Technical Report AMI-004:

Accessible Power-Assist Hospital Bed Back Angle Controller: Review of 2006-2007 RERC-AMI National Student Design Competition

Author: Erin Promersberger, B.S. (M.S. student)

Coordinating Editor: Jack Winters, Ph.D.

Location: Medical Device Accessibility & Usability Laboratory

Current Version: 1.0 (August 2007)

Table of Contents

Executive Summary

The Rehabilitation Engineering Research Center on Accessible Medical Instrumentation (RERC-AMI) has sponsored the national Student Design Competition since 2002.  As part of the 2006-2007 competition, 8 teams designed accessible power-assist hospital bed back angle controllers for 6 fictional clients (“personas”) with various disabilities.  It was specified that the prototype:

A team of judges selected the top three designs.  This report:

introduces a complementary human subjects study by this author as part of a Master’s thesis.

Background

Most examination tables and hospital beds available today are operated through the use of momentary contact switches, which can be difficult for people with physical impairments such as tremor or muscle weakness to use.  The RERC-AMI has documented access barriers for such equipment through a national survey [1], focus groups [2], and usability studies [3].  The motivation is manifold, as reflected in the introductory sections of the design reports provided for this competition.  For instance, one team (St Louis University) noted that in the United States there are 1.5 million people that have Parkinson’s disease [4], 800,000 people that have Cerebral Palsy [5], and 40 million people with arthritis and other rheumatic diseases that cause swelling and stiffness in joints [6]. 
The concept of extended physiological proprioception (EPP) suggests that a person can perceive the static and dynamic characteristics of a device such that with experience over time, it becomes used as if it was an extension of that person.  Simple examples of this concept include the use of a pencil or a tennis racket: the operator of such a device is aware of its location in space without having to look at it, and indeed their control strategy may even depend on contact forces between the device and its environment.  This concept has been used successfully as a prosthesis control technique [7].  In the context of mechanics, a key component of EPP systems is that both force and position (and thus velocity) be transmitted across the human-technology interface, and furthermore in a way that can become predictable to the person. In the context of accessibility, research suggests that EPP may improve the usability of medical devices with powered degrees of freedom such as examination tables and hospital beds [8].
During the 2006-2007 academic year, the overall National Student Design Competition sponsored by the RERC-AMI included 25 teams from 16 universities, competing in three categories. Eight of these engineering senior design teams from across the country competed in the category that is the focus on this report, on designing accessible power-assist hospital bed back angle controllers.  The problem, specific aims, specifications, and clients for the accessible back angle controller problem are shown below (which was provided to all teams):
Problem: Powered hospital beds often use open-loop controls to set the back angle, which typically moves at a constant slow velocity while an operator presses a momentary contact switch. The concept of extended physiologic proprioception (EPP) suggests that users might benefit from a more intuitive approach of manually grasping a handle on the side of the upper portion and using force-assist concepts to set the back angle (maybe with the assistance of anti-gravity mechanisms), in which motion speed increases with the amount of force applied to the handle.
Aim:  Design a reliable, easy-to-use, EPP-based power-assist back angle controller for a prototype platform (or hospital bed or exam table).
Specs: The operating handle should be mounted on side edge of the bed frame, and easy for a caregiver to reach and grasp.  The weight of the back of a patient on the bed will range up to 180 lbs (45% of a 400-lb patient). Operator force will vary normally, in the range from 1 to 20 lbs, but with a caregiver capable of only 5 lbs able to fully operate it (but then at lower speeds). The bed must be stable during and after adjustment, and safe if power is lost.
Clients: Mat, Akiko, Jorge, Lakisha, Sani, Dolores, Tyler
The clients are defined in the chart below:


Table 1. Patient Positioning Aid Personas

Client

Brief Description

Mat

Tal, Blind, experienced small stroke, dislikes technology

Akiko

Low vision, uses infusion pump for treatment

Jorge

Chronic back pain, hard of hearing, mild carpal tunnel syndrome

Lakisha

Parkinson’s disease, limited mobility and dexterity, tremors, experimental infusion therapy

Sani

Partially paralyzed right side, uses infusion pump, vital signs monitored

Dolores

Deaf, severe arthritis, infusions due to heart problems

Tyler

Infusions due to hemophelia, helps Dolores sit up in bed
See http://www.rerc-ami.org/ami/projects/d/2/2/announce/index.aspx for a complete description of these personas and rules for the 2006-2007 competition. The maximum budget for each project was $2,000.

