Abstract
Introduction: Firearm violence now accounts for more than 45,000 deaths annually in America, making it the leading cause of death in the pediatric population. Despite this, primary care clinicians (PCCs) lack screening tools to assess a patient’s future risk of gun violence, hampering the ability to allocate resources that could prevent morbidity and mortality.
Methods: Longitudinal quantitative study designed to evaluate whether the Serious fighting, Friend weapon carrying, community Environment, and firearm Threats (SaFETy) score and other measures could predict gun violence exposure over 6 months in a nonurban primary care (PC) based adolescent-young adult population.
Results: Eighty-eight patients between the ages of 14 to 24 completed the baseline questionnaire and 62 completed the 6-month questionnaire. 45% of participants had a SaFETy score >0. Any score >0 indicates an increased risk of future firearm violence. A strong correlation was found between the SaFETy score and gun violence exposure, and Adverse Childhood Events (ACEs).
Conclusions: This was among the first studies attempting to assess the viability of screening for firearm violence and other associated risk factors in a nonurban PC setting. Our findings suggest that adolescent and young adult PC patients are willing to disclose in a survey topics related to violence and firearms with their PCC. If validated in the PC setting, the SaFETy score could become a crucial tool for PCCs given the ease with which it can be implemented into a routine visit and the nonthreatening, and nonconfrontational question design. Tools such as the SaFETy score can provide PCPs with vital insight into their patients’ past gun violence exposures and future violence risks.
- Accident Prevention
- Adolescent
- Adverse Childhood Experiences
- Exposure to Violence
- Firearms
- Gun Violence
- Interpersonal Violence
- Primary Health Care
- Screening
- Surveys and Questionnaires
Introduction
Firearms injuries are significant contributors to morbidity and mortality in the United States, particularly for adolescents and young adults. Among children and teens ages 1 to 19, firearm injury is the leading cause of death,1,2 while over 40% of firearm decedents are under the age of 35.3 Moreover, for children and teens, the US is the only country among its peers where firearm deaths surpass those attributable to cancer or motor vehicles.4 In 2021, there were 48,830 firearm deaths in the US, of these, nearly 10% (4,752) were children and teens,1 while another estimated 24,770 children and teens suffered nonfatal firearm injuries in 2020.5
The costs for firearm injuries are staggering. Medical spending increases $2,495 per person per month following a nonfatal firearm injury, representing a 402% increase in medical spending compared with individuals not harmed by firearms.6 When extrapolated by the estimated 85,000 annual survivors of firearm related injuries in the US, spending directly attributable to nonfatal firearm injuries in the first year following injury would exceed approximately $2.5 billion nationally.6 Total economic costs of gun violence are estimated to be $557 billion or 2.6% of gross domestic product, when psychiatric/psychological burden for caregivers and indirect costs such as lost productivity and decreased quality of life are added.6
Gun violence exposure is defined as being a victim, witness, or aware of gun violence. It can include being shot, hearing gunshots, or knowing someone who has been shot. Exposure to gun violence is now a common but harmful part of daily life for up to 41% of American youth in higher risk communities. In one study of rural and urban children ages 2 to 17, over a third of participants reported directly seeing gun violence or hearing gunshots in public places in their lifetimes.7 Exposure to gun violence in the community has been linked with mental health symptoms and posttraumatic stress in children and teenagers.8,9 Health care clinicians, namely primary care clinicians (PCCs), can play a role in gun violence prevention by identifying at-risk patients and connecting patients with resources.10 Studies have found patients and clinicians value discussions about firearms (secure storage and changing behavior patterns to reduce risk to children in the home) in clinical settings as a means for injury prevention.11,12 Patients presenting for a medical visit for violent injury are known to be at elevated risk for future violent injury.13 However, there are few validated screening clinical tools available to help PCPs universally measure the risk of firearm violence for all their patients. PCPs need tools which can assess their patients’ risk for future gun violence exposure and guide at-risk individuals to resources for prevention and intervention.
