Dr. Tosi: Which of the three possible explanations for this study’s results do you find most compelling? Is the causality a combination of them, or is there a further possibility not identified by Dr. Wright and his team?Dr. O’Connor: The authors’ results reflect our society. Like all individuals, physicians have biases, including bias based on sex and gender. When we see patients, we must guard against stereotypes.


Published 10/1/2008

Gender bias in TKA: A roundtable discussion Moderated by Laura L. Tosi, MD

Four surgeons discuss unconscious bias—and what can be done about it

Men and women are not always treated equally, especially in the doctor’s office. Although some treatment differences may be appropriate, others may be the result of unrecognized bias. For instance, a man with moderate knee arthritis and a woman with the same degree of disease may not receive a comparable course of treatment, as documented in a recent study appearing in the Canadian Medical Association Journal.

The study, spearheaded by James G. Wright, MD, MPH, of Toronto’s Hospital for Sick Children, found that orthopaedic surgeons are 22 times more likely to recommend total knee arthroplasty (TKA) to a male patient with moderate osteoarthritis (OA) than to a female patient with similar symptoms. Family physicians were found to recommend TKA to males twice as often as they did to females. (If the patient had severe knee OA, however, the recommendation for TKA was the same for both males and females.) A survey of the same physicians prior to the study showed that they were unaware of any gender bias affecting their recommendations for knee replacement.

In the study, “The Effect of Patients’ Sex on Physicians’ Recommendations for Total Knee Arthroplasty,” two standardized patients (one male and one female), both with moderate osteoarthritis of the knee, visited 71 different physicians (33 orthopaedic surgeons and 38 family physicians). Disease severity (based on physical examination and bilateral standing radiographs) was identical for both patients, and the patients had identically scripted scenarios describing their condition and their lifestyle; the only difference was their sex.

The study presented the following potential explanations for the findings:

  • Decisions to recommend TKA may be based on conscious attitudes or overt gender discrimination. It has been observed that some physicians attribute women’s symptoms to emotional rather than physical causes and therefore take them less seriously, even when the degree of disability is greater.
  • Decisions to recommend TKA may be based on unconscious gender bias. “Our study suggests that physicians are susceptible to the same unintentional gender biases that are pervasive in the rest of society,” Dr. Wright explains. “Physicians may not have recommended total knee arthroplasty to the female patient because an unconscious bias resulting from years of experience tells them, or they’ve heard from other physicians, that women do not receive the same benefit from total knee arthroplasty as men.” Statistics show that women historically receive TKA surgery at a more advanced stage of osteoarthritis as men and, due to the progression of disease, have worse surgical outcomes.
  • Discrepancies in recommendations for TKA may be based on the presentation style of male and female patients, despite identical clinical scenarios. Women are more likely to use a narrative style and voice complaints, while men are more likely to describe symptoms factually.

What do you think?
Laura L. Tosi, MD,
director of the bone health program at Children’s National Medical Center (CNMC) and associate professor of orthopaedics and pediatrics at George Washington University in Washington, D.C., served as moderator for a roundtable discussion of the study by four prominent orthopaedic surgeons. Participants included Mary I. O’Connor, MD, chair of the AAOS Women’s Health Issues Advisory Board; Charles L. Nelson, MD, associate professor of orthopaedic surgery at the Hospital of the University of Pennsylvania; Audrey K. Tsao, MD, an orthopaedic surgeon specializing in total joint reconstruction at Sun West Orthopedics, Sun City West, Ariz.; and Thomas P. Vail, MD, chair of orthopaedic surgery at University of California, San Francisco.

Laura L. Tosi,

Unfortunately, this study could not analyze any differences in the recommendations made by male and female orthopaedic surgeons because it included only three female surgeons. But both female and male family practice physicians had the same rate of recommendation for TKA to the male patient (67 percent). Of the 12 female family practice physicians, 42 percent recommended TKA to the female patient, as did 31 percent of the male family practice physicians.

