Michael Schoenbaum

Including information about co-morbidity in estimates of disease burden: results from the World Health Organization World Mental Health Surveys.

BACKGROUNDnThe methodology commonly used to estimate disease burden, featuring ratings of severity of individual conditions, has been criticized for ignoring co-morbidity. A methodology that addresses this problem is proposed and illustrated here with data from the World Health Organization World Mental Health Surveys. Although the analysis is based on self-reports about ones own conditions in a community survey, the logic applies equally well to analysis of hypothetical vignettes describing co-morbid condition profiles.nnnMETHODnFace-to-face interviews in 13 countries (six developing, nine developed; n=31 067; response rate=69.6%) assessed 10 classes of chronic physical and nine of mental conditions. A visual analog scale (VAS) was used to assess overall perceived health. Multiple regression analysis with interactions for co-morbidity was used to estimate associations of conditions with VAS. Simulation was used to estimate condition-specific effects.nnnRESULTSnThe best-fitting model included condition main effects and interactions of types by numbers of conditions. Neurological conditions, insomnia and major depression were rated most severe. Adjustment for co-morbidity reduced condition-specific estimates with substantial between-condition variation (0.24-0.70 ratios of condition-specific estimates with and without adjustment for co-morbidity). The societal-level burden rankings were quite different from the individual-level rankings, with the highest societal-level rankings associated with conditions having high prevalence rather than high individual-level severity.nnnCONCLUSIONSnPlausible estimates of disorder-specific effects on VAS can be obtained using methods that adjust for co-morbidity. These adjustments substantially influence condition-specific ratings.

Risk Factors, Methods, and Timing of Suicide Attempts Among US Army Soldiers

IMPORTANCEnSuicide attempts in the US Army have risen in the past decade. Understanding the association between suicide attempts and deployment, as well as method and timing of suicide attempts, can assist in developing interventions.nnnOBJECTIVEnTo examine suicide attempt risk factors, methods, and timing among soldiers currently deployed, previously deployed, and never deployed at the time this study was conducted.nnnDESIGN, SETTING, AND PARTICIPANTSnThis longitudinal, retrospective cohort study of Regular Army-enlisted soldiers on active duty from 2004 through 2009 used individual-level person-month records to examine risk factors (sociodemographic, service related, and mental health), method, and time of suicide attempt by deployment status (never, currently, and previously deployed). Administrative data for the month before each of 9650 incident suicide attempts and an equal-probability sample of 153u202f528 control person-months for other soldiers were analyzed using a discrete-time survival framework.nnnMAIN OUTCOMES AND MEASURESnSuicide attempts and career, mental health, and demographic predictors were obtained from administrative and medical records.nnnRESULTSnOf the 9650 enlisted soldiers who attempted suicide, 86.3% were male, 68.4% were younger than 30 years, 59.8% were non-Hispanic white, 76.5% were high school educated, and 54.7% were currently married. The 40.4% of enlisted soldiers who had never been deployed (nu2009=u200912u202f421u202f294 person-months) accounted for 61.1% of enlisted soldiers who attempted suicide (nu2009=u20095894 cases). Risk among those never deployed was highest in the second month of service (103 per 100u202f000 person-months). Risk among soldiers on their first deployment was highest in the sixth month of deployment (25 per 100u202f000 person-months). For those previously deployed, risk was highest at 5 months after return (40 per 100u202f000 person-months). Currently and previously deployed soldiers were more likely to attempt suicide with a firearm than those never deployed (currently deployed: OR, 4.0; 95% CI, 2.9-5.6; previously deployed: OR, 2.7; 95% CI, 1.8-3.9). Across deployment status, suicide attempts were more likely among soldiers who were women (currently deployed: OR, 3.4; 95% CI, 3.0-4.0; previously deployed: OR, 1.5; 95% CI, 1.4-1.7; and never deployed: OR, 2.4; 95% CI, 2.3-2.6), in their first 2 years of service (currently deployed: OR, 1.9; 95% CI, 1.5-2.3; previously deployed: OR, 2.2; 95% CI, 1.9-2.7; and never deployed: OR, 3.1; 95% CI, 2.7-3.6), and had a recently received a mental health diagnosis in the previous month (currently deployed: OR, 29.8; 95% CI, 25.0-35.5; previously deployed: OR, 22.2; 95% CI, 20.1-24.4; and never deployed: OR, 15.0; 95% CI, 14.2-16.0). Among soldiers with 1 previous deployment, odds of a suicide attempt were higher for those who screened positive for depression or posttraumatic stress disorder after return from deployment and particularly at follow-up screening, about 4 to 6 months after deployment (depression: OR, 1.4; 95% CI, 1.1-1.9; posttraumatic stress disorder: OR, 2.4; 95% CI, 2.1-2.8).nnnCONCLUSIONS AND RELEVANCEnIdentifying the timing and risk factors for suicide attempt in soldiers requires consideration of environmental context, individual characteristics, and mental health. These factors can inform prevention efforts.

