Transcript Click here to view slides from Dr. Turner`s talk.

Health Disparities and the
Stress Hypothesis
R. Jay Turner, Ph.D
Vanderbilt University
Presented at Montana State
University
February 10, 2012
Stress Process Model
Retrospective self-report, computer
assisted personal interviews
DSM-IV CIDI diagnoses: major depression,
dysthymia, GAD, social phobia, panic disorder, alcohol abuse and
dependence, drug abuse and dependence, PTSD, and antisocial
personality disorder.
PLUS
attention deficit and hyperactive disorder and
Childhood conduct disorder from the Diagnostic
Interview schedule.
Cumulative Adversity and Drug Dependence at Age 18
.06
Conditional Probability of Dependence
.05
.04
.03
.02
.01
.00
0
1
2
3
4
5
6
7
8
Cumulative Adversities
Non-Hispanic White
Cuban
Non-Cuban Hispanic
African American
9
10
Cumulative Adversity and Alcohol Dependence at age 18
.0700
Conditional Probability of Dependence
.0600
.0500
.0400
.0300
.0200
.0100
.0000
1
2
3
4
5
6
7
8
9
Cumulative Adversity Count
Series1
Series2
Series3
Series4
10
11
12
13
FIGURE 3 - DSM-IV Disorder Conditional Onset Risk by Cumulative Exposure
to Adversities among Community-dwelling Young Adults
.06
Estimated Probability of onset
.05
.04
.03
.02
.01
.00
0
1
2
3
4
5
6
7
8
9
10
11
12
Cumulative Adversities
Drug Dependence: Whites
Drug Dependence: African Americans
Psychiatric Disorder: Whites
Psychiatric Disorder: African Americans
13
14
15
TABLE 1.
Mean Standardized Scores on Stress Measures by Gender, Race/Ethnicity
and Socioeconomic Level
Gender
Male
Female
Race/Ethnicity
White
African American
Socioeconomic Level
Upper
Middle
Lower
* p<.05; ** p<.01; *** p<.001
Note: Numbers in parentheses are standard errors.
Recent Life Events
Total Stress Exposure
–.075*
(.967)
.092
(1.033)
.063*
(1.013)
–.077
(.980)
–.092**
(.945)
.112
(1.053)
Difference = .204
–.236***
(.920)
.286
(1.019)
Difference = .522
–.115+
(.938)
.053
(.986)
.063
(1.065)
U-L difference = .178
–.277***
(.894)
.051
(.950)
.227
(1.083)
U-L difference = .504
Table 3. Depressive Symptoms Regressed on Status Characteristics and Five Dimensions of Stress
Exposure (N = 899; 493 non-Hispanic whites and 406 African-Americans)
I
II
III
IV
V
VI
VII
Status Characteristics
Gender (Female=1)
1.570 *** 1.372 *** 1.613 *** 1.241 *** 1.533 *** 1.731 *** 1.533 ***
(.285)
(.277)
(.277)
(.261)
(.282)
(.272)
(.265)
Race/Ethnicity (African-American=1)
1.020 **
(.321)
SES
-.579 *** -.546 *** -.407 *** -.394 *
(.163)
(.158)
(.158)
(.149)
.810 **
(.312)
.559 **
(.311)
.337
(.296)
.751 *
(.311)
.626 *
(.303)
.171
(.294)
-.494 **
(.158)
-.465 **
(.153)
-.312 *
(.148)
Life Stressors
Recent Life Events (past 12 months)
1.075 *** .625 *** .502 *** .927 *** .744 *** .210
(.138)
(.162)
(.140)
(.147)
(.139)
(.157)
Lifetime Major Events
.871 ***
(.172)
Chronic Stressors
.583 **
(.195)
1.594 ***
(.146)
Lifetime Major Discrimination
1.339 ***
(.151)
1.158 ***
(.402)
Daily Discrimination
Constant
R2
Adjusted R2
-.852
(.461)
1.231 *** .783 ***
(.159)
(.167)
2.456
.072 ***
.069
2.637
.131 ***
.127
2.653
.155 ***
.150
2.934
.223 ***
.229
2.624
.139 ***
.134
Table entities represent unstandardized regression coefficient (standard errors).
*, **, ***: p < .05, .01, .001, respectively.
Numbers of subjects prior to weighting are 466 non-Hispanic whites and 433 African-Americans.
