Dataset Information

Using administrative data to identify U.S. Army soldiers at high-risk of perpetrating minor violent crimes.

ABSTRACT: Growing concerns exist about violent crimes perpetrated by U.S. military personnel. Although interventions exist to reduce violent crimes in high-risk populations, optimal implementation requires evidence-based targeting. The goal of the current study was to use machine learning methods (stepwise and penalized regression; random forests) to develop models to predict minor violent crime perpetration among U.S. Army soldiers. Predictors were abstracted from administrative data available for all 975,057 soldiers in the U.S. Army 2004-2009, among whom 25,966 men and 2728 women committed a first founded minor violent crime (simple assault, blackmail-extortion-intimidation, rioting, harassment). Temporally prior administrative records measuring socio-demographic, Army career, criminal justice, medical/pharmacy, and contextual variables were used to build separate male and female prediction models that were then tested in an independent 2011-2013 sample. Final model predictors included young age, low education, early career stage, prior crime involvement, and outpatient treatment for diverse emotional and substance use problems. Area under the receiver operating characteristic curve was 0.79 (for men and women) in the 2004-2009 training sample and 0.74-0.82 (men-women) in the 2011-2013 test sample. 30.5-28.9% (men-women) of all administratively-recorded crimes in 2004-2009 were committed by the 5% of soldiers having highest predicted risk, with similar proportions (28.5-29.0%) when the 2004-2009 coefficients were applied to the 2011-2013 test sample. These results suggest that it may be possible to target soldiers at high-risk of violence perpetration for preventive interventions, although final decisions about such interventions would require weighing predicted effectiveness against intervention costs and competing risks.

SUBMITTER: Rosellini AJ

PROVIDER: S-EPMC5125854 | biostudies-literature | 2017 Jan

REPOSITORIES: biostudies-literature

ACCESS DATA

Publications

Using administrative data to identify U.S. Army soldiers at high-risk of perpetrating minor violent crimes.

Rosellini Anthony J AJ Monahan John J Street Amy E AE Hill Eric D ED Petukhova Maria M Reis Ben Y BY Sampson Nancy A NA Benedek David M DM Bliese Paul P Stein Murray B MB Ursano Robert J RJ Kessler Ronald C RC

Journal of psychiatric research 20160930

Growing concerns exist about violent crimes perpetrated by U.S. military personnel. Although interventions exist to reduce violent crimes in high-risk populations, optimal implementation requires evidence-based targeting. The goal of the current study was to use machine learning methods (stepwise and penalized regression; random forests) to develop models to predict minor violent crime perpetration among U.S. Army soldiers. Predictors were abstracted from administrative data available for all 97 ...[more]

PMID: 27741501

Similar Datasets

Project description:BackgroundHigh rates of mental disorders, suicidality, and interpersonal violence early in the military career have raised interest in implementing preventive interventions with high-risk new enlistees. The Army Study to Assess Risk and Resilience in Servicemembers (STARRS) developed risk-targeting systems for these outcomes based on machine learning methods using administrative data predictors. However, administrative data omit many risk factors, raising the question whether risk targeting could be improved by adding self-report survey data to prediction models. If so, the Army may gain from routinely administering surveys that assess additional risk factors.MethodsThe STARRS New Soldier Survey was administered to 21,790 Regular Army soldiers who agreed to have survey data linked to administrative records. As reported previously, machine learning models using administrative data as predictors found that small proportions of high-risk soldiers accounted for high proportions of negative outcomes. Other machine learning models using self-report survey data as predictors were developed previously for three of these outcomes: major physical violence and sexual violence perpetration among men and sexual violence victimization among women. Here we examined the extent to which this survey information increases prediction accuracy, over models based solely on administrative data, for those three outcomes. We used discrete-time survival analysis to estimate a series of models predicting first occurrence, assessing how model fit improved and concentration of risk increased when adding the predicted risk score based on survey data to the predicted risk score based on administrative data.ResultsThe addition of survey data improved prediction significantly for all outcomes. In the most extreme case, the percentage of reported sexual violence victimization among the 5% of female soldiers with highest predicted risk increased from 17.5% using only administrative predictors to 29.4% adding survey predictors, a 67.9% proportional increase in prediction accuracy. Other proportional increases in concentration of risk ranged from 4.8% to 49.5% (median = 26.0%).ConclusionsData from an ongoing New Soldier Survey could substantially improve accuracy of risk models compared to models based exclusively on administrative predictors. Depending upon the characteristics of interventions used, the increase in targeting accuracy from survey data might offset survey administration costs.

