Evidence-based medicine integrates research findings, clinical expertise, and patient values to provide the highest quality patient care. An orthopaedic outcomes database can assist surgeons in assessing associations between surgical procedures, function, survivorship, complication rates, patient satisfaction, and quality of life as guidance for practice. Outcomes are used to validate techniques or procedures, to benchmark best practices, and to identify areas requiring more clinical research. Tracking the success and failure rates of implants and surgical procedures is important. An outcomes database allows for short-term, midterm, and long-term prospective follow-up studies of specific cohorts and also facilitates studies that are retrospective in design.
Several countries have initiated orthopaedic outcomes registries or databases. The oldest registry is the Swedish Knee Arthroplasty Register that was started in 19751. Other countries have followed with the development of their own joint registries in England, New Zealand, Germany, Finland, Denmark, and Norway2-7.
In 1994, the American Academy of Orthopaedic Surgeons (AAOS) attempted to create an outcomes evaluation system. The Musculoskeletal Outcomes Data Evaluation and Management System did not come to fruition and was terminated in 2000. The attempt was made to collect information on patients undergoing total hip and knee arthroplasty or spine surgery. Difficulties were multifactorial, but were mainly due to inadequate data, record-keeping, or follow-up, which made cleaning the data to extract useful information impossible8.
Although the United States lacks a national orthopaedic registry, several institutions have developed their own databases. For example, Mayo Clinic initiated a database in 1969 that included over 56,000 joint arthroplasties by 19979. The Anderson Orthopaedic Research Institute has over two decades' worth of hip and knee arthroplasty outcomes in its database10, while the Center for Hip and Knee Surgery in Indiana11, the Hospital for Special Surgery in New York, the University of Iowa, and Kaiser Permanente all maintain similar databases. These databases have allowed for extensive follow-up and publication of short-term, midterm, and long-term results associated with various prostheses and procedures.
A database can be helpful in a clinical setting by providing a snapshot of the current clinical practices and a real-time assessment of the associated outcomes. This article describes the planning, development, implementation, and utilization of a comprehensive orthopaedic outcomes program at one institution. We offer a step-by-step guide to initiate an outcomes program as well as methods for analyzing and reporting data.
Pay for Performance
In the United States, rising health-care costs and a rapidly aging population have recently stimulated the development of pay for performance initiatives in an effort to improve the quality of health care that is being delivered. The fundamental principle of pay for performance is to compensate medical providers on the basis of how well they adhere to practice standards. These standards are evidence-based and supported by clinical outcomes. An outcomes program can define and enhance the quality of an orthopaedic practice by providing a mechanism for evidence-based medicine.
Founded in 1924, Scripps Clinic is a multispecialty outpatient facility that provides care to patients at multiple locations throughout San Diego County. Scripps Clinic and its physicians are world-renowned for research-driven care and medical-specialty expertise. The Clinic is an operating unit of Scripps Health, a not-for-profit community-based health-care-delivery network that includes more than 2600 affiliated physicians, five acute-care hospitals, a home-health-care service network, and associated support services. Each year, approximately 5500 orthopaedic surgical procedures are performed at either Scripps Green Hospital or one of the Scripps Clinic's outpatient surgery centers. The Shiley Center for Orthopaedic Research and Education (SCORE) at Scripps Clinic began in 1985 with the mission to improve patient care and quality of life through innovative orthopaedic research and education. SCORE conducts preclinical, biomechanical, clinical, and outcomes research in conjunction with the Division of Orthopaedic Surgery, and it maintains the Scripps Clinic Orthopaedic Outcomes Program, a comprehensive research program that includes several outcomes databases. The initial data collection was directed at capturing data for hip and knee arthroplasties but has expanded to capture information on several other types of orthopaedic procedures performed at the Clinic.
One of us (C.W.C. Jr.), who specializes in lower-extremity reconstruction, has followed the progress of his patients since the late 1970s for outcomes relating to implant longevity, joint function, and the patient's overall well-being and satisfaction with the procedure. The early data on hip and knee arthroplasty procedures were collected via numerous methods, including handwritten lists, Scantron bubble sheets (Scantron, Eagan, Minnesota), and handheld electronic devices, and were stored in many different formats.
