Table 1 presented the sample characteristics for the whole sample and by race-stratification. Significant racial disparities were observed in 15 of the 23 types of ACEs, including two indicators for low family SES, five types of chronic conditions, two types of physical impairments, two childhood behavioral problems, and four types of risky parent behaviors. Specifically, low family SES, severe headache or migraine, vision impairment, grade retention, and father absence were more prevalent among Blacks than among Whites, while a reversed pattern was found for allergies, chronic ear problems, respiratory disorder, stomach problems, head trauma, smoking or other substance use, parent smoking, parent physical abuse, and maternal neglect. Regarding the outcome and mediators, compared to Whites, Blacks have lower cognition score, lower level of education, lower level of physical activity, but higher depressive symptoms. Among the covariates, we observed higher average age among Whites, higher proportion of females among Blacks, and lower rate of marriage among Blacks.
Table 1
Sample characteristics, Health and Retirement Study, United States
| Mean ± SD or % | p-value a |
| All (N = 9,894) | White (N = 7,272) | Black (N = 1,396) | Other (N = 1,224) |
Outcome | | | | | |
Cognition (2014-16) | 21.81 ± 4.86 | 22.49 ± 4.53 | 19.33 ± 5.17 | 19.74 ± 5.33 | 0.000 |
ACE Indicators | | | | | |
Mother low education (< 8 yrs) | 17.62 | 9.74 | 26.02 | 58.00 | 0.000 |
Family relocation due to financial strain | 17.53 | 16.28 | 18.02 | 24.36 | 0.000 |
Allergies | 11.69 | 13.43 | 8.27 | 7.64 | 0.000 |
Chronic ear problems | 9.80 | 11.73 | 3.79 | 8.13 | 0.000 |
Severe headache or migraine | 5.93 | 5.60 | 7.98 | 5.05 | 0.003 |
Respiratory disorder | 14.09 | 15.19 | 12.58 | 10.92 | 0.000 |
Stomach problems | 4.24 | 4.01 | 3.79 | 5.79 | 0.023 |
Heart or circulatory conditions | 3.31 | 3.63 | 2.66 | 2.61 | 0.087 |
Physical disability ( > = 6 months) | 3.92 | 4.09 | 3.55 | 3.33 | 0.345 |
Head trauma or concussion | 9.26 | 10.69 | 4.78 | 5.84 | 0.000 |
Speech impairment | 3.31 | 3.39 | 3.71 | 2.52 | 0.251 |
Vision impairment | 5.69 | 5.27 | 7.82 | 5.14 | 0.002 |
Learning disability | 2.97 | 2.75 | 3.62 | 3.59 | 0.088 |
Depression or other psychological problems | 4.45 | 4.53 | 3.86 | 4.67 | 0.547 |
Smoking or other substance use | 17.60 | 19.18 | 14.12 | 12.17 | 0.000 |
Grade retention | 16.47 | 14.69 | 19.81 | 23.37 | 0.000 |
Trouble with police | 5.93 | 5.93 | 6.87 | 4.86 | 0.096 |
Father present but unemployed for at least 6 months | 19.89 | 20.21 | 17.81 | 20.36 | 0.114 |
Father absent | 7.36 | 4.73 | 17.59 | 11.46 | 0.000 |
Parent smoking | 66.50 | 70.67 | 56.55 | 59.29 | 0.000 |
Parent substance abuse | 17.39 | 17.76 | 16.26 | 16.42 | 0.255 |
Parent physical abuse | 8.79 | 8.96 | 6.89 | 9.93 | 0.015 |
Maternal neglect | 1.62 ± .76 | 1.63 ± .75 | 1.49 ± .71 | 1.72 ± .86 | 0.000 |
Mediators | | | | | |
Education (in years) | 13.00 ± 2.99 | 13.51 ± 2.45 | 12.44 ± 2.78 | 10.62 ± .45 | 0.000 |
Physical activity | 7.96 ± 4.19 | 8.15 ± 4.18 | 7.04 ± 4.12 | 7.85 ± 4.19 | 0.000 |
Smoking | .08 ± .20 | .09 ± .21 | .08 ± .17 | .07 ± .16 | 0.004 |
Depressive symptoms | 1.29 ± 1.87 | 1.12 ± 1.74 | 1.67 ± 2.03 | 1.85 ± 2.21 | 0.000 |
Covariates | | | | | |
Female | 60.91 | 59.76 | 67.19 | 60.54 | 0.000 |
Age | 66.77 ± 9.70 | 67.81 ± 9.65 | 64.23 ± 9.19 | 63.51 ± 9.26 | 0.000 |
Marital Status | | | | | |
