Low-load, prolonged stretch in the treatment of knee flexion contractures in nursing home residents.

The functional consequences of knee flexion contracture (eg, impaired weight-bearing activities, difficulty with bed or chair positioning) make knee flexion contracture a clinically important condition for physical therapists, nursing staff, and patients, although the prevalence of knee flexion contractures in institutionalized elderly populations appears to be low.[1] In a previous study,[1] we concluded that nursing home residents who show a decline in ambulation (decreased distance, increased need for assist) and exhibit resistance to passive knee extension may be more likely to develop increased knee flexion contracture and may benefit from physical therapy. it is not clear, however, what that intervention should be, nor at what intensity or duration it should continue.

In our experience, clinical practice related to knee flexion contractures often includes manual stretching exercise, prolonged stretch through strapping on a tilt table or application of sandbag weights over the distal femur, use of mechanical traction, wearing of bivalved plaster casts, and lying in a prone position with the legs unsupported. Manual stretching and passive range of motion (PROM) exercise have been suggested for management of contractures.[2-6] The general concept is that stretching be done with mild to moderate force for a prolonged time, although specific guidelines for judging intensity and duration are not usually given.

We are aware of only one previous study of physical therapy effectiveness in elderly nursing home residents with knee flexion contractures.[7] Light and colleagues[7] found greater increases in passive extension in knees treated with modified Buck's traction (low load, prolonged stretch) compared with knees treated with passive manual stretching (high load, brief stretch). The Buck's traction was applied for 1 hour, twice a day for 4 weeks, on 11, clients. The weight applied was not reported.

The Dynasplint System[R](*) has been marketed as another method of applying low stretching forces over prolonged periods. Dynasplint knee braces have appeal because clients can remain mobile and comfortable, the splints are easy to apply in a wheelchair or bed, and the splints are approved for Medicare reimbursement. The Dynasplint is constructed of two stainless steel struts (positioned medially and laterally on the extremity) with cloth strups and Velcro[R]([dagger]) closures (applied above and below the joint) (Fig. 1). In the knee extension Dynasplint, knee motion is allowed, but a compression coil spring within the struts resists flexion. This resistance to flexion tends to bias the knee toward a position of maximum extension while the patient is inactive and at rest. The spring tension can be adjusted from 0 (no force resisting flexion) to 12 (approximately 124 kg-cm of torque). The torque output on all Dynasplints is calibrated at 30 degrees of flexion by the manufacturer. Gains in range of motion (ROM) in stiff joints and successful contracture reduction have been reported with use of Dynasplints for the knee,[8] ankle,[9] and elbow,[10-12] although none of these cases were of institutionalized elderly persons.

The purposes of this study were (1) to pilot test the effectiveness of a low-load, prolonged stretch as compared with a traditional program of PROM and manual stretching in decreasing knee flexion contractures in an institutionalized elderly population and (2) to provide a functional and clinical description of our study population.

Method

Design

Longitudinal data were collected before and during treatment with prolonged stretch. Each subject had bilateral knee flexion contractures and could thus serve as his or her own control. The prolonged stretch intervention was done on one side only, five times per week. Because we preferred not to entirely withhold treatment from the opposite leg, we chose to provide a protocol of PROM and manual stretching for both lower extremities. Thus, the prolonged stretch was actually an additional intervention beyond what we considered to be a typical physical therapy intervention for knee flexion contractures. The PROM and manual stretching was done twice a week because this is a usual frequency for maintenance therapy in the nursing home population in the geographic area where the study was conducted. In our study, therefore, we compared prolonged stretching (use of a splint) with PROM and manual stretching.

Subjects

Twenty-eight residents (6 men, 22 women) from three nursing homes agreed to participate in the study. Residents with bilateral knee flexion contractures of 10 degrees or greater were asked to participate and provided with an explanation of the study. Informed consent was obtained from each participant or from a health care representative or guardian if the resident was unable to give consent.

Of the original 28 residents, 18 completed the study, 6 expired, and 4 withdrew. Two residents who withdrew from the study refused to continue the trial of prolonged stretching within the first month of the intervention; in the other 2 cases of withdrawal from the study, family members requested discontinuation of the splint due to perceived discomfort.

