We read the recent article by Saeed et al with great interest.1 Examples of the “Hispanic paradox” (the epidemiologic finding of better health indicators among Hispanics compared with non-Hispanic whites, despite their lower socioeconomic status) include perinatal mortality, life expectancy, and cancer incidence. However, cancer survival is not one of these examples. Most survival studies do not demonstrate an advantage for Hispanic individuals. Instead, they indicate the opposite.2, 3

Population-based analyses using the Survival, Epidemiology, and End Results (SEER) database have 1 major strength: the lack of selection bias because the total population is represented. However, Saeed et al1 excluded subjects without a known birthplace (26% among Hispanics) and those with unstaged cancers (9% of nonsmall cell lung cancers). Both exclusions are sources of selection bias when considering race/ethnicity. Furthermore, as has been demonstrated repeatedly,4, 5 survival analysis by birthplace using registry data is biased because birthplace information is commonly obtained from death certificates, thereby making the outcome of survival analysis (death) intrinsically associated with birthplace.

After adding unstaged cancers and Hispanics with an unknown birthplace to the SEER database used by Saeed et al (SEER 17-registry database [2009] is publicly available),1 the hazard ratios (HRs) were found to be similar for all studied covariates except for Hispanics. The adjusted HR for “All Hispanic whites combined” was 1.026 (95% confidence interval [95% CI], 1.010-1.043). For “US-born Hispanics,” the HR was 1.118 (95% CI, 1.092-1.145); it was 1.000 for “foreign-born Hispanics” (95% CI, 0.972-1.029) and 0.910 for “Hispanics of unknown birthplace” (95% CI, 0.880-0.940). Overall, instead of an advantage noted for Hispanics, a significant disadvantage of 2.6% worse survival for Hispanic whites was found. Based on these data, there is no apparent “Hispanic paradox” for lung cancer survival.

Failure to consider these limitations in registry data when studying survival may produce artificial results and mask true disparities that are important from the public health standpoint. Hispanics now comprise 16% of the US population. Therefore, it is urgent that we improve the methods used to analyze the Hispanic cancer experience.


No specific funding was disclosed.


The authors made no disclosures.


  1. Top of page
  • 1
    Saeed AM, Toonkel R, Glassberg MK, et al. The influence of Hispanic ethnicity on nonsmall cell lung cancer histology and patient survival: an analysis of the Survival, Epidemiology, and End Results database. Cancer. 2012; 118: 4495-4501.
  • 2
    American Cancer Society. Cancer Facts & Figures for Hispanics/Latinos 2009-2011. Atlanta, GA: American Cancer Society; 2009.
  • 3
    Pinheiro PS, Williams M, Miller EA, Easterday S, Moonie S, Trapido EJ. Cancer survival among Latinos and the Hispanic Paradox. Cancer Causes Control. 2011; 22: 553-561.
  • 4
    Lin SS, Clarke CA, O'Malley CD, Le GM. Studying cancer incidence and outcomes in immigrants: methodological concerns. Am J Public Health. 2002; 92: 1757-1759.
  • 5
    Lin SS, O'Malley CD, Clarke CA, Le GM. Birthplace and survival among Asian women diagnosed with breast cancer in cancer registry data: the impact of selection bias. Int J Epidemiol. 2002; 31: 511-513; author reply 513.

Paulo S. Pinheiro MD, PhD*, * Epidemiology and Biostatistics, School of Community Health Sciences, University of Nevada at Las Vegas, Las Vegas, Nevada