For some years, orthopaedic educators have awaited the development
of virtual reality technology in the hope that it might offer a suitable
means of teaching and evaluating skills proficiency. Such a tool
would greatly facilitate resident education while decreasing training
time in the operating room and reducing the likelihood of adverse
technical outcomes. For the practicing orthopaedic surgeon, it would
provide a means of maintaining surgical skills and learning new
techniques. While the promise of virtual reality technology has
been great, in 1996 the American Academy of Orthopaedic Surgeons
(AAOS) evaluated its status and determined that it was too soon
to commit the substantial resources necessary to successfully develop
it as a training and evaluation tool.
The American Board of Orthopaedic Surgery (ABOS) also has been
interested in this technology as a means of evaluating the surgical skills
of candidates for certification and recertification. In 1997, the
Board funded the development of a prototype simulator. This decision was
based on three tenets: first, that in order for this effort to succeed
it must be embraced by the entire orthopaedic community; second, that
the tool must be shown to be valid and reliable; and third, that
before it could be used as an adjunct to certification or recertification
examinations, orthopaedic surgeons must have had ample experience
with it and be confident that it is a realistic and useful surrogate
for actual operative interventions.
An analysis of the case lists submitted by 695 candidates taking
the 1999 part-II oral certification examination revealed that the
most common surgical intervention was knee arthroscopy (12.5 percent
of the total) and that arthroscopic procedures involving all joints (knee,
shoulder, ankle, and wrist) constituted 20 percent of the total.
From the Board's perspective, therefore, knee arthroscopy is not only
an increasingly common surgical procedure but also one that does
not lend itself readily to assessment of the examinee's skill in performing
it.
Arthroscopy of the knee is an ideal model for simulation because
the surgeon is already performing the procedure in a quasi-virtual
environment. The procedure is analogous to remote robotic systems
currently in use by the nuclear power industry and the space program. All
images of the operation are presented to the surgeon on a television
monitor that is separate from the operative field. The procedure is
accomplished through use of long-handled instruments that not only
operate at some distance from the surgeon's hands but also are out of
the surgeon's direct view.
In its requests to developers for proposals, the Board established
the requirements for a realistic simulator that would be acceptable
to orthopaedic surgeons. First, the knee must be part of an articulated
lower extremity that allows the surgeon to manipulate the hip and knee
appropriately; second, it must offer realistic graphics that change
in real time with flexion-extension and varus-valgus maneuvers; and
third, it must provide realistic force feedback so that the operator
can see and feel the structures of the knee.
A prototype simulator was developed by the ABOS and Boston Dynamics
and was presented to the AAOS Council on Education in early 1998
(Fig. 1).
The Council created a Task Force on Virtual Reality, comprising
AAOS and ABOS members, to review the status of virtual reality technology
and the ABOS effort to date and to put forward a plan to embark
on the next steps in the development of a viable simulator.
The Task Force developed a two-step approach. The first step
was to create a simulator that could be used for both training and
assessing a user's proficiency in performing a basic arthroscopic
examination of the knee (Table I). Factors to be assessed would
include the time required to complete the arthroscopic examination
of the knee, the thoroughness of the examination, the tissue damage
incurred during the procedure, and the user's ability to recognize
simulated lesions at any of the sites. For the second step, the
Task Force envisioned implementation of various learning modules
that would be employed to assess the user's proficiency in performing
certain procedures, such as repair of the anterior cruciate ligament
and resection of a torn meniscus.
The Task Force strongly recommended that the graphic representation
of the anatomy of the knee be as realistic as possible. To this
end, it collaborated with the Center for Human Simulation at the
University of Colorado, the developers of the Visible Human Project8-10, to provide exquisitely accurate
graphic representations of the structures of the knee (Fig. 2).
The Virtual Reality Arthroscopic Knee Simulator comprises six
basic elements.
Force feedback: The ability to recreate the
forces and torques experienced in an actual activity is known as haptic
feedback. By monitoring the position of the arthroscope with respect
to a mathematical model of the knee and its surrounding tissues (the
so-called synthetic environment), the system can recreate both a
correct view for the video display and the appropriate forces to
be transmitted back to the surgeon's instruments.
Hardware-controlled knee joint: The surrogate
leg model with built-in sensors connected to the computer should
allow the user to manipulate the leg in the same manner that an
orthopaedic surgeon manipulates an actual patient's leg. This includes
varus and valgus manipulation. Similar devices have been developed
for other arthroscopic simulator systems1,3,11.
Software and data elements: The software replicates
the visual, mechanical, and behavioral aspects of the knee through a
combination of control, modeling, and content programs. The control
program moderates the haptic interface and interacts with the modeling
software to indicate when the user has collided with a surface.