Married | 67.77 | 71.48 | 47.15 | 69.21 | 0.000 |
Divrced/separed/widowed | 28.03 | 25.47 | 42.61 | 26.66 |
Never married | 4.19 | 3.04 | 10.24 | 4.14 |
Baseline cognition | 23.93 ± 3.96 | 24.79 ± 3.44 | 21.41 ± 4.22 | 21.63 ± 4.37 | 0.000 |
* p-values are from between-racial-group F-tests for numeric variables and chi-square tests for categorical variables. |
[Table 1 about here]
The upper panel of Table 2 summarized standardized regression coefficients from the path model linking cumulative ACE and later life cognition for the whole sample. Cumulative ACE was negatively associated with years of education (\(\:beta=\:-1.00,\:p<.001\)) and physical activity (\(\:beta=-.19,\:p<.001\)), but positively associated with smoking (\(\:beta=.27,\:p<.001\)) and depressive symptoms (\(\:beta=.20,\:p<.001\)). Education as an upperstream mediator was positively associated with physical activity (\(\:beta\:=.28,\:p<.001\)), but negatively associated with smoking (\(\:beta\:=-.04,\:p<.001\)) and depressive symptoms (\(\:beta\:=-.08,\:p<.001\)). Cognition was positively associated with education (\(\:beta\:=.16,\:p<.001\)) and physical activity (\(\:beta\:=.05,\:p<.001\)) while negatively associated with smoking (beta =-.07, p < .001) and depressive symptoms (\(\:beta\:=-.16,\:p<.001\)). After adjusting for all mediating pathways, cumulative ACE was still significantly negatively associated with cognition (\(\:beta\:=-.13,\:p<.001\)). Together, the mediators explained over 65% of cumulative ACE’s association with later life cognition (see Table 3 for effects decomposition). While education served as the most significant mediator, physical activity, smoking, and depressive symptoms also played minor roles as mediators in the relationship between ACE and cognitive functioning.
The middle and lower panels of Table 2 presented standardized regression coefficients from the the path models for Whites and Blacks, respectively. The model’s results for the White subsample closely resembled those of the overall sample, with similar patterns of significance and direction in the relationships between ACE, the mediators, and cognition scores. This consistency suggests that the role of ACE in forecasting cognition through these mediators is broadly applicable within this demographic group.
However, the analysis of the Black subsample yielded nuanced differences in the magnitude and, in many instances, the significance of the relationships between ACE, the mediators, and cognition. Specifically, the cumulative ACE was only negatively associated with education, which in turn predicted physical activity, depressive symptoms, and cognition. In predicting cognition, the only significant mediators other than education was depressive symptoms (\(\:beta=-.16,\:p<.05\)). Notably, for Blacks, the direct association between cumulative ACE and cognition was no longer significant after taking into account the mediators. Figure 2 presented the significant pathways linking cumulative ACE and later life cognition for Whites and Blacks separately after removing nonsignificant paths from the models. Overall, the path model is much simpler for Blacks, that is, cumulatie ACE leads to lower levels of education, and lower levels of education predicts lower levels of cognition both directly and indirectly by elevating the depressive symptoms. Education alone explained more than 60% of the total effects of ACE on cognition.