Ninety-three percent of the 28 subjects were already receiving physical therapy before participating in the study, and most (86%) of these subjects came to therapy twice a week (X=2, range = 1-3 times per week). The frequency of diagnoses in major categories in the initial 28 subjects were as follows: organic brain syndrome (680/o), cerebrovascular accident (320/o), Parkinson's disease (21%), arthritis (50%), and hip fracture (25%). The subjects had resided in the current nursing home an average of 3 years (SD=2.55, range=2 months-9 years).

Procedure

Measurement. Descriptive data (age, length of stay in the nursing home, and diagnoses) on each subject were obtained from the medical record. Repeated measurements were made over a 7-month period of PROM (hip extension, knee extension, and ankle dorsiflexion), the torque required to maintain maximum passive knee extension, overall functional level, ambulation and transfer status, cognitive status, and knee pain. In the first month, prior to initiation of treatment, measurement sessions occurred at 2-week intervals in order to obtain three bilateral baseline measurements. Thereafter, measurements were repeated monthly throughout the remainder of the study. The physical therapists performing the measurements were not involved in treatment of the subjects, nor were they aware of the side of the experimental treatments. One of three measuring therapists was assigned exclusively to each study site for the duration of the study.

Range of motion measurements were made using a standard 360-degree goniometer,([double dagger]) which was covered on one side to prevent the therapist from reading a value until after the measurement was made. Subjects were on a therapy mat table for measurements. Hip extension and ankle dorsiflexion were measured with each subject positioned supine. For hip extension measurements, the opposite hip was flexed maximally within the subject's tolerance, and the ipsilateral lower extremity was free to extend over the side edge of the mat. For hip extension, the goniometric center of rotation was placed over the greater trochanter, with the stationary arm of the goniometer aligned with the lateral trunk midline and the movable ann aligned toward the lateral femoral epicondyle. For ankle dorsiflexion measurements, the ipsilateral knee was flexed to minimize length limitations of the two-joint gastrocnemius muscle. The goniometric center of rotation for ankle dorsiflexion was placed just inferior to the lateral malleolus, with the stationary arm of the goniometer aligned toward the head of the fibula and the movable arm parallel to the fifth metatarsal.

Passive knee extension was measured with the subject positioned side lying, with the top leg supported on an elevated board to allow the hip to be in neutral relative to abduction and adduction. In addition, the hip was passively moved into maximum extension and held by an assistant and by a vertical post on the board supporting the leg (Fig. 2). The knee extension measurement was made after the therapist passively extended the knee, attempting three times to reach maximum extension and holding this extension for a count of 10 each time. When this final extension position was reached, an assistant held the extended position as the therapist measured ROM.

To describe the resistance usually encountered in passively extending the knee of subjects with flexion contractures, we used a SPARK hand-held dynamometer([section]) to measure the force required to hold the knee in maximum extension. A mark was made on the lateral aspect of the measured leg, at approximately the level where the Achilles tendon blends with the muscle belly. After maximum knee extension was achieved, an assistant held this extended position by placing the dynamometer perpendicularly on the posterior leg at the level of the skin mark (Fig. 2). The dynamometer force reading was multiplied by the distance from the skin mark to the lateral knee joint line to provide a torque value. The force measurement was taken simultaneously with the measurement of knee extension. After the initial measurement, the knee was allowed to flex, and then the extension procedure was repeated to provide two more force measurements. The measuring therapist was asked to rate the pain associated with passive knee motion (absent/present) based on the subject's verbal response or facial expression.

At the first, third, and last measurement sessions, the treating therapist scored the subject's overall functional level using the Barthel Index.[13] The reliability of this index has been previously reported.[14] The treating therapist also rated the need for assistance with ambulation and transfers each month using a previously described eight-point scale[15] (Tab. 1). Type of assistive ambulatory device and ambulation endurance were also recorded. Staff at each site administered the Short Portable Mental Status Questionnaire[16] to describe the cognitive status of each subject at the first, sixth, and ninth (last) intervals in the study. The reliability and validity of this questionnaire have been reported.[16,17]

[TABULAR DATA 1 OMITTED]

Treatment. Treatment for the knee flexion contractures was initiated at the beginning of the second month of the study and continued through the seventh month. Use of the prolonged stretch was atternately assigned to the right or left knee as each subject entered the study. Treating therapists were asked to stop any ROM or stretching exercises to the lower extremities other than those specified for the study protocol during the study period. Functional training or strengthening exercise could be continued or started if indicated.