It is based upon three-dimensional models of the knee, and it interacts
with the content software to send the appropriate images to the
video display. Image displays of the content are created through
a technique known as volume rendering. This process captures voxels
of information in three planes, much in the way that a two-dimensional
picture captures pixels in just two planes. Volume rendering is
the recommended method for capturing and displaying data, as the
resulting data are consistent with the original form. The content
program is responsible for the actual appearance of the knee on
the video display; it includes knee pathology such as meniscal tears
and chondral defects as well as normal anatomy. It also monitors
task-specific performance, such as the shaving of a torn meniscus
or the capturing of an intra-articular loose body. These programs
also can provide the user with feedback on the work accomplished.
Such feedback can range from the time required to complete a task
to a rating of the quality of the surgical effort.
Arthroscopic equipment: Probes, shavers, and
graspers, visualized on the computer monitor, must work in concert with
force-feedback devices and sensors placed in the leg model to provide
an accurate simulation for the learner. Only the handles of the
tools need be real; the rest of the instrument is generated by the
computer.
Monitor: The monitor should allow for high-resolution display
of all images.
Computer: A desktop workstation with accelerated
three-dimensional graphics should be used to maximize system throughput
for three-dimensional volumetric rendering and display of virtual
reality programming.
The AAOS is partnering with speciality groups to provide content
expertise as well as additional funding to cover the estimated development
costs of approximately $1,000,000. The cost of an individual simulator
will be approximately $30,000 to $40,000, depending on the number
of units purchased by members of the orthopaedic community.
Crucial to the acceptance of this tool is evidence that it is
reliable, sensitive, and validated. It must be able to distinguish
between the novice and the expert arthroscopist, and it also must
be able to document an individual's improvement after repeated use.
It must demonstrate that it has made an impact on proficiency when
the student's performance on the simulator is compared with his
or her performance on the current gold standard - that is, a cadaver
model. The various tasks of the examination/assessment must be quantifiable.
The tool must be validated by showing that the exercise actually
provides a measure of surgical skill rather than a measure of gamesmanship.
Recently, O'Toole et al.6 reported
on the use of a virtual reality simulator for comparison of the
suturing techniques of medical students with those of experienced vascular
surgeons. While they found that the simulator readily differentiated
the superior proficiency of the surgeons with regard to certain
parameters (total tissue damage, time to complete the task, and
amount of instrument motion), they urged that validation studies
be performed to better assess the role of the simulator in education
and training. Reznick7, in an
accompanying editorial, suggested that the real task is to assess
the efficacy of virtual environments in the training and evaluation
of surgeons.
If the arthroscopic simulator is to become a useful teaching
and evaluation tool, we will have to establish construct validity
- that is, the ability to distinguish between novices and experts
performing the same task on the machine. With use of data gathered
from beta testing sites, standards could then be established for
a particular simulation program running on a given simulator. We
expect that initially the minimum level of competence will be the
easiest to establish, whereas an extended period of testing will
be required to establish an advanced-skills level.
The AAOS, through its Task Force on Educational Effectiveness,
will create a validation study of the arthroscopic simulator at
various stages during its development. In the first study, use of
the simulator will be compared with use of fresh cadaveric specimens
for the learning of arthroscopic techniques. The study will be built
around the educational effectiveness model known as the Objective
Structured Assessment of Technical Skills (OSATS), which focuses
on the evaluation of surgical skills with use of structured activity
checklists4.
The goal of a similar validation study, of laparoscopic simulators,
conducted by the Institute for Defense Analyses, was to develop
a classification scheme for the surgical tasks of laparoscopy and
to establish psychomotor performance standards for surgical residents2. The initial work distinguished between
novice and experienced laparoscopists but not between surgeons with
moderate experience (ten to fifty cases) and experts (more than
100 cases)2.
The aviation industry has studied the effects of aircraft-simulator
training on the performance of pilots flying real aircraft. The
transfer effectiveness ratio, developed by Orlanski et al.5, indicates how much time that training
with a flight simulator saves the pilot in the air. For the arthroscopic
simulator, this would be defined as (As)/S, where A is the arthroscopy time
without use of a simulator, As is the arthroscopy time after use
of the simulator, and S is the actual time spent using the simulator. In
aviation, this ratio is 0.48, which means that one hour spent using
the simulator saves about one-half hour in a real aircraft5. Validation studies will be necessary
to determine whether the transfer effectiveness ratio for arthroscopic
simulation compares favorably with the value for aviation.
