 |
 |
 |
|
| home | CME | help | feedback | subscriptions | archive |
|
© 1999 The Journal of Bone and Joint Surgery, Inc.
Commentary - The Fallacy of Short-Term Outcomes Analysis in Pediatric OrthopaedicsOutcomes analysis, which has become the buzzword in orthopaedic surgery over the past ten years, is not just a new name that we've put on our old friend clinical research. What is outcomes analysis? Some clarification is provided by Johnson1, who stated: "A hallmark of the outcomes research movement is the demand for and acceptance of measures of patient perceptions, functional status, emotional health, and quality of life as indicators of the effectiveness of treatment." A new development in medicine, such as a new medication, a new operation, or a new theory, is first met by reluctance, which is followed by overenthusiasm, which is in turn followed by rebellion. Outcomes analysis has been subjected to this sequence (often called the pendulum process) and is currently in the phase of overenthusiasm. We must keep in mind the true purpose of such studies, which is the evaluation of our management process by looking at the functional result (that is, whether the patient perceives a positive value from the management process) as opposed to the objectively measured result. A good example is the fractured femur that is solidly united with perfect length and alignment. This appears to be an excellent result when evaluated on a radiograph, but it may not be a good outcome if muscle adhesions or loss of knee-joint motion renders the limb well aligned but not very useful. Elective procedures, such as total hip replacement in adults, lend themselves beautifully to outcomes analysis. If the patient fills out a functional questionnaire both before the procedure and after an appropriate follow-up interval, a clear-cut benefit (or lack thereof) can be documented. Documentation includes clinical elements such as range of motion, muscle strength, and walking endurance as well as functional elements such as the ability to put on shoes, go up and down stairs, and so on. Clinical documentation involves measurement, but measurements can be tricky. For example, Cobb measurements of spinal deformity are useful, standardized worldwide, and reproducible. When applied to scoliosis, Cobb measurements can be made preoperatively, postoperatively, and during follow-up. Percentage correction can thus be established, allowing us to compare one method of treatment with another. However, in our enthusiasm to reduce everything to numbers, we can fall into the trap of using percentage correction when analyzing deformities in the sagittal plane such as kyphosis or lordosis. I have often heard a surgeon state in a presentation that the average preoperative Cobb measurement in his or her patients with kyphosis was 80 degrees and the average postoperative measurement was 40 degrees and thus a correction of 50 percent had been achieved. Since 40 degrees of thoracic kyphosis is normal and a pathological state was corrected to a normal state, did the surgeon not achieve 100 percent correction? Simply, one cannot use percentage correction for problems in the sagittal plane. The key to most such analyses is that a pathological state exists and the treatment is designed to correct or at least to alter positively that pathological process. The point at which the entire outcomes-analysis system falls apart is when the treatment rendered is designed to prevent rather than to treat a pathological state. A classic example of this problem is the case of an eighteen-month-old girl who is discovered to have a shallow acetabulum related to congenital hip dysplasia. The pediatric orthopaedist performs an innominate osteotomy in order to allow proper development of the hip joint. The purpose of the treatment is to prevent the need for a total hip replacement when the patient is thirty-five years old. Other than the slight limp that led to the diagnosis of hip dysplasia, there was no pathological state; the child had no pain and no limitation of functional motion. A questionnaire could not be used since an eighteen-month-old child cannot fill one out; it is equally absurd to ask the parents to fill out a functional questionnaire when there is no malfunction. How, then, do we measure the outcome of an innominate osteotomy performed on an eighteen-month-old child? Since the purpose of the operation was to prevent a painful, disabling hip condition at the age of thirty-five, the only way that we can measure the outcome is to follow her for 33.5 years and see what her hip is like at that time. Examination of the hip when the patient is sixteen years old can give a good clue as to what the status might be when she is thirty-five years old, but it is no guarantee. Our traditional end-of-growth evaluations in pediatric orthopaedics are more satisfactory, but they do not necessarily reflect the reality of adult years. Another example of this problem is the growing child who has scoliosis. One performs a spinal arthrodesis in a thirteen-year-old girl who has a 50-degree thoracic scoliosis not because she is suffering now but because the goal is to prevent a pathological state, especially a pulmonary function deficit, in the future. At thirteen years old, when the scoliosis is 50 degrees, the patient has no pain, no pulmonary deficit, and only minor aesthetic concerns. Current outcomes instruments (questionnaires) can best show only aesthetic improvement of body shape (at the price of a scar from the operation). Are we then only aesthetic surgeons who will be denied payment for our efforts? No. As spine surgeons, we are looking to the future. It is far easier, far cheaper, and far safer to perform a straightforward correction and arthrodesis in a thirteen-year-old girl than to wait until she is forty years old and then subject her to a combined anterior and posterior approach with less correction, higher risk, and greater morbidity. The only way that one can measure outcomes of preventive treatment is to compare the results with the results of no treatment—that is, the natural history. I recently had the pleasure of talking with a man who had had a posterior spinal arthrodesis for the treatment of a rapidly progressing juvenile idiopathic scoliosis when he was eleven years old. At the time that I spoke with him, he was eighty years old, had only mild low-back pain, and had no respiratory problems. His younger brother had had the same operation, but not at the right time, and he had died at the age of forty-five of cor pulmonale. These are true outcomes studies. However, I probably will not be allowed to publish the results because the series is too small and not prospective. There are so-called purists who now tell us that we need prospective, randomized, controlled clinical trials to test what we are currently doing against an untreated control group (the natural history). This can be ridiculous when the natural history is already well known. We don't need to wait for our control group to die of cor pulmonale. Such a practice would be immoral and unethical. In conclusion, I believe that we should abandon efforts to use outcome instruments as they now exist when trying to analyze the results of preventive treatment. For the evaluation of such results, only long-term analysis (when the patient is an adult) and comparison with established natural history studies can give us the answers that we want.
R. B. Winter, M.D. Twin Cities Spine Center Piper Building, Suite 600 913 East 26th Street Minneapolis, Minnesota 55404
References
  CiteULike  Connotea  Del.icio.us  Facebook  Technorati  Twitter What's this?
This article has been cited by other articles:
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| home | CME | help | feedback | subscriptions | archive | site index |
|
Stanford University Libraries' HighWire Press (TM) assists in the publication of JBJS Online. |