Psychosomatic Medicine Faster Service from Outside North America
HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
This Article
Right arrow Abstract Freely available
Right arrow Full Text (PDF)
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrowRequest Permissions
Citing Articles
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Lauterbach, D.
Right arrow Articles by Rakow, M.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Lauterbach, D.
Right arrow Articles by Rakow, M.
Related Collections
Right arrow Health Psychology
Right arrow PTSD
Right arrow Other Psychiatric Disorders
Psychosomatic Medicine 67:939-947 (2005)
© 2005 American Psychosomatic Society

ORIGINAL ARTICLES

The Relationship Between Posttraumatic Stress Disorder and Self-Reported Health Problems

Dean Lauterbach, PhD, Rajvee Vora, MD and Madeline Rakow, BSW

From the Department of Psychology, Eastern Michigan University, Ypsilanti, Michigan.

Address correspondence and reprint requests to Dean Lauterbach, PhD, 507 Mark Jefferson Hall, Department of Psychology, Eastern Michigan University, Ypsilanti, MI 48197. E-mail dlauterba{at}emich.edu


   ABSTRACT
TOP
 ABSTRACT
INTRODUCTION
 METHODS
 RESULTS
 DISCUSSION
 NOTES
 REFERENCES
 
Objective: There is increasing awareness that posttraumatic stress disorder (PTSD) influences physical health. The current study used the National Comorbidity Survey (NCS) data to determine if the physical health risk imparted by PTSD was an artifact of a broad range of empirically identified confounding variables.

Methods: Participants (n = 5877) were administered a modified version of the DSM-III-R PTSD module from the Diagnostic Interview Schedule (DIS). The prevalence of 14 classes of physical disorders was assessed along with six groups of variables predictive of negative health outcomes (demographics, perception of health, stress, health-related behaviors, insurance coverage, trauma/psychiatric history, and neuroticism).

Results: The PTSD group had a higher frequency of most disorders and scored higher on variables predictive of negative health outcomes. To determine if the health risk imparted by PTSD was an artifact of these confounding variables, a mixed direct–sequential binary logistic regression was computed. After controlling for sex, health perceptions, stress, health-related behaviors, insurance coverage, total trauma exposure, total number of psychiatric diagnoses, and neuroticism, persons with PTSD were more likely to report the presence of a disorder.

Conclusions: Previous scholars have noted that PTSD could affect health outcomes. This study found that after controlling for a wide range of variables predictive of poor health, PTSD was predictive of poor health.

Key Words: posttraumatic stress disorder • health risk factors • National Comorbidity Survey

Abbreviations: PTSD = posttraumatic stress disorder; NCS = National Comorbidity Survey; NVVRS = National Vietnam Veterans Readjustment Survey; DIS = Diagnostic Interview Schedule.


   INTRODUCTION
TOP
 ABSTRACT
 INTRODUCTION
METHODS
 RESULTS
 DISCUSSION
 NOTES
 REFERENCES
 
Post-traumatic stress disorder (1) (PTSD) is a frequently occurring psychiatric illness in the United States. PTSD has been implicated in poor medical and psychiatric outcomes in various trauma-exposed populations, including combat veterans (2–4), help-seeking combat veterans (5,6), and victims of natural disasters (7). Conversely, high users of primary care services are more likely to meet the diagnostic criteria for PTSD (8). Although this finding is not unique to trauma survivors or to persons with PTSD (9), it does suggest that PTSD is associated with a heavy burden of medical illness. The diagnosis of PTSD requires both exposure to a trauma and development of a characteristic set of symptoms. It is possible that trauma exposure alone explains the PTSD–health outcome relationship.

Surprisingly few studies have examined the relationship between physical health and trauma exposure. Those that have done so suggest that trauma exposure has a strong negative effect on physical health. The National Vietnam Veterans Readjustment Study (NVVRS) (4) found that men and women veterans with high war-zone exposure reported poorer health and more health problems relative to era veterans who were not exposed to combat and civilian controls. Similarly, women with a history of sexual victimization report poorer health, more negative health behaviors, more somatic complaints, and poorer perceived health relative to women without such a history (10–12).

Several studies have found PTSD to be an important mediator of the relationship between trauma exposure and health status (13,14). Wolfe et al. (14) examined the unique association between war-zone exposure and PTSD on perceived physical health outcomes in women Vietnam veteran nurses. Robust effects were found for both variables on health outcome when each predictor was considered separately. However, when both exposure and PTSD were included in the models, only PTSD strongly predicted poorer health outcomes. A secondary analysis of the Wolfe et al. data (13) examined the mediating influence of PTSD on self-reported current health and the number of current health problems among women Vietnam veteran nurses. They found that 56% of the total effect of war-zone exposure on current health was indirectly mediated through PTSD. Similarly, for the number of current health problems, only the indirect effect of exposure (76% of the total) was significant.

One of the largest studies to investigate the effect of combat-related PTSD on physical problems merged two datasets (NVVRS and Hawaiian Vietnam Veterans Project) and conducted analyses after adjusting for age, other axis I disorders, and substance abuse (15). Veterans with PTSD reported a greater number of health problems and greater medical utilization than veterans without PTSD.

It is possible that the PTSD–health status link may be mediated by a specific element of the trauma response complex. Specifically, the medical problems seen in persons with this disorder may be a result of sympathetic hyperreactivity, endocrinologic abnormalities, or altered immunologic mechanisms (13). Boscarino and Chang (16) linked PTSD status with particular cardiovascular disorders using electrocardiogram abnormalities.