In my mind, this raises the question of how we are trained in medicine. Perhaps even female physicians are molded through our medical school and residency training to have an unconscious bias with regard to gender-influenced communication style. Some studies have suggested that male patients “speak the language of the doctor” better than female patients.

Dr. Nelson: The factors behind both gender- and ethnic-based disparities are complex and multifactorial. Physicians, like the rest of society, have both conscious and unconscious biases and beliefs as well as prior experiences that dictate how they respond to specific circumstances. A physician may not recommend TKA for a female patient compared with a male with “moderate” osteoarthritis for the following reasons:

  • Belief that the female patient is less likely to agree to have TKA
  • Belief that the female patient is less likely to do well with TKA
  • Belief that the female patient is more accepting of pain and functional limitations of OA and that the male patient will be more insistent that a remedy is needed as soon as possible to restore function and quality of life
  • Belief that the female patient’s pain complaint is exaggerated relative to the male patient or is less physical/mechanical/correctable with TKA
  • Belief the female patient is more likely to have pain related to an emotional issue such as depression or to a condition such as fibromyalgia
  • Perceived greater importance of a male patient’s quality of life, ability to function, or ability to support his family
  • Overt sexism and decreased empathy with the female patient’s pain and disability
  • Desire to protect the female patient from surgery because of perceived surgical pain or the cosmetic aftereffects associated with surgery such as a scar

I believe that some physicians may have one or more of these biases and thus may defer a recommendation of TKA for female patients who have moderate arthritis, when the need for joint replacement is not as clear as with end-stage OA.

Audrey K. Tsao, MD

Many women tend to express themselves in functions that involve care of family or home, whereas men may assume that their physicians automatically understand the significance of a disability. The experience and sensitivity of individual healthcare providers to these subtle nuances can affect their interpretations of individual disability and pain and may have a profound effect on their decisions to recommend TKA as a treatment alternative.

In treatment discussions, female and male patients may also have different attitudes of expectation. The way that many female patients discuss potential treatment options can be interpreted as a lack of desire to commit to a definitive treatment procedure such as TKA. As a result, the physician may recommend nonsurgical treatment. The discussion style of male patients may be more succinct and to the point, which the physician may interpret as the patient’s being more receptive to a definitive surgical procedure. This, in conjunction with the commonly held perception that patient satisfaction with TKA is not as high among women as among men, especially in kneeling and squatting activities, may lead to a bias toward less surgical treatment for women.

Thomas P. Vail, MD

Dr. Tosi: Which of the three possible explanations for this study’s results do you find most compelling? Is the causality a combination of them, or is there a further possibility not identified by Dr. Wright and his team?Dr. O’Connor: The authors’ results reflect our society. Like all individuals, physicians have biases, including bias based on sex and gender. When we see patients, we must guard against stereotypes.
Dr. Tsao: Objective symptoms of pain and disability can often be described in differing functional contexts. It may be that the limitations of function described by female patients are interpreted differently based on the experience of the individual healthcare provider. This bias may be heightened by an expressive justification for considering the condition a disability.
Dr. Vail: Interestingly, women do seem to experience OA in greater numbers than men, based upon statistics from the Centers for Disease Control and Prevention. Additionally, from 1979 to 2002, the rate of knee replacement procedures among patients older than age 65 has increased by 8-fold. The overall incidence of knee replacement has been consistently greater in women than in men.
Dr. O’Connor: I try to look at every patient as if he or she were a member of my family and treat each patient with the same value and importance. But I must admit that certain patients can be difficult. For example I would much prefer to perform a TKA on a thin patient than on an obese patient (and I operate on many obese patients).
Dr. Nelson: The most critical step, from a physician’s point of view, in providing the best possible care to all patients and preventing disparities in health care due to gender, ethnicity, socioeconomic status, religion, obesity, or sexual orientation is self-awareness of individual biases. I spend extra time and effort in communication, patient education, and peer interaction to prevent biases from affecting my decision making.

