Community Firearm Violence Prevention Network (CFVP)

Community Firearm Violence Prevention Network (CFVP) is sponsored by National Institutes of Health (NIH). This research network is designed to develop, implement, and evaluate interventions preventing firearm injury.

## [](https://pmc. ncbi. nlm.

nih. gov/articles/PMC13051374/)Abstract The Community Firearm Violence Prevention Network (CFVP Network), funded by the National Institutes of Health (NIH), supports a network of research projects that develop and test interventions through collaborations with community partners to prevent firearm violence, injury, and mortality. The CFVP Network presents a unique opportunity to accelerate the science of preventing firearm injuries.

The data harmonization workgroup of the CFVP Network led the process of aligning studies across the three unique inaugural network projects, with particular attention to how the CFVP Network could address current gaps in the science.

The goal of the data harmonization workgroup was to align study measures, assessment timelines, and data management and archival processes across projects to enable robust cross-project analyses that accelerate the science of preventing firearm injuries. To accomplish this goal, the workgroup established the infrastructure to facilitate cross-project data collection, data sharing and archiving, and analyses.

Among the three inaugural network projects, the workgroup’s process resulted in harmonizing two assessment timepoints (baseline and one year post-implementation) and 60 constructs (with 31 identical standardized constructs). These harmonized products provide opportunities for novel analyses across the network projects.

We expect that the harmonized study infrastructure developed through this process will catalyze future research focused on preventing firearm injury, including and extending beyond CFVP Network projects. The CFVP data harmonization workgroup’s process can serve as a model for future networks that seek to build the science in a particular area. ## [](https://pmc.

ncbi. nlm. nih.

gov/articles/PMC13051374/)Why Harmonize Network Projects? Even the most rigorous standalone study is limited in its ability to advance the field. Findings from an individual study might not generalize to other contexts, key constructs may not be measured, or sample sizes may not be large enough to detect effects (in either the full sample or important sub-groups).

While meta-analyses may use retrospective harmonization to evaluate results from similar studies, they are often constrained by critical differences, such as incompatible measures, differing assessment timelines, and the inability to access original data.

Prospective harmonization, or the process of aligning studies with a similar focus during the initial design phase, has the potential to overcome these challenges and promote scientific advancements to enhance community health and safety. Harmonizing assessment timelines, measures, and data management plans across studies with distinct populations, interventions, and geographies provides opportunities for robust cross-project analyses.

These cross-project analyses benefit from having (1) higher power to detect effects; (2) uniform or comparable measures across studies; (3) opportunities to evaluate for heterogeneous effects by population, intervention, and/or geography; and (4) the ability to test key risk and protective factors that may moderate outcomes or serve as key mechanisms of change to focus future interventions.

Harmonizing studies is part of a larger movement toward scientific collaboration, team science, and open science (Ducharme et al. , 2021; Ridenour et al. , 2023; Volkow et al.

, 2018). One approach for conducting cross-project analyses is integrative data analysis (IDA). Researchers have recognized that harmonizing constructs enables data pooling from multiple studies to conduct IDA (Hussong et al.

, 2013). While IDA bears similarities to other data pooling methods (e.g., meta-regression (Baker et al. , 2009)), this approach emphasizes psychometric analyses to harmonize measurements across participants to enable pooled analyses.

The goal is to create commensurate scores at the factor- (i.e., construct) rather than item-level. Adopting this method provides numerous benefits. First, IDA can provide greater coverage across developmental ages.

For instance, researchers can pool individual-level data from longitudinal studies that sample different age ranges and use IDA to gain a broader perspective of a phenomenon across human development. Second, IDA can increase statistical power and create larger counts of rare behaviors (e.g., youth firearm aggression).

Third, IDA allows for evaluating result replication across studies: researchers can evaluate whether the association between two constructs differs in magnitude between studies and examine potential factors (e.g., sampling) contributing to discrepant results. Curran and colleagues offer an illustration of data harmonization and IDA (Curran et al. , 2018).

## [](https://pmc. ncbi. nlm.

nih. gov/articles/PMC13051374/)Establishing the Data Harmonization Workgroup ### CFVP Data Harmonization Workgroup The CFVP data harmonization workgroup provides a structure for the coordinating center and network projects to align and harmonize data (Fig. 1).