Each subject received PROM exercise to both lower extremities twice a week by on-site physical therapists trained in a standardized protocol. The exercise protocol was agreed on by four therapists with geriatric experience and was designed to give a manual stretch into hip-knee extension and ankle dorsiflexion. Each leg was moved into extension and held for 10 seconds at the point of maximum resistance. This procedure was repeated until maximum knee extension was reached and then held for 1 minute. The leg was flexed and then moved into this sustained position of extension two more times. Five repetitions of ankle dorsiflexion with the knee extended were each held for 5 seconds.

A trained physical therapy aide applied the splint to the assigned leg five times a week, after initial fitting by the treating therapist. Wearing time of the splint was progressed from 1 hour to 3 hours by the end of the first week. We chose this duration because we wanted the physical therapy staff to directly apply and remove the splints. This application and removal of the splints necessitated coordination with nursing schedules for bathing and cares. The tension setting on the splint was initially 0 and progressed to 6 (62.2 kg-cm) between weeks 2 and 5 of the study. In our initial trials with subjects, we found that many nursing home residents could not tolerate settings greater than 6. To simplify data analysis, we chose a tension setting of 6 as a maximum for all subjects. The aide kept a log of subject tolerance and response to the splint. The Dynasplint fit and use were regularly monitored by the physical therapist involved in treatment. All the subjects in two of the nursing homes were checked for fit by the designer of the splint, who also owns the company that makes the splint. Some padding and cuff changes were made subsequent to his visit.

We monitored the treatment protocol, splint fit, and the measurement protocol on a weekly basis. We were not, however, involved directly with treatment or measurement.

Data Analysis

The Statistical Package for the Social Sciences-PC+ software program([parallel]) was used for data analysis. We did t tests on knee, ankle, and hip ROM; torque; pain; Batthel Index scores; transfers; ambulation status; and ambulation endurance to compare those subjects who completed the study and those subjects who withdrew or expired. Paired t tests were also done to identify differences between sides in either ROM or torque at the beginning of the study. A repeated-measures multivariate analysis of variance (MANOVA) and paired t tests were done to establish whether any differences existed between sides at any interval or on either side across time in knee ROM. Longitudinal knee extension data on each subject were graphed to allow study of individual variation from the group data.

Rater Reliability

Ambulation/transfers. Two physical therapists experienced in geriatrics independently rated the ambulation and transfer abilities of 16 nursing home residents who were not involved in the study. The intraclass correlation coefficient (ICC[2,1])[18] showed good intertater reliability for this assessment tool (transfers=.982, ambulation level=.960, assistive ambulatory device=1.00, ambulation endurance=1.00).

Range of motion. To test the intrarater reliability of the three measuring therapists on knee ROM measurements, each therapist repeated ROM measures on 8 to 10 lower extremities. A complete set of study measurements was made, the subject was allowed to briefly relax the extremity, and then the repeat measurements were made within one session. The ICC (2,1) showed good intrarater rehability of repeated measurements of ROM at the knee (.98 on the right, .99 on the left). Repeated knee extension measurements varied an average of 2 degrees range=0[degrees]15[degrees]).

Variance in Repeated Torque Measures

Repeated measures of the torque required to maintain passive knee extension showed high reproducibility within a given measurement session (N=25-28). Intraclass correlation coefficients 2,1) for repeated torque measures within each of the three baseline intervals were as follows: baseline 1--right side=.729, left side=.778; basefine 2--right side=.828, left side=.874; baseline 3--right side=.903, left side=.872. An average of the three repeated torque measures at any given interval was used in the data analysis.

Variance in Repeated Baseline Measures

A previous study on knee flexion contractures[1] led us to conclude that there may be some degree of variance in measurement of contractures over time (especially in the presence of resistance to passive motion) unrelated to the reliability of the measuring therapist. Three consecutive baseline measures of ROM and torque were made at 2-week intervals prior to the initiation of treatment to identify the stability of these variables. Table 2 presents ICC values for baseline measures repeated over 1 month. We used an average of the three baseline measures of ROM and torque in our analysis of the longitudinal data.