In 2002, an outcomes data team was developed to revise and update the data-collection tools, to standardize the data-collection process, and to purchase a dedicated database software program to store the data. After the database was purchased and institutional review board approval was obtained, all new surgical procedures were entered into the database prospectively; data from the past were retrospectively entered or imported into the database. The outcomes program was expanded in 2005 to include additional data team members, and an outcomes research manager with a statistics background was hired to oversee the daily activities of the program. Since then, outcomes databases for several other subspecialties within the Division of Orthopaedic Surgery have been developed.
All patients scheduled for selected surgeries are invited to participate in outcomes data collection; 99% of patients participate and sign an institutional review board consent and Health Insurance Portability and Accountability Act (HIPAA) authorization. Clinical and radiographic data collected by the health-care team and questionnaires completed by patients include data from all phases of study, from the preoperative visit through the appropriate follow-up visits. We maintain outcomes data on the following procedures:Hip and knee arthroplastyFresh osteochondral allografting of the knee or ankleShoulder arthroplastyAnterior cruciate ligament reconstructionFinger arthroplastyGreat-toe arthroplastySpine surgery
Hip and knee arthroplasty
Fresh osteochondral allografting of the knee or ankle
Shoulder arthroplasty
Anterior cruciate ligament reconstruction
Finger arthroplasty
Great-toe arthroplasty
Spine surgery
On the basis of our experience with the development and implementation of an outcomes program, we have formulated a ten-step guide12, which is summarized in Figure 1, that others can follow to initiate a similar program at their institution.
Source of Funding
Our outcomes program is funded by the surgeons in the Division of Orthopaedic Surgery and has no outside funding.
The following is a description of each of the Scripps Clinic Orthopaedic Outcomes Program databases, including a short description of the procedure for each surgical population, the data collected, the questionnaires used, and examples of how we utilize the outcomes data at our institution.
Hip and Knee Arthroplasty
Procedures: Total hip arthroplasty, hemiarthroplasty, total knee arthroplasty, unicompartmental knee arthroplasty, and revision hip or knee arthroplastyData collected: Demographics, pain and pain medication use, joint range of motion, surgical details, implant type and size, complications, and general well-being of the patientQuestionnaires used: Harris hip score, Knee Society clinical rating system13, Knee Injury and Osteoarthritis Outcome Score (KOOS)14, Hip Disability and Osteoarthritis Outcome Score (HOOS)15, and Short Form-12 (SF-12)16
Procedures: Total hip arthroplasty, hemiarthroplasty, total knee arthroplasty, unicompartmental knee arthroplasty, and revision hip or knee arthroplasty
Data collected: Demographics, pain and pain medication use, joint range of motion, surgical details, implant type and size, complications, and general well-being of the patient
Questionnaires used: Harris hip score, Knee Society clinical rating system13, Knee Injury and Osteoarthritis Outcome Score (KOOS)14, Hip Disability and Osteoarthritis Outcome Score (HOOS)15, and Short Form-12 (SF-12)16
The hip and knee arthroplasty database was started in 1977 by a single surgeon (C.W.C. Jr.) and has expanded to include all five lower-extremity surgeons within the Division of Orthopaedic Surgery. To date, the database includes information on over 13,000 hip and knee procedures; approximately 1200 are added annually. The database has allowed for many short-term, midterm, and long-term prospective follow-up studies as well as studies that are retrospective in design. Examples of the types of reports and analyses utilizing the outcomes data are shown in Figures 2 and 3. Research areas have included implant survivorship, surgical approach, incision length, utility of intraoperative radiographs, and postoperative treatment modalities. Recent interest in pain has stimulated studies looking at effects of anesthesia type, postoperative analgesics, pain-management processes, and complementary and alternative medicine techniques to reduce pain after surgery. This interest has led to nursing and allied health studies.