The workgroup includes individual network project representatives nominated by the site principal investigator, members of the coordinating center’s data and methods core, and a NIH project scientist. The CFVP Network coordinating center tasked the data harmonization workgroup with harmonizing study timelines, measures, and data management plans across network projects.

The inaugural UG3/UH3 network projects are located at the University of Mississippi Medical Center (UMMC) in Jackson, Mississippi ; the University of Chicago (UC) in Chicago, Illinois; and George Washington University (GW) in the District of Columbia. Each network project principal investigator elected one to three team members to join the data harmonization workgroup to support study alignment across the CFVP Network.

The UG3/UH3 funding mechanism involves (1) an initial UG3 phase of a milestone-driven developmental study that demonstrates sufficient preparation, feasibility, and capacity for implementing an intervention, and (2) a UH3 phase that involves implementing and evaluating the interventions or strategies planned or developed in the UG3 phase (National Institutes of Health, 2022).

The NIH funded these research projects as cooperative agreements, such that NIH scientific and program staff assist, guide, coordinate, or participate in project activities. Network projects will pilot the harmonized measures as part of their UG3 phases, with any changes for the UH3 phase approved by the NIH through consultation with the coordinating center.

### University of Mississippi Medical Center The “Mississippi Violence Injury Prevention (MS VIP) Project” will evaluate the impact of multiple community-based firearm injury prevention programs on firearm crimes, injuries, and mortality, and their individual, neighborhood, and city-wide economic effects.

The team involves researchers, medical professionals, and leaders from community-based organizations engaged in violence prevention work in Jackson, Mississippi. These include the People’s Advocacy Institute, Mississippi Public Health Institute, Strong Arms of Jackson, and Operation Good. A two-year community-engaged process has facilitated the selection and design of the interventions.

The MS VIP team will implement the interventions in a three-year stepped wedge cluster randomized trial. Based on the cluster randomization and year, patients who live in the Jackson metropolitan area and are treated in the UMMC Emergency Department for a firearm injury will be offered participation.

The interventions include community-based violence interruption, cash assistance, and post-traumatic stress disorder treatment, as well as longitudinal follow-up of health and social determinants of health. The proposed units of randomization are empirically derived geographic clusters identified using the geocoded addresses of gun violence injury victims treated at UMMC from 2013 to 2021 and k-means cluster analysis.

### University of Chicago The “Harmonizing Hospital-Based Violence Intervention Programs with a Novel Medical-Legal Partnership (MLP)” study evaluates a bedside legal assistance intervention for addressing social and economic root causes of violence. The study team is implementing the program, known as Recovery Legal Care, among patients hospitalized for violent injury at the University of Chicago Medicine Trauma Center.

The trauma center is located in a dense urban area on the south side of Chicago with high economic hardship (University of Chicago Medicine, 2022). The program screens and assesses patients for health-harming legal needs (HHLNs) at the bedside and refers them to an on-site legal team from Legal Aid Chicago, a provider of free civil legal services in Cook County, Illinois.

The legal team can address a broad number of HHLNs, including access to public benefits (e.g., Supplemental Nutrition Assistance Program, Temporary Assistance for Needy Families), consumer debt, criminal record expungement, and housing and employment insecurity. The research team is working closely with Legal Aid Chicago and an established Hospital-Based Violence Intervention Program (HVIP) to evaluate this novel HVIP-MLP model.

The effectiveness of this novel model will be evaluated using an individually randomized group treatment trial to compare outcomes among participants (ages 14–64 years) receiving legal care (HVIP-MLP) and those receiving usual care (HVIP only). Currently, legal care has limited capacity which enables an experimental study design while still providing services to the maximum number of patients possible.

Key outcome measures are violence victimization and aggression, including firearm behaviors, post-traumatic stress disorder symptoms, perceived stress, and future expectations. Participant data will also be matched to administrative databases from the Chicago Police Department and Chicago Public Schools to assess for differences in police involvement and school engagement.

### George Washington University ### Data and Methods Core of the Coordinating Center The Institute for Firearm Injury Prevention at the University of Michigan leads the CFVP Network through the coordinating center.