Table 2
Standardized regression coefficients and associated standard errors from the pathway model linking cumulative ACE and later life cognition, Health and Retirement Study, United States a
| Mediators b | | Outcome c |
| Education | Physical Activity | Smoking | Depressive Symptoms | | Cognition |
All (N = 9,894) | | | | | |
Cumulative ACE d | -1.00*** | -0.19*** | 0.27*** | 0.20*** | | -0.13*** |
[0.03] | [0.05] | [0.04] | [0.03] | | [0.05] |
Adulthood Mediators | | | | | | |
Education | | 0.28*** | -0.04*** | -0.08*** | | 0.16*** |
| [0.02] | [0.01] | [0.01] | | [0.01] |
Physical Activity | | | | | | 0.05*** |
| | | | | [0.01] |
Smoking | | | | | | -0.07*** |
| | | | | [0.02] |
Depressive Symptoms | | | | | | -0.16*** |
| | | | | [0.02] |
White (N = 7,272) | | | | | |
Cumulative ACE d | -0.91*** | -0.22*** | 0.28*** | 0.20*** | | -0.13* |
[0.03] | [0.06] | [0.04] | [0.03] | | [0.05] |
Adulthood Mediators | | | | | | |
Education | | 0.39*** | -0.07*** | -0.08*** | | 0.14*** |
| [0.02] | [0.01] | [0.01] | | [0.02] |
Physical Activity | | | | | | 0.05*** |
| | | | | [0.01] |
Smoking | | | | | | -0.07** |
| | | | | [0.02] |
Depressive Symptoms | | | | | | -0.15*** |
| | | | | [0.03] |
Black (N = 1,396) | | | | | |
Cumulative ACE d | -0.86*** | -0.14 | 0.29 | 0.17 | | -0.11 |
[0.09] | [0.14] | [0.15] | [0.12] | | [0.19] |
Adulthood Mediators | | | | | | |
Education | | 0.32*** | -0.03 | -0.15*** | | 0.24*** |
| [0.04] | [0.02] | [0.02] | | [0.05] |
Physical Activity | | | | | | 0.05 |
| | | | | [0.03] |
Smoking | | | | | | -0.02 |
| | | | | [0.06] |
Depressive Symptoms | | | | | | -0.16* |
| | | | | [0.07] |
Note. a Full Information Maximum Likelihood estimator was used to deal with missing data; b Models adjusted for gender, race, age, and marital status; c Model adjusted for gender, race, age, marital status, and baseline cognitive functioning; Standard errors in parentheses; * p < 0.05, ** p < .01, *** p < .001 |
Table 3
Decomposition of total and indirect effects of cumulative ACE on later life cognition, Health and Retirement Study, United States
Effects Components | All a (N = 9,894) | White b (N = 7,272) | Black b (N = 1,396) |
Total effect of cumulative ACE on cognition | -0.38*** | -0.35*** | -0.39* |
| [0.05] | [0.05] | [0.19] |
Total indirect effect of cumulative ACE on cognition | -0.25*** | -0.23*** | -0.28*** |
| [0.02] | [0.02] | [0.05] |
Indirect Effects Components | | | |
Cumulative ACE → Education → Cognition c | -0.19*** | -0.17*** | -0.24*** |
[0.02] | [0.02] | [0.05] |
Cumulative ACE → Physical Activity → Cognition | -0.01** | -0.01** | -0.01 |
[0.00] | [0.00] | [0.01] |
Cumulative ACE → Smoking → Cognition | -0.02*** | -0.02** | -0.01 |
[0.01] | [0.01] | [0.02] |
Cumulative ACE → Depressive Symptoms → Cognition | -0.03*** | -0.03*** | -0.03 |
[0.01] | [0.01] | [0.02] |
Note. a Estimation adjusted for gender, race, age, marital status, and baseline cognitive functioning; b Estimations adjusted for gender, age, marital status, and baseline cognitive functioning; c The indirect effect of cumulative ACE on cognition via education includes the following components: Cumulative ACE → Education → Cognition; Cumulative ACE → Education → Physical Activity → Cognition; Cumulative CA → Education → Smoking → Cognition; Cumulative CA → Education→ Depressive Symptoms → Cognition; Standard errors in parentheses; * p < 0.05, ** p < .01, *** p < .001 |
[Table 2 about here]
[Figure 2 about here]
The measurement model for the cumulative ACE suggested variations in the impact of specific types of ACE across racial groups. Figure 3 illustrates these differences by presenting the statistically significant standardized regression coefficients from the measurement models for the White and Black samples. Each regression coefficient quantifies the relative contribution (i.e., weight) of a specific ACE to the cumulative ACE score, indicating the varied influence of individual ACEs. For both racial groups, learning problems, grade retention, smoking or other substance use, and mother low education stood out as significant contributors to the cumulative ACE score. However, the magnitude of contributions from these ACEs differed between groups. For instance, learning problems carried a greater weight in the cumulative ACE score for the Black sample compared to the White sample. In contrast, grade retention held a greater weight in the cumulative ACE score among Whites. Physical disabilities emerged as a significant ACE factor exclusively for the Black sample, while parental substance abuse, parental smoking, father absence, and father joblessness emerged as additional significant factors only for Whites. Notably, allergies contribute negatively to the cumulative ACE score for both groups, indicating that this condition was inversely associated with other ACE variables and predicted subsequent outcomes in a direction opposite to that of the cumulative ACE. Together, these regression coefficients indicate that various types of ACEs hold differing prognostic values for cognitive functioning across different racial groups.
[Figure 3 about here]