In a study of the long-term effects of trauma exposure and PTSD on immune system function, Boscarino and Chang (17) compared leukocyte and lymphocyte status of the following three groups of male Vietnam "theater" veterans: 1) current partial PTSD, 2) current depressive disorders, and 3) neither PTSD nor depressive disorders 20 years after military service. Groups were equated on variables linked to health status, including intelligence, race, age, income, education, type of enlistment, Vietnam volunteer status, region of birth, cigarette smoking, illicit drug use, body mass index, and alcohol consumption. PTSD-positive veterans are more likely to have adjusted leukocyte and T-cell counts above the normal range and higher mean adjusted leukocyte, lymphocyte, T-cell, and CD4 cell counts. Results also indicate reactive delayed cutaneous hypersensitivity, which suggests the presence of highly sensitized T-cell lymphocytes. The PTSD-positive men also had abnormally high CD4 and CD8 T-cell lymphocyte counts. These findings indicate that the PTSD-positive veterans have a highly active and reactive immune system. When taken together, these data suggest that the effects of trauma exposure and PTSD on health status are not limited to self-report data, cardiovascular disorders, and do not reflect transient symptomatology.

The PTSD–health status relationship may be mediated by likelihood to engage in health-enhancing behaviors and/or elevations in ambient stress. For example, cumulative abuse experiences have been associated with poorer health behaviors (e.g., smoking, poor diet, binge drinking) and higher levels of reported physical symptoms (20). There are clear data that engagement in various behaviors and level of stress are predictive of physical health. For example, educating patients with chronic diseases about exercise, diet, communication with physicians, sleep management, and cognitive symptom management produced positive results on physical symptoms (18). Other research supports the benefits of a healthy diet on physical health. A prospective study examining the relationship between diet and health outcomes among older adults (mean, 73 years old) (19) found that participants with healthy dietary habits lived the longest and had the most years of healthy life.

Despite the growing evidence that PTSD is strongly related to poor health, only a few investigators have explored this relationship in civilian populations like victims of sexual trauma (21,22), crime victims (7), and political refugees (23). No studies could be located that examined these issues in a nonclinical population. Most studies have been conducted on war veterans using VA medical and mental health services (5,6,8,14–16).

Although a number of studies have examined the incidence of medical illness among those with PTSD, there are multiple routes to chronic illness, including elevations in stress, inadequate engagement in health-enhancing behaviors, and inadequate insurance coverage. The current study was conducted to examine the relationships between lifetime history of PTSD, recent history of medical illness, and the following variables: 1) demographic variables, 2) perceptions of health, 3) ambient stress, 4) health-related behaviors, 5) presence/comprehensiveness of healthcare coverage, 6) lifetime history of trauma exposure, 7) number of axis I disorders (excluding PTSD), and 8) elevations in neuroticism. The current study examined these issues in a national sample representative of the United States.

It was hypothesized that individuals with PTSD will have an elevated incidence of medical illnesses (particularly stress-related disorders), perceive their health to be worse than the general population, report greater levels of ambient stress, engage in fewer health-related behaviors, and have less comprehensive healthcare coverage than persons without PTSD.


   METHODS
TOP
 ABSTRACT
 INTRODUCTION
 METHODS
RESULTS
 DISCUSSION
 NOTES
 REFERENCES
 
Overview of the National Comorbidity Survey
The National Comorbidity Survey (NCS) (24) was a stratified, multistage, area probability sampling study of the noninstitutionalized civilian population of the 48 contiguous states. The first phase was a diagnostic interview, which was administered to all participants (n = 8098). The second phase was a risk factor interview and was administered to a subsample (n = 5877) of participants. The participants for the second phase consisted of the following: 1) all part 1 respondents between 15 and 24 years of age, 2) all older part 1 respondents who were positive on initial questions assessing one or more diagnostic sections, and 3) a one-in-six subsample of the remaining part 1 respondents. There were equal numbers of persons in each of the following age categories: 1) 25 to 34, 2) 35 to 44, and 3) 45 to 54. There was an oversampling of persons between the ages of 15 and 24.

Weighting of Data
Based on previous evidence that survey nonrespondents have higher rates of psychiatric disorder than respondents, a supplemental nonresponse survey was carried out. In this supplemental survey, a random sample of initial nonrespondents was offered a financial incentive to complete a short form of the diagnostic interview. A higher rate of psychiatric illness was found in this sample. Consequently, a nonresponse adjustment weight was constructed for the main survey data to compensate for this systematic nonresponse. A second weight was used to adjust for variation in probabilities of selection both within and between households. A third weight was used to adjust the data to approximate the national demographic characteristics.

Analyses
SPSS Version 10.0.1 was used to compute mean group differences on perceptions of health, ambient stress level, health-related behaviors, lifetime history of trauma exposure, number of axis I disorders (excluding PTSD), and neuroticism. SPSS was also used to compute nonlinear associations ({chi}2) between PTSD status and demographic variables, health status, and healthcare coverage. As a result of the nature of the sampling strategy, the results were weighted to adjust for variation in probabilities of selection both within and between households, nonresponse bias, and poststratification. The Bonferroni test was used to control for familywise alpha inflation. Data analysis is best conceptualized as a three-step process.

The first set of analyses compared those with and those without PTSD on: 1) demographic variables, 2) perceptions of health, 3) ambient stress, 4) engagement in health-related behaviors, 5) availability of insurance coverage, 6) lifetime history of trauma exposure and number of axis I disorders, and 7) neuroticism. This initial portion of the data analysis resulted in a subset of variables related to PTSD group status. The second set of analyses compared those with and those without a self-reported disease on the same predictors. Thus, the second stage in the analysis identified the variables predictive of disease status (presence/absence). The final step examined whether PTSD status is a significant predictor of disease status after controlling for significant covariates identified in steps 1 and 2.