Table 2. Intraclass Correlation Coefficients for Repeated Baseline Measurements of Torque and Range of Motion N= 24)

                        Right           Left
                        Side            Side
Torque(a)              .528            .380
Knee extension         .897            .884
Hip extension          .678            .647
Ankle dorsiflexion     .612            .592
(a) An average of the three repeated torque
values at each baseline interval was used for
this calculation.

Results

Descriptive Data

Table 3 describes the 18 subjects who completed the study. Sixteen of these subjects were female. Fourteen (780/o) of the subjects had 8 to 10 errors on the Short Portable Mental Status Questionnaire, indicating severely impaired intellectual function; 2 subjects showed moderately impaired function (5-7 errors); 1 subject showed mildly impaired function; and 1 subject had no errors (intact intellectual function). All of, the subjects had Barthel Index scores below 60 (100 possible points), which Granger and colleagues[14] consider an indication of serious limitations in personal care independence. Thirteen subjects had scores below 40 (unquestionable severe disability), and 12 subjects scored below 20 (total dependence). Twelve subjects (71%) were nonambulatory, and the remainder required supervision (1 subjet), minimal assistance (3 subjects), or moderate assistance (1 subject) to ambulate. None of the subjects were independent in ambulation. All of the ambulatory subjects used a walker for support but had varied endurance when ambulating from 7.6 to 91.4 m (25-300 ft). Most of the subjects were totally dependent for transfers (530/o) or required maximum assistance to pivot (240/o). Only 1 subject was independent with transfers, and the remainder required minimal to moderate assistance to pivot. Knee pain during passive extension was rated as present in 9 knees (500/o) on the side that received prolonged stretching and in 11 knees (610/o) on the side that received only PROM and manual stretching at the initiation of the study.

[TABULAR DATA 3 OMITTED]

Statistics Analysis

The t tests showed that, at the initiation of the study, the subjects who completed the study (n=18) were not different from those who withdrew or expired (n = 10) in ROM at the knee, ankle, or hip; torque; pain; Barthel Index scores; transfer ability; ambulation status; and ambulation endurance. Paired t tests also demonstrated no difference between sides in knee, hip, or ankle ROM or torque prior to initiation of treatment.

Table 4 shows the t-test results for the side that received prolonged stretching and the side that received only PROM and manual stretching over the duration of the study. The t tests showed no difference in knee ROM between sides at any measurement interval. The repeated-measures MANOVA showed no difference between sides in knee ROM at any interval and no difference in knee ROM between intervals for either side.

[TABULAR DATA 4 OMITTED]

Individual Data

At the beginning of the study, 10 of the 18 subjects had differences of 10 degrees or greater between the side that received prolonged stretching and the side that received only PROM and manual stretching (Tab. 5). These differences were in both directions. Similarly, at the end of the study, 8 subjects had differences of 10 degrees or greater between knees, with differences occurring in both directions. The differences in extension between knees were of similar magnitude and in the same direction at the beginning and end of the study for almost all subjects.

[TABULAR DATA 5 OMITTED]

At the end of the study, 58% of the knees (12 knees receiving prolonged stretching, 9 knees receiving only PROM and manual stretching) showed a gain in extension, 330/o of the knees (5 knees receiving prolonged stretching, 7 knees receiving only PROM and manual stretching) showed a loss of extension, and 80/o of the knees (1 knee receiving prolonged stretching, 2 knees receiving only PROM and manual stretching) showed no change. Most of the end-study gains and losses in extension (790/o), however, were less than 10 degrees, and 580/o of the gains and losses were 5 degrees or less.

Two subjects showed bilateral changes greater than 10 degrees, one with bilateral gains (subject 8) and the other with bilateral losses (subject 16) (four knees). Only one subject (subject 11) showed what might be considered a positive treatment effect.