Product Table

The product table compares key features of the eight designs entered in the design competition.

The first two rows give the name and manufacturer (design team) of the device.  This is followed by cost information.  Many teams gave an expected implementation cost in addition to the development cost, so both were included in the table.

The next four rows refer to the components used in each team’s design: bed, handle, power, and control.  The subsequent five rows give the following information on the performance of the device, when it was available: maximum force/torque, motor speed, maximum speed of device, angle range, and weight limit.

There are then eleven rows that are based on the problem statement of the design competition.  The first eight are yes/no answers; the last three are ranked on a scale of 1-5 in which a “5” completely satisfies the criterion and a “1” does not satisfy the criterion.  The criteria are: ability to support 180 lbs, safety in case of power outage, EPP-based for patient (patient must receive both force and position feedback from the device), EPP-based for caregiver (caregiver must receive both force and position feedback from the device), handle mounted appropriately, speed proportional to input force, allow operator force of 1-10 lbs, 5 lbs can fully operate, ease of use (1-5), stability during movement (1-5), and stability maintaining position (1-5).  The method of maintaining position is also specified.

The last row is populations to address in the future.  If a device is not accessible to people with a specific impairment, a suggestion is made on how the device can be improved to be more usable to that population.  For a list of impairment abbreviations see Appendix A.

Survey of Prototypes

Columbia University

Columbia University designed the ProprioAngle Controller.  This design utilizes a speed control motor unit purchased through Oriental Motors and a Hill-Rom Advanta hospital bed, donated by a nearby hospital.  The motor can output 50 lb-in of torque at speeds ranging from 20-480 rpm.  The Advanta bed uses a worm gear with a gear reduction ratio ranging from 1:20 to 1:300.  The handle, made of 1/2” PVC wrapped in vinyl, pivots to mimic the movement of the bed.  It is mounted on the side of the bed to allow the operator to know both speed and position at any point, thus satisfying EPP.  There is a contact switch on the backside of the handle to prevent motion if the handle is accidentally bumped.  Compression springs (k=3.8 lbs/in) are used to oppose the force applied by the user.  Each spring is connected to a potentiometer; a switch is used to determine which potentiometer controls the signal to the motor.  A questionaire was used for evaluation of the device.  The team reported that all users gave high marks to all questions (including comfort of handle, ease of use, and perceived safety) [9].

Photo of
Figure 1. Columbia University Prototype

Photo of

Figure 2. Close-up of Columbia's Handle
Indiana University – Purdue University Indianapolis

Indiana University-Purdue University Indianapolis (IUPUI) designed a system using the Actionjac 90VDC linear actuator.  This actuator is capable of providing 750 lbf at 0.6 in/sec and has a self-locking mechanism, which would hold the bed in place in the case of a power outage.  A joystick is connected to two extension springs to provide force feedback to the operator and to allow variable speed.  Additionally, a control button would be placed at the caregiver station to allow remote control of the bed.  This team reported that a team of biomedical engineers reviewed the final design and was satisfied with the outcome [10].

Photo of
Figure 3. Testing conducted by IUPUI

Milwaukee School of Engineering

The team from Milwaukee School of Engineering (MSOE) incorporated a standing-assist system to allow patients to better position themselves for standing and keep them from falling backwards as they stand (see Figures 4, 5).  A small joystick is used to control both the seatback and the standing assist system.  The joystick is connected to potentiometers, whose signals are fed into an RHV series Pulse Width Modulator control, purchased from RAE Corporation.  This controller was used to vary the motor speed based on the input force from the user.  Testing was performed, but no specifics were given in this team’s report [11].

