although population-based incidence data are not available prior to 1975. Trends over the past three decades are very similar by sex. Declines in mortality through 2000 are attributed to improvements in treatment (12%), changing patterns in CRC risk factors (35%), and screening (53%).43 However, screening likely played an even larger role in more recent trends given its steep increase since 2000.52 Rapid declines in CRC death rates of about 3% per year from 2002 to 2012 slowed to 2% per year from 2012 to 2017. Age-specific mortality trends Like incidence, CRC mortality trends vary by age (Figure 7). Among older adults, decades of rapid declines have slowed, from 1% annually during 2004-2013 to 0.6% during 2013-2017 in those ages 50-64 years and from 3.3% to 2.6%, respectively, in those ages 65 and older. In contrast, CRC death rates have increased in individuals younger than 50 years of age by 1.3% per year since 2004. Racial/ethnic mortality trends CRC death rates in whites began a slow decline in the early 1970s that accelerated over time. In contrast, death rates in blacks increased from the early 1970s until 1990, then decreased sluggishly during the 1990s before matching the decline in whites in the early 2000s (Figure 8). As a result of these divergent trends, although CRC death rates in blacks were 10% lower than those in whites in the early 1970s, they were almost 50% higher in 2005. The widening racial disparity was largely driven by trends for distant-stage disease, which declined in whites while remaining stable in blacks through the mid-2000s.53 About half of the racial disparity in mortality is attributed to a combination of less screening and lower stage-specific survival rates among blacks.30 Since the early 2000s, CRC death rates have declined consistently by 1.8% per year in Hispanics and APIs and by 2.8% per year in blacks; however, rates were stable in AIANs during this time, and in whites declines slowed from 2.5% per year during Rates are age adjusted to the 2000 US standard population. Incidence rates are adjusted for reporting delays and exclude appendix. White and black race are not mutually exclusive from Hispanic ethnicity. Source: Incidence – SEER program, 2019. Mortality – NCHS, 2019. ©2020, American Cancer Society, Inc., Surveillance Research Rate per 100,000 population Figure 8. Trends in Colorectal Cancer Incidence (1975-2016) and Mortality (1970-2017) Rates by Race, US Incidence, blacks Mortality, blacks Mortality, whites Mortality, Hispanics Incidence, whites Incidence, Hispanics a result, the black-white gap has slowly begun to narrow. How does colorectal cancer occurrence vary by state? The geographic pattern of CRC has changed dramatically over the past several decades. In contrast to the 1970s and 1980s, when the burden was highest across the Northeast and lowest in the South,54 today it is highest in parts of the South, Midwest, and Appalachia and lowest in the West and Northeast. Current incidence rates range from 49 (per 100,000) in Kentucky to 30 in Utah, while death rates range from 18 in Mississippi and West Virginia to 11 in Connecticut and Utah (Figure 9). This shift is consistent with the racial and socioeconomic crossover in disease burden that occurred during the latter half of the 20th century because of changes in dietary and smoking patterns, as well as differences in access to early detection and high-quality treatment.55 For example, CRC mortality among residents of poor counties was 20% lower than that among residents of affluent counties in the early 1970s, but is currently 30% to 40% higher.54, 56 Geographic patterns are generally similar for blacks and whites, particularly for mortality, highlighting the importance of socioeconomic status over race in cancer disparities.57 Table 2 shows state-level incidence and death rates by race/ethnicity. Consistent with overall incidence, rates in NHWs and blacks are lowest in the West and highest in the South and Midwest. However, among Hispanics there is no clear pattern, perhaps reflecting geographic heterogeneity within this population in terms of place of birth and duration of residence, both of which influence CRC risk. Although data for AIANs are too sparse to provide by state, a recent study found that