ABSTRACT
Importance: The months after psychiatric hospital discharge are a time of high risk for suicide. Intensive postdischarge case management, although potentially effective in suicide prevention, is likely to be cost-effective only if targeted at high-risk patients. A previously developed machine learning (ML) model showed that postdischarge suicides can be predicted from electronic health records and geospatial data, but it is unknown if prediction could be improved by adding additional information. Objective: To determine whether model prediction could be improved by adding information extracted from clinical notes and public records. Design, Setting, and Participants: Models were trained to predict suicides in the 12 months after Veterans Health Administration (VHA) short-term (less than 365 days) psychiatric hospitalizations between the beginning of 2010 and September 1, 2012 (299â¯050 hospitalizations, with 916 hospitalizations followed within 12 months by suicides) and tested in the hospitalizations from September 2, 2012, to December 31, 2013 (149â¯738 hospitalizations, with 393 hospitalizations followed within 12 months by suicides). Validation focused on net benefit across a range of plausible decision thresholds. Predictor importance was assessed with Shapley additive explanations (SHAP) values. Data were analyzed from January to August 2022. Main Outcomes and Measures: Suicides were defined by the National Death Index. Base model predictors included VHA electronic health records and patient residential data. The expanded predictors came from natural language processing (NLP) of clinical notes and a social determinants of health (SDOH) public records database. Results: The model included 448â¯788 unique hospitalizations. Net benefit over risk horizons between 3 and 12 months was generally highest for the model that included both NLP and SDOH predictors (area under the receiver operating characteristic curve range, 0.747-0.780; area under the precision recall curve relative to the suicide rate range, 3.87-5.75). NLP and SDOH predictors also had the highest predictor class-level SHAP values (proportional SHAP = 64.0% and 49.3%, respectively), although the single highest positive variable-level SHAP value was for a count of medications classified by the US Food and Drug Administration as increasing suicide risk prescribed the year before hospitalization (proportional SHAP = 15.0%). Conclusions and Relevance: In this study, clinical notes and public records were found to improve ML model prediction of suicide after psychiatric hospitalization. The model had positive net benefit over 3-month to 12-month risk horizons for plausible decision thresholds. Although caution is needed in inferring causality based on predictor importance, several key predictors have potential intervention implications that should be investigated in future studies.
Subject(s)
Suicide Prevention , Suicide , Humans , Suicide/psychology , Patient Discharge , Inpatients , AftercareABSTRACT
The COVID-19 pandemic has heightened social isolation and loneliness. There is a lack of consensus on rating scales to measure these constructs. Our objectives were to identify commonly used loneliness and social isolation scales over the last two decades and test their user characteristics. 7928 articles were searched in PubMed/MEDLINE, CINAHL, Web of Science, and APA PsychINFO databases. 41 articles were included based on study criteria. Among fourteen scales reported, UCLA 3-item loneliness scale was found to be most commonly used. The scale is specifically developed for telephone use and is the fastest taking less than a minute for self-administration.
Subject(s)
COVID-19 , Pandemics , Humans , Loneliness , SARS-CoV-2 , Social IsolationABSTRACT
Recidivism, repeated criminal behavior after conviction and correction of prior offenses, is a costly problem across the nation. However, the contribution of empathy in determining the risk of recidivism has received limited attention, although lack of empathy has been related to antisocial personality disorder in various studies. Studies linked testosterone to aggression, antisocial behavior, and criminality, and evidence support hormonal connections between empathy and aggression. Adult male prison inmates convicted of violent or nonviolent offenses were included in a cross-sectional study of empathy, antisocial behavior, salivary testosterone, and recidivism. Subjects underwent criminal history, Empathy Quotient, Levenson Self-Report Psychopathy Scale, Beck Depression Inventory, Spielberger State-Trait Anxiety Inventory, and salivary testosterone assays. Bivariate analyses indicated multiple correlations between variables. Multivariate modeling analyses found a significant relationship between self-reported conviction number and psychopathy scale score (p = 0.013). These preliminary results suggest avenues of investigation of factors contributing to recidivism risk.