Photo of
Figure 4. Computer model of MSOE design


Photo of
Figure 5. MSOE Prototype

San Diego State University

San Diego State University’s design involved a linear actuator and a pulse width modulation (PWM) controller (Figure 6).  An extension spring was used to provide resistance (force feedback) to the operator.  The device can be fully operated using less than 5 pounds of force.  The force on the handle was measured and found to be within 5% of the expected value, based on the analysis.  The maximum load was also tested [12].

Photo of
Figure 6. San Diego State University Prototype

St. Louis University

St. Louis University’s design incorporated two DC servo motors and gear trains recycled from a Microsoft Sidewinder force-feedback joystick.  The new joystick was mounted to the side rail of a Hill-Rom Model 820 adjustable hospital bed (Figure 7).  A 12VDC electric gear motor was coupled to the existing AC motor to enable variable speed.  A microcontroller was used to interface all components.  The back angle is positioned by an actuator and measured using the Accustar electronic inclinometer.  Velocity of the bed is dependent upon the direction of movement, the bed angle, and force applied to the joystick.  Following testing it was determined that the bed was able to raise and lower with no load and able to support 180 pounds without moving, but the bed was unable to raise and lower with the 180 pound load.  Testing was also performed on the joystick, force feedback, and integration of the system [13]. 

Photo of
Figure 7. St. Louis University Prototype

Texas A&M University

Texas A&M University used a Micromax 3-phase AC motor to provide power, and an AC-Drive to control the motor.  A bed was donated by a nursing home.  This team designed a C++ program to work with a USB joystick.  This team reported that the seatback system shown in Figure 8 works for patients weighing more than 400 pounds, but no specific information about the testing was included in the final report [14].

Photo of
Figure 8. Texas A&M University Prototype
University of Connecticut

The team from the University of Connecticut designed a custom-made bed.  A DC motor was used with a pulse width modulator to allow for variable speed.  Torsion springs were used to provide resistance in the handle (Figure 9), and the motor operated on a telescoping jack (Figure 10).  A microcontroller converted the analog potentiometer signal (from the handle) into digital.  A safety button was incorporated to eliminate unintentional movement.  As shown in Figure 11, the bed was tested for blindness by covering the subject’s eyes with a blind-fold, one-handed operation by tying one of the subject’s arms down (handle can be moved to either side of the bed), and weakness/arthritis by having the subject operate with only the pinky finger [15]. 
Photo of
Figure 9. University of Connecticut Handle Design

 

Photo of
Figure 10. University of Connecticut Prototype

Photo of
Figure 11. University of Connecticut Testing (Blind)

  

University of Wisconsin - Madison

The team at University of Wisconson-Madison purchased an old Hill-Rom bed to begin their project.  A three-phase AC induction motor was purchased to provide power, and a Variable Frequency Drive (VFD) was used to control the AC voltage frequency.  This team provided two options for caregivers: direct control over motor speed and direction through the joystick, and “cruise control” buttons to allow the caregiver to choose a desired angle (0, 30, 60) without having to apply continual force.  An angle sensor was used to provide feedback.  The device shown in Figure 12 was tested, but no specific information was provided in the written report [16].

Photo of

Figure 12. University of Wisconsin-Madison Testing

Evaluation of Designs

In June 2007 a group of three judges evaluated each design, and awards were given to the highest-scoring designs.  The design by the team from Connecticut won first place in the competition with the Adjustable Back Angle Controller.  Strengths of this design are the use of a scissor jack to move the bed, a control switch to prevent unintentional movement, successful implementation of EPP for a person on the bed, and the extensive testing performed.  The report was also strong, with the design well documented. 
The design by the team from MSOE won second place.  The strengths of their design included the innovative standing-assist device and the sound mathematical analysis. 
The University of Wisconsin at Madison won third palce.  A key strength of this design is the ability to use two different control methods.