Measures
Posttraumatic Stress Disorder
Posttraumatic stress disorder was assessed using a modified version of the Revised Diagnostic Interview Schedule (DIS) for PTSD (25). At the outset of the PTSD section, participants were queried about the presence of the following 10 event categories: 1) direct combat experience in a war; 2) life-threatening accident; 3) fire, flood, or natural disaster; 4) witnessing someone being badly injured or killed; 5) rape; 6) sexual molestation; 7) serious physical attack or assault; 8) physical abuse as a child; 9) seriously neglected as a child; and 10) threatened with a weapon, held captive, or kidnapped. Earlier versions of the DIS underestimated the prevalence of PTSD in the general population (26) and in Vietnam veteran populations (27,28). Consequently, some have suggested caution be used when using the DIS to diagnose PTSD until modifications are made (29). To enhance respondent comfort and accuracy of reporting, respondents received a numbered listing of the events. In all interviews, event type was referenced by a number. As a result, participants did not have to verbalize the type of trauma they had experienced. To obtain an index of lifetime history of trauma exposure, each event was dummy-coded and summed to yield a total score (mean, 1.25; standard deviation [SD], 1.56; range, 0–10).

Medical History
To assess history of medical illness, participants were asked to read a list of disorders and indicate whether they have experienced any of them within the past 12 months. The presence of each of the following disorders (and classes of disorders) was assessed: 1) severe arthritis, rheumatism, or other bone or joint diseases; 2) severe asthma, bronchitis, emphysema, tuberculosis, or other lung problems; 3) AIDS; 4) blindness, deafness, or severe visual or hearing impairment; 5) high blood pressure or hypertension; 6) diabetes or high blood sugar; 7) heart attack or other serious heart trouble; 8) severe hernia or rupture; 9) severe kidney or liver disease; 10) lupus, thyroid disease, or other autoimmune disorders; 11) multiple sclerosis, epilepsy, or other neurologic disorders; 12) chronic stomach or gallbladder trouble; 13) stroke; and 14) ulcer.

Perceptions of Health
Two items assessed participants’ perception of their health. Respondents were asked to rate their overall physical health and compare their health with other people their age. Ratings were made on five- and three-point Likert scales, respectively, with higher scores reflecting the perception that personal health is poor in absolute terms and relative to others. Scores ranged from one to five for participants ratings of their own health and from one to three for their ratings of their health relative to others. Participants rated their health between good and very good (mean, 2.26; SD, 0.95) and saw their physical health about the same as others (mean, 1.92; SD, 1.35). There was a significant relationship between these two indices of health perception (r = 0.41, p < .001).

Ambient Stress Level
Seven items assessed ambient stress. These items were obtained from the "home and work" section of the University of Michigan’s revised version of the Composite International Diagnostic Interview (CIDI) (30–32). For each item, respondents rated how often the event occurred on a four-point Likert scale (1 = often, 2 = sometimes, 3 = rarely, and 4 = never). The following indices of ambient stress were assessed: 1) things at home create tension on the job, 2) things at work create tension at home, 3) demands of family interfere with your job, 4) demands of your job interfere with your family life, 5) when at work, how often do you think about things at home, 6) when at home, do you think about things going on at work, and 7) perception of inadequate time to do a good job both at home and at work. The mean values for ambient stress ranged from 1.87 (SD, 1.32) for item seven to 2.32 (SD, 1.35) for item one. Thus, for all ambient stress items, participants on average indicated that they occurred sometimes. The value for coefficient alpha was 0.96, indicating substantial interitem association.

Health-Related Behaviors
Engagement in the following three health-related behaviors was assessed: how often they eat balanced meals, how many hours they sleep in a typical 24-hour period, and how often they get physical exercise. Values indicate that participants ate three balanced meals per day between "sometimes" and "most of the time" (mean, 2.85; SD, 1.19), got regular exercise between "most of the time" and "always" (mean, 1.52; SD, 0.78), and slept an average of 7.19 hours per night (SD, 1.33). These three indices of health behaviors were not substantially interrelated ({alpha} = –0.15).

Healthcare Coverage
Six items assessed healthcare coverage. Initially, participants were asked to indicate whether they had coverage by a health insurance plan other than Medicaid (yes/no) and whether they had coverage by Medicaid, welfare, or any other public assistance that pays for part of their medical care (yes/no). For those who had insurance coverage, they were asked to indicate whether their insurance covered all, part, or none of the following: 1) hospital stay for a general medical illness; 2) hospital stay for a mental health, drug, or alcohol problem; 3) care in a doctor’s office for a general medical illness; and 4) mental health, drug, or alcohol care outside of a hospital. The majority of the sample (81.3%) had health insurance, but a sizable subgroup (27.5%) indicated that Medicaid or some other form of public assistance paid for part of their medical care. Among those with insurance, complete coverage for a medical hospitalization, psychiatric hospitalization, office visits, or outpatient psychiatric care was reported by only 44.4%, 30.7%, 26.1%, and 17.4%, respectively.

Neuroticism
Level of neuroticism was assessed by a 10-item unipolar scale (33) designed to assess reliable markers of emotional stability. Participants read a list of words that people use to describe themselves and indicated how well each word describes the way they generally are. Items were rated on a four-point Likert scale with the following anchors: 1 = very, 2 = somewhat, 3 = a little, and 4 = not at all. For ease of interpretation, this scale was reverse-scored so that higher scores correspond with greater levels of neuroticism. The following 10 trait descriptors were included on the scale: 1) tense, 2) nervous, 3) temperamental, 4) irritable, 5) envious, 6) unstable, 7) discontented, 8) insecure, 9) emotional, and 10) high-strung. The mean item score across all items was 1.87, indicating that, on average, item endorsement was between "a little" and "not at all." These indices of neuroticism were substantially interrelated ({alpha} = 0.88). Scores for these 10 items were summed to yield a total neuroticism score (mean, 18.61; SD, 5.59).

