From 1999 to 2018, the NHANES included 49,720 adults aged between 20 and 85 years, containing information about PIR and cancer or malignancy (Fig. 1). The general characteristics of these adults are detailed in the Table 1. Among the 12 811 participants in the high-income group, 6553 (51.2%) were men, 6258 (48.8%) were women. Among the 26 484 participants in the middle-income of the population, 12 857 (48.5%) were men, 13 627 (51.5%) were women. Among the 10 425 participants in the low-income group, 5844 (56.1%) were women, 4581 (43.9%) were men. The overall prevalence of lung cancer was 0.3% (n = 137), breast cancer was 1.5% (n = 727), esophagus cancer was 0.1% (n = 30), stomach cancer was 0.1% (n = 41), colon and rectum cancer were 0.8% (n = 383) and liver cancer was 0.1% (n = 35).
Table 1
Characteristics of Study Participants, 1999-2018.
Characteristics | | No. (weighted %) | p value |
Total (N=49720) | Family income to poverty ratio ≤1.0 (n =10425) | Family income to poverty ratio 1.0-4.0 (n =26484) | Family income to poverty ratio ≥4.0 (n = 12811) |
Mean (SD) age, y | 49(34, 64) | 44(30,62) | 50(34,67) | 50(37,62) | <0.001 |
Sex | | | | | <0.001 |
Men | 23991(48.3) | 4581(43.9) | 12857(48.5) | 6553(51.2) | |
Women | 25729(51.7) | 5844(56.1) | 13627(51.5) | 6258(48.8) | |
Race/ethnicity | | | | | <0.001 |
Non-Hispanic white | 22557(45.4) | 3218(30.9) | 11790(44.5) | 7549(58.9) | |
Non-Hispanic black | 10315(20.7) | 2577(24.7) | 5651(21.3) | 2087(16.3) | |
Mexican American | 8407(16.9) | 2642(25.3) | 4756(18.0) | 1009(7.9) | |
Other | 8441(17.0) | 1988(19.1) | 4287(16.2) | 2166(16.9) | |
Marital status | | | | | <0.001 |
Married | 25839(52.5) | 3605(34.9) | 13616(51.9) | 8618(67.9) | |
Not married | 23392(47.5) | 6713(65.1) | 12609(48.1) | 4070(32.1) | |
Health insurance | | | | | <0.001 |
Covered | 39650(79.8) | 6652(63.9) | 20866(78.9) | 12132(94.7) | |
Not covered | 10021(20.2) | 3752(36.1) | 5596(21.1) | 673(5.3) | |
Education levels | | | | | <0.001 |
Less than high school | 13232(26.6) | 5004(48.1) | 7359(27.8) | 869(6.8) | |
High school diploma or GED certificate | 11491(23.1) | 2472(23.8) | 7031(26.6) | 1988(15.5) | |
Greater than high school | 24938(50.2) | 2926(28.1) | 12061(45.6) | 9951(77.7) | |
Citizenship status | | | | | <0.001 |
US citizenship | 42926(86.5) | 7914(76.2) | 22968(86.8) | 12044(94.1) | |
Non-US citizenship | 6724(13.5) | 2473(23.8) | 3491(13.2) | 760(5.9) | |
BMI, kg/m2 | | | | | |
<25.0 | 13909(29.8) | 2960(30.4) | 7147(28.8) | 3802(31.5) | |
25.0-29.9 | 15699(33.6) | 3061(6.6) | 8319(33.5) | 4319(35.8) | |
≥30.0 | 17052(36.5) | 3731(38.3) | 9369(37.7) | 3952(32.7) | |
Drinking status | | | | | <0.001 |
Non-drinker | 6142(17.6) | 1707(25.1) | 3440(18.9) | 995(10.0) | |
༜2drinks/d | 10229(29.2) | 1362(20.0) | 5206(28.6) | 3661(36.8) | |
≥2drinks/d | 18616(53.2) | 3744(54.9) | 9584(52.6) | 5288(53.2) | |
Smoking status | | | | | <0.001 |
Non-smoker | 26923(73.8) | 5140(61.6) | 14128(72.0) | 7655(82.8) | |
Former smoker | 1094(3.00) | 536(6.4) | 995(5.1) | 373(4.0) | |
Current smoker | 8475(23.2) | 2760(29.9) | 4497(22.9) | 1218(13.2) | |
Leisure time physical activity | | | | | <0.001 |
Never | 22427(45.1) | 5814(55.8) | 12533(47.3) | 4080(31.8) | |
Moderate | 13615(27.4) | 2181(20.9) | 6910(26.1) | 4524(35.3) | |
Vigorous | 13678(27.5) | 2430(23.3) | 7041(26.6) | 4027(32.8) | |
Continuous variables are expressed as mean standard deviation (SD), while classified variables are expressed as numbers and their proportions. We use Chi-square test for classified variables, one-way ANOVA for normal continuous variables and Kruskar-Wallis test for skewed continuous variables. |
Abbreviations: GED, General Educational Development; BMI, body mass index. |
The main demographic differences between the three groups included marital status and educational levels. Most of the high-income population is married (8618 [67.9%]) and has a college degree or above (9951 [77.7%]). Less than half of the low-income participants (3605 [34.9%) were married, and only a small proportion (2926 [28.1%) had a college degree or higher. The race/ethnicity composition also varied between the three groups. Among the high-income group, 7,549 participants (58.9%) were self- identified White, 2,087 participants (16.3%) were Black, and 1,009 (7.9%) were Mexican. In low-income populations, the percentage of Whites was lower (3 218 [30.9%]) and was higher in Blacks (2 577 [24.7%] or higher for Mexican (2 642 [25.3%]).