Recommendations

It may be beneficial to incorporate some of the innovative ideas from other schools into Connecticut’s design.  Some of the unique features included:

The design competition focused mostly on the physical aspects of the design of a human-technology interface that met the specifications, and less on the types of controller designs that might be effective for persons with a diversity of abilities.  As part of a Master’s thesis, this author is working on a complementary study focusing on the controls side of this problem, specifically evaluating user performance with various impedance controller algorithms, where the users include persons with disability.  The flowchart of Figure 13 illustrates this study.


block diagram of design of impedance controller
Figure 13. Flowchart of Proposed Human Subjects Study

This human subjects study involves approximately 15 people with disabilities.  Each subject will be asked to move a custom-made seatback to the following positions:
90° à 0° à 10° à -35° à 10° à 0° à 90° (Seatback is horizontal at 0°)
This sequence will be repeated three times for each of the following conditions (16 combinations):

Questionnaire and video data will be analyzed using the Mobile Usability and Accessibility Lab (MUA-Lab) System [17] that includes use of the MVTA video analysis software [18] to evaluate upper extremity movements made while performing the experiment.  A combination of performance and preferences will be used to evaluate each controller.
Results obtained from this study could be implemented for most of the designs from this competition, improving the effectiveness of EPP for the device.

References

[1] Winters, J.M.W., Story, M.F., Barnekow, K., Kailes, J,.I., Premo, B., Schwier, E., Danturthi, R.S., & Winters, J.M. (2007). Results of a national survey on accessibility of medical instrumentation. In Winters, J.M. and Story, M.F., eds., Medical Instrumentation: Accessibility and Usability Considerations. Boca Raton, FL: CRC Press, 13-28.

[2] Kailes, J.I. (2007). The patient’s perspective on access to medical equipment. In Winters, J.M. and Story, M.F., eds., Medical Instrumentation: Accessibility and Usability Considerations. Boca Raton, FL: CRC Press, 1-12.
[3] Lemke, M.R. (2005). The evaluation of three alternative methods for understanding biomechanical aspects of medical device accessibility. Unpublished master’s thesis, Marquette University, Milwaukee, WI.

 [4] "National Parkinson Foundation." 30 July 2007 <http://www.parkinson.org/NETCOMMUNITY/Page.aspx?&pid=201&srcid=201>. [St. Louis University]

[5] “National Institute of Neurological Disorders and Stroke.” 30 July 2007 < http://www.ninds.nih.gov/index.htm>. [St. Louis University]

[6] Strange, Carolyn J. "Coping with Arthritis in Its Many Forms." June 1997. FDA. 30 July 2007 <http://www.fda.gov/fdac/features/296_art.html>. [St. Louis University]

[7] Doubler PhD, James A., and Dudley S. Childress. (1984) "An Analysis of Extended Physiological Proprioception as a Prosthesis-Control Technique." Journal of Rehabilitation Research 2: 5-18.

[8] Winters, J.M. (2007). Future Possibilities for Interface Technologies that Enhance Universal Access to Healthcare Devices and Services. In Winters, J.M. and Story, M.F., eds., Medical Instrumentation: Accessibility and Usability Considerations. Boca Raton, FL: CRC Press, 321-340.

[9] “ProprioSystems”. Columbia University. 29 July 2007 <http://www.columbia.edu/~ra2125/>.

[10] “Accessible Power-Assist Hospital Bed Back Angle Controller”. Indiana University-Purdue University Indianapolis. 29 July 2007 <http://www.engr.iupui.edu/me/courses/designbed/index/index.html>.

[11] “Hospital Bed Controller”. Milwaukee School of Engineering. 29 July 2007 <http://myweb.msoe.edu/~kumpaty/RERConAMI/indexX.html>.

[12] “Accessible Power-Assist Hospital Bed Back Angle Controller”. San Diego State University. 29 July 2007 <http://sdsu-me490.tripod.com/>.

[13] “Bed Back & Beyond”. St. Louis University. 29 July 2007 <http://pages.slu.edu/org/macartrx/>.

[14] “Accessible Power-Assist Hospital Bed Back Angle Controller”. Texas A&M University. 29 July 2007 <http://pages.slu.edu/org/macartrx/>.

[15] “Adjustable Back Angle Controller”. University of Connecticut. 29 July 2007 <http://www.bme.uconn.edu/sendes/Spring07/Team8/index.htm>.

[16] “Hospital Bed Back Angle Controller”. University of Wisconsin – Madison. 29 July 2007 <http://homepages.cae.wisc.edu/~bcnelson/>.