The coordinating center includes multiple cores to guide the CFVP Network, including the data and methods core which provides expert consultation and technical assistance to network projects on design, methods, measures, and analyses. The core also facilitates cross-project data collection, management, harmonization, linkage, analysis, and archiving activities.

The data and methods core includes researchers with expertise in content areas including violence and firearm injury prevention, substance use, mental health, and associated risk/protective factors, and methodological expertise in psychometrics, epidemiology, administrative and spatial data collection and analyses, and data archiving.

Unifying terminologies and definitions for harmonizing study timelines, measures, and data management plans guided the data harmonization workgroup. We defined _harmonization_ (adapted from Ridenour and colleagues (Ridenour et al. , 2023)) as the process of identifying or developing common measured constructs, assessment timelines, and data management plans across studies.

The products of these harmonization processes were harmonized constructs, harmonized assessment timelines, and harmonized data management plans. The data harmonization workgroup parsed measures into measurement _domains_, _constructs_, and _items_ (see Fig. 2 for illustration).

Measurement _domains_ are broad categories, such as mental health. These domains contain one or more measurement _constructs_ that analysts operationalize as variables, such as depression or anxiety symptoms. _Measurement items_ are the specific scales, questions, and prompts that study participants respond to that inform the quantification of measurement constructs, such as the Patient Health Questionnaire (PHQ) (Kroenke et al.

, 2001) or Generalized Anxiety Disorder questionnaire(GAD) (Spitzer et al. , 2006). Each measurement construct contains one or more measurement items.

The harmonization process occurred at the _measurement construct_ level. Harmonized products do not need to be identical products. Rather, harmonized products should sufficiently overlap to provide an avenue for advanced statistical methods to ask novel research questions across projects with the potential for informing policy and practice in diverse settings (Ridenour et al.

, 2023). For example, network projects need not have identical assessment timelines, but the workgroup sought two uniform assessment points (i.e., at baseline plus one additional point) across all project sites to facilitate cross-project analyses. Similarly, harmonized constructs do not need to be identical or include across all network projects.

Rather, harmonized constructs are constructs adopted by two or more network projects; furthermore, harmonized constructs may vary between network projects. For example, some response options for harmonized constructs varied between network projects to reflect local contexts.

For other constructs, some network projects selected measurement items from a larger scale (e.g., Patient Health Questionnaire-2), while other sites included the entire scale (e.g., Patient Health Questionnaire-8). Some harmonized measures were identical across all network projects. The data harmonization workgroup defined these special cases of completely harmonized constructs as _standardized_ constructs (Ridenour et al.

, 2023). The data harmonization workgroup iteratively organized constructs into a tiered structure during the measurement harmonization process. At the conclusion of this process, harmonized construct tiers ranged from 1 to 3 (Table 1).

Tier 1 constructs were standardized, with collective agreement that all network projects should include the associated items in their entirety. Tier 2 constructs were harmonized, whereby all network projects agreed to include associated items, with allowable variations in response options to accommodate the local context and/or allowable exclusion of item subsets.

Tier 3 constructs were strongly recommended, with at least two of the three inaugural network projects including the construct. ### Harmonizing Assessment Timeline One of the first activities of the data harmonization workgroup was to harmonize study assessment timelines across network projects, as these influence wording of measurement items and opportunities for cross-project analyses.

For example, if one or more project sites conduct assessments every three months, measurement items with a recall period of six months would be inappropriate for measuring change since the previous assessment. Following discussions within the data harmonization workgroup, all three inaugural projects elected to conduct assessments at baseline and 12 months post-intervention implementation.

Two inaugural network projects (UMMC and UC) elected to include additional assessments at two–three and six months post-intervention implementation. Harmonizing measures across network projects was an iterative process.

The data harmonization workgroup (1) collected proposed and identified additional constructs whose inclusion could facilitate novel contributions to preventing firearm violence; (2) prioritized constructs and identified opportunities for harmonization; (3) identified, adapted, and/or created candidate measures to align with network projects and fill existing gaps in the science; and (4) assigned construct tier levels.

In each phase, the data and methods core defined constructs so that the workgroup—with members from various disciplines—had a shared understanding of the considered measures and could communicate across disciplines (National Academies of Sciences, Engineering & Medicine, 2022). To harmonize measures, the workgroup met ten times from January to June 2023, representing twenty hours of working meetings.