Overall Prevalence of Cancer Disease by Income Group
The prevalence of lung cancer was lower in high-income participants than in middle-income participants (0.15% [n= 19] vs 0.35% [n= 92], p <0.001) (Fig. 2A). We found an inverse relationship between income levels and breast cancer. For the low-income stratum, the prevalence of breast cancer was 1.12% [n = 117], but the number of adults in the middle (1.48% [n = 391], p = 0.009) and high-income levels (1.71% [n = 219], p <0.001) has increased (Fig. 2B). We found no statistically significant relationship between income levels and the prevalence of esophagus cancer, stomach cancer, colon and rectum cancer or liver cancer (Fig. 2C-F).
Trends in Cancer Disease Prevalence
In the high-income group, the prevalence of cancer disease decreased between 1999-2008 and 2009-2018. The prevalence of lung cancer decreased from 0.172% (n = 11) in 1999-2008 to 0.124% (n = 8) in 2009-2018 (p = 0.878); esophagus cancer from 0.627% (n = 4) in 1999-2008 to 0.016% (n = 1) in 2009-2018 (p = 0.217); colon and rectum cancer 0.736% (n = 47) in 1999-2008 to 0.700% (n = 45) in 2009-2018 (p = 0.808) and liver cancer 0.078% (n = 5) in 1999-2008 to 0.047% (n = 3) in 2009-2018 (p = 0.506). In contrast, the prevalence of breast cancer increased from 1.614% (n = 103) in 1999-2008 to 1.805 (n = 116) in 2009-2018 (p = 0.278) and stomach cancer from 0.031% (n = 2) in 1999-2008 to 0.046 (n = 3) in 2009-2018 (p = 1) (Fig. 3 and Supplemental Fig. 1).
In the middle-income group, lung cancer prevalence decreased from 0.378% (n = 49) in 1999-2008 to 0.318% (n = 43) in 2009-2018 (p = 0.878); the prevalence of stomach cancer decreased from 0.116% (n = 15) in 1999-2008 to 0.096% (n = 13) in 2018 (p = 0.621). In contrast, breast cancer prevalence increased from 1.421% (n = 184) in 1999-2008 to 1.529% (n = 207) in 2009-2018 (p = 0.463); the prevalence of esophagus cancer increased from 0.046% (n = 6) in 1999-2008 to 0.096% (n = 13) in 2009-2018 (p = 0.131); colon and rectum cancer prevalence non-significantly increased from 0.842% (n = 109) in 1999-2008 to 0.864% (n = 117) in 2009-2018 (p = 0.840) and liver cancer prevalence slightly increased from 0.077% (n = 10) in 1999-2008 to 0.089% (n = 12) in 2009-2018 (p = 0.747) (Fig. 3 and Supplemental Fig. S1).
Below the federal poverty level, the prevalence of lung cancer increased from 0.241% (n = 11) in 1999-2008 to 0.256% (n = 15) in 2009-2018 (p = 0.878); the prevalence of breast cancer increased from 0.986% (n = 45) in 1999-2008 to 1.229% (n = 72) in 2009-2018 (p = 0.242); the prevalence of esophagus cancer increased from 0.044% (n = 2) in 1999-2008c to 0.068% (n = 4) in 2009-2018 (p = 0.702); the prevalence of colon and rectum cancer increased from 0.547% (n = 25) in 1999-2008 to 0.683% (n = 40) in 2009-2018 (p = 0.384) and liver cancer prevalence slightly increased from 0.044% (n = 2) in 1999-2008 to 0.051% (n = 3) in 2009-2018 (p = 1). Conversely, the prevalence of stomach cancer decreased from 0.088% (n = 4) in 1999 to 0.068% (n = 4) in 2009-2018 (p = 0.736) (Fig. 3 and Supplemental Table 1).