Cost

The cost of using the Dynasplint includes splint rental or purchase; therapist time for initial evaluation and splint fitting; therapist time for at least monthly monitoring of ROM, functional status, and splint fit; and possibly physical therapy aide time for daily application and removal of the splint (approximately 15 minutes per day per nursing home resident). In nursing homes in which restorative nursing care is assigned to specially trained nursing assistants, the task of applying and removing the splint could be done by the nursing assistants and not require a therapy charge.

The Medicare benefits of our clients covered the cost of renting the Dynasplint at the time of the study. In 1991, at the time of data collection, the Medicare allowable charge for rental of the Dynasplint was $150 per month and $721 for purchase of the unit. Current Medicare regulations require that functional gains be made and documented in order to receive continued reimbursement for the Dynasplint.

Discussion

Our study does not support the use of the splint for 3 hours per day, 5 days per week, at an average setting of 6 (62 kg-cm) as an effective treatment for decreasing knee flexion contractures in elderly nursing home residents any more than a traditional twice-a-week program of PROM and manual stretching. Our study, however, must be viewed as a pilot study because the number of subjects in our sample lacked statistical power (beta=.90, N = 18) to differentiate fully whether the splint, at this dosage, is ineffective. Post hoc power analysis indicates that 160 subjects would be needed in this study to conclude, with 800/o confidence, that no change occurred with the experimental treatment with the effect size we observed.

We believe there were multiple strengths in the study design, which give the results clinical relevance (ie, each subject had bilateral contractures and served as his or her own control, treatment was randomly assigned to either leg, the subjects showed wide variation in the initial severity of the contractures, the study was conducted at three different sites, different therapists performed treatment and measurement, and the investigators did not do the actual treatment or data collection). In addition, on average, there was no difference in knee ROM between the side that received prolonged stretching and the side that received only PROM and manual stretching prior to adding the treatment.

Recognizing that group averages may not be helpful in guiding individual treatment decisions, we felt that study and presentation of the individual subject data was important. The group data showed no difference in extension between knees at any interval. Individual analysis showed that more than half of the subjects had differences of 10 degrees or greater between sides at the beginning of the study. Bilateral knee flexion contractures appear as likely to be of different magnitudes as to be the same. Available diagnoses did not explain this side-to-side difference in knee flexion contractures, except possibly in 2 subjects who had a history of hip fracture on the side of the greater knee flexion contracture. Other subjects with unilateral diagnoses, such as hip fracture (n=3) or hemiparesis (n=5), did not have greater knee flexion contractures on the side of the unilateral problem.

In interpreting the clinical significance of change in knee extension across time, measurement reliability must be considered. The high reliability of our raters in measuring knee extension and the good reproducibility of knee extension when measured across our baseline intervals gave us confidence in our measurers and gave an indication of the stability of knee extension prior to beginning the treatment. Because repeated measures of knee extension within one session could differ as much as 15 degrees, gains or losses greater than 15 degrees can be considered true change. This cutoff point for true change is probably very conservative, because the average difference between repeated measures of knee extension was only 2 degrees.

Interestingly, in a similar study, Light and colleagues7 also used 15 degrees as the test for change in knee extension over a 4-week intervention period. Using the 15-degree criterion, only three subjects showed change in knee extension over time in our study. Subject 8 had a bilateral gain in knee extension, subject 16 had a bilateral loss of knee extension, and subject 15 gained extension on the side that received PROM and manual stretching only. Like many of the subjects in the study, these three subjects were nonambulatory and had severe cognitive impairment, Barthel index scores below 20 points, and poor transfer skills (subjects 8 and 16 were totally dependent, and subject 15 could pivot only with maximum assistance). Of these three subjects, only subject 15 improved in any measure of function, and this improvement was of questionable significance (ie, at the final measurement interval, his ambulation rating progressed from nonambulatory to ambulatory with maximum assistance of two persons). Like die three subjects who met the 15-degree test for change in knee extension, subject 11 (with what might be considered a positive treatment effect) was nonambulatory and totally dependent for transfers, showed severe cognitive impairment, had a Barthel index score of 0, and showed no functional improvement over time. The bilateral change of similar magnitude in subjects 8 and 16 suggests that change, in either direction, may be occurring in spite of treatment. For most of the subjects, gains or losses in knee extension were of similar magnitude and direction, regardless of the similarity in knee extension between sides at the beginning of the study.

