Department ofVeterans Affairs

Journal of Rehabilitation Researchand Development Vol . 30 No. 4, 1993Pages 436–442

CLINICAL REPORT

Research Device to Preproduction Prototype: A ChronologyW. Edwin Langbein, PhD and Linda Fehr, MS, Rehabilitation Research and Development Center,Edward Hines, Jr . VA Hospital, Hines, IL 60141

Abstract—This paper describes the evolution of theWheelchair Aerobic Fitness Trainer (WAFT), a wheel-chair ergometer developed for determination of aerobiccapacity and the diagnosis of coronary artery disease inlower limb disabled persons whose mobility dependsprimarily on the manual wheelchair . The device wasoriginally developed for use in research studies todetermine peak exercise capacity in persons with spinalcord injuries and other lower limb disabilities and toformulate associated graded exercise stress test protocols.In subsequent research, the device was incorporated intoa specially designed testing station for the detection ofcoronary artery disease in persons who cannot adequatelyundergo treadmill or cycle ergometry testing because oflower limb disabilities . Based on the usefulness of thedevice for both rehabilitation and diagnostic purposes,the WAFT has been brought into the technology transferprocess of the Department of Veterans Affairs Rehabilita-tion Research and Development Service . Under a contractwith Packer Engineering, Inc ., Naperville, IL, develop-ment of a preproduction version of the device and sixunits for field evaluation has commenced. Thepreproduction prototype of the WAFT has incorporatednumerous improvements over the original device andpromises to expand the potential for future research,rehabilitation, and diagnostic applications.

For further information, contact : W. Edwin Langbein, Edward HinesJr. VA Hospital, P .O . Box 20, Hines, IL 60141 . Phone : (708)343-7200Ext. 3526.The initial research and development was supported by the Departmentof Veterans Affairs Rehabilitation Research and Development Service,Washington, DC and by a grant from the Vaughan Chapter, ParalyzedVeterans of America.' Wheelchair Graded Exercise Test for Persons with Lower LimbDisabilities . Final Report on the Department of Veterans AffairsRehabilitation R&D Project No . B398-RZ. W.E . Langbein and M .H.Hwang, 1991 .

Key words : aerobic fitness, cardiorespiratory health,coronary artery disease, mobility impairment, spinal cordinjury, WAFT, wheelchair ergometer.

INTRODUCTION

A major focus of our research has been thecardiorespiratory health of persons with lower limbdisabilities . Access to a viable wheelchair ergometerthat could be used to simulate the full range ofexertional stress imposed by manual wheelchairpropulsion was a necessity . From the very begin-ning, the decision was made that the form ofexercise used in testing and conditioning must bephysiologically, biomechanically, and psychologi-cally specific to the primary mode of travel used bypersons with lower limb disabilities . We also feltthat the exercise test conditions for determination ofaerobic capacity and the diagnosis of coronaryartery disease for persons with lower limb disabili-ties should be equivalent to those currently offeredto the nondisabled . This is especially importantgiven that many wheelchair-dependent personsadopt a sedentary lifestyle . This fact was confirmedby data collected from 51 apparently healthy malespinal cord injured volunteers who participated inresearch conducted in our laboratory . Among thesesubjects only 10°7o reported engaging in any form ofregular physical activity greater than that requiredfor activities of daily living . Such lack of physicalactivity leads to significant decrements in physicalfitness and an increased risk of cardiovasculardisease.

436

437

CLINICAL REPORT :

Research Device to Preproduction Prototype

In response to the foregoing considerations, theWheelchair Aerobic Fitness Trainer (WAFT) wasdesigned and constructed at the Edward Hines, Jr.VA Hospital, Rehabilitation Research and Develop-ment Center (Rehab R&D) in Hines, IL . Specifica-tions used in the design and construction of the firstversion of the device included patient safety, dura-bility, appearance, ease of access and operation, andfootprint similar to that of a commercial treadmill.In addition, the WAFT was intended to accommo-date an individual's personal wheelchair, be accessi-ble by the wheelchair user with minimal assistance,and provide flexibility in setting workloads so thatthe degree of cardiovascular stress could be readilycontrolled, thereby enabling the testing and condi-tioning of persons with lower limb disabilities whopossess widely varied exercise capacities.

