Letter to the Editor: Editorial: Beware of Studies Claiming that Social Factors are “Independently Associated” with Biological Complications of Surgery

To the Editor,

As two young physicians and clinician scholars at the beginning of our orthopaedic surgery residencies, we have the great honor and privilege to learn to care for those with a wide range of musculoskeletal pathologies and to assist in research efforts to answer some of our specialty’s most vexing questions. Further, we feel an additional responsibility to lead by our actions, attitudes, and words to combat racism, diversify orthopaedic surgery, and care for all patients equally [1]. To accomplish these lofty but necessary goals, we believe we must be more careful in our language and evaluation of analyses, especially in the scholarly work impacting daily clinical orthopaedic care.

Many of the warnings urging appropriate language and understanding of the limitations of data in orthopaedic surgery clinical research are articulated eloquently by Seth S. Leopold MD in his editorial, “Beware of Studies Claiming that Social Factors are ‘Independently Associated’ with Biological Complications of Surgery” [2]. After reading this piece, we were inspired to reflect and assess our own research efforts and realized this was a good opportunity for personal and professional growth. We hope others take a similar pause to consider how they would handle the issues Dr. Leopold’s editorial raised in their scholarly contributions as well. Through this kind of introspection, we can help our specialty ensure that the clinical research we perform results in better care for all patients.

The editorial reinforced that many commonly used social characteristics evaluated, such as insurance status or race, cannot cause biological complications, such as wound infections [2]. We agree that the use of these variables in clinical research to guide care decisions is imperfect at best and dangerous at worst, perhaps feeding into implicit and even explicit bias in many care scenarios. For this reason, we think that clinician scientists should take extreme caution when reporting research findings to ensure that readers do not inappropriately associate social characteristics with biological outcomes and then carry that bias forward into their own scholarly work, policy initiatives, and clinical care.

However, we want to emphasize that we still believe the collection and inclusion of these variables in clinical research are quite valuable. Why? Because the variables we would prefer to include—such as measures of health disparities (difference in access to appropriate care), structural racism, systemic oppression, implicit bias, and generational trauma—are simply not readily available. Therefore, patient factors such as insurance status and race, among others, can be considered proxy variables for unknown—yet incredibly important—metrics that may actually impact biological outcomes. This general idea is consistent with the thoughts of Simkin and colleagues [4] in their letter to the editor in response to this editorial. However, we feel that it is of utmost importance to move beyond faulty surrogate measures, whenever possible, to include the true variables we want to measure and address in our care delivery policies [3].

In the end, our hope and expectation are that those dedicated to tackling race-associated health outcomes disparities are willing to continue their research with the data available, while also taking on the challenge of better defining covariates over time. Crucially, whenever imperfect substitutes are utilized, authors should acknowledge this in the limitations sections of their papers.

We are heartened by the efforts being made across medicine, including within orthopaedic surgery, to promote diversity and inclusion, as well as tackle systemic barriers to equitable health care for all. Clinical research plays a crucial role in this movement; therefore, we must all take responsibility to ensure our data and the conclusions drawn from our work are as accurate and appropriately worded as possible. We pledge to the following two principles and hope our fellow trainees, as well as seasoned orthopaedic surgeons, will make similar commitments: (1) When including social factors in clinical research, we vow to acknowledge explicity that they are proxy variables for the true characteristics of interest. Examples include health disparities (difference in access to appropriate care), structural racism, systemic oppression, implicit bias, and generational trauma. (2) As we continue to conduct clinical research aimed at improving musculoskeletal care for all patients, we will include direct measures of health disparities, structural racism, and systemic oppression, among other similar variables, whenever possible and/or they are available for use.