To the Editor:
I read with great interest the article by Silber et al.1 reporting a statistically significant increase in operative times between black and white Medicare patients who were very closely matched for age, sex, procedure, comorbidities, hospital, risk score, and body mass index. Particularly striking is the finding that when the difference in procedure times was greater than 30 min, black patients were significantly more likely to have the longer procedure time (a worse surgical outcome).
The literature documenting racial disparities in health and health care is extensive, particularly in the primary care and public health arenas. Large, systematic studies looking at racial disparities in surgical care and outcomes have been far fewer and have recently concentrated on surgical volume of the hospitals attended as the chief cause of disparity. A large study in 2005 using the Medicare database confirmed earlier findings that blacks are consistently more likely to die after major surgery and attributed this mortality difference mainly to low surgical procedure volumes at the hospitals attended.2 A study in 2006 of racial disparity in surgical complications between black and white patients based on New York State hospital discharge data found that these differences were due mainly to comorbidities and hospitals attended,3 and a study in 2010 matched for comorbidities and essentially replicated those findings.4
Yet after matching for comorbidity and hospital, a racial disparity in operative time, another clinically significant surgical outcome, still persists. What explains this? The authors posit but did not match for ecological factors. Their study would have been helped enormously by matching for income and education. They also speculate about racial disparity in who performs the surgery (attending vs. resident surgeon) but admit that these data cannot be captured from Medicare claims data or chart abstracts. A study examining racial disparity in operative times between similarly matched Medicare patients at nonteaching hospitals could be designed to indirectly address that question. Questions such as these must be asked and answered before we can hope to discern the causes of unfair disparities, which threaten our patients’ health. This article models a well-designed study for doing just that.
The author declares no competing interests.