Dataset Information

Developing a practical suicide risk prediction model for targeting high-risk patients in the Veterans health Administration.

ABSTRACT: OBJECTIVES:The US Veterans Health Administration (VHA) has begun using predictive modeling to identify Veterans at high suicide risk to target care. Initial analyses are reported here. METHODS:A penalized logistic regression model was compared with an earlier proof-of-concept logistic model. Exploratory analyses then considered commonly-used machine learning algorithms. Analyses were based on electronic medical records for all 6,360 individuals classified in the National Death Index as having died by suicide in fiscal years 2009-2011 who used VHA services the year of their death or prior year and a 1% probability sample of time-matched VHA service users alive at the index date (n = 2,112,008). RESULTS:A penalized logistic model with 61 predictors had sensitivity comparable to the proof-of-concept model (which had 381 predictors) at target thresholds. The machine learning algorithms had relatively similar sensitivities, the highest being for Bayesian additive regression trees, with 10.7% of suicides occurred among the 1.0% of Veterans with highest predicted risk and 28.1% among the 5.0% of with highest predicted risk. CONCLUSIONS:Based on these results, VHA is using penalized logistic regression in initial intervention implementation. The paper concludes with a discussion of other practical issues that might be explored to increase model performance.

SUBMITTER: Kessler RC

PROVIDER: S-EPMC5614864 | biostudies-literature | 2017 Sep

REPOSITORIES: biostudies-literature

ACCESS DATA

Publications

Developing a practical suicide risk prediction model for targeting high-risk patients in the Veterans health Administration.

Kessler Ronald C RC Hwang Irving I Hoffmire Claire A CA McCarthy John F JF Petukhova Maria V MV Rosellini Anthony J AJ Sampson Nancy A NA Schneider Alexandra L AL Bradley Paul A PA Katz Ira R IR Thompson Caitlin C Bossarte Robert M RM

International journal of methods in psychiatric research 20170704 3

<h4>Objectives</h4>The US Veterans Health Administration (VHA) has begun using predictive modeling to identify Veterans at high suicide risk to target care. Initial analyses are reported here.<h4>Methods</h4>A penalized logistic regression model was compared with an earlier proof-of-concept logistic model. Exploratory analyses then considered commonly-used machine learning algorithms. Analyses were based on electronic medical records for all 6,360 individuals classified in the National Death Ind ...[more]

PMID: 28675617

Similar Datasets

Project description:ImportanceThe Veterans Health Administration (VHA) serves a population of veterans with a high prevalence of comorbid health conditions and increased risk for suicide.ObjectiveTo replicate the findings of a previous study and assess whether exposure to angiotensin receptor blockers (ARBs) is associated with differential suicide risk compared with angiotensin-converting enzyme inhibitors (ACEIs) among veterans receiving VHA care.Design, setting, and participantsThis nested case-control design included all suicide decedents from 2015 to 2017 with a VHA inpatient or outpatient encounter in the prior year and with either an active ACEI or ARB prescription in the 100 days prior to death. Using a 4:1 ratio, controls were matched to cases by age, sex, and hypertension and diabetes diagnoses. Controls were alive at the time of the death of the matched case, had a VHA encounter within the previous year, and had either an active ACEI or ARB medication fill within 100 days before the death of the matched case.ExposuresAn active ACEI or ARB prescription within 100 days before the death of the case.Main outcomes and measuresCases were suicide decedents from 2015 to 2017 per National Death Index search results included in the Veteran Affairs/Department of Defense Mortality Data Repository.ResultsAmong 1309 cases, the median (interquartile range [IQR]) age was 68 (60-76) years and among 5217 controls, the median (IQR) age was 67 (60-76) years, and 1.9% of veterans in both groups were female. ARBs were received by 20.2% of controls and 19.6% of cases; ACEIs were received by 79.8% of controls and 80.4% of cases. The crude suicide odds ratio for ARBs vs ACEIs was 0.966 (95% CI, 0.828-1.127). Controlling for covariates, the adjusted odds ratio for ARBs was 0.985 (95% CI, 0.834-1.164). Sensitivity analyses using only those covariates that differed significantly between groups, restricting to veterans ages 65 and older, dropping matching criteria, and adjusting for the quantity and temporal proximity of ACEI and ARB exposure in the 100 days prior to the index date, had consistent findings.Conclusions and relevanceThis case-control study did not identify differences in suicide risk by receipt of ARBs vs ACEIs in analyses specific to veterans receiving VHA care in contrast with findings from the referent study.