One tool is the Serious fighting, Friend weapon carrying, community Environment, and firearm Threats (SaFETy) score, a 4-item scoring system designed and validated in the Flint Youth Injury studies in 2017.14 The SaFETy score was found to predict 24-month gun violence exposure in drug-using adolescents and young adults presenting to an urban emergency department (ED). Gun violence was defined as any victimization or perpetration with a firearm, firearm injury, or firearm death. Importantly, the score was effective at predicting firearm violence among those who did not present for a violent injury. However, that single-site study was among substance-using children and emerging adults (age 14 to 24) presenting to the ED, and, outside of one published study,15 the SaFETy score has not been measured in primary care (PC). Thus, questions remain about the usability of the SaFETy score in PC.
In this report, we detail findings from a longitudinal study in PC practices in North Carolina. Study objectives were: 1) to demonstrate the viability of asking adolescents/young adults SaFETy score screening questions in a PC setting; and 2) to assess this population's gun violence exposure and the ability to anticipate future risk using the SaFETy score as well as other existing tools.
Methods
Study Design and Setting
This study was a longitudinal quantitative study designed to evaluate whether the SaFETy score and other selected measures predict gun violence exposure over 6 months in a nonurban primary-care based adolescent-young adult population. The study was approved by the American Academy of Family Physicians (AAFP) Institutional Review Board. The study was originally conceived as an in-office study, but following the outbreak of the COVID-19 pandemic, the study design was modified to facilitate non–in person communication and study execution.
Four nonurban PC practices in North Carolina, who were also participants in the North Carolina MedServe16 program, participated in the study. MedServe fellows at the participating clinics were responsible for study implementation and execution. Eligible patients were between the ages of 14 to 24 and were patients at one of the participating practices. Ineligible patients were those with inability to consent/assent and complete study tasks due to severe mental health diagnosis or developmental disability.
Eligible patients at each participating clinic were identified by querying the clinic’s patient database and assigning each eligible patient a unique study ID. The list of eligible patient names was randomly sorted to assure random sample selection. Study team members at each practice contacted and recruited patients to participate in the study by phone call, by using the practice’s patient portal messaging system, or by speaking with the patient in-person if the patient was in the clinic office. Patients who responded to portal recruitment messages were scheduled for a follow-up recruitment phone call. Patients who expressed interest in the study were verbally consented over the phone or in-person. For eligible patients 14 to 17 years of age, the study team first contacted the patient's parent/legal guardian for consent and then the patient was contacted, recruited to participate in the study, and assented. Patients were enrolled from October 2021–February 2022.
Enrolled patients were asked to complete baseline and 6-month surveys in Qualtrics (Provo, Utah). Patients could complete surveys online, via a uniquely generated Qualtrics survey link delivered by e-mail, or verbally (over the phone) with a member of the study team. The study team monitored survey completions and sent up to 3 reminders to patients to complete their surveys. Baseline survey responses were collected at enrollment. For 6-month follow-up survey responses, patients were initially contacted using their preferred communication platform (phone, e-mail/portal), then the study team sent up to 3 reminders to complete their survey. Six-month survey responses were collected from April 2022 to August 2022. Participants received a $10 gift card for completing each survey.
Measures
We measured the SaFETy score, violence exposures, firearm behaviors, adverse childhood experiences (ACEs), retaliatory attitudes, resiliency, and peer influences (Table 1). In total, 126 items were measured, including 9-items capturing participant demographics. See Table 1 for constructs and measures used and Appendix 1 for questions and response options for all described items.
Measures Used
Statistical Analyses
Descriptive statistics, including means, standard deviations, and counts, were completed for all survey questions. The SaFETy score, ACES, and gun violence exposure (6-month exposure, lifetime exposure, firearm behaviors) were calculated using methods described elsewhere. Due to the Likert scales with nonnormal distributions of the data, we used nonparametric statistics for analysis on all data. Mann-Whitney U test were used to test for group differences between baseline and 6 months. Wilcoxon Signed Rank tests were used to test for individual differences between baseline and 6 months. We also completed Spearman correlations to determine relationships among the calculated scores (SaFETy score, ACES score, and Gun Violence Exposure score) between baseline and 6 months. Partial Spearman correlations were calculated correcting for gender, age, or gun present. These 3 were chosen a priori due to previous literature. A 2-sided α of 0.05 was used. SPSS 27 (Armonk, NY) was used to complete all analyses.