The WAFT evolved through several genera-tions, incorporating improvements in both mechani-cal and electronic features and addressing suchdesign challenges as roller configuration, method ofcreating progressive resistance, and mechanisms forwheelchair loading/unloading. The next sectionpresents a brief description of the last version of theWAFT. It is this version that has been used in ourlaboratory over the past 5 years (see Figure 1 andFigure 2) ; a more complete description appearselsewhere (1).

DEVICE DESCRIPTION

Wheelchair Aerobic Fitness Trainer : DevelopmentPhase One

The user/patient mounts the WAFT by backinghis/her wheelchair up a pair of ramps which aretilted at an 8° angle during loading and unloading(Figure 1) . The loading ramps are adjustable fordifferent sizes of standard, lightweight, and sportwheelchairs . The wheelchair is backed up the rampsuntil the rear wheels rest on three rollers built intoeach ramp . The rollers were configured to accept a61-cm-diameter wheel and provide an even distribu-tion of the weight of chair and rider duringwheeling . A lever on each ramp is pushed forwardto level the ramps and form a stable platform . Whenthe ramps are in the down position (Figure 1), theindentations used to position the front casters areclosed and the forward-most roller is locked so it

Figure 1.Wheelchair Aerobic Fitness Trainer—Phase One shown with theramps down in preparation for loading the wheelchair.

cannot turn . In the level position (Figure 2), the rearwheelchair wheels turn freely on the rollers and thefront wheels are secured in the open indentations.

Each rear wheel of the wheelchair is cradledand rotates on three weighted rollers, 16 cm long by10 cm in diameter (Figure 2). Variable brakingresistance is applied to each rear wheel via amagnetic eddy current braking assembly . To makethe brakes effective at speeds commonly used indaily wheelchair propulsion, it was necessary tomake modifications to the braking assemblies (1).Seven manually-selected resistance settings are avail-able, and the braking units may be disengaged sothat a wheelchair can be "free-wheeled" on the

Figure 2.Wheelchair Aerobic Fitness Trainer—Phase One shown inposition for exercise with computer and prototype echocardio-graphy imaging table .

438

Journal of Rehabilitation Research and Development Vol . 30 No. 4 1993

device. An inertial flywheel is factory-mounted onthe axle of each braking assembly . Based on thequalitative judgments of a number of manualwheelchair users who tested the configuration, theinertial flywheel, weighted rollers, and magneticeddy current brakes appear to produce conditionsthat simulate actual manual wheelchair propulsion.The resistance applied to each wheel may bemanipulated independently to accommodate theobjectives and physical abilities of the wheelchairuser . Workloads are created by (a) increasing theresistance setting; (b) increasing speed at a givenresistance setting; or (c) increasing both resistancesetting and speed.

A 286 IBM-compatible computer is interfacedto the WAFT to monitor wheel speed from which itcalculates power output, distance traveled, andprojected wheelchair heading (Figure 2) . The com-puter also times the duration of each exercise stageand provides storage of individual performancerecords . Elapsed time, speed of each wheel, andprojected heading are displayed on the color moni-tor of the computer in both digital and graphic formas motivational feedback to the user/patient . Wheelspeed for each wheel is also displayed on a pair ofsmall analog panel meters.

A series of speed-torque calibration experimentsenabled us to specify the relationships among wheelspeed, resistance setting, and the power used over aninterval of time (1) . The power output relations wereused to develop exercise test protocols to accommo-date differing levels of cardiorespiratory fitness (see"Research/Clinical Application of Device," p. 439).These protocols and the WAFT are currently beingused in an ongoing study to evaluate the clinicalusefulness of a new technique for the detection ofcoronary artery disease (CAD) in persons whocannot adequately undergo treadmill or cycleergometry testing because of lower limb disabilities.In this study, a specially designed imaging table isadjoined to the WAFT (Figure 2), and wheelchairergometry graded exercise testing is combined withdigital two-dimensional echocardiography.

