In the early 1980s, orthopaedic surgeons made the transition
from reviewing patient cohorts on paper to storing patient data
in computerized databases. These databases have proven to be useful
tools for documenting patient care, performing clinical research,
and fostering continuing education. Patient databases will play
an even more important role in the future as orthopaedic surgeons
and software developers further exploit the power of the Internet.
In this paper, we review our experience with use of a patient database
at a single institution, the strides being taken to create a registry
that can be used by many centers, and the potential to accelerate
these efforts through the widespread use of the Internet.
Patient databases store information collected from instruments
designed to measure components of patient care. We have used a system model
(Fig. 1) to
consider these individual components. In the model, patient care
comprises the continuous and recurring cycle of evaluation, decision, and
intervention, and the interactions among these three elements. Each
of these components can be measured independently. Validity, reliability,
and sensitivity are measures of evaluation. Appropriateness is a
measure of the decision, and outcome is a measure of the intervention.
Satisfaction is a measure of patient care as a whole.
To demonstrate the use of one of these measures with a patient
database, consider the example of a study that we conducted to determine
the appropriateness of total hip replacement. We measured the change
in the physical scales of the Short Form-36 (SF-36) from a point
immediately before hip replacement surgery to two years after surgery
in a consecutive group of 200 patients1.
The patient population was divided into three groups - high, middle,
and low - on the basis of preoperative function (Fig. 2). The patients
who were in the low and middle-function groups preoperatively demonstrated
dramatic improvements in function after total hip replacement; however,
patients with relatively high preoperative function did not improve
substantially. The frequency of patients who reported dissatisfaction
was also higher in this group.
These results suggest that patients who are relatively high-functioning
require more intensive preoperative educational efforts to align their
expectations with those of the surgeon. These patients in particular
must understand the limits as well as the promise of hip replacement
surgery before deciding whether to undergo it. This example demonstrates
how measurement of function and satisfaction as part of a patient
database provides insight that can lead to improved patient care.
The Total Joint Replacement Registry at Brigham and Women's Hospital
in Boston has amassed a database with information on more than 15,000
patients and more than 20,000 joint replacement procedures performed
in the last three decades. This database comprises patient demographics,
histories, findings on physical examination, intraoperative findings, information
on complications, patient outcomes, and patient-satisfaction data
that were collected prospectively with informed patient consent.
The annual operating budget of the Registry is currently $350,000,
with external grants accounting for $150,000 of the total.
The Registry has been a substantial resource for attending surgeons,
residents, and fellows. Since 1990, the Registry has accommodated more
than 1500 data requests for clinical research projects. These projects
have generated 110 publications as well as 140 presentations at the
annual meeting of the American Academy of Orthopaedic Surgeons.
The Registry database also has been a valuable resource for collaboration
with visiting scientists and a resource for collaborative studies
with other departments at our institution, such as Rheumatology,
the Multipurpose Arthritis Center, and Internal Medicine2-6,8. We could not have conducted
these studies without this database of prospectively collected information.
In recent years, the database has made it possible for the Department
of Orthopaedic Surgery to be a leader at Brigham and Women's Hospital
in implementing new patient-education programs, clinical pathways, and
cost-efficiency programs.
Through the operation of the Registry, we have learned many lessons
regarding the successful implementation of a patient database. Most
importantly, the reason or reasons for constructing a patient database
must be clear and specific; otherwise, a great deal of energy and
money may be expended collecting data that will never be used. Is
the reason for developing a patient database to perform retrospective
and/or prospective studies, to assist in influencing patient care,
or to document patient outcomes? The more reasons that there are,
the more complex, time-consuming, expensive, and, in the end, difficult
the task.
There are practical limits to the quantity and frequency of data
collection. Our experience suggests that patient encounters for
obtaining outcomes information should be limited to ten minutes.
The number of encounters during which the patient is asked to provide
information also must be limited. We have found that patient cooperation
declines when the assessment of outcomes for the same condition
occurs more than once a year. The number of patients for whom substantial
data are collected must be targeted to the research question being
asked. It is impractical to collect large amounts of data from all
patients visiting a busy orthopaedic clinic.