Lifetime History of Psychiatric Disorders
Lifetime history of a broad range of Diagnostic and Statistical Manual of Mental Disorders, 3rd Revision, Revised axis I psychiatric disorders was assessed using a modified version of the CIDI. The CIDI assessed the following disorders: 1) alcohol abuse, 2) alcohol dependence, 3) drug abuse, 4) drug dependence, 5) major depressive disorder, 6) dysthymic disorder, 7) bipolar disorder, 8) mania, 9) generalized anxiety disorder, 10) social phobia, 11) simple phobia, 12) panic disorder, and 13) agoraphobia. Field trials of the CIDI conducted by the World Health Organization indicate that there is good interrater agreement (34,35), diagnoses are stable over time (36,37) and valid (38–40). To obtain an index of the lifetime history of psychiatric illness, each disorder was dummy-coded and the resulting values summed to yield a total score. Values ranged from zero to 10 (mean, 1.14; SD, 1.63).

Participants
The sample was composed of participants in the part II subsample (n = 5877). The unweighted groupings identified 591 persons (weighted n = 429) with PTSD. Table 1 lists descriptive data for the entire sample and those with a lifetime history of PTSD.


View this table:
[in this window]
[in a new window]
  		TABLE 1. Demographic Characteristics of the Sample

 

Initial analyses compared those with and without PTSD on five demographic variables that are thought to be related to health status: highest educational level achieved (measured from 1–17 years), income level (self-reported income range), age, sex, and marital status. The groups differed significantly on educational level and income level but not on age. The PTSD group was significantly lower on educational level and income level. Persons in income range 15 (i.e., PTSD+ group) had an annual income before taxes between $15,000 and $17,499 and persons in income range 16 (PTSD–) had an income range between $17,500 and $19,999. The groups also differed on sex but not on marital status. Women were more likely than men to have a history of PTSD.


   RESULTS
TOP
 ABSTRACT
 INTRODUCTION
 METHODS
 RESULTS
DISCUSSION
 NOTES
 REFERENCES
 
Relationship Between Posttraumatic Stress Disorder Group Status and the Presence of Chronic Physical Health Diagnoses
First, those with and those without PTSD were compared on the presence of 14 physical disorders assessed by the NCS. Table 2 lists the results of a series of chi-squared tests. Examination of the last row of Table 2 reveals that persons with PTSD were twice as likely to report at least one of these disorders. Examination of the within-group percentages reveals that the PTSD+ group was at increased risk to report virtually all disorders. It is possible that those with and those without PTSD differ on important variables that may be predictive of the presence/absence of various disease processes. The next set of analyses compared the two focal groups on hypothesized predictors of health status.


View this table:
[in this window]
[in a new window]
  		TABLE 2. Relationship Between PTSD Group Status and Prevalence of Physical Disorders

 

Relationship Between Posttraumatic Stress Disorder Group Status and Perceptions of Health, Ambient Stress, Health-Related Behaviors, Insurance Coverage, Number of Traumas, Number of Psychiatric Disorders, and Neuroticism
The first analysis compared those with and those without PTSD on perceptions of health. The groups differed significantly in (t [482.11]1 = –7.03, p < .0005) and relative to others (t [468.66] = –5.82, p < .0005). The PTSD group rated their health more poorly both in absolute terms (PTSD+ mean, 2.60; PTSD– mean, 2.23) and relative to others (PTSD+ mean, 2.37; PTSD– mean, 1.88).

Persons with a history of PTSD were higher than those without a history of PTSD on all indices of ambient stress. Table 3 lists values for each type of stress among those with and without PTSD. These values clearly indicate elevations in stress at home and work among those with PTSD.


View this table:
[in this window]
[in a new window]
  		TABLE 3. Relationship Between PTSD Group Status and Ambient Stressors

 

The two focal groups differed on two indices of health-related behaviors. Persons with PTSD were less likely to eat balanced meals (t [5875] = –4.62, p < .0005) and exercise (t [481.23] = –4.42, p < .0005) (see Table 3).

Group status was also associated with healthcare coverage. Examination of the within-group percentages revealed that the PTSD group was less likely to have non-Medicaid insurance ({chi}2 [1, n = 5810] = 37.31, p < .0005) and was correspondingly more likely to report being on Medicaid or welfare ({chi}2 [1, n = 1083] = 20.22, p < .0005). The PTSD group was less likely to have insurance that covered the costs for care in a doctor’s office for a general medical illness (t [375.75] = 4.01, p < .0005).

Lifetime history of PTSD was also related to total number of traumas experienced (t [459.45] = –23.95, p < .0001), total number of psychiatric diagnoses (t [457.13] = –14.44, p < .0001), and neuroticism (t [475.55] = –12.58, p < .0001). Persons with a lifetime history of PTSD had experienced more traumatic events (PTSD+ = 3.46, PTSD– = 1.08), more psychiatric diagnoses (PTSD+ = 2.66, PTSD– = 1.02), and had higher neuroticism scores (PTSD+ = 22.35, PTSD– = 18.32).

Relationship Between Disease Status and Demographic Risk Factors, Health Perceptions, Ambient Stress, Health-Related Behaviors, Insurance Coverage, Number of Traumas, Number of Psychiatric Disorders, and Neuroticism
Those with and those without a self-reported disease did not differ in education or income level but did differ in age (t [2271.01] = –15.76, p < .0001) (disease+ mean = 37.09; disease– mean = 31.93). The groups also differed on sex ({chi}2 [1, n = 5876] = 4.93, p < .05) and marital status ({chi}2 [1, n = 5877] = 7.84, p < .005). Persons with a history of some physical disorder were more likely to be women (women = 52.60%, men = 47.40%) and more likely to be married (disease+ group, married = 63.01%, disease– group, married = 58.81%). Disease status was significantly related to both indices of perception of health (absolute health t [2035.82] = –16.94, p < .0005; health relative to others t [1913.73] = –12.70, p < .0005). Persons in the disease group rated their health more poorly both in absolute terms (disease+ mean = 2.66; disease– mean = 2.13) and relative to others (disease+ mean = 2.37; disease– mean = 1.77). The two health perception items were each converted to z-scores and summed to yield a composite index. Disease status was related to four indices of ambient stress. Table 4 lists values for each type of stress among those with and without a self-reported disease. These values clearly indicate elevations in stress at home and work. The four ambient stress items that were significant predictors of both PTSD group status and disease group status were each converted to z-scores and summed to yield a composite index (alpha = 0.93).2 Disease status was related to two health-related behaviors. Persons with some disease were less likely to exercise (t [2077.73] = 7.06, p < .0005) and slept less (t [2130.90] = 4.31, p < .0005) (see Table 4).