[17] Winters, J.M., Rempel, D., Story, M.F., Lemke, M., Barr, A., Campbell, S. and Danturthi, S. The Mobile Usability Lab Tool for Accessibility Analysis of Medical Devices: Design Strategy and Use Experiences. In Medical Instrumentation: Accessibility and Usability Considerations (Winters, J.M., Story, M.F., eds.), Chapter 13, pp. 173-190, CRC Press, Boca Raton, 2007.

[18] Yen , T.Y., Radwin, R.G., Usability Testing by Multimedia Video Task Analysis, In Medical Instrumentation: Accessibility and Usability Considerations (Winters, J.M., Story, M.F., eds.), Chapter 12, pp. 159-172, CRC Press, Boca Raton, 2007.


Appendix A: User Abilities

User Abilities (Impairments) Reference:
Cognitive – CG
Vision, Blind – VB
Vision, Partial – VP
Hearing, Deaf – HD
Hearing, Partial – HP
Reaching (forward, above head, or over shoulder) – R
Standing – ST
Manipulation of hand – MH
Manipulation of feet – MF
Postural control of head – PH
Postural control of arms – PA
Postural control of legs – PL
Postural control of torso – PT
Walking/mobility - MO

Appendix B: Text Descriptions of Figures

Text Description for Figure 1
ALT Text: Columbia University Prototype.
Summary: This figure shows the prototype designed by Columbia University.
Long Description: This figure shows a hospital bed in a typical setting.  The moving portion of the bed (seatback) is positioned at an angle of approximately 45 degrees.  The customized handle (created from ½” PVC pipe) is surrounded by casing and attached to the side rail on the moving portion of the bed.  There are cupboards and a sink in the background.

Text Description for Figure 2
ALT Text: Close-up of Columbia’s Handle.
Summary: This figure shows a closer view of the handle designed by Columbia University.
Long Description: This figure shows a closer view of the handle designed by Columbia University.  It is a piece of ½” PVC pipe that is attached at one end with a pin joint, allowing it to rotate about that point.  At the other end there are compression springs to resist motion (and provide force feedback) if the handle is moved up or down.

Text Description for Figure 3
ALT Text: Testing conducted by IUPUI.
Summary: This figure shows the testing of the prototype designed by IUPUI.
Long Description: This figure shows the prototype created by IUPUI during testing.  It is a hospital bed with a polka dot sheet.  The angle of the moving portion of the bed (seatback) is at approximately 45 degrees.  A person in a suit is sitting in the bed with his left hand just above the handle, which is mounted near his left hip.  The handle has a white ball at the top to make it easier to grab.

Text Description for Figure 4
ALT Text: Computer model of MSOE design.
Summary: This figure shows a computer of the MSOE design.
Long Description: This figure shows a computer model of the design created by MSOE.  In this bed there is a hinge running the length of the bed, allowing one side to move up like a futon.  There are also three hinges across the width of the bed.  One is for the seatback, and the other two allow for adjustability under the legs and feet. 

Text Description for Figure 5
ALT Text: MSOE prototype.
Summary: This figure shows the prototype designed by MSOE.
Long Description: This figure shows the prototype designed by MSOE.  The bed is flat and there is no mattress.  The operator, shown from the waist down, is holding a small, black joystick which is used to control the seatback as well as the futon (explained in the previous figure).

Text Description for Figure 6
ALT Text: San Diego State University Prototype.
Summary: This figure shows the prototype designed by San Diego State University.
Long Description: The prototype designed by San Diego State University is shown in the next figure.  This shows a side view of a hospital bed without a mattress.  The seatback is at an angle of about 60 degrees from horizontal (closer to sitting up than lying down).  The custom-made joystick is on the floor next to the bed.

Text Description for Figure 7
ALT Text: St. Louis University Prototype.
Summary: This figure shows the prototype designed by St. Louis University.
Long Description: This figure illustrates the prototype designed by St. Louis University.  The joystick is shown in the center of the picture and is mounted on the right side of the bed near the side rail.  There are a mattress and a pillow on the bed.