Project description:Studies have suggested that sickle cell trait elevates the risks of exertional rhabdomyolysis and death. We conducted a study of sickle cell trait in relation to these outcomes, controlling for known risk factors for exertional rhabdomyolysis, in a large population of active persons who had undergone laboratory tests for hemoglobin AS (HbAS) and who were subject to exertional-injury precautions.We used Cox proportional-hazards models to test whether the risks of exertional rhabdomyolysis and death varied according to sickle cell trait status among 47,944 black soldiers who had undergone testing for HbAS and who were on active duty in the U.S. Army between January 2011 and December 2014. We used the Stanford Military Data Repository, which contains comprehensive medical and administrative data on all active-duty soldiers.There was no significant difference in the risk of death among soldiers with sickle cell trait, as compared with those without the trait (hazard ratio, 0.99; 95% confidence interval [CI], 0.46 to 2.13; P=0.97), but the trait was associated with a significantly higher adjusted risk of exertional rhabdomyolysis (hazard ratio, 1.54; 95% CI, 1.12 to 2.12; P=0.008). This effect was similar in magnitude to that associated with tobacco use, as compared with no use (hazard ratio, 1.54; 95% CI, 1.23 to 1.94; P<0.001), and to that associated with having a body-mass index (BMI; the weight in kilograms divided by the square of the height in meters) of 30.0 or more, as compared with a BMI of less than 25.0 (hazard ratio, 1.39; 95% CI, 1.04 to 1.86; P=0.03). The effect was less than that associated with recent use of a statin, as compared with no use (hazard ratio, 2.89; 95% CI, 1.51 to 5.55; P=0.001), or an antipsychotic agent (hazard ratio, 3.02; 95% CI, 1.34 to 6.82; P=0.008).Sickle cell trait was not associated with a higher risk of death than absence of the trait, but it was associated with a significantly higher risk of exertional rhabdomyolysis. (Funded by the National Heart, Lung, and Blood Institute and the Uniformed Services University of the Health Sciences.).

Project description:IntroductionUnderstanding how soldiers die after separation from military service, particularly those who die shortly after separating from service, may help to identify opportunities to ease transitions to civilian life.Materials and methodsMortality data were analyzed for male U.S. Army soldiers who died within 18 months of their separation from service (from 1999 to 2011). Descriptive statistics were calculated for natural, accidental, suicidal, homicidal, undetermined, and legal/operation of war deaths and were stratified by age, component, time in service, and type of discharge. Crude and age-adjusted mortality rates, standardized for all years using the 2004 Regular Army population, were also calculated. The Public Health Review Board of the U.S. Army Public Health Center approved this study as Public Health Practice.ResultsOf the 1,884,653 male soldiers who separated from service during the study period, 3,819 died within 18 months of separation. A majority of all separations were Reserve or National Guard (58%), and 62% of decedents were in the Reserve or National Guard. Deaths from natural causes (38%) were the most common, followed by accident deaths (34%), suicides (20%), homicides (6%), undetermined deaths (2%), and legal/operation of war deaths (<1%). For overall mortality, age-adjusted rates were higher among the male U.S. population when compared to male soldiers who had separated from the Army.ConclusionsThe time period immediately following separation from the Army presents a unique challenge for many soldiers. Developing more effective pre-separation prevention programs that target specific risks requires knowing the causes of death for natural deaths, suicides, and accidents. Over half of all deaths occurring shortly after separation from service are preventable. Continued surveillance of specific causes of preventable deaths following separation can improve pre-separation prevention programs and transitions to post-service care.

Project description:BackgroundAlthough the majority of active duty U.S. Army soldiers are full-time personnel in the Active Component (AC), a substantial minority of soldiers on active duty are in the Reserve Components (RCs). These "citizen-soldiers" (Army National Guard and Army Reserve) represent a force available for rapid activation in times of national need. RC soldiers experience many of the same stressors as AC soldiers as well as stressors that are unique to their intermittent service. Despite the important role of RC soldiers, the vast majority of military mental health research focuses on AC soldiers. One important goal of the Army Study to Assess Risk and Resilience in Servicemembers (Army STARRS) is to address this gap. Here we examine predictors of suicide attempts among activated RC soldiers.MethodsThis longitudinal, retrospective cohort study used individual-level person-month records from Army and Department of Defense administrative data systems to examine socio-demographic, service-related, and mental health predictors of medically documented suicide attempts among activated RC soldiers during 2004-2009. Data from all 1103 activated RC suicide attempters and an equal-probability sample of 69,867 control person-months were analyzed using a discrete-time survival framework.ResultsEnlisted soldiers comprised 84.3% of activated RC soldiers and accounted for 95.7% of all activated RC suicide attempts (overall rate = 108/100,000 person-years, more than four times the rate among officers). Multivariable predictors of enlisted RC suicide attempts included being female, entering Army service at age ≥ 25, current age < 30, non-Hispanic white, less than high school education, currently married, having 1-2 years of service, being previously deployed (vs. currently deployed), and history of mental health diagnosis (particularly when documented in the previous month). Predictors among RC officers (overall rate = 26/100,000 person-years) included being female and receiving a mental health diagnosis in the previous month. Discrete-time hazard models showed suicide attempt risk among enlisted soldiers was inversely associated with time in service.ConclusionsRisk factors for suicide attempt in the RCs were similar to those previously observed in the AC, highlighting the importance of research and prevention focused on RC enlisted soldiers in the early phases of Army service and those with a recent mental health diagnosis.

Dataset Information

Using administrative data to identify U.S. Army soldiers at high-risk of perpetrating minor violent crimes.

Publications

Using administrative data to identify U.S. Army soldiers at high-risk of perpetrating minor violent crimes.

Similar Datasets

OmicsDI is part of the ELIXIR infrastructure

Tweets