Predicting suicides after outpatient mental health visits in the Army Study to Assess Risk and Resilience in Servicemembers (Army STARRS)

The 2013 US Veterans Administration/Department of Defense Clinical Practice Guidelines (VA/DoD CPG) require comprehensive suicide risk assessments for VA/DoD patients with mental disorders but provide minimal guidance on how to carry out these assessments. Given that clinician-based assessments are not known to be strong predictors of suicide, we investigated whether a precision medicine model using administrative data after outpatient mental health specialty visits could be developed to predict suicides among outpatients. We focused on male nondeployed Regular US Army soldiers because they account for the vast majority of such suicides. Four machine learning classifiers (naive Bayes, random forests, support vector regression and elastic net penalized regression) were explored. Of the Army suicides in 2004–2009, 41.5% occurred among 12.0% of soldiers seen as outpatient by mental health specialists, with risk especially high within 26 weeks of visits. An elastic net classifier with 10–14 predictors optimized sensitivity (45.6% of suicide deaths occurring after the 15% of visits with highest predicted risk). Good model stability was found for a model using 2004–2007 data to predict 2008–2009 suicides, although stability decreased in a model using 2008–2009 data to predict 2010–2012 suicides. The 5% of visits with highest risk included only 0.1% of soldiers (1047.1 suicides/100u2009000 person-years in the 5 weeks after the visit). This is a high enough concentration of risk to have implications for targeting preventive interventions. An even better model might be developed in the future by including the enriched information on clinician-evaluated suicide risk mandated by the VA/DoD CPG to be recorded.

Mental health treatment among soldiers with current mental disorders in the Army Study to Assess Risk and Resilience in Service Members (Army STARRS)

A representative sample of 5,428 nondeployed Regular Army soldiers completed a self-administered questionnaire (SAQ) and consented to linking SAQ data with administrative records as part of the Army Study to Assess Risk and Resilience in Service members. The SAQ included information about prevalence and treatment of mental disorders among respondents with current Diagnostic and Statistical Manual of Mental Disorders, Fourth Edition (DSM-IV) internalizing (anxiety, mood) and externalizing (disruptive behavior, substance) disorders. 21.3% of soldiers with any current disorder reported current treatment. Seven significant predictors of being in treatment were identified. Four of these 7 were indicators of psychopathology (bipolar disorder, panic disorder, post-traumatic stress disorder, 8+ months duration of disorder). Two were sociodemographics (history of marriage, not being non-Hispanic Black). The final predictor was history of deployment. Treatment rates varied between 4.7 and 71.5% depending on how many positive predictors the soldier had. The vast majority of soldiers had a low number of these predictors. These results document that most nondeployed soldiers with mental disorders are not in treatment and that untreated soldiers are not concentrated in a particular segment of the population that might be targeted for special outreach efforts. Analysis of modifiable barriers to treatment is needed to help strengthen outreach efforts.

Childhood adversity, adult stress, and the risk of major depression or generalized anxiety disorder in US soldiers: a test of the stress sensitization hypothesis.