Given the commercial viability of the device forboth rehabilitation and diagnostic purposes, a Re-quest for Evaluation was submitted to the Depart-ment of Veterans Affairs, Rehabilitation Researchand Development Service, Technology Transfer Sec-tion (Baltimore, MD) to bring the WAFT into theServices's technology transfer process . Following

review by a panel of clinicians, the request wasapproved, contract performance specifications weredeveloped, and a request for competitive proposalswas announced in appropriate publications . Aftercompetitive proposals were received, Packer Engi-neering, Inc ., Naperville, IL, was awarded a con-tract for development of a preproduction version ofthe device and six units for field evaluation . Theknowledge gained regarding design and functionduring Phase One development of the WAFT servedas the basis for the contract specifications for thepreproduction prototypes . Packer's project engi-neers have made extensive modifications to theoriginal design of the WAFT while meeting orexceeding the contract specifications.

Wheelchair Aerobic Fitness Trainer : DevelopmentPhase Two

The preproduction prototype WAFT utilizes anew dolly system designed to self-adjust to thewheelchair front caster width and to capture andimmobilize the casters (see Figure 3) . The wheelchairuser backs his/her chair up and over a 10° plateuntil the front casters contact guide bars whichmove two parallel slotted plates into position so thatthe casters drop into the slots . Several sets ofinterchangeable plates are provided to fit variouscaster sizes . The casters are strapped to the plateswith Velcro TM straps as an additional safety precau-tion and to keep the chair securely anchored duringexercise . The caster plates are attached to trackslocated on each side of the ergometer frame . As thewheelchair user backs up a 5 .5° main ramp, thecasters, secured in the plates, follow a straight-linecourse up the ramp until the chair is level and therear wheels drop into indentations that house theergometer rollers . A two-position momentary switchoperates a linear actuator which locks the positionof the dolly at each side of the ergometer . To ensurethat wheel and roller meet at a 90° angle regardlessof wheel diameter and camber, a second switch-activated linear actuator brings the rollers forwardunder the wheels . To unload the wheelchair, thesteps in the loading sequence are reversed ; the speedof descent must be controlled by the patient.

The preproduction prototype WAFT uses a pairof weighted aluminum rollers, 6 .5 inches in diameterand 10 .5 inches in length. Computer-controlledmagnetic particle brakes are used to create a variableresistance. A 386, 33 MHz IBM-compatible corn-

439

CLINICAL REPORT:

Research Device to Preproduction Prototype

r

Figure 3.Wheelchair Aerobic Fitness Trainer—Phase Two shown fromthe front with caster dolly in loading position, rollers disen-gaged, computer monitor moved aside for loading.

puter (software by Computer Application Special-ists, San Diego, CA) is interfaced to the WAFT.Based on the parameters of an operator-specifiedprotocol, the computer controls the timing of eachexercise stage and sends a controlling signal to theparticle brakes to set the desired resistance at thebeginning of each stage . The computer also samplesspeed and direction signals for each wheel . Thesesignals are produced by digital encoders mounted onthe axle of each roller.

The WAFT's computer and keyboard aremounted at the side of the ergometer for easy accessby the clinician. The monitor is attached to aflexible arm to allow optimal positioning of thescreen for the exercising patient (see Figure 3 andFigure 4) . The computer stores (by patient name andidentification number) the test/conditioning proto-col for each exercise session, including duration ofexercise, target speed, and resistance level for eachexercise stage . The clinician/technician may inter-

Figure 4.Wheelchair Aerobic Fitness Trainer—Phase Two shown fromthe side with caster dolly in locked position, computer monitorpositioned for patient viewing ; clinician access to keyboard andactuator control switches shown.

rupt an exercise session for any reason with akeystroke and may resume after a pause of anylength without loss of information.

