AAOS Bulletin - August, 2006

Evidence-based practice and ethical decision making

Helping to reduce unintended bias in clinical decision making

By William C. Watters III, MD

Both the public and regulatory bodies are taking a hard look at the ethics of medical practitioners and the possible influence of relationships beyond the traditional physician-patient relationship.1-3 No matter how benign these relationships might seem to the practitioner—and they could range from a free dinner to a consulting position with a manufacturer, or ownership of stock in a drug company whose products are frequently prescribed by the practitioner—questions can arise about whether they introduce bias into the clinical decision-making process. In fact, they can. Recent research shows that something as benign as a medical company’s sponsorship of grand rounds can influence physician prescribing habits.4

To encourage high ethical standards in medical practice, most state credentialing bodies require health care providers to present proof that they have completed a certain amount of ethics training annually. In Texas, where I practice, the requirement is one hour of ethics-related CME for physicians and surgeons each year. Little information exists on the effectiveness of such demands in sustaining the ethical integrity of the clinical decision-making process, but as Columbia University professor and aphorist Mason Cooley once said, “Reading about ethics is about as likely to improve one’s behavior as reading about sports is to make one into an athlete.”

EBM and the Maine Lumbar Spine Study

Although taking a course in ethics can help instill ethical principles, these must be practiced daily. Practicing evidence-based medicine (EBM) is an important means of reducing unintended bias in clinical decision making and contributing to ethical decision making. Because a practicing physician’s (or surgeon’s) privileged role in society demands a high ethical framework to support that practitioner’s clinical decision making, most practitioners wish to, try to and, in fact, think they are making the best possible decisions with their patients. Nonetheless, all physicians are subject to subtle influences and bias that can shift their decision-making process out of its appropriate ethical context into something potentially more self-serving than patient-serving.

As an example, consider the Maine Lumbar Spine Study,5 a prospective study that examined 655 patients with lumbar herniated nucleus pulposus (HNP), or spinal stenosis. The study used a small area analysis to develop three distinct service areas in the state, based on patterns of hospital admissions. A variety of validated outcome measures was used during the follow-up on 250 surgically treated patients, which occurred every three months for up to four years, with a mean of two years. The results showed that surgical treatment of HNP was superior to nonoperative measures.

Although surgeons often quote this study as support for the surgical treatment of a lumbar HNP, a closer reading uncovers other important findings. The three service areas had significantly different operative rates for HNP (p < .001) and these differences, which varied threefold between the lowest and highest areas, did not relate to the population in each area. Most important, when the surgeons in the study were questioned, they all felt that they had used similar indications and had similar outcomes as other surgeons in the study.

However, the patients from the area with the lowest operative rate had significantly better outcomes than those who lived in areas with higher operative rates. Furthermore, the patients in areas with a higher surgical rate had less severe symptoms prior to their surgeries. These findings illustrate the impact that subtle influences had on the decision-making practices of the surgeons in the areas with higher operative rates.

When participating surgeons were informed of the study’s findings, they did not dispute the findings but asked what they could do to rectify the disparities. They then acted upon these recommendations.

The Maine Lumbar Spine Study demonstrates that physicians—even when they act in good faith—can still be biased in their clinical decision making. Yet when physicians are presented with data on best care, they will modify their practices and improve their decision making. Though the authors didn’t intend it, this study nicely embodies a currently accepted definition of evidence-based practice (EBP): “The integration of the best research evidence with the practitioner’s expertise and the patient’s values.”6

A closer look at EBP

The three components of this definition of EBP deserve closer scrutiny.

Clinical expertise derives from the physician’s education and training before entering practice, the experience gained while in practice and the physician’s continuing efforts at education through reading and course work. But this expertise has limits.

The longer a physician is in practice, the more likely it is that much of his or her training will be proven incomplete or even wrong. Although experience can be a great educator, investigations in learning have found that the brain is vulnerable to remembering and valuing many events and experiences more as a function of their uniqueness than as a function of their usefulness. Thus, the unexpected diagnosis or unusual presentation of a disease in a past patient can influence the physician’s future diagnostic conduct out of proportion to the likelihood of its ever being encountered again. Finally, the rapidly expanding knowledge base in all clinical fields can no longer be realistically absorbed by any one person, no matter how dedicated he or she might be to continuing education.

Respecting patient values is also an important part of EBP. Each patient brings his or her own knowledge of the condition, as well as a unique social experience and a set of beliefs, to the physician-patient relationship. But patients continue to rely heavily on the recommendations of their physicians in making diagnostic and treatment choices.

The third part of this definition is a critical, new component: the use of the best current clinical evidence. Not all clinical research evidence is used to answer a clinical question in EBP, just the best evidence available. The best evidence is determined by reading and rating the clinical literature into hierarchical levels of evidence and accepting only the highest-rated evidence available for a particular clinical decision.

The practitioner can identify best available evidence by using evidence-based guidelines, such as those being developed by the AAOS, and by reading evidence-based articles such as those in The Journal of Bone and Joint Surgery. Using the best research evidence in EBP leads to “best practices” in clinical medicine and reduces the undue influence of factors outside the physician-patient relationship. Consequently, EBP includes a self-correcting mechanism that can reduce bias in clinical decision making and promote the practice of ethical medicine.

References

1. Armstrong D. “Delicate operation: how a famed hospital invests in device it uses and promotes.” Wall Street Journal, December 12, 2005:A1.

2. Rundle R, Hensley S. “Backfire: J&J’s new device for spine surgery raises questions: artificial disk aims to help body’s natural movement; some see risk if it slips.” “Big money riding on this.” Wall Street Journal, July 7, 2001:A1.

3. Abelson R, Petersen M. “An operation to ease back pain bolsters the bottom line too.” New York Times, December 31, 2003.

4. Dana J, Lowenstein G. A social science perspective on gifts to physicians from industry. JAMA 2003;290:252-255.

5. Keller R, Atlas S, Soule D, Singer D, Deyo R. The Relationship Between Rates and Outcomes of Operative Treatment for Lumbar Disc Herniation and Spinal Stenosis. JBJS 1999;81-A:752-762.

6. Straus S, Richardson W, Glasziou P, Haynes B. Evidence-Based Medicine. Churchill Livingston, 2005.

William C. Watters III, MD, is chairman of the AAOS Guidelines Oversight Committee. He can be reached at spinedoc@pdq.net


Close Archives | Previous Page