During this time, the data and methods core held an additional twenty-four internal meetings (representing forty-seven hours of working meetings) to discuss prioritized constructs and potential measures to present to network projects. The data and methods core dedicated additional time between meetings reviewing the literature and compiling measures for the workgroup’s review.

### Phase 1: Collecting Proposed and Identifying Additional Constructs The first step in harmonizing measures involved network projects providing the data and methods core with the constructs and items originally proposed in their NIH grant applications. The data and methods core reviewed project sites’ proposed measures, identifying areas of overlap between projects.

Table 2 lists the measures that network projects proposed before the measure harmonization process; the only constructs common across the inaugural network projects were demographic items.

### Phase 2: Prioritizing Constructs and Identifying Opportunities for Harmonization The data and methods core grouped all potential constructs (both from the network projects and those identified by the core) into eight domains: (1) demographics (e.g., age, gender, race, and ethnicity); (2) social risk factors (e.g., food insecurity, discrimination including racism); (3) mental health (e.g., stress, depression); (4) substance use (e.g., illicit drug use, misuse of prescription drugs); (5) physical health (e.g., general health, pain intensity); (6) non-firearm violence (e.g., victimization, community violence exposure); (7) firearms and firearm violence (e.g., firearm carriage, firearm violence exposure); and (8) protective factors (e.g., social cohesion, relational social capital).

The data and methods core presented these domains to network projects during data harmonization workgroup meetings to identify opportunities for harmonizing or standardizing measures across studies, typically focusing on constructs within a single domain each meeting.

The data and methods core provided an overview of a specific domain and its corresponding constructs during workgroup meetings for network project representatives to discuss. This presentation and collaborative approach allowed the workgroup to prioritize specific constructs within a domain for standardization or harmonization.

### Phase 3: Identifying, Adapting, and Creating Candidate Measures After identifying constructs that network projects were interested in harmonizing or standardizing, the data and methods core identified several measurement options per construct. When identifying candidate measurement options, the data and methods core prioritized measures used in national samples (including measures from the PhenX toolkit (Hamilton et al.

, 2011)) and consulted peer-reviewed literature to assess measures’ psychometric characteristics, opportunities for adaptations, and alternative options. The data and methods core would then present one to three measurement options per construct during workgroup meetings where network projects would discuss and determine which measurement options they were interested in standardizing or harmonizing.

If the measurement options presented in the workgroup meeting did not meet the network projects’ needs, the data and methods core identified additional measures to present at the following workgroup meeting. In some instances, the data and methods core adapted existing scales to suit network projects’ needs (e.g., by making wording changes to enhance clarity or improve cultural or developmental relevance).

This process stemmed from network projects suggesting changes to accommodate their focal population and context. In one example, the workgroup made a simple modification to the Friends’ Delinquent Behavior – Peer Deviancy Scale (Thornberry et al. , 1994) to account for digital friendships that project sites believed may be important to respondents—particularly adolescents.

Specifically, the workgroup revised the prompt to ask participants how many friends they “_interact with_” instead of “_see_. ” The data and methods core maintained a parallel measures codebook that documented changes the workgroup made to existing measures.

### Phase 4: Assigning Construct Tier Level Once network projects determined whether and how to harmonize a specific construct, the workgroup assigned the construct a tier-level, 1–3 (see Table 1 for tier descriptions). Table 2 notes which constructs network projects planned to assess before and after the harmonization process.

Overall, the data harmonization workgroup harmonized sixty constructs, including thirty-one tier 1, sixteen tier 2, and thirteen tier 3 constructs, including newly created measures for three constructs. Importantly, the sixteen tier 2 constructs provide unique IDA opportunities.

All network projects agreed to include these constructs, but the item wording for tier 2 constructs may vary between projects, and/or some project sites are asking item subsets for some tier 2 constructs. IDA can help us better understand how variations in items do, or do not, map on to the overall construct they are seeking to measure.

For instance, even if a subset of items is not shared across all projects, researchers can treat these items as missing data and handle them through maximum likelihood estimation.