Measures of functional change are also important in interpreting the meaningfulness of change in knee extension across time. A change in ROM is of questionable value without a coincident improvement in function. Subject 15 is the only subject who improved in ambulation, but this improvement was only one rating level, from nonambulatory to requiring maximum assistance. Only two subjects improved in transfer ability during the study. Subject 1 improved from a score of 1 to 2 within the first month of intervention, and subject 10 improved from 2 to 3 within the fourth month. These improvements in transfer ability cannot be explained in terms of treatment or change in knee extension. With only one subject showing improvement in function along with gain in knee extension (and that gain on the side that received PROM and manual stretching), we find that neither the average data nor the individual data are encouraging for use of the experimental treatment protocol.

Our findings do not concur with those of Light et al,[7] although our study sample and design seem similar to theirs. Use of Buck's traction for 2 hours per day clearly resulted in gains in knee extension in their subjects, whereas 3 hours of splint use per day did not result in gains for our subjects. It may be that a static stretch is more effective than the dynamic prolonged stretch. Another factor that may account for each of the gains made in our subjects is the maximum tension setting used on the splints. Our maximum spring tension setting of 6, chosen for subject wearing tolerance, corresponds to a torque of 62.2 kg-cm; this torque is well below the mean and median torques required to maintain maximum knee extension in our subjects (Tab. 3). It is likely that higher splint tension settings and longer wearing times are necessary to achieve increases in knee extension in this population. Hepburn[8] reported an average of 490/o improvement in knee extension in four subjects (mean age=50 years) after 11 weeks of Dynasplint use. The average tension setting was 8, and wearing time was 8 to 12 hours per day in Hepburn's subjects. Whether elderly nursing home residents with severe cognitive impairment, who would rely on multiple caregivers for splint application, could tolerate and comply with these increases in torque and wearing time seems questionable. The question may be worth exploring, however, in individual case studies.

Another previous study[19] also suggests that stretching is only effective over a long period of time. Tardieu et al[18] studied children with cerebral palsy with plantar-flexor contractures. They found that children who showed no progression in contracture over a 7-month period were getting at least 6 hours of stretch per day to the soleus muscle. This stretch was achieved through daily activity as well as a therapy program; thus, it was sporadic over a 24-hour period. This required length of stretch certainly casts doubt on the possible benefit of ROM and manual stretch exercise at any frequency per week.

Conclusion

Use of the Dynasplint for 3 hours per day, 5 days per week, 'at a tension setting of 6 (62.2 kg-cm torque) did not clearly result in reduction of knee flexion contracture or improvement in function in this pilot study of 18 elderly nursing home residents with bilateral knee flexion contructures.

Further research is needed to assess the effectiveness of other interventions for managing knee flexion contractures in the geriatric population. This research needs to address variations in dosage for low-load, prolonged stretch methods such as use of the Dynasplint. The benefit of these interventions cannot be measured in terms of knee ROM alone, but also in terms of improved function.

Acknowledgments

We thank the physical therapy staff and administration of Shared Therapeutic Services Inc for assisting and supporting us in completion of this study; the nursing and social services staff of the three nursing homes for their cooperation; Don Neuman, Phd, PT, Marquette University, Milwaukee, Wis, for his consultation concerning this study; and Sheryl Kelber, Biostatistician, and Carol Porth, Phd, RN, School of Nursing, University of Wisconsin-Milwaukee, for their consultation on the analysis.

We also thank George Hepburn, PT, Dynasplint Systems Inc, for his suggestions, assistance, and training of three therapists in appropriate use and fit of the Dynasplint, and Dynasplint Systems Inc for provision of Dynasplints in two cases and for all costs above Medicare reimbursement (no supplemental insurances were billed).

(*) Dynasplint Systems Inc, 645 Baltimore Annapolis Blvd, Severna Park, MD 22146-3923. ([dagger]) Velcro USA Inc, 406 Brown Ave, PO Box 5218, Manchester, NH 03108. ([double dagger]) Fred Sammons Inc, PO Box 32, Brookfield, IL 60513-0032. ([section]) SPARK Instruments and Academics Inc, PO Box 5123, Coralville, TA 52241. ([parallel]) SPSS Inc, 444 N Michigan Ave, Chicago, IL 60611.