View this table:
[in this window]
[in a new window]
  		TABLE 4. Relationship Between Disease Group Status and Ambient Stressors

 

Disease status was associated with healthcare coverage, and the findings mirror those found for PTSD group status. Examination of the within-group percentages revealed that the disease+ group was less likely to have non-Medicaid insurance ({chi}2 [1, n = 5809] = 5.04, p < .05) and was more likely to report being on Medicaid or welfare ({chi}2 [1, n = 1083] = 15.36, p < .0005). The disease+ group was less likely to have insurance that covered the costs for care in a doctor’s office for a general medical illness (t [1967.08] = –4.10, p < .0005).

Lifetime disease history was also related to number of traumas experienced (t [1911.57] = –10.96, p < .0001), number of psychiatric diagnoses (t [1994.26] = –9.38, p < .0001), and neuroticism (t [2122.98] = –7.46, p < .0001). Persons with a lifetime history of some disease had experienced more traumatic events (disease+ = 1.71, disease– = 1.11), received more psychiatric diagnoses (disease+ = 1.54, disease– = 1.02), and had higher neuroticism scores (disease+ = 19.64, disease– = 18.29).

Relationship Between Health Status and Posttraumatic Stress Disorder After Controlling for All Predictors
The final analysis examined the relationship between PTSD (presence/absence) and health status (absence/presence of a disease) after controlling for other predictors of health status. Based on the previous analyses, the following variables were found to be related to PTSD group status and health group status: sex, health perceptions, ambient stress, exercise, insurance (not Medicaid), Medicaid/welfare, whether insurance covers regular office visits, number of traumas, number of psychiatric diagnoses, and neuroticism. Two additional baseline demographic variables (education and income) were included in the equation because they were related to PTSD status and may serve as confounders between PTSD caseness and health status. A mixed direct–sequential binary logistic regression was performed with health status as the dependent variable. The order of entry was as follows: block 1—sex, education, income, block; 2—composite index of health perceptions, composite index of ambient stress, exercise, insurance (not Medicaid),3 and whether insurance covers regular office visits; block 3—number of traumas, number of psychiatric diagnoses, and neuroticism; and block 4—PTSD group status (presence/absence). Within each block, variables were entered simultaneously. A test of the full model with all 12 predictors against a constant-only model was significant ({chi}2 [10, n = 5877] = 390.14, p < .0001) (Table 5). Using the Wald criterion, two variables from block 2 were significant predictors of presence of a medical disorder: health perceptions composite (Wald = 190.53; odds ratio [OR], 1.38) and insurance coverage of office visits (Wald = 9.90; OR, 0.81; p < .001 for both effects). Two variables from block 3 were significant predictors: number of traumatic events (Wald = 37.40; OR, 1.17) and number of diagnoses (Wald = 6.26; OR, 1.06; p < .001 for both effects). PTSD group status was a significant predictor of the presence of a medical disorder (Wald = 2.20; OR, 1.24; p < .05).


View this table:
[in this window]
[in a new window]
  		TABLE 5. Relationship Between Disease Group Status and All Predictors

 

Although the Wald statistic is frequently used to assess the significance of each predictor variable, some have expressed concern about its use (41). Therefore, as an additional test of individual variables, a series of nested measurement models were tested in which one predictor was removed and decrement in model fit was assessed (likelihood ratio). The final column in Table 5 lists the results. A significant decrement in model fit was found when the following variables were omitted from the equation: education, income, composite index of health perceptions, ambient stress, exercise, quality of insurance coverage (coverage of office visits), number of axis I diagnoses, lifetime history of trauma exposure, neuroticism, and PTSD.

Overall prediction success was 77.0%. Among those with no disease, prediction success was 97.0%, and among those with a disease prediction, success was 15.2%. This represents a slight improvement in predictive accuracy over simply predicting disease absence (76.1%). To identify the magnitude of the relationship between obtained disease status and disease status predicted by the model, Tau-b was computed. The value for Tau-b was 0.05 (p < .0005). A test statistic of 0.05 indicates that the error rate in classification has been reduced by 5.0% over what would be expected by chance. Thus, although there is a significant relationship between predicted and obtained disease status, the magnitude of that relationship is weak.


   DISCUSSION
TOP
 ABSTRACT
 INTRODUCTION
 METHODS
 RESULTS
 DISCUSSION
NOTES
 REFERENCES
 
This study found a converging relationship between PTSD and negative health outcomes. Previous research has suggested that the presence of a psychiatric disorder (e.g., PTSD) is associated with worse overall physical health. Persons with a lifetime history of PTSD were at greater risk to report a wide range of physical disorders. Some of these disorders may be "stress-related," including asthma, hypertension, and ulcers. However, it should be noted that history of PTSD was also related to presence of several disorders not typically considered stress-related, including visual/hearing impairments, diabetes, kidney/liver disease, lupus/thyroid/autoimmune diseases, epilepsy/neurologic disorders, and stomach/gallbladder disorders.