Prospectively collected data from multiple centers offer the
promise of providing sufficiently large cohorts to serve as an early-warning
system for implant failures and to formulate research questions
far earlier than would be possible at any individual institution. The
Hip Society has initiated a pilot project, at ten participating
centers, to begin to construct a multicenter clinical and radiographic
database. The initial project would require only a limited data
set.
The hypothesis of the first study to be conducted is that the
impact of dislocation following total hip arthroplasty on the patient's
perception of outcome is greater than surgeons have appreciated.
With informed patient consent, participating centers will collect
data preoperatively, at surgery, and at follow-up, and they will
send digitized radiographs to the Hip Society Registry to be electronically
integrated with the clinical data. The study will enroll all patients
undergoing primary total hip arthroplasty during a one-year period
at the ten participating centers. We have projected that this cohort
will be sufficient to answer the research questions posed.
Simultaneously, the Hip Society Registry will perform a pilot
audit that records a limited data set to document the date of implantation and
the types of components that are being removed during revision total
hip replacement8 at these few
centers over a period of one year.
The Internet affords the opportunity to outsource the management
of patient databases by using the World Wide Web and ubiquitous Web-browser
software. Off-site databases are growing rapidly in all fields.
Now, a data-management company can maintain the hardware, software,
and data at its external location, while a personal computer and
browser are all that the orthopaedic surgeon needs in order to enter
and access his or her data.
In 1999, we decided to transfer the management of the Brigham
Total Joint Replacement Registry to an off-site data-management
company. Reciprocal data transfers will be conducted through the
World Wide Web. We realized immediate cost-savings by relinquishing
expensive rental space and reducing the number of full-time employees
from 7.5 to three. We budgeted the total cost for the year 2000
to be the same as that for 1998, but we anticipate that, in succeeding
years, the cost of maintaining the Registry will be less because of
outsourcing.
The Web provides a unique opportunity for multicenter registry
databases to flourish. The Hip Society Registry will be available
to participants online so that real-time interactions can occur;
these interactions would have been virtually impossible without
the Web. The Web also allows orthopaedic departments and practices
to participate more easily in multicenter "best practice" groups
to study the effectiveness of programs.
In addition to single-center and multicenter registries and studies,
the Web is transforming the training and certification processes
in medicine through its influence on patient databases. For example,
one requirement of the American Board of Orthopaedic Surgery's examination
process is that candidates submit to the Board a six-month case
list of operative procedures. In previous years, this task has proven
arduous for candidates using pen and paper or a disk-based system.
This year, candidates for the part-II certification examination were
able to electronically record their cases in an ongoing manner and
then to submit them to the Board through the Web. The resulting
patient database, consisting of more than 92,000 cases submitted
by more than 700 candidates, provides a snapshot of orthopaedic
surgery practice in the United States during this six-month window
of time.
Finally, patient databases and the World Wide Web now offer residency
programs the possibility of automatically generating the information
needed for the Residency Review Committee's Program Information
Form. Residents tracking their cases with use of online databases
generate the statistics required for the report, and program directors
are able to follow, in real time, both the experience that their
programs are providing to residents and the performance of residents
in their programs.
The World Wide Web is enabling patient databases to reach their
full potential to provide timely and accurate information to orthopaedic
surgeons. Its impact will continue to grow for years to come.
Note: We are indebted to Maurice E. Müller, M.D., for his counsel
and collaboration in the development and implementation of many
of the concepts expressed in this paper. We also gratefully acknowledge
the Maurice E. Müller Foundations of Switzerland and North America
for their support of the Brigham Total Joint Replacement Registry;
the contributions of Sonya van der Meer, John Kwon, and Tad Bergman
to the success of the Brigham Total Joint Replacement Registry;
and the contributions of Beth Imrem of Data Harbor, Incorporated,
to the Hip Society Registry. Finally, we acknowledge and thank the
Council on Research of the American Academy of Orthopaedic Surgeons,
the Musculoskeletal Evaluation and Research Institute, and the Orthopaedic
Research and Education Foundation for their support of the Hip Society
Registry.
Address for J. J. Harrast and R. Poss: Department of
Orthopaedic Surgery, Brigham and Women's Hospital, 75 Francis Street,
Boston, Massachusetts 02115-6110. E-mail address for J. J. Harrast:
jjharrast@ msn.com. E-mail address for R. Poss: rposs@partners.org.