The collection of outcomes data across a population of patients who have undergone total hip arthroplasty or total knee arthroplasty primarily offers the ability to determine, with significance, correlations between treatment methods and the impact on disease, quality of life, and success rates. Secondly, long-term data analysis can provide information on beneficence to the patient, which is inextricably tied to quality in health care. Finally, the power generated by carefully conducted research analysis provides the orthopaedic surgeon with a decision-making framework for selecting the methodology that will offer the best possible outcome and the highest quality of care for the patient.
Fresh Osteochondral Allografting of the Knee or Ankle
Procedures: Osteochondral allografting and revisionsData collected: Demographics, pain, joint range of motion, activity levels, surgical details, graft size and location, complications, and general well-being of the patientQuestionnaires used: Knee Society clinical rating system, KOOS, International Knee Documentation Committee (IKDC)17, 18-point modified Merle d'Aubigné and Postel scale18, 100-point Ankle Assessment19, and the American Orthopaedic Foot and Ankle Society (AOFAS) Ankle-Hindfoot Scale20
Procedures: Osteochondral allografting and revisions
Data collected: Demographics, pain, joint range of motion, activity levels, surgical details, graft size and location, complications, and general well-being of the patient
Questionnaires used: Knee Society clinical rating system, KOOS, International Knee Documentation Committee (IKDC)17, 18-point modified Merle d'Aubigné and Postel scale18, 100-point Ankle Assessment19, and the American Orthopaedic Foot and Ankle Society (AOFAS) Ankle-Hindfoot Scale20
Over the past quarter century, fresh osteochondral allografting has emerged as a viable biologic reconstructive option capable of restoring mature hyaline cartilage in acquired osteoarticular defects. The fundamental paradigm of fresh osteochondral allografting is the transplantation of mature orthotopic hyaline cartilage along with viable chondrocytes that are capable of maintaining their metabolic activity and sustaining the surrounding collagen matrix.
The allograft database maintains outcomes on all knee and ankle allografts that have been performed at Scripps Clinic. The data on approximately sixty procedures per year are collected prospectively from a single surgeon. Medical records from the surgeon's previous practice were added to the database retrospectively. Currently, the database contains data on 661 procedures from 1983 to the present. The average patient age is thirty-six years, and 57% of the patients are men.
Recently, the database was used to prepare multiple presentations, both nationally and internationally. Presented material included an outcomes study of osteochondral allograft transplantation for bipolar lesions of the knee (Figs. 4 and 5), lesions of the talus, and treatment of steroid-associated osteonecrosis of the femoral condyles. All of these presentations indicated that osteochondral allografting is a reasonable treatment option for such indications. Also, a comparison study was completed of patients in whom cytotoxic antibodies developed and patients who had no immune response postoperatively. The conclusion was that patients with a large graft area (>10 cm2) were thirty-six times more likely to have the development of antibodies; however, further study is needed to understand how immunologic responses influence the survival of osteochondral allograft transplants.
We plan to use the database to compare the outcomes of knee osteochondral allograft with other methods of focal cartilage lesion repair, such as autologous chondrocyte implantation, osteoarticular transfer system, and microfracture. We would also like to assess the utility of osteochondral allograft in delaying the need for total knee arthroplasty in a young patient population. We plan to compare the outcomes of ankle osteochondral allograft with ankle arthroplasty and to assess how it may provide an alternative to arthrodesis or fusion.