During exercise, the SVGA color monitor onthe computer provides the wheelchair user withnumeric and graphic feedback . The display showscurrent exercise stage and resistance setting, actualversus target speeds for each wheel, accumulateddistance, actual versus target expenditure ofkilocalories, elapsed time for the current stage, totalelapsed time for the session, and an indication ofprojected wheelchair heading (see Figure 5) . At theconclusion of exercise, an exercise performancereport is stored under the patient's name and may berecalled for review or printing by the clinician at anytime. The report includes duration, brake setting,average wheel speed, and power output for eachexercise stage.

RESEARCH/CLINICAL APPLICATION OFDEVICE

Empirical evidence used to specify the operatingcharacteristics of the WAFT was generated from aseries of projects . In the first project, the results ofsuccessive calibration experiments were used toestablish two wheelchair exercise test protocols . Oneprotocol was intended for patients with lower limbdisabilities and symptoms of coronary artery diseaseand/or apparently healthy persons with low initial

440

Journal of Rehabilitation Research and Development Vol . 30 No . 4 1993

Figure 5.Wheelchair Aerobic Fitness Trainer—Phase Two motivationalfeedback screen showing exercise stage duration and brakesetting, actual and target speeds for each wheel, projectedwheelchair heading, cumulative and target kilocalories, distancetraveled and elapsed time for current stage, and total forsession.

levels of physical fitness ; a second for apparentlyhealthy persons with average to above-average initiallevels of physical fitness . Rate of incremental workloads distinguished the protocols.

Phase One Wheelchair Aerobic Fitness Trainer vs.Arm Crank Ergometry

The purpose of the second project was to (a)establish the validity of the WAFT for determiningthe peak exercise capacity of a heterogeneous sampleof persons with spinal cord injuries and other lowerlimb disabilities ; (b) compare the peak metabolicresponses of persons with lower limb disabilitieswho completed stress test protocols using continu-ous arm crank ergometry (ACE) and continuous andintermittent wheelchair ergometry (WCE) ; and, (c)analyze the differentiated ratings of perceived exer-tion given by lower limb disabled subjects duringdifferent modes of upper body exercise . Fifty-onemales, 17 to 69 years old—with quadriplegia, para-plegia, amputation(s), or lower limb fractures—wereassigned to three experimental groups : upper-levelinjury (ULI), C5 - T3 (n = 12); mid-level injury(MLI) T4 - T 10 (n =19); and lower-level injury(LLI), below T t® or lower limb fractures (n=20).All subjects completed three maximal graded exer-cise test protocols, one intermittent and one contin-

uous on the Phase One WAFT, and one continuouson an arm crank ergometer.

Experimental group means for peak poweroutput (p< 0 .0001) and tidal volume (p< 0 .009)were significantly greater during ACE than WCE,whereas peak measures of minute ventilation(p< 0.036) and breathing frequency (p< 0.0001)were significantly greater during WCE than ACE.There were no significant mean differences betweencontinuous and intermittent WCE stress test proto-cols . Significant between-group differences werefound for (a) peak measures of oxygen uptake ; (b)minute ventilation ; (c) tidal volume ; (d) heart rateand rate pressure product for continuous WCE andACE; and oxygen uptake, minute ventilation, tidalvolume, breathing frequency, heart rate, and ratepressure product for the WCE intermittent protocol.

Correlation coefficients between peak measuresof oxygen uptake, minute ventilation, and heart ratefor WCE and ACE were 0 .91, 0 .85, and 0.95,respectively (p<0 .001) . In addition, 88 to 90 percentof the total variance between differentiated RPE forcontinuous WCE and ACE was accounted for inthese analyses . This finding and the strength of theassociation observed between the criterion ACEexercise test and the experimental WCE was inter-preted as empirical evidence that the Phase OneWAFT and graded exercise stress test protocolprovide a valid measure of both the physiologicaland psychophysiological factors that contribute tothe peak functional capacity of persons with lowerlimb disabilities.

The measured increases in metabolic equiva-lents (METS) with increasing stages of the wheel-chair stress test follow the general principles forexercise testing proposed by the American HeartAssociation and American College of Sports Medi-cine (2,3) . Additionally, a substantial portion of thisdata has been used to set normative standards foraerobic fitness for persons with spinal cord injuriesand other lower limb disabilities .'