Furthermore, by implementing tests of measurement invariance and differential item functioning (e.g., moderated non-linear factor analysis), researchers can ascertain whether the hypothesized factor being measured is in fact consistent across the network projects. Thus, researchers will be able to estimate factor scores across multiple studies based on a common metric with the sixteen tier 2 constructs.

### Harmonizing Data Management Plans The data and methods core created a principal codebook that included all measurement items for tier 1–3 constructs and individual network project codebooks that included the harmonized measures relevant to their specific project.

In an iterative process, the data and methods core shared the principal codebook with network projects, who provided feedback and questions, resulting in a refined baseline codebook for the inaugural network projects and follow-up codebooks for three months and six months (UC, UMMC), as well as twelve months (GW, UC, UMMC).

The data and methods core provided network projects with details regarding item response option coding, how to calculate summary scores in cases where multiple items inform values for a single construct, and adaptations to existing measures and items, where applicable.

The CFVP Network is collaborating with the Interuniversity Consortium for Political and Social Research (ICPSR) at the University of Michigan to house the network project data and codebooks to allow for future analyses of project data. The data and methods core is providing ongoing technical assistance to network projects regarding codebooks and will support data upload to ICPSR when projects reach that stage.

Integral to this process are ongoing conversations with ICPSR to understand the technical requirements and possibilities of the repository. Each network project will provide their data to ICPSR at the conclusion of their studies.

### Integrating Future Studies into the Data Harmonization Process The infrastructure created from the work between the inaugural three network projects and the data and methods core will support future projects funded by the NIH as members of the CFVP Network. As the coordinating center onboards a new cohort of network projects, the data and methods core will reform a data harmonization workgroup with new project representatives.

The core will provide projects with the inaugural cohort’s principal codebook as a starting point, as adopting salient constructs and corresponding measures may provide opportunities for cross-project analyses that explore whether findings generalize not only across space, but also across time.

Similar to the process for the first three network projects, the data and methods core will review project proposals and preliminary measures to identify areas of overlap and potential for harmonization among new cohorts. When these harmonization opportunities fall within an existing tier 1–3 construct, the workgroup will consider if the existing measurements meet project needs.

We also anticipate that new opportunities for data harmonization may emerge, at which point the data harmonization workgroup will resume the process of identifying and modifying measures to harmonize across new network projects, adding new items to an updated principal codebook.

### Lessons Learned and Recommendations for Harmonization Processes #### Invest Time and Resources Harmonizing studies takes time, resources, and planning to bring all involved parties together early and often in the project planning process.

As previously noted, the process of harmonizing the inaugural three CFVP Network Projects included twenty hours of data harmonization workgroup meetings, forty-seven hours of data and methods core meetings, plus time between working meetings during which data and methods core team members would identify measures to propose and network projects would review proposed measures with other team members.

Two full-time data and methods core staff members with expertise in data management supported this harmonization process. The overall process from reviewing preliminary measures to distributing codebooks to network project teams lasted 12 months.

The CFVP Network anticipates a return on this investment when projects collect data and are able to conduct cross-project analyses, as harmonization can support efficiently building the evidence base while capitalizing on between-study heterogeneity (Ridenour et al. , 2023). The presently described process was substantial, harmonizing sixty constructs across three studies.

The requisite time and resources will vary depending on the number of studies and constructs to harmonize; research networks interested in harmonizing studies in a specific research area should plan accordingly. #### Develop Shared Terminology and Definitions The data harmonization workgroup developed shared vocabulary to describe the harmonization process (see Unifying Terminology, above) and the harmonized constructs.

We expect shared terminologies and definitions will support and accelerate collaborative cross-project analyses within the CFVP Network (National Academies of Sciences, Engineering & Medicine, 2022). Network project investigators, data and methods core faculty, and NIH representatives came from various academic disciplines and, as such, defined similar constructs in different ways.

Developing shared construct definitions was key to successfully presenting and harmonizing measures across studies. #### Clarify Priorities and Assumptions In the process of discussing domains, constructs, and items during data harmonization workgroup meetings, network projects considered and verbalized their relationships with communities and their implications for selecting measures.

The various perspectives among the data harmonization workgroup fostered a space that helped foreground projects’ priorities and assumptions about their focal contexts and populations.

