Integrating topic modeling and word embedding to characterize violent deaths

Content-Aware Human Mobility Pattern Extraction

Extracting meaningful patterns of human mobility from accumulating trajectories is essential for understanding human behavior. However, previous works identify human mobility patterns based on the spatial co-occurrence of trajectories, which ignores the effect of activity content, leaving challenges in effectively extracting and understanding patterns. To bridge this gap, this study incorporates the activity content of trajectories to extract human mobility patterns, and proposes acontent-aware mobility pattern model. The model first embeds the activity content in distributed continuous vector space by taking point-of-interest as an agent and then extracts representative and interpretable mobility patterns from human trajectory sets using a derived topic model. To investigate the performance of the proposed model, several evaluation metrics are developed, including pattern coherence, pattern similarity, and manual scoring. A real-world case study is conducted, and its experimental results show that the proposed model improves interpretability and helps to understand mobility patterns. This study provides not only a novel solution and several evaluation metrics for human mobility patterns but also a method reference for fusing content semantics of human activities for trajectory analysis and mining.

Artificial Intelligence and the National Violent Death Reporting System: A Rapid Review

As the awareness on violent deaths from guns, drugs, and suicides emerges as a public health crisis in the United States, attempts to prevent injury and mortality through nursing research are critical. The National Violent Death Reporting System provides public health surveillance of US violent deaths; however, understanding the National Violent Death Reporting System's research utility is limited. The purpose of our rapid review of the 2019-2023 literature was to understand to what extent artificial intelligence methods are being used with the National Violent Death Reporting System. We identified 16 National Violent Death Reporting System artificial intelligence studies, with more than half published after 2020. The text-rich content of National Violent Death Reporting System enabled researchers to center their artificial intelligence approaches mostly on natural language processing (50%) or natural language processing and machine learning (37%). Significant heterogeneity in approaches, techniques, and processes was noted across the studies, with critical methods information often lacking. The aims and focus of National Violent Death Reporting System studies were homogeneous and mostly examined suicide among nurses and older adults. Our findings suggested that artificial intelligence is a promising approach to the National Violent Death Reporting System data with significant untapped potential in its use. Artificial intelligence may prove to be a powerful tool enabling nursing scholars and practitioners to reduce the number of preventable, violent deaths.

Limitations of Sexual Orientation and Gender Identity Information as Reported in the National Violent Death Reporting System

The National Violent Death Reporting System (NVDRS) is a Centers for Disease Control and Prevention (CDC) restricted-access database detailing precipitating circumstances to U.S. violent deaths. In 2013 and 2015, the CDC added codes denoting sexual orientation and gender identity (SOGI) and sex of partner. In the past decade, researchers have leveraged NVDRS data to document SOGI-related patterns and characteristics of violent death including suicide. Yet, there are substantial limitations to NVDRS SOGI information that should be considered in responsible reporting by researchers and informed assessment by reviewers. In this perspective, we summarize some of these challenges and offer recommendations for using NVDRS SOGI data responsibly.

Gendered Patterns in Manifest and Latent Mental Health Indicators Among Suicide Decedents: 2003-2020 National Violent Death Reporting System (NVDRS)

Objectives. To investigate differences in the documentation of mental health symptomology between male and female suicide decedents in the 2003-2020 US National Violent Death Reporting System (NVDRS). Methods. Using information on 271 998 suicides in the 2003-2020 NVDRS, we evaluated precoded mental health-related variables and topic model-derived latent mental health themes in the law enforcement and coroner or medical examiner death narratives compiled by trained public health workers. Results. Public health records of male compared with female suicides were less likely to include notations of mental health conditions or treatment interventions. However, topic modeling of death summaries revealed that male suicide decedents were more likely to evidence several subclinical cognitive and emotional indicators of distress. Conclusions. Suicide death records vary by gender, both in recorded evidence for mental health conditions at time of death and in accompanying narratives describing proximal circumstances surrounding these deaths. Our findings hint that patterns of subclinical mental health changes among men might be less well captured in commonly used mental health indicators, suggesting that prevention efforts may benefit from measures that also target assessment of subclinical distress. (Am J Public Health. 2024;114(S3):S268-S277. https://doi.org/10.2105/AJPH.2023.307427).

Identifying Witnessed Suicides in National Violent Death Reporting System Narratives

There is increasing attention to suicides that occur in view of others, as these deaths can cause significant psychological impact on witnesses. This study illuminates characteristics of witnessed suicides and compares characteristics of these deaths to non-witnessed suicides. We develop a codable definition of what constitutes witnessed (vs. non-witnessed) suicide. Our data include a sample of 1200 suicide descriptions from the 2003-2017 National Violent Death Reporting System (NVDRS). We first developed criteria to identify probable cases of witnessed suicide. The coding scheme achieved 94.5% agreement and identified approximately 10% (n = 125) of suicides as witnessed. Next, we examined differences between witnessed and non-witnessed suicides in demographics, manner of death, and social/environmental factors using bivariate Chi-squared tests, multivariate logistic regression, and ANOVA. Witnessed suicide decedents were significantly more likely than non-witnessed suicide decedents to be male, younger, and members of a sexual minority, and to have died in living spaces by means of a firearm. Two thirds of witnesses were strangers to the decedents, while 23.2% were romantic partners or ex-partners of the decedents. Our coding method offers a reliable approach to identify witnessed suicides. While witnessed suicides are relatively infrequent, these deaths have profound impact on witnesses. Articulating the features of witnessed suicides may contribute to identifying potential risk mitigation strategies.

Policy instruments facilitate China's COVID-19 work resumption

Governments worldwide have announced stimulus packages to remobilize the labor force after COVID-19 and therefore to cope with the COVID-19-related recession. However, it is still unclear how to facilitate large-scale work resumption. This paper aims to clarify the issue by analyzing the large-scale prefecture-level dataset of human mobility trajectory information for 320 million workers and about 500,000 policy documents in China. We model work resumption as a collective behavioral change due to configurations of capacity, motivation, and policy instruments by using qualitative comparative analysis. We find that the effectiveness of post-COVID-19 recovery stimulus varied across China depending on the fiscal and administrative capacity and the policy motivation of the prefecture. Subnational fiscal and procurement policies were more effective for the wholesale and retail sector and the hotel and catering sector, whereas the manufacturing and business services sectors required more effort regarding employment policies. Due to limited prefectural capacity and wavering policy motivation, the simultaneous adoption of fiscal, employment, and procurement policy interventions endangered post-COVID-19 work resumption. We highlight the necessity of tailored postcrisis recovery strategies based on local fiscal and administrative capacity and the sectoral structure.

Detecting intimate partner violence circumstance for suicide: development and validation of a tool using natural language processing and supervised machine learning in the National Violent Death Reporting System

BACKGROUND

Intimate partner violence (IPV) victims and perpetrators often report suicidal ideation, yet there is no comprehensive national dataset that allows for an assessment of the connection between IPV and suicide. The National Violent Death Reporting System (NVDRS) captures IPV circumstances for homicide-suicides (<2% of suicides), but not single suicides (suicide unconnected to other violent deaths; >98% of suicides).

OBJECTIVE

To facilitate a more comprehensive understanding of the co-occurrence of IPV and suicide, we developed and validated a tool that detects mentions of IPV circumstances (yes/no) for single suicides in NVDRS death narratives.

METHODS

We used 10 000 hand-labelled single suicide cases from NVDRS (2010-2018) to train (n=8500) and validate (n=1500) a classification model using supervised machine learning. We used natural language processing to extract relevant information from the death narratives within a concept normalisation framework. We tested numerous models and present performance metrics for the best approach.

RESULTS

Our final model had robust sensitivity (0.70), specificity (0.98), precision (0.72) and kappa values (0.69). False positives mostly described other family violence. False negatives used vague and heterogeneous language to describe IPV, and often included abusive suicide threats.

IMPLICATIONS

It is possible to detect IPV circumstances among singles suicides in NVDRS, although vague language in death narratives limited our tool's sensitivity. More attention to the role of IPV in suicide is merited both during the initial death investigation processes and subsequent NVDRS reporting. This tool can support future research to inform targeted prevention.

Examining Intimate Partner Violence-Related Fatalities: Past Lessons and Future Directions Using U.S. National Data

Purpose

Among homicides in the United States, intimate partners kill almost 50% of female and 10% of male victims. Intimate partner violence (IPV) also contributes to an estimated 6% of suicides. These trends suggest that opportunities for IPV interventions prior to the fatalities may have been missed. Thus, researchers must investigate the context and circumstances of IPV-related fatalities to inform effective prevention strategy development. There are two primary national fatality databases that can be used to examine such factors: the National Violent Death Reporting System (NVDRS, homicide and suicides); and the Uniform Crime Reporting-Supplementary Homicide Reports (UCR-SHR, homicides). These datasets include data on many IPV-related violent deaths but are limited by variations in data quality.

Method

This critical review summarizes opportunities and challenges when examining IPV-related fatalities using these national datasets. To document how the current literature is conceptualizing IPV, a rapid review on IPV-related homicide and suicide articles was performed (2019-2022). Missingness analyses were conducted to describe limitations in key dataset variables.

Results

These datasets enable tracking IPV-related fatalities nationally over time. However, issues with the operationalization of variables that record IPV circumstances, particularly in the UCR-SHR, and high levels of missingness represent significant barriers to research. Novel methodologies can optimize the use of these datasets.

Conclusion

National-level datasets enable researchers to examine IPV-related fatalities, evaluate policy differences between states, and monitor trends and disparities. This research can inform key recommendations for interventions to prevent IPV-related fatalities.

Supplementary Information

The online version contains supplementary material available at 10.1007/s10896-022-00487-2.