Shoulder Arthroplasty
Procedures: Total shoulder arthroplasty, reverse total shoulder arthroplasty, hemiarthroplasty, and revisionsData collected: Demographics, pain and pain medication use, joint range of motion, surgical details, implant type and size, complications, and general well-being of the patientQuestionnaires used: Constant and Murley shoulder assessment21, American Shoulder and Elbow Surgeons (ASES) subjective shoulder scale22, and Disabilities of the Arm, Shoulder and Hand (DASH)23
Procedures: Total shoulder arthroplasty, reverse total shoulder arthroplasty, hemiarthroplasty, and revisions
Data collected: Demographics, pain and pain medication use, joint range of motion, surgical details, implant type and size, complications, and general well-being of the patient
Questionnaires used: Constant and Murley shoulder assessment21, American Shoulder and Elbow Surgeons (ASES) subjective shoulder scale22, and Disabilities of the Arm, Shoulder and Hand (DASH)23
The total shoulder arthroplasty and reverse total shoulder arthroplasty database started in 2006. Approximately forty patients a year are added from one participating surgeon. Sixty percent of the patients are women, and their average age is seventy-one years. Initial analyses of pain and joint function have been evaluated to compare preoperative and postoperative scores. Patients who had total shoulder arthroplasty and patients who had reverse total shoulder arthroplasty have shown improvements in Constant and Murley and ASES scores (Figs. 6 and 7). Mean pain scores improved by 87% for patients who had total shoulder arthroplasty and by 57% for patients who had reverse total shoulder arthroplasty. We plan to continue collecting data on these patients in order to better quantify the difference in outcomes in these two surgical populations.
Anterior Cruciate Ligament Reconstruction
Procedures: Anterior cruciate ligament reconstruction and revisionsData collected: Demographics, pain, joint range of motion, surgical details, implant type and size, complications, general well-being of the patient, and KT-1000 arthrometer measurementsQuestionnaires used: Tegner activity score24, IKDC, and KOOS
Procedures: Anterior cruciate ligament reconstruction and revisions
Data collected: Demographics, pain, joint range of motion, surgical details, implant type and size, complications, general well-being of the patient, and KT-1000 arthrometer measurements
Questionnaires used: Tegner activity score24, IKDC, and KOOS
The purpose of anterior cruciate ligament reconstruction is to stabilize the tibiofemoral joint and restore a high level of function to the knee. A variety of choices with regard to grafts, methods of fixing grafts, and postoperative rehabilitation are available. Knowledge of patient outcomes after anterior cruciate ligament reconstruction can be valuable in determining better forms of fixation, physical therapy, and biologic treatments.
The anterior cruciate ligament database was started in 2007. Approximately 175 procedures are added each year. We plan to use the outcomes data to compare the use of single-bundle grafts and double-bundle grafts during surgery. We also plan to study the revision rates and functional outcomes that have been associated with the new biologic and synthetic materials to determine if they reflect an improvement over the rates and outcomes that have been reported after metal screw fixation. Another topic we plan to study is how anterior cruciate ligament reconstruction may be an early risk factor for the need for total knee arthroplasty.
Finger Arthroplasty
Procedures: Proximal interphalangeal arthroplasty, carpometacarpal arthroplasty, and revisionsData collected: Demographics, pain, joint range of motion, surgical details, implant type and size, grip and pinch strength, complications, and general well-being of the patientQuestionnaires used: DASH and SF-12
Procedures: Proximal interphalangeal arthroplasty, carpometacarpal arthroplasty, and revisions
Data collected: Demographics, pain, joint range of motion, surgical details, implant type and size, grip and pinch strength, complications, and general well-being of the patient
Questionnaires used: DASH and SF-12
Tendon interposition arthroplasty with ligament reconstruction is a technique in which the flexor carpi radialis tendon is harvested and passed through drill holes at the metacarpal base. The short and long-term follow-up after this procedure has shown favorable results, with an increase in grip and pinch strength, an increase in the ease of performing activities of daily living, and a subjective decrease in pain.
The finger arthroplasty database was started in 2007, and twelve to fifteen procedures are added annually. We plan to use the database to compare the success rates associated with different surgical techniques and implant types, (e.g., ligament reconstruction and tendon interposition with insertion of a pyrocarbon implant or with insertion of both silicone and pyrocarbon implants).
