Impact of Medicaid Expansion on Pancreatic Cancer: An Examination of Sociodemographic Disparity in 1-Year Survival

J Am Coll Surg. 2022 Jan 1;234(1):75-84. doi: 10.1097/XCS.0000000000000018.

Abstract

Background: This study examined the effect of Medicaid expansion on 1-year survival of pancreatic cancer for nonelderly adults. We further evaluated whether sociodemographic and county characteristics alter the association of Medicaid expansion and 1-year survival.

Study design: We obtained data from the Surveillance Epidemiology and End-Results dataset on individuals diagnosed with pancreatic cancer from 2007 to 2015. A Difference-in-Differences model compared those from early-adopting states to non-early-adopting states, before and after adoption (2014), while taking into consideration sociodemographic and county characteristics to estimate the effect of Medicaid expansion on 1-year survival.

Results: In the univariable Difference-in-Differences model, the probability of 1-year survival for pancreatic cancer increased by 4.8 percentage points (ppt) for those from Medicaid expansion states postexpansion (n = 35,347). After adjustment for covariates, the probability of 1-year survival was reduced to 0.8 ppt. Interestingly, after multivariable adjustment the effect of living in an expansion state on 1-year survival was similar for men and women (0.6 ppt for men vs 1.2 ppt for women), was also similar for Whites (2.6 ppt), and was higher in those of other races (5.9 ppt) but decreased for Blacks (-2.0 ppt). Those who were insured (-0.1 ppt) or uninsured (-2.2 ppt) experienced a decrease in the probability of 1-year survival; however, those who were covered by Medicaid at diagnosis experienced an increase in the probability of 1-year survival (7.4 ppt).

Conclusions: Medicaid expansion during or after 2014 is associated with an increase in the probability of 1-year survival for pancreatic cancer; however, this effect is attenuated after adjustment for sociodemographic characteristics. Of note, the positive association was more pronounced in certain categories of key covariates suggesting further inquiry focused on these subgroups.

Publication types

  • Research Support, N.I.H., Extramural
  • Research Support, Non-U.S. Gov't

MeSH terms

  • Adult
  • Female
  • Humans
  • Insurance Coverage
  • Male
  • Medicaid*
  • Medically Uninsured
  • Pancreatic Neoplasms*
  • Patient Protection and Affordable Care Act
  • United States / epidemiology
  • White People