Previous scholars have noted that the symptoms of PTSD could be proximal or secondary causes of physical morbidity by undermining proper nutrition or health behaviors (42). This study also found that there is a convergence of variables that, when taken together, constitute a powerful set of risk factors. Relative to persons without PTSD, those with PTSD had less education (half year of schooling), lower income ($2000 less per year), higher levels of ambient stress at home and work, less frequent engagement in health-promoting behaviors (eating well and exercising), and less access to health insurance. The PTSD group was more likely to be covered by Medicaid or welfare. Furthermore, among those in the PTSD group who had insurance other than Medicaid or welfare, their insurance did not cover doctors’ visits. This study found that persons with PTSD rated their health status less favorably than those without PTSD. Most importantly, this study found that the health risk imparted by the presence of PTSD was not an artifact of these confounding variables. This combination of variables represents a significant risk for both acute and chronic health outcomes. It should be noted, however, that although the overall classification rate of the logistic regression was adequate (77.0% correct), it is extremely high among those with no disease (97% correct) and very low among those with the disease (15.2% correct). Thus, the resulting model is highly specific but insensitive. At issue is the incremental validity (43) of the "test" or model for predicting disease status. Incremental validity refers to the increase in correct classifications when using the test and is optimized when the base rate approaches 50%. In the current study, the base rate for disease presence was 23.9%. To illustrate the importance of this issue, when the base rate for a disorder is 10%, using a moderately valid test (validity coefficient = 0.6) produces an increase in incremental validity of 11%. Using the same instrument in a population in which the base rate is 50% yields an increase in incremental validity of 25% (44).

The phrase "medical cost offset" refers to the hypothesized decrease in medical costs that would result from the treatment of psychiatric problems (45). A decrease in medical expenses that exceeds the cost of the psychological intervention is a significant offset. A metaanalysis examining the magnitude of this medical cost offset (45) found that 90% of studies reported a decrease in medical utilization after psychological intervention. The treatment groups reported a reduction in healthcare utilization of 15.7%, and the control group reported an increase of 12.3%. The overall effect size for 40 treatment-comparison group articles was 0.34. Given that those with PTSD were at risk for a wide range of physical disorders and concurrently less likely to have access to private healthcare coverage, it follows that healthcare coverage is likely to fall to emergency room facilities.

Several methodological issues should be considered when interpreting the results of this study. Data for this study were collected as a part of a larger study, the objective of which was to assess a wide variety of physical and psychological disorders as well as their demographic correlates using a large, normative nonclinical population. All data were collected by interview, and there were no objective measures of health status such as medical records. Therefore, findings of association between adverse health outcomes and PTSD may be best conceptualized as reflecting an association between PTSD and self-reported physical health outcomes. It should be noted, however, that substantial agreement (kappa = 0.80) has been found between medical report of heart disease and self-report (46).

The research clearly indicates that PTSD frequently co-occurs with a wide range of axis I disorders, including depression (47,48) and substance abuse (49). The current study did not create a "pure" PTSD group or a "pure" comparison group with no psychiatric disorder. Although this would provide important information on the independent relationship between PTSD and health-related outcomes, it also obscures the combined effect of these naturally co-occurring disorders. Most previous research using VA samples has found that persons with multiple psychiatric disorders use services at a greater rate than persons with a single diagnosis (3,5,15), although this increased risk has not been universally found (9). Investigators are encouraged to examine the relationship between health outcomes and the presence of pure disorders and specific combinations of disorders.

Lastly, this study focused on the relationship between lifetime history of PTSD and physical health. It is important to note that this group is composed of people with a history of PTSD from which they have at least partially recovered and people who currently have PTSD. Thus, the findings of this study can best be generalized to other mixed groups. In addition, no attempt was made to examine the relationship between chronicity of PTSD symptoms and health outcomes. It is worth noting that the majority (n = 238) of those who developed PTSD did so before age 18. Thus, although the sample ranged in age from 15 to 54, more than half of those who developed PTSD did so before the age of 18. Given the age of the sample (mean, 33.18), it is likely that the symptoms have been present for some time.

In conclusion, this study found that persons with a lifetime history of PTSD fared significantly worse than the comparison group on a wide range of variables related to physical health. They had lower educational and income levels, higher ambient stress, and engaged in health-promoting behaviors less frequently. These factors can partially explain the increased prevalence of chronic physical disorders among those with a history of PTSD. Persons with PTSD report higher rates of healthcare needs and are less likely to have adequate health insurance. Prospective studies examining these issues are pivotal to understand better the mechanisms linking PTSD, physical health, and stress.


   NOTES
TOP
 ABSTRACT
 INTRODUCTION
 METHODS
 RESULTS
 DISCUSSION
 NOTES
REFERENCES
 
1 Levene’s test for homogeneity of variance revealed that this assumption had been violated for both contrasts. Therefore, t tests using separate variance estimates were computed. Back

2 Alpha was computed based on weighted values of raw scores, not z-scores. Back

3 It should be noted that PTSD caseness and health status caseness were associated with both insurance (not Medicaid) and Medicaid/welfare. However, when both insurance indices are entered into logistic regression, they yield a constant (i.e., only persons who reported not having insurance [not Medicaid] endorsed the item pertaining to Medicaid/welfare). Therefore, only the item pertaining to the presence of insurance was entered into the equation. Back

Portions of this paper were presented at the 19th annual conference of the International Society for Traumatic Stress Studies, Chicago, Illinois.

