TABLE 4-1 Multiple Logistic Regression Analysis of Major Depression in Past 12 Months Among Males

 

Odds ratio (95% C.I.a)

p-value

Interaction of father sad for 2+ weeks during childhood with father had drug/alcohol problems during childhood (overall p = 0.01)

 

 

Father sad, drug/alcohol problems vs. neither

0.37 (0.07, 1.95)

0.24

Father not sad, drug/alcohol problems vs. neither

1.73 (0.95, 3.15)

0.07

Father sad, no drug/alcohol problems vs. neither

3.65 (1.20, 11.14)

0.02

Mother sad for 2+ weeks during childhood

1.58 (0.79, 3.15)

0.20

Closeness of father: summary scoreb

1.39 (1.08, 1.80)

0.01

Race/ethnicity (overall p = 0.07)

 

 

African-American vs. white

0.75 (0.32, 1.75)

0.51

Hispanic vs. white

0.85 (0.37, 1.94)

0.70

Other vs. white

3.64 (1.23, 10.71)

0.02

Any PTSD before age

18 3.23 (1.36, 7.69)

0.008

NOTES: PTSD = Posttraumatic stress disorder; N = 759, 65 with major depression, 694 without; model p-value = 0.003; generalized R2 = 0.10; C-statistic = 0.65.

aConfidence interval.

bFor a one standard deviation difference; higher values denote less closeness.

SOURCE: Tabulations based on the National Comorbidity Survey-Replication (see http://www.icpsr.umich.edu/CPES/).

ratio 3.17) and statistically significant (p < 0.0001) predictor of major depression in the past 12 months (Table 4-2), as was experience of trauma before the age of 18 (odds ratio of 3.41, p < 0.0001). Additional statistically significant factors in the model for females predictive of major depression in the past year were lower levels of closeness to the mother during childhood (p = 0.005), greater parental neglect during childhood (p = 0.047), a paradoxical effect of fewer social problems among the respondents’ fathers (p = 0.047), increased depression with older age at the time of the interview (p = 0.03) and having never been married compared to having been married or co-habiting (p = 0.02).

These findings underscore the importance of examining the effects of paternal risk factors as well as maternal risk factors for psychopathology in children and that comprehensive models are required to properly quantify the effects of these risk factors, many of which are intercorrelated and may show effect moderation. Further studies of this type are needed, especially those in which more detailed information on personal trauma and neglect



The National Academies | 500 Fifth St. N.W. | Washington, D.C. 20001
Copyright © National Academy of Sciences. All rights reserved.
Terms of Use and Privacy Statement