No doubt important and subtle factors subconsciously come into play when a physician makes the decision to operate. The presentation and demeanor of the patient, the physician’s interpretation of the patient’s clinical picture, and the rapport that the physician and patient are able to achieve at the first visit are among these factors. Although orthopaedic surgeons generally agree on the appropriate indications for surgery, deciding when to operate remains firmly in the realm of the physician-patient relationship and lacks standardization.

Dr. Tosi: Recognizing that we all have individual biases, what strategies do you employ in your practice to minimize your personal prejudices and keep them from interfering with patient care?

Mary I. O’Connor, MD

That doesn’t mean that I don’t believe that obese patients wouldn’t benefit from the surgery, but perhaps my own bias permeates my interactions. I could be sending obese patients some relatively negative signals as compared to thin patients. I am sensitive to that and try to focus on the obese patient’s pain and not body size. I also work hard with my support team to remind them that our goal to provide the best care we can to every patient, regardless of body size.

I suspect that we, as ortho­paedic surgeons, have similar unconscious biases relative to female patients. We worry that women will be more difficult to take care of, more sensitive to pain, or want to talk too much when we visit while on rounds. Overcoming our biases is not easy.

Charles L. Nelson, MD

Unconscious biases are more difficult to address. I believe the best way to identify unconscious biases is to question patients, coworkers, and peers who are not dependent upon you and who can honestly share their perceptions of your biases. Additionally, taking implicit association tests, as Dr. Wright suggested in the study, is an effective method of assessing for any biases.

Dr. Tsao: I try to look at any objective data such as pain scores and diagrams first. I also try to gather any information about changes in patient activity, such as sleep patterns. This enables me to use the patient as the primary control and to gauge function in an individualized manner that accounts for both gender and cultural differences.

Expressions of pain associated with disability are also important, but this should be done in conjunction with family members who may be present and can give a clearer idea of the degree of disease. A woman may minimize her pain when discussing it with her physician, but her family can often assist in describing the severity of disability that she may not express. My staff can often elicit this information as well.

Dr. Vail: In my own practice, I review the objective measures of arthritis through the physical exam, radiographs, and a clinical history to establish whether a knee replacement is appropriate to consider. I then move to a discussion about “the decision to operate.” This part of the discussion requires input from the patient about quality of life, expectations of surgery, and risks versus benefits. Involving family members in this part of the discussion is often helpful.

Dr. Tosi: Obviously, to help address the healthcare disparities identified in the Wright study, the AAOS must continue to remain patient-focused. So, where do we go from here?

Dr. O’Connor: Continued education related to gender and racial healthcare disparities is essential to improve the care of our patients. We also need to emphasize that our research efforts and reporting of clinical studies should include analysis based on patient sex and race so we can learn more about where we are now and how we can improve. The 2010 AAOS Musculoskeletal Healthcare Disparities research symposium, which I am cochairing with Dr. Nelson and Carlos Lavernia, MD, is a step in the right direction.

Dr. Nelson: I believe the AAOS has a responsibility to repeatedly educate members on the existence of musculoskeletal health disparities; to emphasize better self-awareness, improved cultural competence, and enhanced communication skills of members; and to encourage and support research to better understand these disparities. Progress in this area will only be made by repeatedly highlighting these disparities and the importance of eliminating them.

Dr. Vail: Promoting clinical research and professional discussion of these issues will elevate the level of sophistication in decision-making among orthopaedic surgeons. Despite abundant information regarding the objective measures of arthritis and the indications for joint replacement, substantial regional, ethnic, socioeconomic, and gender variations exist in the incidence of TKA.

Dr. Tsao: We still don’t understand the reasons for gender disparities in surgery. We need research into behavioral differences so that we can develop long-term recommendations for changes in delivery of care and establish definitive changes in general clinical practice.

Borkhoff CM, Hawker GA, Kreder HJ, Glazier RH, Mahomed NN, Wright JG. The effect of patients’ sex on physicians’ recommendations for total knee arthroplasty. CMAJ 2008;178:681-687.