DOI:10.1097/01.psy.0000188572.91553.a5


   REFERENCES
TOP
 ABSTRACT
 INTRODUCTION
 METHODS
 RESULTS
 DISCUSSION
 NOTES
 REFERENCES
 

  1. American Psychiatric Association. Diagnostic and Statistical Manual of Mental Disorders. Washington, DC: American Psychiatric Association; 2000.
  2. Beckham JC, Taft CT, Vrana SR, Feldman ME, Barefoot JC, Moore SD, Mozley SL, Butterfield MI, Calhoun PS. Ambulatory monitoring and physical health report in Vietnam veterans with and without chronic posttraumatic stress disorder. J Trauma Stress 2003;16:329–35.[CrossRef][Medline]
  3. Kramer TL, Booth BM, Han X, Williams DK. Service utilization and outcomes in medically ill veterans with posttraumatic stress and depressive disorders. J Trauma Stress 2003;16:211–9.[CrossRef][Medline]
  4. Kulka RA, Schlenger WE, Fairbank JA, Hough RL, Jordan BK, Marmar CR, Weiss DS. Trauma and the Vietnam War Generation: Report of Findings From the National Vietnam Veterans Readjustment Study. New York: Brunner/Mazel; 1990.
  5. Calhoun PS, Bosworth HB, Grambow SC, Dudley TK, Beckham JC. Medical service utilization by veterans seeking help for posttraumatic stress disorder. Am J Psychiatry 2002;159:2081–6.[Abstract/Free Full Text]
  6. Cradock-O’Leary J, Young AS, Yano EM, Wang M, Lee ML. Use of general medical services by VA patients with psychiatric disorders. Psychiatr Serv 2002;53:874–8.[Abstract/Free Full Text]
  7. Norris FH, Kaniasty KZ, Scheer DA. Use of mental health services among victims of crime: frequency, correlates, and subsequent recovery. J Consult Clin Psychol 1990;58:538–47.[CrossRef][Medline]
  8. Deykin EY, Keane TM, Kaloupek D, Fincke G, Rothendler J, Siegfried M, Creamer K. Posttraumatic stress disorder and the use of health services. Psychosom Med 2001;63:835–41.[Abstract/Free Full Text]
  9. Druss BG, Rosenheck RA. Use of medical services by veterans with mental disorders. Psychosomatics J Consult Liaison Psychiatry 1997;38:451–8.[Abstract/Free Full Text]
  10. Goodman LA, Koss MP, Russo NF. Violence against women: physical and mental health effects: II. Research findings. Appl Prev Psychol 1993;2:79–89.[CrossRef]
  11. Kimerling R, Calhoun KS. Somatic symptoms, social support, and treatment seeking among sexual assault victims. J Consult Clin Psychol 1994;62:333–40.[CrossRef][Medline]
  12. Waigandt A, Wallace DL, Phelps L, Miller DA. The impact of sexual assault on physical health status. J Trauma Stress 1990;3:93–102.
  13. Friedman MJ, Schnurr PP. The relationship between trauma, posttraumatic stress disorder, and physical health. In: Friedman MJC, Dennis S, eds. Neurobiological and Clinical Consequences of Stress: From Normal Adaptation to Posttraumatic Stress Disorder. Philadelphia: Lippincott Williams & Wilkins; 1995:507–24.
  14. Wolfe J, Schnurr PP, Brown PJ, Furey J. Posttraumatic stress disorder and war-zone exposure as correlates of perceived health in female Vietnam war veterans. J Consult Clin Psychol 1994;62:1235–40.[CrossRef][Medline]
  15. Schnurr PP, Friedman MJ, Sengupta A, Jankowski MK, Holmes T. PTSD and utilization of medical treatment services among male Vietnam veterans. J Nerv Ment Dis 2000;188:496–504.[Medline]
  16. Boscarino JA, Chang J. Electrocardiogram abnormalities among men with stress-related psychiatric disorders: implications for coronary heart disease and clinical research. Ann Behav Med 1999;21:227–34.[Medline]
  17. Boscarino JA, Chang J. Higher abnormal leukocyte and lymphocyte counts 20 years after exposure to severe stress: research and clinical implications. Psychosom Med 1999;61:378–86.[Abstract/Free Full Text]
  18. Lorig KR, Sobel DS, Ritter PL, Laurent D, Hobbs M. Effect of a self-management program on patients with chronic disease. Eff Clin Pract 2001;4:256–62.[Medline]
  19. Diehr P, Beresford SAA. The relation of dietary patterns to future survival, health, and cardiovascular events in older adults. J Clin Epidemiol 2003;56:1224–35.[CrossRef][Medline]
  20. Mcnutt L-A, Carlson BE, Persaud M, Postmus J. Cumulative abuse experiences, physical health and health behaviors. Ann Epidemiol 2002;12:123–30.[CrossRef][Medline]
  21. Fry R. Adult physical illness and childhood sexual abuse. J Psychosom Res 1993;37:89–103.[CrossRef][Medline]
  22. Koss MP. The underdetection of rape: methodological choices influence incidence estimates. J Soc Issues 1992;48:61–75.
  23. Drozdek B, Noor AK, Lutt M, Foy DW. Chronic PTSD and medical services utilization by asylum seekers. J Refug Stud 2003;16:202–11.[Abstract]
  24. Kessler RC, McGonagle KA, Zhao S, Nelson CB. Lifetime and 12-month prevalence of DSM-III-R psychiatric disorders in the United States: results from the National Comorbidity Study. Arch Gen Psychiatry 1994;51:8–19.[Abstract/Free Full Text]
  25. Breslau N, Davis GC, Andreski P, Peterson E. Traumatic events and posttraumatic stress disorder in an urban population of young adults. Arch Gen Psychiatry 1991;48:216–22.[Abstract/Free Full Text]
  26. Helzer JE, Robins LN, McEvoy L. Posttraumatic stress disorder in the general population: findings of the Epidemiologic Catchment Area survey. N Engl J Med 1987;317:1630–4.[Abstract]