A third project focused on the hemodynamicresponses of lower limb disabled subjects to WCEand ACE. Table 1 presents age, peak rate pressureproduct (RPP), and peak heart rate values for 76lower limb disabled patients tested in the authors'

'Wheelchair Graded Exercise Test for Persons with Lower Limb Disa-bilities . Final Report on the Department of Veterans Affairs RehabR&D Project No . B398-RZ . W .E . Langbein and M .H . Hwang, 1991.

Deportment of Veterans Affairs

MPHto

5

®

3 .0

3 .5LEFT RIGHT

DISTANCE ELAPSED

Stage

0.05 MilesTo :.a1

CALORIES-50.0

-25 .0

NameBRAKE

LOW LOWRIGH

NumberMNMMMIMII

20 .4

441

CLINICAL REPORT :

Research Device to Preproduction Prototype

Table 1.Mean (± standard deviation) for age, rate pressure product (RPP) (RPP = heart rate x systolicblood pressure), and peak heart rate during maximal wheelchair and arm crank exercise performedby apparently healthy persons with lower limb disabilities.

Wheelchair ErgometryParameter

Upper Level

Lower LevelInjuries (n = 16)

Injuries (n = 55)and Fractures

(n = 5)

Arm Crank ErgometryUpper Level

Lower LevelInjuries (n= 15)

Injuries (n=42)and Fractures

(n = 5)

Age

32 (3: 11 .0)

41 (± 10 .9)

RPP

11912 (± 3308)

25281 (± 4690)

Heart Rate

120 (± 20)

166 (± 22)

31 (± 10 .7)

41 (±10 .9)

11383 (±3071)

24851 (±4166)

118 (±20)

167 (±22)

Upper Level Injuries C 5 -T2 ; Lower Level Injuries T3 -L5 and Fractures

laboratory . 2 Peak RPP were slightly greater forWCE exercise tests as compared to ACE exercisetests, indicating equal or greater potential forinducing myocardial ischemia . It should be notedthat peak RPP is probably underestimated for bothWCE and ACE because blood pressure was mea-sured immediately upon cessation of exercise.

We are currently conducting pilot wheelchairexercise echocardiography (WC EX + ECHO) usinga unique testing station developed at the HinesRehab R&D Center (see Figure 2) . Symptom-limitedwheelchair graded exercise tests have been com-pleted by 41 lower limb disabled patients withknown or suspected CAD (Table 2) . The mean RPPand heart rates shown in Table 2 were derived fromdata on all patients who took the wheelchair exercisetest . As a result, the variability was large . Under thecondition, "exercise tests with angiography," pa-tients were frequently stopped by the supervisingcardiologist due to signs and/or symptoms wellbefore reaching predicted maximal values . Addition-ally, data included patients who were, at the time oftesting, taking medications that lowered blood pres-sure and/or heart rate during exertion (i .e., betareceptor antagonists and calcium channel blockers).

Within two months of the WCE exercise test,14 patients underwent coronary angiography . Ofthese 14 individuals, two had angiography prior toWCE testing and were referred to our laboratory forevaluation of antianginal therapy . Thirteen of the 14angiograms showed significant CAD with �- 50

2 Unpublished additional observations made as part of an extension ofRehab R&D Project No . B398-RZ, not included in Final Report .

percent reduction in luminal diameter of at least onemajor epicardial coronary artery . Of these 14 cases,there were 10 true positive, one true negative, twonondiagnostic, and one false negative exercise tests.Thus, the overall diagnostic accuracy was 79 per-cent . The patient with a false negative test was beingtested for evaluation of antianginal therapy . Reviewof the Hines VA Hospital patient records revealedno cardiac events in those patients (16 negative and11 nondiagnostic) not completing angiographic as-sessments (follow-up 1 to 12 months) . These prelim-inary results from our WCE exercise test comparefavorably with the findings for lower body exerciseas reviewed by DeTrano and Froelicher (4) . Basedon these observations, we have concluded that oursystem of exercise testing may offer a viable alterna-tive to more costly pharmacologic stress testingprocedures for diagnostic evaluation of selectedpatients with lower limb disabilities.