BACKGROUNDnThe stress sensitization theory hypothesizes that individuals exposed to childhood adversity will be more vulnerable to mental disorders from proximal stressors. We aimed to test this theory with respect to risk of 30-day major depressive episode (MDE) and generalized anxiety disorder (GAD) among new US Army soldiers.nnnMETHODSnThe sample consisted of 30 436 new soldier recruits in the Army Study to Assess Risk and Resilience (Army STARRS). Generalized linear models were constructed, and additive interactions between childhood maltreatment profiles and level of 12-month stressful experiences on the risk of 30-day MDE and GAD were analyzed.nnnRESULTSnStress sensitization was observed in models of past 30-day MDE (χ2 8 = 17.6, p = 0.025) and GAD (χ2 8 = 26.8, p = 0.001). This sensitization only occurred at high (3+) levels of reported 12-month stressful experiences. In pairwise comparisons for the risk of 30-day MDE, the risk difference between 3+ stressful experiences and no stressful experiences was significantly greater for all maltreatment profiles relative to No Maltreatment. Similar results were found with the risk for 30-day GAD with the exception of the risk difference for Episodic Emotional and Sexual Abuse, which did not differ statistically from No Maltreatment.nnnCONCLUSIONSnNew soldiers are at an increased risk of 30-day MDE or GAD following recent stressful experiences if they were exposed to childhood maltreatment. Particularly in the military with an abundance of unique stressors, attempts to identify this population and improve stress management may be useful in the effort to reduce the risk of mental disorders.

Risk factors for accident death in the u.s. Army, 2004-2009

BACKGROUNDnAccidents are one of the leading causes of death among U.S. active-duty Army soldiers. Evidence-based approaches to injury prevention could be strengthened by adding person-level characteristics (e.g., demographics) to risk models tested on diverse soldier samples studied over time.nnnPURPOSEnTo identify person-level risk indicators of accident deaths in Regular Army soldiers during a time frame of intense military operations, and to discriminate risk of not-line-of-duty from line-of-duty accident deaths.nnnMETHODSnAdministrative data acquired from multiple Army/Department of Defense sources for active duty Army soldiers during 2004-2009 were analyzed in 2013. Logistic regression modeling was used to identify person-level sociodemographic, service-related, occupational, and mental health predictors of accident deaths.nnnRESULTSnDelayed rank progression or demotion and being male, unmarried, in a combat arms specialty, and of low rank/service length increased odds of accident death for enlisted soldiers. Unique to officers was high risk associated with aviation specialties. Accident death risk decreased over time for currently deployed, enlisted soldiers and increased for those never deployed. Mental health diagnosis was associated with risk only for previous and never-deployed, enlisted soldiers. Models did not discriminate not-line-of-duty from line-of-duty accident deaths.nnnCONCLUSIONSnAdding more refined person-level and situational risk indicators to current models could enhance understanding of accident death risk specific to soldier rank and deployment status. Stable predictors could help identify high risk of accident deaths in future cohorts of Regular Army soldiers.

Including information about co-morbidity in estimates of disease burden

BACKGROUNDnThe methodology commonly used to estimate disease burden, featuring ratings of severity of individual conditions, has been criticized for ignoring co-morbidity. A methodology that addresses this problem is proposed and illustrated here with data from the World Health Organization World Mental Health Surveys. Although the analysis is based on self-reports about ones own conditions in a community survey, the logic applies equally well to analysis of hypothetical vignettes describing co-morbid condition profiles.nnnMETHODnFace-to-face interviews in 13 countries (six developing, nine developed; n=31 067; response rate=69.6%) assessed 10 classes of chronic physical and nine of mental conditions. A visual analog scale (VAS) was used to assess overall perceived health. Multiple regression analysis with interactions for co-morbidity was used to estimate associations of conditions with VAS. Simulation was used to estimate condition-specific effects.nnnRESULTSnThe best-fitting model included condition main effects and interactions of types by numbers of conditions. Neurological conditions, insomnia and major depression were rated most severe. Adjustment for co-morbidity reduced condition-specific estimates with substantial between-condition variation (0.24-0.70 ratios of condition-specific estimates with and without adjustment for co-morbidity). The societal-level burden rankings were quite different from the individual-level rankings, with the highest societal-level rankings associated with conditions having high prevalence rather than high individual-level severity.nnnCONCLUSIONSnPlausible estimates of disorder-specific effects on VAS can be obtained using methods that adjust for co-morbidity. These adjustments substantially influence condition-specific ratings.