The specific constructs prioritized by network projects were influenced by considerations ranging from historical tensions between academic institutions and communities, expected participant demographics, anecdotal and qualitative feedback from focus testing/pilot participants and community partners, and assumptions about how and when violence occurs in these communities.

In identifying these priorities among the project sites, the workgroup could balance community-specific interests with other research concerns and propose alternative measurements. These conversations enabled the entire CFVP to approach firearm injury prevention research in a way that maximized both contextual, cultural relevance and scientific rigor.

#### Identify and Understand the Limits of the Data Harmonization Process It is not feasible to harmonize every construct that each unique project is interested in investigating. Such an approach would yield participant questionnaires and interviews that would be too long to feasibly administer. As such, networks must decide which constructs to harmonize and which constructs to leave for individual projects to consider separately.

Although the inaugural network projects had many areas of overlapping interests, they also had research questions unique to their sites. The data harmonization workgroup was not able to harmonize every construct that each individual site was interested in including. As such, network projects had to critically consider constructs important to their sites that the harmonized constructs did not adequately address.

In these circumstances, it was at the discretion of individual network projects to include non-harmonized constructs in their own projects in addition to the harmonized constructs from the CFVP Network. ## [](https://pmc. ncbi.

nlm. nih. gov/articles/PMC13051374/)Conclusions The CFVP Data Harmonization Workgroup’s process can serve as a model for future networks that seek to build the science in a particular focal area.

The CFVP Network harmonization process originally included three network projects, and resulted in overlapping assessment timepoints, sixty harmonized measures, and a unified data management plan. We expect the study infrastructure developed through this process to support and catalyze future research on preventing firearm injuries, including and extending beyond future CFVP Network Projects. ## [](https://pmc.

ncbi. nlm. nih.

gov/articles/PMC13051374/)Supplementary Material **Supplementary Information** The online version contains supplementary material available at https://doi. org/10. 1007/s11121-024-01723-5.

## [](https://pmc. ncbi. nlm.

nih. gov/articles/PMC13051374/)Funding This network study was supported by the Community Firearm Violence Prevention (CFVP) Coordinating Center (U24HD111315, Carter/Zimmerman) from the National Institutes of Health (NIH)/_Eunice Kennedy Shriver_ National Institute of Child Health and Human Development (NICHD).

Individual projects were supported by NIH/NICHD (UG3HD111325, Zakrison) and NIH/National Institute on Minority Health and Health Disparities (NIMHD) (UG3MD018296, Edberg; UG3MD018298, Zhang/Kutcher/Vearrier). This manuscript is the sole responsibility of the authors and does not represent the official view of the National Institutes of Health or of the CFVP Network. **Conflict of Interest** The authors declare no competing interests.

**Ethics Approval** This process paper did not involve human subjects research. **Consent to Participate** Not applicable. ## [](https://pmc.

ncbi. nlm. nih.

gov/articles/PMC13051374/)References * . Baker, Michael White, Cappelleri, Kluger, & Coleman (2009). From the Health Outcomes P and Economics (HOPE) collaborative group.

Understanding heterogeneity in meta-analysis: The role of meta-regression. _International Journal of Clinical Practice_, 63(10), 1426–1434. doi: 10.

1111/j. 1742-1241. 2009.

02168. x [DOI] [PubMed] [Google Scholar] * . Cunningham, Carter, Ranney, _et al.

_ (2019). Prevention of firearm injuries among children and adolescents: Consensus-driven research agenda from the firearm safety among children and teens (FACTS) consortium. _JAMA Pediatrics_, 173(8), 780–789.

doi: 10. 1001/jamapediatrics. 2019.

1494 [DOI] [PMC free article] [PubMed] [Google Scholar] * . Curran, Cole, Giordano, Georgeson, Hussong, & Bauer (2018). Advancing the study of adolescent substance use through the use of integrative data analysis.

_Evaluation and the Health Professions_, 41(2), 216–245. doi: 10. 1177/0163278717747947 [DOI] [PMC free article] [PubMed] [Google Scholar] * .

Ducharme, Wiley, Mulford, Su, & Zur (2021). Engaging the justice system to address the opioid crisis: The Justice Community Opioid Innovation Network (JCOIN). _Journal of Substance Abuse Treatment_, 128, 108307.

doi: 10. 1016/j. jsat.