References

[1] Mollinger LA, Steffen TM. Knee flexion contractures in institutionalized elderly: prevalence, severity, stability, and related variables. Phys Ther. 1993;73:437-446. [2] Kottke FJ, Pauley DL, Ptak RA. The rationale for prolonged stretching for correction of shortening of connective tissue. Arch Phys Med Rehabil. 1966;47:345-352. [3] Wynn Parry CB. Stretching. In: Basmajian JV, ed. Manipulation, Traction, and Massage. 3rd ed. Baltimore, Md: Williams & Wilkins; 1985:157-171. [4] Kottke FJ, Lehmans JF. Krusen's Handbook of Physical Medicine and Rehabilitation. 4th ed. Philadelphia, Pa: WB Saunders Co; 1990. [5] Cherry DB. Review of physical therapy alternatives for reducing muscle contracture. Phys Ther. 1980;60:877-881. [6] Yarkony GM. Prevention and management of contractures. In: Kaptan PE, ed. The Practice of Physical Medicine. Springfield, Ill: Charles C Thomas, Publisher; 1984:526-537. [7] Light KE, Nuzik S, Personius W, Barstrom A. Low-load prolonged stretch vs high-load brief stretch in treating knee contractures. Phys Ther. 1984;64:330-333. [8] Hepburn GR. Case studies: contracture and stiff joint management with Dynasplint. J Orthop Sports Phys Ther. 1987;8:498-504. [9] Richard RL, Jones LM, Miller SF, Finley RK. Treatment of exposed bilateral achilles tendons with use of the Dynasplint: a case report. Phys Ther. 1988;68:989-991. [10] MacKay-Lyons M. Low-load, prolonged stretch in treatment of elbow flexion contractures secondary to head trauma: a case report. Phys Ther. 1989;69:292-296. [11] Hepburn GR, Crivelli KJ. Use of elbow Dynasplint for reduction of elbow flexion contractures: case study. J Oythop Sports Phys Ther. 1984;5:269-274. [12] Richard RL. Use of the Dynasplint to correct elbow flexion burn contracture: a case report. J Burn Care Rehabil. 1986;7:151-152. [13] Mahoney FI, Barthel DW. Functional evaluation: the Barthel Index. Md State Med J. February 1965:61-65. [14] Granger CV, Albrecht GL, Hamilton BB. outcome of comprehensive medical rehabilitation: measurement by PULSES Profile and the Barthel Index. Arch Phys Med Rehabil. 1979;60:145-154. [15] Steffen TM, Meyer AD. Physical therapists' notes and outcomes of physical therapy. Phys Ther. 1985;65:213-217. [16] Pfeiffer E. A Short Portable Mental Status Questionnaire for the assessment of organic brain deficit in elderly patients. J Am Geriatr Soc. 1975;23:433-441. [17] Smyer MA, Holland BF, Jonas EA. Validity study of the Short Portable Mental Status Questionnaire for the elderly. J Am Geriatr Soc. 1979;27:263-269. [18] Shrout PE, Fleiss J. Intraclass correlations: uses in assessing rater reliability. Psychol Bull. 1979;86:420-428. [19] Tardieu C, Lespargot A, Tabary C, Bret MD. For how long must the soleus muscle be stretched each day to prevent contracture? Dev Med Childveurol. 1988;30:3-10.

TM Steffen, PhD, PT, is Director of Physical Therapy, Concordia University, 1'2800 N Lake Shore Dr, Mequon, WI 53092 (USA) ([e-mail]). Address all correspondence to Dr Steffen.

LA Mollinger, PT, is Physical Therapist, Omni Therapy, 1810 Kensington Dr, Waukeshil, Wl 53188.

This research was supported in part by a grdnt from the Foundation for Physical Therapy inc.

The study was approved by the utilization review committees of Clement Manor Health Center, Greenfield, WI, and Villa Clement Nursing Home, Milwaukee, WI, and by the Medical Director of Shorehaven Health Center, Oconomowoc, WI.

This article was submitted March 24, 1995, and was accepted June 12, 1995.