Text Description for Figure 8
ALT Text: Texas A&M University Prototype.
Summary: This figure shows the prototype designed by Texas A&M University.
Long Description: This figure shows the prototype designed by Texas A&M University, which appears to be in a garage or a workshop.  The bed is a bare metal frame that is painted red with the word “AGGIES” printed at the top in white.  This bed has four sections (similar to that designed by MSOE) to allow more adjustment under the legs.  The joystick, which is a commercially available computer joystick, is on the floor next to the bed.

Text Description for Figure 9
ALT Text: University of Connecticut Handle Design.
Summary: This figure shows a close-up of the handle designed by University of Connecticut.
Long Description: This figure shows a close view of the handle designed by University of Connecticut.  There are two drawings side by side.  The drawing on the left side is a Back view.  There are three labels: torsion springs, potentiometer, and handle.  The drawing on the left side is a side view and shows the same components. 

Text Description for Figure 10
ALT Text: University of Connecticut Prototype.
Summary: This figure shows the prototype designed by University of Connecticut.
Long Description: This figure shows the prototype designed by University of Connecticut.  The seatback is positioned at approximately 45 degrees, with a scissor jack holding it in place.  The custom made bed has a padded surface and there is a dust ruffle around the bottom frame.  The handle is located on the right side of the bed near where the patient’s hip would be.  The top of the handle is red to make it easier to find.

Text Description for Figure 11
ALT Text: University of Connecticut Testing (Blind).
Summary: This figure shows the University of Connecticut’s protocol for testing the usability of their device by people with blindness.
Long Description: This figure shows the same prototype described in figure 11, except there is a person in the bed.  The person has a blindfold on to simulate use by people with blindness.  The person’s right hand is on the handle and the seatback is positioned at approximated 45 degrees.

Text Description for Figure 12
ALT Text: University of Wisconsin – Madison Prototype.
Summary: This figure shows the prototype designed by University of Wisconsin - Madison.
Long Description: This figured shows the prototype designed by University of Wisconsin – Madison.  There is a person in the bed, shown from mid-thigh up.  There are rails on both sides of the seatback.  The control is located on the left side near the person’s hip.  There is a joystick controller at the top with a large red knob, as well as three buttons illustrating the seatback horizontal, at 45 degrees, and vertical.

Text Description for Figure 13
ALT Text: Flowchart of Proposed Human Subjects Study.
Summary: This figure shows a flowchart that illustrates the proposed human subjects study.
Long Description: This figure illustrates the study proposed by this author.  It is a series of boxes and arrows illustrating the pathway of a signal.  Moving from the top left corner in a clockwise manner there are boxes labeled as follows: “Person”, “Force Sensor”, “A/D, Digital Filter”, “Impedance Controller”, “D/A, Amplifier, Filtering”, “Motor and Linkage System”, and “Data Collection”.  The “Impedance Controller” box is actually four rectangular boxes placed on top of each other labeled (from top to bottom): “Impedance Controller 1”, “Impedance Controller 2”, “Impedance Controller 3”, and “Impedance Controller 4”.  There are solid and dotted arrows connecting the boxes.  There is a solid arrow with a switch between the “Person” and the “Force Sensor” labeled with the word “Force”.  There is a solid line connecting “Force Sensor” to “A/D, Digital Filter” with a dotted line extending to “Data Collection”.  There is a solid arrow from “A/D, Digital Filter” to a bracket which branches to each of the four impedance controllers.  There is a solid line to Impedance Controller 1 and switches to the other three controllers.  There is a bracket at the other end which combines the four impedance controllers to a single path.  A solid arrow connects this bracket to “D/A, Amplifier, Filtering”.  A solid arrow connects “D/A, Amplifier, Filtering” to “Motor and Linkage System”.  There are there arrows extending from “Motor and Linkage System”.  Solid arrows go to the Person (labeled “Position, Velocity”), and to the bracket before the four impedance controllers.  A dotted line extends to data collection.

team]

Valid XHTML 1.0! Level Triple-A conformance icon, W3C-WAI Web Content Accessibility Guidelines 1.0