Great-Toe Arthroplasty
Procedures: Great-toe arthroplasty and revisionsData collected: Demographics, pain, joint range of motion, activity levels, surgical details, implant type and size, complications, general well-being of the patientQuestionnaires used: AAOS Foot and Ankle Module25, AOFAS hallux-metatarsophalangeal-interphalangeal scale20, and SF-12
Procedures: Great-toe arthroplasty and revisions
Data collected: Demographics, pain, joint range of motion, activity levels, surgical details, implant type and size, complications, general well-being of the patient
Questionnaires used: AAOS Foot and Ankle Module25, AOFAS hallux-metatarsophalangeal-interphalangeal scale20, and SF-12
Hallux rigidus, which affects the first metatarsophalangeal joint, can be a debilitating condition. The metatarsophalangeal joint bears one-third of the total body weight during normal gait. When this joint is damaged, the patient experiences not only pain but a shift from the normal gait pattern. Total joint prostheses have been on the market for more than thirty years. Unfortunately, many of the early constrained models were associated with biologic reactions that required removal of the implant, leaving the patient with a large osseous deficit, while other models would fail at the hinge over a period of years, thus necessitating removal and fusion of the joint.
The great-toe arthroplasty database was started in 2008, and approximately fifteen procedures will be added each year. Many great-toe arthroplasty studies have shown questionable outcomes with relatively short-term follow-up. We plan to assess this procedure with regard to its ability to decrease pain and increase range of motion and to also assess whether great-toe arthroplasty is an effective alternative to fusion.
Spine Surgery
Procedures: Lumbar and cervical proceduresData collected: Demographics, pain, joint range of motion, activity levels, surgical details, implant type and size, complications, general well-being of the patientQuestionnaires used: Scoliosis Research Society-30 (SRS-30) questionnaire26, Oswestry Disability Index (ODI)27, and Neck Disability Index (NDI)28
Procedures: Lumbar and cervical procedures
Data collected: Demographics, pain, joint range of motion, activity levels, surgical details, implant type and size, complications, general well-being of the patient
Questionnaires used: Scoliosis Research Society-30 (SRS-30) questionnaire26, Oswestry Disability Index (ODI)27, and Neck Disability Index (NDI)28
The spine presents a complex challenge for orthopaedists due to the prevalence of back pain, the comorbidities associated with back pain, and the ambiguity of treatment outcomes. More than 70% of individuals will experience acute lower back pain during their lives29. Surgical intervention for chronic lower back pain includes fusion surgeries (anterior and posterior lumbar interbody fusion, and posterolateral instrumented fusion) and nonfusion surgeries (decompression, dynamic stabilization, and total disc replacement). In 2006 alone, more than 340,000 patients underwent spinal fusion surgery in the United States30. Despite this large number of patients undergoing spine surgery, relatively little is known regarding patient outcomes after surgical treatment of common conditions such as spinal stenosis, disc herniation, and degenerative spondylolisthesis.
The spine outcomes database is currently in development and will focus on clinical outcomes of lumbar and cervical fusion techniques, lumbar and cervical decompression techniques, and novel technology in minimally invasive spine surgery. We will collect surgical outcomes on pain, quality of life, surgical approach, screw fixation technique, implant type, rate of fusion, and complications.
We plan to compare various surgical approaches and techniques as well as the choices that are made regarding the use of fusion, implants, and fixation. The variability of clinical outcomes and fusion rates after lumbar interbody fusion surgery and the variable accuracy of the current radiographic methods that are used to diagnose the development of complications or the need for fusion provide ample opportunity for research and development of better surgical implant design and techniques. The advent of new intervertebral fusion devices, fixation implants, and biologically based therapeutic agents (biologics) provides incredible research opportunities for the improvement of evidence-based approaches to surgical spine care. We will be studying these modalities in a systematic way. New techniques are being developed for minimally invasive approaches to lumbar fusion, and we are studying the optimal use of bone-graft alternatives for cervical and lumbar fusions. As imaging technology progresses, we will correlate and modify our outcomes measurements in conjunction with improvements in noninvasive diagnostic tools. We also plan to study the use of less-rigid fixation systems in posterolateral lumbar fusion models, in hopes of improving implant retention and decreasing the occurrence of disease in adjacent segments. In addition, with use of the various outcome measures that have been discussed in this paper, lumbar decompression through laminoforaminotomy will be compared with decompression through traditional laminectomy.