Subject(s)
Antisocial Personality Disorder/psychology , Empathy , Prisoners , Recidivism , Risk Assessment , Testosterone/analysis , Adult , Cross-Sectional Studies , Humans , Male , Multivariate Analysis , Saliva/chemistrySubject(s)
Pregnancy Complications/psychology , Substance-Related Disorders/complications , Adult , Amphetamine-Related Disorders/complications , Benzodiazepines/urine , Emergency Service, Hospital , Female , Humans , Legislation, Drug , Mandatory Reporting , Marijuana Abuse/complications , Methamphetamine/urine , Pregnancy , Pregnancy Complications/therapy , Public Policy , Substance-Related Disorders/therapy , United StatesABSTRACT
The study aimed to examine the course of obsessive-compulsive disorder (OCD) across pregnancy and its impact on obstetric and neonatal outcomes. Women enrolled prior to 20-week gestation in a prospective, observational study. The Structured Clinical Interview for DSM-IV was completed to obtain lifetime Axis I diagnoses. A total of 56 women with OCD were followed at 1 to 3-month intervals through 52 weeks postpartum. Each visit, the Yale-Brown Obsessive Compulsive Scale (YBOCS), clinical assessment, and medication/exposure tracking were performed. Obstetric and neonatal data were abstracted from the medical record. In subjects with OCD, associations between perinatal obsessive-compulsive symptoms (OCSs) and outcomes were examined. Additionally, outcomes were compared to 156 matched psychiatric patients without OCD. Maternal age inversely correlated with the YBOCS scores across the study period (ß = -0.5161, p = .0378). Cesarean section was associated with increased OCSs in the postpartum period compared to vaginal delivery (ß = 5.3632, p = 0.043). No associations were found between severity of perinatal obsessions or compulsions and any specific obstetric or neonatal complications. Subjects without OCD had higher frequency of fetal loss compared to mothers with OCD (χ (2) = 4.03, p = 0.043). These novel prospective data fail to identify an association of OCSs with adverse outcomes. In contrast, there is an association of delivery method and younger maternal age with increased postnatal symptoms of OCD. Psychiatric subjects without OCD may have a higher risk of miscarriage and intrauterine fetal demise compared to subjects with OCD.
Subject(s)
Obsessive-Compulsive Disorder/diagnosis , Obsessive-Compulsive Disorder/psychology , Postpartum Period/psychology , Pregnancy Complications/diagnosis , Pregnancy Complications/psychology , Adult , Diagnostic and Statistical Manual of Mental Disorders , Female , Humans , Obsessive-Compulsive Disorder/epidemiology , Parturition , Pregnancy , Pregnancy Complications/epidemiology , Pregnancy Outcome , Pregnancy Trimester, Second , Prospective Studies , Psychiatric Status Rating Scales , Risk Factors , Severity of Illness Index , Socioeconomic Factors , United States/epidemiologyABSTRACT
Helping elderly people to live independently within their homes for as long as possible, before transitioning to higher levels of care, can significantly reduce healthcare expenditures. However, achieving this vision requires continuous monitoring of the condition of elderly adults within their homes. In particular, activity, gait velocity, movement, and location of elderly adults are critical biomarkers for healthy aging. We present a prototype integrating a wearable location-tracking sensor with back-end cloud-based data processing, thereby enabling real-time tracking and analysis of a large number of people simultaneously. The resulting vertically-integrated prototype provides a basic infrastructure for future work, including new products and services that offer real-time monitoring and early disease diagnosis to help elderly people live independently for as long as possible.
Subject(s)
Activities of Daily Living , Freedom , Aged , Humans , InternetABSTRACT
We describe a low-cost wearable system that tracks the location of individuals indoors using commonly available inertial navigation sensors fused with radio frequency identification (RFID) tags placed around the smart environment. While conventional pedestrian dead reckoning (PDR) calculated with an inertial measurement unit (IMU) is susceptible to sensor drift inaccuracies, the proposed wearable prototype fuses the drift-sensitive IMU with a RFID tag reader. Passive RFID tags placed throughout the smart-building then act as fiducial markers that update the physical locations of each user, thereby correcting positional errors and sensor inaccuracy. Experimental measurements taken for a 55 m × 20 m 2D floor space indicate an over 1200% improvement in average error rate of the proposed RFID-fused system over dead reckoning alone.