.
Summary of Measurements
Participant B
Participant B was a 93-year-old woman who had a CVA 14 years previously. She had been a resident at an extended care facility for 3 years and was not currently receiving physical therapy. Relevant past medical history included diagnoses of non-insulin-dependent diabetes mellitus, hypertension, and osteoarthritis. Although a diagnosis of dementia was not in her medical record, she was unable to provide any information about her medical history and was usually not oriented to place or time. Her medications included acetaminophen, aspirin, Peri-Colace,‖ glipizide, Prozac,# and Darvocet. She had hallux valgus and bilateral foot deformities, requiring custom-made accommodative foot orthoses and extra depth shoes. She also had moderate hearing loss and a history of cataract surgery. She required assistance to perform activities of daily living, including dressing and bathing. We observed that transfers seemed unsafe and required supervision because of decreased cognition and impulsivity. Based on our observation of gait, there was marked trunk flexion throughout the gait cycle, short step length bilaterally, and difficulty with foot clearance during the swing phase bilaterally. She used a rolling walker or pushed her wheelchair to assist with ambulation. She propelled herself with her feet while sitting in her wheelchair, and this was her primary mode of mobility. Her 10-m walking time (while pushing her wheelchair) was 30.3 seconds. Her average step length was 22.9 cm on the right and 23.6 cm on the left. Her step length ratio was 0.97, and she scored 5/12 on the gait portion of the Tinetti Gait and Balance Assessment.
Because of cognitive and memory deficits, we were unable to gather meaningful information from participant B about her difficulties with ambulation. From her medical record and caregivers, we discovered that she had a history of falls, with the last fall occurring approximately 1 month prior to data collection for this case report. She was unable to negotiate a step with or without the use of a handrail. Her balance difficulties were reflected in her score of 24/56 on the Berg Balance Scale. Table 1 gives a summary of pre-intervention performance.
Intervention
Body Weight Support System and Treadmill
The BWS system used was the Lite Gait I** (Figure). The Lite Gait I was selected for intervention because we believe it is especially useful for gait training with patients with neurologic involvement. The harness and 2 points of BWS, in our opinion, provide needed postural support along with the weight support. It is equipped with locking wheels that allow BWS ambulation training with a treadmill as well as overground. The system consists of an adult-sized frame with an adjustable handrail. A harness, which connects to an overhead frame via adjustable straps, supports the participant about the lower abdomen and pelvis. Padded groin straps keep the harness in place.
Lite Gait I body weight support system.
A bilateral symmetry scale mounted within the frame displays the amount of weight (in pounds) supported by the device at any particular moment. The amount of weight supported in standing at the initiation of each session, as measured by the bilateral symmetry scale, was monitored and documented. The treadmill used was the Proform Crosswalk,†† which had a 35.6- × 114.3-cm (14- × 45-in) walking surface and ranged in speed from 0.1 mph (0.04 m/s) to 10 mph (4.44 m/s).
Overview of Intervention
Body weight support ambulation training occurred 2 to 3 times per week over a 6- to 7-week period. In each session, there were 4 bouts of ambulation. The first 3 bouts consisted of ambulation on the treadmill with BWS, and the fourth bout consisted of overground ambulation with BWS. Bouts were separated by a 5-minute rest interval. The duration of each bout was limited by participant tolerance (we stopped if the participant requested a rest) or to a maximum of 10 minutes. At initiation of the intervention protocol, 40% of each participant's body weight was supported during BWS gait training. This percentage of BWS was selected based on research that demonstrated that support greater than 40% actually interfered with heel-ground contact for some patients.11 Participants were started on the treadmill at 0.5 mph (0.22 m/s).
Body weight support consisted of 3 levels: 40%, 20%, and 0% of body weight. Treadmill speed also consisted of 3 levels: 0.5 mph (0.22 m/s), 0.75 mph (0.33 m/s), and 1.0 mph (0.44 m/s). Guidelines based on achieving adequate gait technique and endurance were established to direct decisions about progression through these levels. Progression was accomplished by decreasing BWS or increasing treadmill speed. Gait was judged by observation. We considered the participants as having an adequate gait technique if they were able to support weight on the affected limb without buckling, along with somewhat consistent and symmetrical step length, cadence, and foot clearance. If adequate gait technique was achieved, the percentage of BWS was decreased to the next level during the next intervention day. We defined adequate endurance as occurring when the participants could ambulate more than 5 minutes in 2 of 3 treadmill bouts. If adequate endurance was achieved, speed was increased to the next level during the next intervention day. During the final week of intervention, BWS was decreased to 0% to transition the participant toward ambulating without BWS.