The ABOS and the AAOS have brought the concept of an arthroscopic
knee simulator to the developmental stage. If the simulator is shown
to be effective and valid, it should have wide-ranging positive
effects on the education and training of residents and the continuing surgical
proficiency of practicing orthopaedic surgeons. Additionally, it
should assure the public that orthopaedic surgeons are educated and
trained in the most rigorous real and surrogate environments. The
simulator should be the first tool with which our profession can
begin to grapple with the complex subject of surrogate-skills education.
Address for R. Poss: Department of Orthopaedic Surgery,
Brigham and Women's Hospital, 75 Francis Street, Boston, Massachusetts 02115-6110.
E-mail address: rposs@partners.org.
Address for J. D. Mabrey: Department of Orthopaedics,
University of Texas Health Science Center at San Antonio, 7703 Floyd
Curl Drive, San Antonio, Texas 78284-7774. E-mail address: mabrey@uthscsa.edu.
Address for S. D. Gillogly: 3200 Downwood Circle, Suite
530, Seeabue, Georgia 30327.
Address for J. R. Kasser: Children's Hospital Medical
Center, 300 Longwood Avenue, Boston, Massachusetts 02115. E-mail
address: kasser@a1.tch.harvard.edu.
Address for H. J. Sweeney: 1144 Wilmette Avenue, Wilmette,
Illinois 60091.
Address for B. Zarins: Massachusetts General Hospital,
15 Parkman Street, Suite 514, Boston, Massachusetts 02114. E-mail
address: bertram@mgh.harvard.edu.
Address for W. E. Garrett, Jr.: The University of North
Carolina School of Medicine, Burnett-Womack Building, Room 236,
Campus Box 7055, Chapel Hill, North Carolina 27599. E-mail address:
bill_garrett@med.unc.edu.
Address for W. D. Cannon: University of California at
San Francisco Medical Center, Level 1, 500 Parnassus Avenue, San
Francisco, California 94143. E-mail address: dcannon@ortho1.ucsf.edu.
Encarnacao, J. L.: VR Arthroscopy
Training Simulator. Fraunhofer Institut für Graphische Datenverarbeitung,
www.igd.fhg. de/igd-a4/flyers/medicine/arthro/, 1998.
Johnston, R.; Bhoyrul, S.; Way,
L.; Satava, R.; McGovern, K.; Fletcher, J. D.; Rangel, S.; and Loftin,
R. B.: Assessing a Virtual Reality Surgical Skills Simulator. Virtual
Environment Technology Laboratory (University of Houston-NASA/JSC),
Houston, Texas, www.vetl.uh.edu/surgery/mmvr4b.html, 1995.
Logan, I. P.; Wills, D. P. M.; Avis, N. J.; Mohsen, A. M. M. A.; and Sherman, K. P.: Virtual environment knee arthroscopy training system. Soc. Comput. Simulat., Simulation Series,28(4): 17-22, 1996.28(4)17
1996
Martin, J. A.; Regehr, G.; Reznick, R.; MacRae, H.; Murnaghan, J.; Hutchison, C.; and Brown, M.: Objective Structured Assessment of Technical Skills (OSATS)
for surgical residents. British J. Surg.,84: 273-278, 1997.84273
1997
Orlanski, J.; Dahlman, C. J.; Hammon,
C. P.; Metzko, J.; Taylor, H. L.; and Youngblut, C.: The Value of
Simulation for Training. Alexandria, Virginia, Institute for Defense
Analyses, 1994.
O'Toole, R. V.; Playter, R. R.; Krummel, T. M.; Blank, W. C.; Cornelius, N. H.; Roberts, W. R.; Bell, W. J.; and Raibert, M.: Measuring and developing suturing technique with a virtual
reality surgical simulator. J. Am. Coll. Surgeons,,189: 114-127, 1999.189114
1999
Reznick, R. K.: Virtual reality surgical simulators: feasible but valid?. J. Am. Coll. Surgeons,189: 127-128, 1999.189127
1999
Spitzer, V.; Ackerman, A. J.; Scherzinger, A. L.; and Whitlock, D.: The Visible Human Male: a technical report. J. Am. Med. Inform. Assn.,3: 118-130, 1996.3118
1996
Spitzer, V. M.: The visible human: a new language for communication in
health care education. Caduceus,13: 42-48, 1997.1342
1997
[PubMed]
Spitzer, V. M., and Whitlock, D. G.: The Visible Human Dataset: the anatomical platform for
human simulation. Anat. Rec,253: 49-57, 1998.25349
1998
[PubMed]
Trowbridge, E. A., and Hollands,
R. J.: A virtual reality training tool for the arthroscopic treatment
of knee disabilities. Read at the European Conference on Disability,
Virtual Reality and Associated Technologies '96. Maidenhead, Berkshire,
United Kingdom, July 8-10, 1996.