One-third of colorectal cancer survival disparities a product of treatment gaps
Click Here to Manage Email Alerts
Treatment disparities mediated about one-third of OS differences observed between white and Black patients with stage IV colorectal cancer, according to a retrospective cohort study presented at Gastrointestinal Cancers Symposium.
This finding may be an underestimate because of the limited granularity of available treatment mediator variables, according to Douglas S. Swords, MD, MS, surgical oncology fellow at The University of Texas MD Anderson Cancer Center.
“These results suggest that the oncology community should invest resources into developing interventions aimed at eliminating socioeconomic treatment disparities and measure whether such interventions improve survival,” Swords said in a press release.
Swords and colleagues sought to estimate the proportion of survival disparities in stage IV colorectal cancer that are mediated by treatment disparities. Treatment mediators they considered included chemotherapy, metastasectomy and treatment at more than one Commission on Cancer-accredited facility, which researchers distinguished from other social determinants of health that might mediate survival, such as social/family support, housing and food insecurity. They also did not analyze primary site surgery as a treatment mediator due to the potential for immortal time bias.
Researchers used the 2010 to 2016 National Cancer Database to identify 70,773 patients aged 18 to 80 years with stage IV colorectal cancer.
OS served as the study’s primary outcome.
Researchers considered a patient’s socioeconomic status (SES) to be low if they lived in a quartile 1 ZIP code-level for income and education, and high if they lived in a quartile 4 ZIP code for income and education. All other patients were deemed in the middle.
“In pretty much any cancer you look at, patients with low socioeconomic status have shorter survival,” Swords said during his virtual presentation. “Not only that, but low SES patients also often have lower treatment rates, and low SES Black and Hispanic patients often have lower treatment rates that low SES white patients.”
Researchers used inverse odds weighting mediation analyses, which allow for the simultaneous analysis of multiple mediators, to calculate the proportion mediated (PM) for individual and collective mediators in two analyses: SES as the exposure regardless of race, and the intersection of SES with race for Black patients and white patients.
In the first analysis, the use of chemotherapy increased from 86.9% among those with low SES to 89.3% for those with middle SES and 91.5% for those with high SES. Researchers reported similar increases for use of metastasectomy (13.1% vs. 15.1% vs. 18.1%) and treatment at more than one Commission on Cancer-accredited facility (10.9% vs. 13.7% vs. 17.4%).
Five-year OS appeared highest for those with high SES (21.8%), followed by those with middle SES (18.2%) and low SES (16.3%; P < .001).
Based on these data, researchers calculated a PM of 8.9% (95% CI, 5.1-12.8) for chemotherapy for low SES patients, meaning use of chemotherapy mediated 8.9% of the survival differences between low and high SES patients. The PMs reached 15.6% (95% CI, 11.5-19.7) for metastasectomy, 10.6% (95% CI, 7.8-13.5) for treatment at more than one Commission on Cancer-accredited facility, and 29.2% (95% CI, 23.5-34.8) for a composite of these treatment mediators. The results were similar when researchers compared middle vs. high SES patients.
In the second analysis, researchers found the composite PM was 28.2% (95% CI, 21.1-35.4) for Black patients with low SES vs. high SES, whereas it was only 19% (95% CI, 13.9-24) for white patients with low SES vs. high SES.
In a subgroup analysis of patients with isolated liver metastases, researchers calculated a composite PM of 41.6% (95% CI, 28.4-54.8) for low SES and 34% (95% CI, 23.5-55.5) for middle SES vs. high SES.
There are three take-home messages from these findings, Swords said.
“First, treatment disparities partially mediate SES-based survival disparities in stage IV colorectal cancer,” he said. “These are likely underestimates of the true proportion mediated, because we lacked granular variables on many important details, like number of chemotherapy regimens, type of regimen, supportive care, etc. Also, the proportion mediated was larger for Black patients, and this is because they experience larger treatment disparities.
“Second, these data suggest that policies that address treatment disparities could be expected to narrow survival gaps, although probably not eliminate them,” he added. “And, finally, I want to highlight that nontreatment-related pathways are also important, and we should pay just as much attention to developing interventions that address these mediators.”
Perspective
Back to Top
Michael J. Hall, MD, MS
Swords and colleagues used mediation analysis to examine the fraction of heath care disparities attributable to race and SES among Black patients and white patients treated for colorectal cancer, with ZIP code used as a surrogate for SES. They found variability in 5-year OS by SES (low SES, 16.3% vs. high SES, 21.8%), as well as in use of chemotherapy (low SES, 86.9% vs. high SES, 91.5%), metastasectomy (low SES, 13.1% vs. high SES, 18.1%) and treatment at a comprehensive center (low SES, 10.9% vs. high SES, 17.4%).
Treatment disparities mediated approximately 20% to 35% of the OS disparity in low and middle SES patients, whereas race mediated 20% to 35% of the OS disparity among Black patients and 14% to 30% of the OS disparity among white patients.
Overall, these analyses demonstrate the power of pairing this statistical approach with a well-annotated clinical data resource like the National Cancer Database to estimate the individual contribution of factors, such as SES and race, to survival outcomes, providing insight into the magnitude with which predictors that have been associated with poorer health care outcomes, such as low SES and Black race, influence survival both independently and together.
Collapse
Swords DS, et al. Abstract 13. Presented at: Gastrointestinal Cancers Symposium (virtual meeting); Jan. 15-17, 2021.
Click Here to Manage Email Alerts