During the overground bouts of ambulation, the amount of BWS was consistent with the amount provided during the treadmill bouts on that day. Speed of ambulation during overground walking was primarily determined by the participant. We discovered, however, that we needed to assist the participants with moving and steering the BWS device during overground ambulation in the hallway. We attempted to propel the device at the self-selected speed of the participants and instructed them to inform us if we were going too fast.
During treadmill and overground walking, verbal cueing and manual guidance were given. For example, to facilitate a longer step length for improved symmetry, the participant might have been encouraged to “step out” or we might have encouraged a better swing forward by guiding her lower extremity from toe-off to heel-strike. The blood pressure and heart rate of each participant were monitored prior to intervention, between bouts, and at the end of each session. Guidelines for cessation of intervention included complaints of light-headedness, confusion, or dyspnea; onset of angina; excessive blood pressure changes (systolic blood pressure greater than 220 mm Hg, diastolic blood pressure greater than 110 mm Hg); and inappropriate bradycardia (drop in heart rate of greater than 10 beats per minute). Intervention was never terminated based on these criteria, but always by participant request secondary to fatigue. These criteria are in accordance with the American College of Sports Medicine criteria for termination of an inpatient exercise session.25
Participant A
Participant A completed 21 days of intervention over a 7-week period. She attended each session as scheduled. Overground training was not completed on 4 scheduled days due to conflicting scheduling of activities at the facility and inaccessibility to the hallway. Intervention was progressed in accordance with our guidelines (decrease BWS if adequate gait technique is achieved; increase speed if adequate endurance is achieved). The time for ambulation bouts ranged from 1 minute 46 seconds to 5 minutes 51 seconds. Each bout of ambulation was stopped if the participant reported fatigue. The drops in average intervention time on days 5, 8, 12, and 21, as noted in Table 2, correlated with progression to a more demanding level. For example, BWS was decreased from 40% to 20% on day 8, and average intervention time decreased from 4 minutes 51 seconds to 2 minutes 48 seconds. We noted that increased demand on day 8 resulted in decreased gait symmetry, as evidenced by a shorter left step length.
Participant A: Levels of Intervention and Description of Performance at Each Levela
During BWS overgound ambulation, the participant determined the speed of ambulation. This was in contrast to BWS treadmill ambulation when speed of ambulation was externally driven. Overground BWS ambulation speed increased across treatment sessions, as evidenced by mean speeds of 0.33 mph (0.15 m/s) at level 1 and 0.74 mph (0.33 m/s) at level 5 (Tab. 2).
Participant B
Participant B completed 17 days of intervention over a 6-week period. Three bouts of overground ambulation were not completed due to conflicting scheduling at the facility and the participant's refusal of a scheduled intervention one time during the training period. The times for bouts of ambulation ranged from 1 minute 43 seconds to 6 minutes 56 seconds. Ambulation was stopped if the participant complained of fatigue.
The protocol for progression of intervention was altered on 2 occasions (Tab. 3). The percentage of BWS was reduced from 40% initially to 20% on day 2 of intervention without achieving adequate gait technique according to our pre-established criteria. We prematurely reduced BWS because of her tendency to collapse into the harness with 40% of BWS. In addition, the speed of the treadmill was increased to 0.75 mph (0.44 m/s) on intervention day 14 without attaining adequate symmetry. We made this change because she had not achieved the symmetry requirements to decrease BWS or increase speed after 4 weeks of intervention. This change in speed assisted with more appropriate foot placement and step length.
Participant B: Levels of Intervention and Description of Performance at Each Levela
Throughout the 6 weeks of intervention, we used verbal cues and manual cues at the shoulders and hips to encourage upright posture. Cues were necessary during both BWS treadmill and overground ambulation.
As with participant A, speed of ambulation was primarily determined by participant B during bouts of overground ambulation. Her speed increased across intervention from a mean speed of 0.28 mph (0.12 m/s) at level 1 to a mean speed of 0.67 mph (0.30 m/s) at level 3. Level 4 consisted of one day of training with no overground ambulation due to a scheduling conflicting at the facility.
Outcomes
Participant A
The measurements reflecting participant A's post-intervention performance are presented in Table 1. Her ambulation improved, as indicated by all measurements. She demonstrated improved balance, as indicated by her Berg Balance Scale scores (26/56 pretest to 40/56 posttest). Her walking improved, as demonstrated by her scores on the Tinetti Gait and Balance Assessment. Her 10-m walking time (with rolling walker) improved from 0.35 m/s to 0.65 m/s (46% improvement), and her right and left step lengths improved by 35% and 41%, respectively. Our observation of participant A's gait indicated a more symmetrical gait pattern than before intervention. We believe that this observation was substantiated by an improved step length ratio. In our view, however, excessive lateral sway of the trunk remained bilaterally during single-leg stance.