2.582
.185 ***
.181
2.872
.262 ***
.255
TABLE 3. Depressive Symptomatology (CES-D) Regressed on Social Status and Stress Process Components: Toronto, 1990–1991
1
Intercept
Female
Age
Age2
Previously-married
Never-married
SES
24.744***
(4.320)
2.643***
(.623)
–.739**
(.237)
.008*
(.003)
3.342**
(1.064)
1.198
(.808)
–1.060**
(.299)
Stress
2
23.766***
(4.199)
2.149***
(.568)
–.701**
(.227)
.008**
(.003)
1.339
(1.004)
1.176
(.727)
–.682*
(.294)
3.968***
(.289)
3
44.327***
(4.221)
2.226***
(.546)
–.597*
(.228)
.006†
(.003)
3.803***
(.978)
1.872**
(.688)
–.175
(.267)
4
59.348***
(4.226)
2.234***
(.595)
–.598**
(.210)
.006*
(.003)
3.491***
(.866)
.647
(.693)
–.712**
(.266)
5
14.286**
(4.148)
2.096**
(.612)
–.592*
(.228)
.006*
(.003)
3.909***
(.994)
1.891*
(.799)
–.713*
(.294)
6
30.444***
(4.488)
2.196***
(.580)
–.728**
(.233)
.008*
(.003)
2.994**
(1.035)
1.038
(.778)
–1.417***
(.280)
23.123***
(4.136)
3.105***
(.599)
–.619**
(.230)
.006*
(.003)
1.324
(.861)
–.083
(.723)
–.877**
(.283)
–.767***
(.060)
Mastery
–1.326***
(.081)
Self Esteem
Emotional Reliance
.543***
(.071)
–.618***
(.105)
Assertion of Autonomy
Social Support
R2
7
.104
.276
.258
.266
.143
.141
–2.829***
(.247)
.182
8
47.003***
(4.584)
1.556**
(.482)
–.447*
(.189)
.004†
(.003)
1.733*
(.757)
1.114†
(.629)
–.091
(.251)
2.799***
(.287)
–.359***
(.050)
–.683***
(.079)
.302***
(.051)
–.250**
(.076)
–.931**
(.264)
.443
† p <.10 *p < .05 **p < .01 ***p < .001
Note: Unstandardized OLS regression coefficients; standard errors are in parentheses; data weighted to adjust for respondent selection bias; standard errors adjusted for
clustered sampling design using Stata survey regression procedure. N = 1338.
• 1. To evaluate the stress hypothesis and assess the
utility of an elaborated “stress process” model for
explaining race and socioeconomic position (SES)
differences in health and changes in health over time,
where:
• a) health status is estimated in multiple ways that
address the problem of misclassification.
• b) variations in stress exposure are estimated more
comprehensively than in prior research.
• c) the direct, mediating and moderation effects of an
extended array of social and personal coping resources
and of contextual and cultural factors are examined in
the context of a) and b) above.
•
• 2. To confirm and extend epidemiological estimates of
race and SEP differences in the prevalence of health
problems, variously and collectively defined, while
addressing the misclassification problem.
• 3. To evaluate the interrelationships of differing
dimensions of health, variously estimated, and their risk
significance for one another, in cross section and over
time.
• 4. To describe the distribution of each risk factor
identified across race, SEP, age and gender while
examining antecedent factors that put individuals and
groups at differential risk for acquiring such risk factors.
• 5. To obtain and preserve samples to support
subsequent studies (not here proposed) on geneenvironment interactions relating to factors hypothesized
to be of risk or protective significance.
Table 1. Co-occurrence of Lifetime DSM-IV Psychiatric and Substance
Dependence Disorders (No Temporal Ordering)
Substance Dependence c
OR e
N
%
95% CI
Comorbid Depression (Across and/or Within Category)
Pure Depression
185
130
47.2
18.5
6.8***
1.8*
4.7 - 10.0
1.1 - 3.0
Comorbid PTSD (Across and/or Within Category)
Pure PTSD
140
65
45.0
20.8
7.0***
2.8**
4.6 - 10.7
1.4 - 5.5
Comorbid Anxiety Disorder a
75
41.5
5.0***
2.9 - 8.5
Pure Anxiety Disorder
22
15.0
1.3
0.4 - 3.9
Comorbid Personality Disorder b
287
43.4
6.3***
4.5 - 8.7
Pure Personality Disorder
179
16.4
1.5†
0.9 - 2.4
Table 2. Temporally Ordered Associations between Lifetime DSM-IV Psychiatric and Substance
Dependence Disorders
Substance Dependence b
Comorbid Depression
Pure Depression
OR d
N
%
162
39.7
5.0***
3.3 - 7.4
12.6
1.1
0.6 - 2.0
6.7***
121
Comorbid PTSD
76
44.0
Pure PTSD
43
20.0
Comorbid Anxiety Disorder a
70
37.5
4.2***
Pure Anxiety Disorder
20
10.5
0.6
3.0**
95% CI
3.9 - 11.3
1.3 - 6.9
2.4 - 7.4
0.1 - 2.9
Table 4. Effects of Stress Exposure on Associations between Lifetime DSM-IV Psychiatric and
Substance Dependence Disorders
Substance Dependence b
Not Ordered
OR d
Comorbid Depression
2.4***
Ordered
95% CI
d
OR
1.5 - 3.8
1.5†
0.9 - 2.6
0.7
0.3 - 1.4
Pure Depression
1.2
0.6 - 2.2
Comorbid PTSD
1.7*
1.0 - 3.1
1.5
0.7 - 3.1
Pure PTSD
1.1
0.4 - 2.7
1.7
0.6 - 4.5
Comorbid Anxiety Disorder a
1.7
0.8 - 3.4
1.3
Pure Anxiety Disorder
1.3
0.4 - 4.1
0.8
0.8 - 2.1
0.1 - 3.6