Trends in the Association Between Income Group and Cancer Disease
Adjusting the models for demographic variables, the odds of reporting lung cancer were reduced in the highest resource population over time (odds ratio [OR], 0.431; 95%CI, 0.257-0.723; p = 0.001). Conversely, the richest participants had higher odds of reporting breast cancer (OR, 1.203; 95%CI, 1.001-1.446; p = 0.049), while no significant change was observed in esophagus cancer (OR, 0.511; 95%CI, 0.181-1.446; p = 0.206), stomach cancer (OR, 0.496; 95%CI, 0.182-1.352; p = 0.171), colon and rectum cancer (OR, 0.891; 95%CI, 0.686-1.157; p = 0.386) or liver cancer (OR, 0.627; 95%CI, 0.266-1.475; p = 0.285) (Supplemental Table 1-6). When cancer risk factors were included in the model, the odds of high-income group reporting lung cancer remained low over time (OR, 0.452; 95%CI, 0.234-0.875; p = 0.019), but no statistically significant change in the odds of reporting breast cancer (OR, 1.127; 95%CI, 0.899-1.412; p = 0.300) (Supplemental Table 7-12).
Over time, those in the middle-income level had higher odds of reporting lung cancer (OR,1.047; 95% CI, 0.657-1.668; p = 0.848), breast cancer (OR,1.041; 95% CI, 0.832-1.303; p = 0.725) and colon and rectum cancer (OR,1.061; 95% CI, 0.793-1.420; p = 0.689) but this difference was not statistically significant. In contrast, these participants were less likely to report esophagus cancer (OR,0.959; 95% CI, 0.344-2.677; p = 0.937), stomach cancer (OR,0.686; 95% CI, 0.304-1.545; p = 0.363) and liver cancer (OR,0.627; 95% CI, 0.266-1.475; p = 0.285), but the difference was not statistically significant (Supplemental Table 1-6). When cancer risk factors were included in the model, the risk trend had not changed and the difference was still not statistically significant (Supplemental Table 7-12).
Association Between Cancer Disease and Other Variables
Both logistic regression analysis models suggest that, in general, older age is associated with an increased likelihood of reporting cancer disease. The ORs of cancer disease ranged from 4.729 (95% CI, 1.330-16.822) to 11.776 (95% CI, 1.513-91.651) for participants aged 40 to 59 years and from 11.525 (1.319-100.709) to 38.696 (20.953-71.466) for people 60 years or older compared with the youngest age group (20-39 years). Conversely, men had a largely higher probability of cancer cancer disease than women (OR ranged from 1.448 [95% CI, 1.017-2.061] to 3.730 [95% CI, 1.567-8.881]), except for breast cancer 0.003 (95% CI, 0.001-0.011) and liver cancer 0.982(95% CI, 0.497-1.939).
Married vs nonmarried individuals had a lower probability of reporting a cancer disease (OR ranged from 0.194 [95% CI, 0.049-0.767] to 0.935 [95% CI, 0.765-1.141]), health insurance covered vs. not covered participants had higher odds of reporting cancer disease (OR ranged from 2.330 [95% CI, 1.570-3.460] to 8.152 [95% CI, 1.073-61.923]), and those with US citizenship had higher probability of reporting breast cancer compared with those without US citizenship (model 1: OR, 1.591 [95% CI, 1.041-2.431]; model 2: OR, 1.729 [95% CI, 1.005-2.974]) (Supplemental Table 1-12).
In the first model, the association between race/ethnicity and cancer diseases was mixed, which included only demographic variables. Compared to Black participants, White participants had a higher probability of reporting lung cancer (OR, 1.269; 95% CI, 0.829-1.943), breast cancer (OR, 1.539; 95% CI, 1.244-1.905), esophagus cancer (OR, 2.787; 95% CI, 0.946-8.211), colon and rectum cancer (OR, 1.497; 95% CI, 1.137-1.970) and liver cancer (OR, 1.311; 95% CI, 0.517-3.321) and a lower probability of reporting stomach cancer (OR, 0.660; 95% CI, 0.307-1.421) (Supplemental Table 1-6). The second model, which included cancer risk factors, yielded similar but more pronounced results. Compared with black participants, Hispanic and Mexican participants had a lower possibility of reporting cancer diseases (OR ranged from 0.528 [95% CI, 0.271-1.028] to 0.842 [95% CI, 0.551-1.287]) but not esophagus cancer (OR, 1.645; 95% CI, 0.140-19.279) or stomach cancer (OR, 1.493; 95% CI, 0.376-5.937) (Supplemental Table 7-12).
In the first model, an inverse correlation was found between level of education and the probability of reporting cancer disease. People with a high school diploma or general education development (GED) certificate (OR ranged from 0.497 [95% CI, 0.213-1.160] to 0.948 [95% CI, 0.604-1.489]) followed by those with a college degree or above (OR ranged from 0.421 [95% CI, 0.191-0.924] to 0.822 [95% CI, 0.635-1.064]) are the least likely to protect cancer disease than those without a high school diploma or GED certificate (Supplemental Table 1-6). When cancer risk factors were included in the second model, both groups had a generally lower probability of reporting cancer disease with higher education (Supplemental Table 7-12).