Following intervention, participant A stated that “she felt like she was walking faster and was more conscious about taking larger steps.” According to her sister-in-law, participant A was able to get in and out of the car without assistance and was able to ambulate greater distances on community visits. The nursing staff, however, reported no “dramatic change” in her amount of walking or willingness to participate in activities.
Participant B
Measurements reflecting participant B's post-intervention performance are presented in Table 1. Minimal improvements in gait and balance were evident. The greatest improvements were a 22% improvement in 10-m walking time (while pushing her wheelchair) and improvements in step length. Step length improved more on the right (35%) than on the left (25%), which accounts for the decrease in step length ratio. Our observation of gait led us to believe that there were inconsistent improvements in step length and foot clearance bilaterally.
Although participant B was unable to provide useful information about her performance or the intervention, both her daughter and members of the nursing staff stated that she was more social and alert during the day throughout the duration of the intervention.
Discussion
Two patients with chronic deficits secondary to stroke demonstrated improved ambulation performance and balance following BWS treadmill and overground training as measured with the Berg Balance Scale, Tinetti Gait and Balance Assessment, 10-m timed walk test, and measure of step length. These results are consistent with those of previous studies evaluating BWS systems for retraining gait in patients following a CVA.9,14–18 Participant B displayed less improvement on the tests than participant A, but both participants demonstrated faster 10-m walking times and improved step lengths. Although performance on the tests administered improved, real-world functional performance improved minimally, especially for participant B.
Prior to intervention, participant A ambulated at a speed of 0.35 m/s (0.79 mph). After intervention, she ambulated at 0.65 m/s (1.46 mph). According to Robinett and Vondran,26 the minimum safe speed for street crossing is 0.5 m/s (1.12 mph), whereas Perry et al27 suggested that the speed needed to safely cross the street in the time allotted by crosswalk signals is 0.6 m/s (1.34 mph). By achieving a walking speed of 0.65 m/s (1.46 mph), participant A surpassed these minimal requirements. Her increased walking speed may be an important gain with implications for function.
Participant A also had a 14-point improvement in her balance, as measured with the Berg Balance Scale. Her pretest score of 26 on the Berg Balance Scale improved to 40 at the posttest measurement. Shumway-Cook et al28 suggested that Berg Balance Scale scores of community-dwelling well elderly people were associated with their fall risk, with better scores predicting a lower probability of falling. The relationship between score and risk, however, is believed to be nonlinear. A perfect score of 56/56 was reported to be associated with a 10% risk of falling, with scores below 36 predicting a 100% fall risk.28 Thus, as scores decrease, the probability of falls increases rapidly, with a score of 40 calculated as being associated with a 95% risk of falling.28 Although it is tempting to speculate that participant A's 14-point improvement in Berg Balance Scale score is meaningful, it may represent only a slight decrease in her fall risk. Research is needed to investigate whether BWS ambulation training of longer duration or intensity can improve balance to a level that would substantially reduce fall risk.
Patients with chronic CVA were selected as participants to limit the influence of extraneous variables such as spontaneous recovery and concomitant therapeutic interventions. Although it has been documented that a large part of physical improvement following stroke is attained in the first 6 months,29 our 2 participants were able to make gains years after their strokes. This work supports the tenet that even patients with chronic CVA may be capable of making progress.
There has been some discussion that BWS gait on a treadmill is not task specific for ambulation.30 We acknowledge that BWS is not task specific, but we agree with other authors that BWS training is task oriented.14,17,18,31 Often it is not possible to attain task specificity secondary to the degree of physical involvement with patients with CVA; therefore, task-oriented training needs to be emphasized. In an attempt to provide maximum task orientation, we incorporated BWS ambulation training both on a treadmill and overground. Body weight support overground ambulation was both feasible and well tolerated by participants. In this protocol, we allowed participants to self-select their overground walking speed. It is possible, however, that trying to maintain an overground speed equal to or exceeding that achieved on the treadmill may have been more beneficial.
It is unlikely that either of these participants would have worked as hard without the impetus provided by the treadmill. Several authors14,32,33 have suggested that the treadmill may compel increased effort by informing patients about performance (speed, duration), by providing motivating challenges to the patient, and by demanding maintenance of speed.
Differences in results between participants were observed, with participant A showing more improvements than participant B. These differences might have been influenced by several factors. First, the participant's level of cognition may have affected the outcome. Participant B had occasional days of agitation and confusion that may have negatively affected her performance. Motivation also varied between participants. Participant A verbalized a desire to improve her pattern of ambulation, and she was willing to take part in any activities that might improve her functional status. In contrast, participant B appeared to be satisfied with her functional level. In addition, premorbid diagnoses and prior activity status could have influenced the effectiveness of intervention. Although both participants were ambulatory, participant A ambulated more frequently and greater distances throughout the day, whereas participant B used a wheelchair as her primary mode of mobility within the facility. In addition, participant B had other complications prior to initiation of intervention. She had bilateral foot deformities and diabetes, which may have resulted in decreased awareness of lower-extremity performance during gait.
