Research
Current Projects
Mental Health Under Surveillance:
Neighborhood Voices NYC
Neighborhood Voices is a community-based study examining how surveillance technologies affect the daily lives and mental well-being of New Yorkers. Using surveys and data analysis, the project measures how residents perceive safety, privacy, and trust as monitoring tools expand citywide.
PTSD Risk Score Recalibration (VA Focus)
Early testing shows that existing VA PTSD screening tools miss a significant share of true cases while also over-flagging many false positives. Our study is recalibrating the risk score by incorporating behavioral, clinical, and contextual data to reduce error rates. The goal is a tool that identifies veterans at risk with greater accuracy and fairness.
Fentanyl “Lean” Pilot Intervention
Pilot data confirm that fentanyl-induced sedation episodes (“lean”) are not only dangerous but often go untreated until collapse. We are testing a targeted behavioral-health intervention that helps bystanders and frontline providers recognize episodes early and respond safely. The pilot will provide evidence on feasibility and potential for scaling.
Bias Detection in Immigration Adjudication
Preliminary analysis of adjudication records suggests that outcomes vary depending on applicant background and case officer workload, beyond the legal merits of the case. We are developing a behavioral-bias detection model to systematically flag these patterns, with the goal of making adjudication more transparent and equitable.
Completed Research
Housing as Suicide Prevention for U.S. Veterans
Evidence revealed that stable housing reduces suicide risk by providing both safety and continuity of care. Results underscore housing policy as a frontline suicide-prevention strategy, highlighting the urgent need to treat housing security as part of public health.
Bias and Fairness in PTSD Risk Prediction Models
Examination of emerging PTSD risk algorithms revealed consistent bias risks, particularly against diverse veteran groups. Findings support the adoption of bias-detection audits before deployment, ensuring predictive models are transparent, equitable, and reliable.
Loneliness, Cognitive Lapses, and Suicide Risk in Older Adults
Ongoing fieldwork in Brownsville, Brooklyn, is tracking daily patterns of loneliness, memory lapses, and mood among older adults living alone. Early observations show that cognitive lapses often precede spikes in suicidal thoughts. By combining digital tools with in-person check-ins, the project is testing practical ways to detect warning signs before crisis points.
Recalibrating PTSD Screening and Prediction
Analysis of current VA screening tools showed they frequently miss true PTSD cases while also producing unnecessary false positives. Findings demonstrated that integrating behavioral and contextual markers can sharply improve accuracy, ensuring more veterans are correctly identified and treated.
Intergenerational Impact of Parental Personality Disorders
The review found that daughters raised by parents with borderline or narcissistic personality disorders are at significantly higher risk for depression, anxiety, and long-term identity difficulties. Results point to the importance of targeted prevention and intervention programs for at-risk families.