2021. 108307 [DOI] [PMC free article] [PubMed] [Google Scholar] * . Hamilton, Strader, Pratt, _et al.

_ (2011). The PhenX Toolkit: Get the most from your measures. _American Journal of Epidemiology_, 174(3), 253–260.

doi: 10. 1093/aje/kwr193 [DOI] [PMC free article] [PubMed] [Google Scholar] * . Hussong, Curran, & Bauer (2013).

Integrative data analysis in clinical psychology research. _Annual Review of Clinical Psychology_, 9, 61–89. doi: 10.

1146/annurev-clinpsy-050212-185522 [DOI] [PMC free article] [PubMed] [Google Scholar] * . Jay (2023). Structural racism and long-term disparities in youth exposure to firearm violence.

_JAMA Network Open_, 6(5), e2312425–e2312425. doi: 10. 1001/jamanetworkopen.

2023. 12425 [DOI] [PMC free article] [PubMed] [Google Scholar] * . Kroenke, Spitzer, & Williams (2001).

The PHQ-9: Validity of a brief depression severity measure. _Journal of General Internal Medicine_, 16(9), 606–613. doi: 10.

1046/j. 1525-1497. 2001.

016009606. x [DOI] [PMC free article] [PubMed] [Google Scholar] * . National Academies of Sciences, Engineering, and Medicine.

(2022). _Ontologies in the Behavioral Sciences: Accelerating Research and the Spread of Knowledge_. National Academies Press.

[PubMed] [Google Scholar] * . National Institutes of Health. (2022).

_Funding Opportunity: Research on Community Level Interventions for Firearm and Related Violence, Injury and Mortality Prevention (CLIF-VP) (UG3/UH3 Clinical Trial Optional)_. https://grants. nih.

gov/grants/guide/pa-files/PAR-22-115. html [Google Scholar] * . Ridenour, Cruden, Yang, _et al.

_ (2023). Methodological strategies for prospective harmonization of studies: Application to 10 distinct outcomes studies of preventive interventions targeting opioid misuse. _Prevention Science_, 24, 16–29.

doi: 10. 1007/s11121-022-01412-1 [DOI] [PMC free article] [PubMed] [Google Scholar] * . Schmidt, Rauk, Cunningham, Zimmerman, Roche, & Carter (2022).

How multi-disciplinary research centers and networks can advance the science of firearm injury prevention. _The Annals of the American Academy of Political and Social Science_, 704(1), 242–266. [Google Scholar] * .

Spitzer, Kroenke, Williams, & Löwe (2006). A brief measure for assessing generalized anxiety disorder: The GAD-7. _Archives of Internal Medicine_, 166(10), 1092–1097.

doi: 10. 1001/archinte. 166.

10. 1092 [DOI] [PubMed] [Google Scholar] * . Thornberry, Lizotte, Krohn, Farnworth, & Jang (1994).

Delinquent peers, beliefs, and delinquent behavior: A longitudinal test of interactional theory. _Criminology_, 32(1), 47–83. [Google Scholar] * .

Tucker, Iwelunmor, Abrams, _et al. _ (2021). Accelerating adolescent HIV research in low-income and middle-income countries: Evidence from a research consortium.

_AIDS_, 35(15), 2503–2511. doi: 10. 1097/QAD.

0000000000003049 [DOI] [PMC free article] [PubMed] [Google Scholar] * . University of Chicago Medicine. (2022).

_Community Health Needs Assessment, 2021–2022_. https://issuu. com/communitybenefit-ucm/docs/ucmc-chna-2021-2022?

fr=sNTc0NTE0ODc0MDM [Google Scholar] * . Volkow, Koob, Croyle, _et al. _ (2018).

The conception of the ABCD study: From substance use to a broad NIH collaboration. _Developmental Cognitive Neuroscience_, 32, 4–7. doi: 10.

1016/j. dcn. 2017.

10. 002 [DOI] [PMC free article] [PubMed] [Google Scholar] * . WISQARS.

(2023). _Fatal injury reports, 2000–2020_. Centers for Disease Control and Prevention, National Center for Injury Prevention and Control; 2020.

Accessed January 25, 2023. https://wisqars. cdc.

gov/fatal-reports [Google Scholar] ## Supplementary Materials