Project description:ImportanceUnderstanding the extent to which population-level suicide risk screening facilities follow-up and engagement in mental health treatment is important as engaging at-risk individuals in treatment is critical to reducing suicidal behaviors.ObjectiveTo evaluate mental health follow-up and treatment engagement in the Veterans Health Administration (VHA) following administration of the Columbia-Suicide Severity Rating Scale (C-SSRS) screen, a component of the VHA's universal suicide risk screening program.DesignThis cross-sectional study used data from VA's Corporate Data Warehouse.Settings140 VHA Medical Centers.ParticipantsPatients who completed the C-SSRS screen in ambulatory care between October 1, 2018-September 30, 2020.ExposureStandardized suicide risk screening.Main outcomes and measuresMental health follow-up (one or more visits within 30 days of C-SSRS screening) and treatment engagement (two or more visits within 90 days of C-SSRS screening) were examined.Results97,224 Veterans in Fiscal Year 2019 (FY19) (mean age 51.4 years; 86.8% male; 64.8% white, 22.4% African-American) and 58,693 Veterans in FY20 (mean age 49.6 years; 85.5% male; 63.4% white, 21.9% African-American) received the C-SSRS screen. Across FYs, a positive C-SSRS screen was associated with increased probability of mental health follow-up and treatment engagement. Patients who were not seen in mental health in the year prior to screening had the greatest increase in probability of mental health follow-up and engagement following a positive screen (P<0.001). For FY19, a positive C-SSRS screen in non-mental health connected patients was associated with an increased probability of follow-up from 49.8% to 79.5% (relative risk = 1.60) and engagement from 39.5% to 63.6% (relative risk = 1.61). For mental health-connected patients, a positive C-SSRS screen was associated with a smaller increase in probability of follow-up from 75.8% to 87.6% (relative risk = 1.16) and engagement from 63.3% to 76.4% (relative risk = 1.21). Results for FY20 were similar.Conclusions and relevanceIdentification of suicide risk through population-level screening was associated with increased mental health follow-up and engagement, particularly for non-mental health connected patients. Findings support the use of a standardized, comprehensive suicide risk screening program for managing elevated suicide risk in a large healthcare system.

Project description:ImportanceThe Veterans Health Administration (VHA) implemented a national clinical program using a suicide risk prediction algorithm, Recovery Engagement and Coordination for Health-Veterans Enhanced Treatment (REACH VET), in which clinicians facilitate care enhancements for individuals identified in local top 0.1% suicide risk tiers. Evaluation studies are needed.ObjectiveTo determine associations with treatment engagement, health care utilization, suicide attempts, safety plan documentation, and 6-month mortality.Design, setting, and participantsThis cohort study used triple differences analyses comparing 6-month changes in outcomes after vs before program entry for individuals entering the REACH VET program (March 2017-December 2018) vs a similarly identified top 0.1% suicide risk tier cohort from prior to program initiation (March 2014-December 2015), adjusting for trends across subthreshold cohorts. Subcohort analyses (including individuals from March 2017-June 2018) evaluated difference-in-differences for cause-specific mortality using death certificate data. The subthreshold cohorts included individuals in the top 0.3% to 0.1% suicide risk tier, below the threshold for REACH VET eligibility, from the concurrent REACH VET period and from the pre-REACH VET period. Data were analyzed from December 2019 through September 2021.ExposuresREACH VET-designated clinicians treatment reevaluation and outreach for care enhancements, including safety planning, increased monitoring, and interventions to enhance coping.Main outcomes and measuresProcess outcomes included VHA scheduled, completed, and missed appointments; mental health visits; and safety plan documentation and documentation within 6 months for individuals without plans within the prior 2 years. Clinical outcomes included mental health admissions, emergency department visits, nonfatal suicide attempts, and all-cause, suicide, and nonsuicide external-cause mortality.ResultsA total of 173 313 individuals (mean [SD] age, 51.0 [14.7] years; 161 264 [93.1%] men and 12 049 [7.0%] women) were included in analyses, including 40 816 individuals eligible for REACH VET care and 36 604 individuals from the pre-REACH VET period in the top 0.1% of suicide risk. The REACH VET intervention was associated with significant increases in completed outpatient appointments (adjusted triple difference [ATD], 0.31; 95% CI, 0.06 to 0.55) and proportion of individuals with new safety plans (ATD, 0.08; 95% CI, 0.06 to 0.10) and reductions in mental health admissions (ATD, -0.08; 95% CI, -0.10 to -0.05), emergency department visits (ADT, -0.03; 95% CI, -0.06 to -0.01), and suicide attempts (ADT, -0.05; 95% CI, -0.06 to -0.03). Subcohort analyses did not identify differences in suicide or all-cause mortality (eg, age-and-sex-adjusted difference-in-difference for suicide mortality, 0.0007; 95% CI, -0.0006 to 0.0019).Conclusions and relevanceThese findings suggest that REACH VET implementation was associated with greater treatment engagement and new safety plan documentation and fewer mental health admissions, emergency department visits, and suicide attempts. Clinical programs using risk modeling may be effective tools to support care enhancements and risk reduction.

Project description:BackgroundFollow-up visits for very preterm infants (VPI) after hospital discharge is crucial for their neurodevelopmental trajectories, but ensuring their attendance before 12 months corrected age (CA) remains a challenge. Current prediction models focus on future outcomes at discharge, but post-discharge data may enhance predictions of neurodevelopmental trajectories due to brain plasticity. Few studies in this field have utilized machine learning models to achieve this potential benefit with transparency, explainability, and transportability.MethodsWe developed four prediction models for cognitive or motor function at 24 months CA separately at each follow-up visits, two for the 6-month and two for the 12-month CA visits, using hospitalized and follow-up data of VPI from the Taiwan Premature Infant Follow-up Network from 2010 to 2017. Regression models were employed at 6 months CA, defined as a decline in The Bayley Scales of Infant Development 3rd edition (BSIDIII) composite score > 1 SD between 6- and 24-month CA. The delay models were developed at 12 months CA, defined as a BSIDIII composite score < 85 at 24 months CA. We used an evolutionary-derived machine learning method (EL-NDI) to develop models and compared them to those built by lasso regression, random forest, and support vector machine.ResultsOne thousand two hundred forty-four VPI were in the developmental set and the two validation cohorts had 763 and 1347 VPI, respectively. EL-NDI used only 4-10 variables, while the others required 29 or more variables to achieve similar performance. For models at 6 months CA, the area under the receiver operating curve (AUC) of EL-NDI were 0.76-0.81(95% CI, 0.73-0.83) for cognitive regress with 4 variables and 0.79-0.83 (95% CI, 0.76-0.86) for motor regress with 4 variables. For models at 12 months CA, the AUC of EL-NDI were 0.75-0.78 (95% CI, 0.72-0.82) for cognitive delay with 10 variables and 0.73-0.82 (95% CI, 0.72-0.85) for motor delay with 4 variables.ConclusionsOur EL-NDI demonstrated good performance using simpler, transparent, explainable models for clinical purpose. Implementing these models for VPI during follow-up visits may facilitate more informed discussions between parents and physicians and identify high-risk infants more effectively for early intervention.

Dataset Information

Developing a practical suicide risk prediction model for targeting high-risk patients in the Veterans health Administration.

Publications

Developing a practical suicide risk prediction model for targeting high-risk patients in the Veterans health Administration.

Similar Datasets

OmicsDI is part of the ELIXIR infrastructure

Tweets