  27. Gerardi R, Keane TM, Penk WE. Utility: sensitivity and specificity in developing diagnostic tests of combat-related posttraumatic stress disorder (PTSD). J Clin Psychol 1989;45:691–703.[Medline]
  28. Kulka RA, Schlenger WE, Fairbank JA, Bough RL, Jordan BK, Marmar CR, Weiss DS. National Vietnam Veterans Readjustment Study (NVVRSI: Readjustment, Current Status, and Initial P7SD Prevalence Estimates. Washington, DC: Veterans Administration; 1988.
  29. Litz BT, Penk WE, Gerardi RJ, Keane TM. Assessment of posttraumatic stress disorder. In: Saigh PA, ed. Posttraumatic Stress Disorder: A Behavioral Approach to Assessment and Treatment. Boston: Allyn and Bacon; 1992:50–84.
  30. Kessler RC, Wittchen H-U, Abelson J, McGonagle KA, Schwarz N, Kendler KS, Knauper B, Zhao S. Methodological studies of the Composite International Diagnostic Interview (CIDI) in the US National Comorbidity Survey. Int J Methods Psychiatr Res 1998;7:33–5.[CrossRef]
  31. Kessler RC, Wittchen H-U, Abelson J, Zhao S. Methodological issues in assessing psychiatric disorders with self-reports. In: Stone AA, Turkkan JS, eds. The Science of Self-Report: Implications for Research and Practice. Mahwah, NJ: USum Associates; 2000:229–55.
  32. Composite International Diagnostic Interview (CIDI), version 1.0. Geneva: World Health Organization; 1990.
  33. Goldberg LR. The development of markers for the big-five factor structure. Psychol Assess 1992;4:26–42.
  34. Wittchen H-U, Robins LN, Cottler LB, Sartorius N, Burke JD, Regier D. Cross-cultural feasibility, reliability and sources of variance of the Composite International Diagnostic Interview (CIDI). Br J Psychiatry 1991;159:645–53.[Abstract/Free Full Text]
  35. Cottler LB, Robins LN, Grant BF, Blaine J, Towle LH, Wittchen HU, Sartorius N. The CIDI-core substance abuse and dependence questions: cross-cultural and nosological issues. Br J Psychiatry 1991;159:653–8.[Abstract/Free Full Text]
  36. Wacker HR, Battegay R, Mullejans R, Schlosser C. Using the CIDI-C in the general population. In: Setefanis CN, Rabavilas AD, Soldatos CR, eds. Psychiatry: A World Perspective. Amsterdam: Elsevier Science Publishers; 1990.
  37. Semier G, von Cranach M, Wittchen H-U. Comparison between the Composite International Diagnostic Interview and the Present State Examination. Report to the WHO/ADAMHA Task Force on Instrument Development. Geneva; 1987.
  38. Wittchen H-U, Burke JD, Semler G, Pfister H. Recall and dating of psychiatric symptoms: test–retest reliability of time-related symptom questions in a standardized psychiatric interview. Arch Gen Psychiatry 1989;46:437–43.[Abstract/Free Full Text]
  39. Janca A, Robins LN, Cottler LB, Early TS. Clinical observation of assessment using the Composite International Diagnostic Interview (CIDI): an analysis of the CIDI field trials-wave II at the St Louis site. Br J Psychiatry 1992;160:815–8.[Abstract/Free Full Text]
  40. Wittchen HU, Zhao S, Abelson JM, Abelson JL, Kessler RC. Reliability and procedural validity of UM-CIDI DSM-III-R phobic disorders. Psychol Med 1996;26:1169–77.[Medline]
  41. Menard S. Applied Logistic Regression Analysis. Thousand Oaks, CA: Sage Publications; 1995.
  42. Shalev A, Bleich A, Ursano RJ. Posttraumatic stress disorder: somatic comorbidity and effort tolerance. Psychosomatics J Consult Liaison Psychiatry 1990;31:197–203.[Abstract/Free Full Text]
  43. Sechrest L. Incremental validity: a recommendation. Educ Psychol Meas 1963;23:153–8.[CrossRef]
  44. Anastasi A, Urbana S. Psychological Testing. Upper Saddle River, NJ, Prentice Hall; 1997:146–8.
  45. Chiles JA, Lambert MJ, Hatch AL. Medical cost offset: a review of the impact of psychological interventions on medical utilization over the past three decades. In: Cummings NAOD, William T, eds. The Impact of Medical Cost Offset on Practice and Research: Making It Work for You: A Report of the First Reno Conference on Medical Cost Offset Healthcare utilization and Cost Series, vol 5. Reno, NV: US Context Press; 2002.
  46. Haapanen N, Miilunpalo S, Pasanen M, Oja P, Vuori I. Agreement between questionnaire data and medical records of chroni diseases in middle-aged and elderly Finnish men and women. Am J Epidemiol 1997;762–9.
  47. Franklin CL, Zimmerman M. Posttraumatic stress disorder and major depressive disorder: investigating the role of overlapping symptoms in diagnostic comorbidity. J Nerv Ment Dis 2001;189:548–51.[CrossRef][Medline]
  48. Green BL, Lindy JD, Grace MC, Leonard AC. Chronic posttraumatic stress disorder and diagnostic comorbidity in a disaster sample. J Nerv Ment Dis 1992;180:760–6.[Medline]
  49. Kessler RC, Sonnega A, Bromet EJ, Hughes M, Nelson CB. Posttraumatic stress disorder in the National Comorbidity Survey. Arch Gen Psychiatry 1995;52:1048–60.[Abstract/Free Full Text]




This Article
Right arrow Abstract Freely available
Right arrow Full Text (PDF)
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrowRequest Permissions
Citing Articles
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Lauterbach, D.
Right arrow Articles by Rakow, M.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Lauterbach, D.
Right arrow Articles by Rakow, M.
Related Collections
Right arrow Health Psychology
Right arrow PTSD
Right arrow Other Psychiatric Disorders

HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS