Measuring interpersonal firearm violence: natural language processing methods to address limitations in criminal charge data

OBJECTIVE

Firearm violence constitutes a public health crisis in the United States, but comprehensive data infrastructure is lacking to study this problem. To address this challenge, we used natural language processing (NLP) to classify court record documents from alleged violent crimes as firearm-related or non-firearm-related.

MATERIALS AND METHODS

We accessed and digitized court records from the state of Washington (n = 1472). Human review established a gold standard label for firearm involvement (yes/no). We developed a key term search and trained supervised machine learning classifiers for this labeling task. Results were evaluated in a held-out test set.

RESULTS

The decision tree performed best (F1 score: 0.82). The key term list had perfect recall (1.0) and a modest F1 score (0.65).

DISCUSSION AND CONCLUSION

This case report highlights the accuracy, feasibility, and potential time-saved by using NLP to identify firearm involvement in alleged violent crimes based on digitized narratives from court documents.

Household food insecurity and associations with hemoglobin A1c and acute diabetes-related complications in youth and young adults with type 1 diabetes: The SEARCH for diabetes in youth study

AIMS

To examine, among youth and young adults (YYA) with type 1 diabetes (T1D), the association of household food insecurity (HFI) with: 1) HbA1c and 2) episodes of diabetic ketoacidosis (DKA) and severe hypoglycemia.

METHODS

HFI was assessed using the U.S. Household Food Security Survey Module in SEARCH for Diabetes in Youth participants with T1D between 2016 and 2019. Linear and logistic regression models adjusted for age, diabetes duration, sex, race, ethnicity, clinic site, parent/participant education, household income, health insurance, and diabetes technology use.

RESULTS

Of 1830 participants (mean age 20.8 ± 5.0 years, 70.0 % non-Hispanic White), HbA1c was collected for 1060 individuals (mean HbA1c 9.2 % ± 2.0 %). The prevalence of HFI was 16.4 %. In the past 12 months, 18.2 % and 9.9 % reported an episode of DKA or severe hypoglycemia, respectively. Compared to participants who were food secure, HFI was associated with a 0.33 % (95 % CI 0.003, 0.657) higher HbA1c level. Those with HFI had 1.58 (95 % CI 1.13, 2.21) times the adjusted odds of an episode of DKA and 1.53 (95 % CI 0.99, 2.37) times the adjusted odds of an episode of severe hypoglycemia as those without HFI.

CONCLUSIONS

HFI is associated with higher HbA1c levels and increased odds of DKA in YYA with T1D.

Race, structural racism and racial disparities in firearm homicide victimisation

OBJECTIVES

To identify an approach in measuring the association between structural racism and racial disparities in firearm homicide victimisation focusing on racism, rather than race.

METHODS

We examined associations of six measures of structural racism (Black/white disparity ratios in poverty, education, labour force participation, rental housing, single-parent households and index crime arrests) with state-level Black-white disparities in US age-adjusted firearm homicide victimisation rates 2010-2019. We regressed firearm homicide victimisation disparities on four specifications of independent variables: (1) absolute measure only; (2) absolute measure and per cent Black; (3) absolute measure and Black-white disparity ratio and (4) absolute measure, per cent Black and disparity ratio.

RESULTS

For all six measures of structural racism the optimal specification included the absolute measure and Black-white disparity ratio and did not include per cent Black. Coefficients for the Black-white disparity were statistically significant, while per cent Black was not.

CONCLUSIONS

In the presence of structural racism measures, the inclusion of per cent Black did not contribute to the explanation of firearm homicide disparities in this study. Findings provide empiric evidence for the preferred use of structural racism measures instead of race.

Rural-Urban Variation in the Association of Adolescent Violence and Handgun Carrying in the United States, 2002-2019

Importance

Adolescent handgun carrying is associated with increased risk of firearm-related violence. Most evidence on adolescent handgun carrying is from urban areas, but these findings may not generalize to rural areas.

Objective

To examine differences in associations of adolescent interpersonal violence with handgun carrying across the rural-urban continuum.

Design, Setting, and Participants

This cross-sectional study used nationally representative data from the US National Survey on Drug Use and Health among adolescents aged 12 to 17 years from 2002 to 2019 to estimate time-varying prevalence ratios (PRs) and prevalence differences (PDs) between interpersonal violence and handgun carrying across the rural-urban continuum. Analyses were conducted in April to July 2022.

Exposures

Any past-year serious fighting, group fighting, and attacking with intent to harm.

Main Outcomes and Measures

Any past-year handgun carrying. Associations were estimated within county rural-urban strata using the US Department of Agriculture's Rural-Urban Continuum Codes.

Results

In each year, the sample included a weighted count of almost 25 million adolescents, with 50.9% (95% CI, 50.2%-51.6%) males and 24.7% (95% CI, 23.8%-25.6%) Hispanic adolescents, 13.5% (95% CI, 12.8%-14.2%) non-Hispanic Black adolescents, and 51.8% (95% CI, 50.8%-52.8%) non-Hispanic White adolescents in 2019. More rural counties had less racial and ethnic diversity. For example, 81.1% (95% CI, 75.9%-85.4%) of adolescents were non-Hispanic White in the most rural counties vs 43.1% (95% CI, 41.7%-44.6%) of adolescents were non-Hispanic White in the most urban counties in 2019. Adolescent handgun carrying increased over time, with the largest increases in the most rural counties, where the prevalence of adolescent handgun carrying increased from 5.2% (95% CI, 3.8%-7.0%) in 2003 to 12.4% (95% CI, 8.9%-16.9%) in 2019. PRs for the association of violence and handgun carrying were greater in more urban counties. For example, in the most urban counties in 2019, adolescents involved in a group fight had 3.7 (95% CI, 2.9-4.8) times the prevalence of handgun carrying vs those not involved in a group fight; this PR was 3.1 (95% CI, 1.6-5.6) in the most rural counties. PDs were similar and, in some cases, larger in rural areas. For example, in the most urban counties in 2019, handgun carrying prevalence was 7.5% (95% CI, 5.7%-9.5%) higher among adolescents who were involved in a group fight compared with those who were not; this PD was 21.8% (95% CI, 8.2%-37.8%) in the most rural counties, where handgun carrying was more common.

Conclusions and Relevance

This cross-sectional study found that associations of interpersonal violence with handgun carrying were stronger in relative terms in urban areas than in rural areas; however, a higher percentage of rural than urban adolescents carried handguns, resulting in a greater absolute prevalence of handgun carrying associated with violence in rural areas than in urban areas. These findings suggest opportunities for preventing handgun carrying-related harms may differ between rural and urban communities.

Firearm purchasing and storage during the COVID-19 pandemic

To better understand motivations behind purchase and storage of firearms during the COVID-19 pandemic, we used Amazon Mechanical Turk to conduct an online survey of individuals who did and did not purchase a firearm since 1 January 2020 in response to COVID-19. The survey was fielded between 1 and 5 May 2020. We asked about motivations for purchase, changes in storage practices and concern for themselves or others due to COVID-19. There were 1105 survey respondents. Most people who purchased a firearm did so to protect themselves from people. Among respondents who had purchased a firearm in response to COVID-19 without prior household firearm ownership, 39.7% reported at least one firearm was stored unlocked. Public health efforts to improve firearm-related safety during COVID-19 should consider increasing access to training and framing messages around the concerns motivating new firearm purchase.