SUMMARY AND CONCLUSIONS

The Phase Two WAFT incorporates severalimprovements relative to its predecessor . The inno-vative caster capture system and self-centering roll-ers make the updated device accessible to almost anywheelchair except three-wheeled racing chairs ; futuregenerations will be adaptable for racing chairs.Given the relatively low incline of the loading ramp,our pilot tests revealed that, with the exception ofsome persons with quadriplegia, most subjects wereable to mount and dismount the WAFT with aminimum of assistance. Subject safety is maximized

442

Journal of Rehabilitation Research and Development Vol . 30 No . 4 1993

Table 2.Mean (± standard deviation) for age, peak rate pressure product (RPP = heart rate x systolicblood pressure), peak heart rate, percentage of predicted maximal heart rate [(220-age) peak heartrate] during symptom limited wheelchair exercise (WCE) performed by persons with lower limbdisabilities and suspected CAD.

ParametersWCE Test and

Angiography (n = 14)

WCE Test andAngiography with

NondiagnosticTests Removed

(n =12)

WCE Test WithoutAngiography (n = 27)

Age (yr)

RPP

Peak Heart Rate

olo Predicted HRn,a,,

64 (1:6 .6)

64 (±6 .8)

65 (±8 .0)

20920

129

83

(± 6726)

(± 27)

(± 16)

22398

134

86

(± 6033)

(±26)

(-1-16)

21132

128

82

(+ 5520)

(±24)

(± 14)

by the electronically interlocked system of switches(i .e ., the rollers cannot be disengaged until the dollyassembly is in locked position) . Though theseswitches must be operated by a clinician or techni-cian, the exercise session itself, once initiated, iscompletely computer-controlled and no further in-tervention is required . Finally, the braking assemblyused in the Phase One WAFT configuration offered7 manually-selected resistance settings . While thesoftware incorporated in the updated WAFT allowsthe clinician/technician a choice of only threeresistance settings (low, medium, and high), thesedesignations may be recalibrated at any time via thecomputer to produce any useful combination ofincreasing resistance levels (i .e ., protocols employ aselection of discrete stages along a continuum ofworkloads) . This feature allows the exercise proto-cols to be tailored to a wide variety of capabilities . Italso has great future potential for research in theinvestigation of optimal test protocols and theevaluation of physiological responses as they relateto a workload ramp.

As progress on the Phase Two WAFT contin-ues, the Phase One device remains in use . Increasingnumbers of WAFT exercise tests have been re-quested by clinicians whose lower limb disabledpatients present symptoms of coronary artery dis-ease. Moreover, previously undocumented CAD wasfound in several apparently healthy individuals whovolunteered for our research . Early intervention inthese patients may have profound effects on long-

term quality of life and future medical expenses . Ina relatively short time the WAFT, with exercisestress echocardiography, has had a significant im-pact on the health care provided to veterans.

ACKNOWLEDGMENTS

Charles J . Robinson, DSc, Department of RehabilitationScience and Technology, School of Health and Rehabili-tation Sciences, University of Pittsburgh, Pittsburgh, PA,was responsible for the electronic design of the Phase OneWAFT. Lutz Kynast, Atlanta VA Rehab R&D Center,Decatur, GA, was responsible for its mechanical designand fabrication.

REFERENCES

1. Langbein WE, Robinson CJ, Kynast L, Fehr L . Calibra-tion of a new wheelchair ergometer : the wheelchairaerobic fitness trainer . IEEE Trans Rehab Eng 1993:1 :49-58.

2. The Preventive and Rehabilitative Exercise Committee ofthe American College of Sports Medicine . Guidelines forgraded exercise testing and prescription . 4th ed . Philadel-phia : Lea & Febiger, 1986.

3. Fletcher GF, Froelicher VF, Hartley LH et al . Exercisestandards: a statement for health professionals from theAmerican Heart Association . Circulation 1990: 82:2286-322.

4. DeTrano R, Froelicher VF .Exercise testing: uses andlimitations considering recent studies . Prog CardiovascDis 1988 :31 :173-204.