Although differences existed between participants, one of the more favorable outcomes was observed and reported behavioral change. Participant A repeatedly reported that she tried to work on improving her walking on her own. She seemed pleased with the progress made in her walking. Staff and family reported that participant B was more interactive and engaged in conversations more frequently after the intervention sessions; however, her improved alertness might be attributed to the increased attention she was receiving during the period of data collection.
We believe that this case report is the first to attempt to describe a specific protocol using a BWS system with patients following stroke. In an effort to describe the clinical decision-making processes used to progress our participants, we defined adequate endurance as the ability to ambulate more than 5 minutes in 2 of 3 treadmill bouts. If the participant was able to meet this criterion, treadmill speed was increased during the next session. We defined adequate technique as the ability to support weight through the involved lower extremity and to demonstrate consistency and symmetry of step length, cadence, and foot clearance (as judged by observational analysis). If adequate technique was achieved, BWS was decreased for the next session.
These definitions of endurance and technique, in our view, were helpful as guidelines for progressing our participants, but it was evident that clinical judgment played an important part in decision making. For example, participant B did not demonstrate the symmetry or endurance to decrease BWS or increase speed after 12 days of level 2 intervention. We made the decision to increase her walking speed from 0.5 to 0.75 mph on day 14 to determine whether the increased speed might lead to a more symmetrical gait. Not only was she able to walk with more appropriate foot placement and step length at the faster speed, but she demonstrated greater mean walking time on the treadmill. Thus, although guidelines were useful, we believe there also was a place for the use of therapist judgment.
This case report is the first to involve very old participants with multiple pathologies in BWS training. In spite of their age and frailty, the intervention was well tolerated by both participants.
The participants described in this case report were involved only in BWS ambulation training. It may be that they could have attained better outcomes if other modes of therapy would have been used in conjunction with BWS gait training. For example, participant A might have benefited from strengthening exercises aimed at decreasing her sway during gait.
One potential advantage of the BWS system is a decreased need for personnel to provide a safe environment for gait training. Although one therapist can safely provide BWS training with some patients, 2 people may be required to attach the harness and to facilitate the desired gait pattern with patients with greater involvement.
We believe that our case report generates several ideas for future research studies. It would be interesting to assess how long improvement would be maintained by adding a delayed posttest. Improvements in the suggested protocol would also be beneficial. Other aspects to be studied include comparing rehabilitation outcomes with and without the use of overground ambulation and inclusion of a measure to determine whether observed improvements are a result of learning or solely due to better endurance.
Conclusions
For these 87- and 93-year-old participants, BWS overground ambulation was both feasible and well tolerated. These 2 participants with chronic deficits secondary to stroke made l improvements following BWS treadmill and overground training. Guidelines along with clinical judgment were important in designing patient progression. More research is necessary to further our understanding of the use of BWS systems in the rehabilitative process of patients following stroke.
Footnotes
-
All authors provided concept/project design, writing, data collection and analysis, and project management. Dr Miller provided fund procurement. The authors thank Sister Sharon Bierman and Cindy Yeich for assistance with providing participants and facilities. They also thank Dr Elizabeth Domholdt and Dr Rebecca Porter for their helpful comments on previous versions of the manuscript.
These case reports were presented at the Combined Sections Meeting of the American Physical Therapy Association; February 11–15, 1998; Boston, Mass.
The protocol for these case reports was approved by the Committee on Research Involving Human Participants at the University of Indianapolis.
This work was supported by a Krannert School of Physical Therapy Research Grant.
-
↵* Acutred-Shutrack Inc, 2900 Vassar, PO Box 20189, Reno, NV 89515-0189.
-
↵† AstraZeneca Pharmaceuticals LP, 1800 Concord Pike, Wilmington, DE 19850-5437.
-
↵‡ Eli Lilly and Company, Lilly Corporate Center, Indianapolis, IN 46285.
-
↵§ Pharmacia & Upjohn, 100 Rte 206 N, Peapack, NJ 07977.
-
↵‖ Shire US Inc, 7900 Tanners Gate Dr, Suite 200, Florence, KY 41042.
-
↵# Dista Products Co, Div of Eli Lilly and Company, Lilly Corporate Center, Indianapolis, IN 46285.
-
↵** Mobility Research, LLC, 2211 W First St, Tempe, AZ 85281.
-
↵†† ICON Health & Fitness Inc, 1500 S 1000 W, Logan, UT 84321.
- Received September 11, 2000.
- Accepted April 16, 2001.
- Physical Therapy
References
Issue highlights
