Skip to main content
. 2022 Jul 31;19:101185. doi: 10.1016/j.ssmph.2022.101185

Table 2.

Fine & Gray method for proportional hazard models presented as Sub-Distribution Hazard ratios (SHR) and 95% confidence intervals (CI) for the association between allostatic load and cancer death accounting for competing risks of all-cause mortality, among 41,218 NHANES participants with 2,559 cancer-related deaths, and 8,988 competing deaths.

No. Cancer Deaths (%)a
No.All-cause Deaths (%)a
Mean Survival Months (SE)b
Sub-Distribution Hazard Ratio (SHR) and 95% Confidence Interval (CI)
Unadjustedb Model 1b Model 2b Model 3b
Risk among All Adults (N = 41,218)
Allostatic Load
 Low Allostatic Load 785 (3.02) 2,315 (7.85) 359.44 (0.41) 1.00 (Referent) 1.00 (Referent) 1.00 (Referent) 1.00 (Referent)
 High Allostatic Load 1,774 (7.71) 6,673 (26.84) 335.45 (0.69) 2.40 (2.21–2.61) 1.28 (1.18–1.40) 1.21 (1.10–1.33) 1.14 (1.04–1.26)
Risk among Non-Hispanic White Adults (N = 19,078)
Allostatic Load
 Low Allostatic Load 453 (3.35) 1,424 (8.85) 343.12 (0.65) 1.00 (Referent) 1.00 (Referent) 1.00 (Referent) 1.00 (Referent)
 High Allostatic Load 957 (8.50) 3,783 (29.48) 322.45 (1.19) 2.28 (2.04–2.55) 1.32 (1.17–1.49) 1.28 (1.13–1.45) 1.18 (1.03–1.34)
Risk among Non-Hispanic Black Adults (N = 9,095)
Allostatic Load
 Low Allostatic Load 156 (3.26) 331 (6.43) 339.69 (0.83) 1.00 (Referent) 1.00 (Referent) 1.00 (Referent) 1.00 (Referent)
 High Allostatic Load 460 (6.92) 1,500 (22.88) 326.24 (1.16) 2.26 (1.88–2.71) 1.06 (0.87–1.31) 1.02 (0.83–1.26) 1.08 (0.87–1.34)
Risk among Hispanic Adults (N = 11,399)
Allostatic Load
 Low Allostatic Load 163 (1.68) 482 (3.84) 363.41 (0.59) 1.00 (Referent) 1.00 (Referent) 1.00 (Referent) 1.00 (Referent)
 High Allostatic Load 323 (3.84) 1,221 (14.99) 344.67 (1.15) 2.64 (2.18–3.18) 1.14 (0.93–1.40) 1.15 (0.93–1.41) 1.03 (0.84–1.28)
p-value for interactionc 0.29 0.33 0.27 0.37

Model 1 is adjusted for age.

Model 2 is adjusted for age, time period, and sociodemographic factors including sex, race (only in unstratified analysis), PIR, and education.

Model 3 is adjusted for age, time period, sociodemographic factors, and health factors including current smoker status, having ever diagnosed with cancer, ever diagnosed with congestive heart failure, ever diagnosed with heart attack.

a Percentages are weighted.

b Models are un-weighted.

c Interaction term between race/ethnicity and allostatic load on association with cancer death determined by Wald Chi-Square.