Results
In total, 269 patients were contacted, of whom, 126 (46.8%) consented to participate, 35 (13.0%) declined, and 108 (40.1%) were unresponsive, meaning they did not respond to multiple outreach attempts to participate in the study. At baseline, 88 patients completed the survey (69.8% of consented). At 6 months, 62 patients completed the survey (49.2% of consented/70.5% of baseline completers).
The majority were female (69.3%), aged 18 to 24 (77.3%), not Hispanic or Latinx (73.6%), completed high school or GED (62.5%), and living at their parents’ home (64.8%). Almost half were white (44.7%) and 36.5% were Black or African American; almost half indicated they achieved mostly B’s in school (45.5%). The majority do not have a gun in the house (56.8%), do not have easy access to a gun (61.4%), and have not taken a gun safety course (79.5%; Table 2).
Demographics
SaFETy scores ranged from 0 to 6 with an average of 0.71 ± 1.25 at baseline and 0.64 ± 0.91 at 6 months (Table 2). The majority of patients had a SaFETy Score of 0 at both time points (Baseline: 60.0%; 6 Months: 52.5%); however, 40% of patients had a SaFETy Score of 1 or greater at baseline and 47.5% at 6 months (Appendix 2). There were no significant differences between baseline and 6 months for the SaFETy Score nor individual components (Table 3).
Questions Included in Scores
ACEs scores ranged from 0 to 7 with an average of 2.40 ± 1.85 at baseline and 2.07 ± 1.79 at 6 months (Table 2). At baseline, 82.7% had at least one ACE; at 6 months, 78.3% had at least one ACE. There was not a significant difference in score nor the individual components between baseline and 6 months (Table 3).
Gun violence exposure scale scores ranged from 0 to 7 with an average of 1.05 ± 1.32 at baseline and 1.48 ± 1.74 at 6 months (Table 2). At baseline, 54.3% and, at 6 months, 62.1% had at least one lifetime gun violence exposure. There was not a significant difference in score nor the individual components between baseline and 6 months. (Table 3). Knowledge of someone who has been murdered is a specific type of gun violence exposure. A high percentage of participants indicated they had known someone who was murdered in their lifetime (Baseline: 23.2%; 6 Months: 19.7%) (Table 3).
Very strong correlations, with no correction (r > 0.7), were observed between ACEs at both time points (r = 0.825) and between Gun Violence Exposure at both time points (r = 0.728). The SaFETy Score at 6 months was moderately correlated with the SaFETy Score at Baseline, ACEs at Baseline, and Gun Violence Exposure at Baseline (r = 0.492, r = 0.354, r = 0.374, respectively). When corrected for gender, age, or gun present at home, the SaFETy Score at Baseline was more highly correlated with the SaFETy Score at 6 months (r = 0.581, r = 0.578, r = 0.582, respectively). ACEs at Baseline and SaFETy Score at 6 months were also moderately correlated when corrected (Gender: r = 0.429, Age: r = 0.414, Gun Present: r = 0.421). However, the correlation strength decreases when correcting for gender, age, and gun present between the Gun Violence Exposure at Baseline and SaFETy Score at 6 months (r = 0.352, r = 0.342, r = 0.349, respectively). Only one correlation was not significant - between the SaFETy Score at Baseline and ACEs at 6 months (Table 4). Retaliatory attitudes results, which were secondary to the firearm-specific outcomes and excluded from Table 3 due to length, are available in Appendix 1.
Spearman Correlations between Scores
Discussion
Overall Findings
This pilot study demonstrates individuals ages 14 to 24 can be asked about firearm violence and other at-risk behaviors and exposures in a nonurban PC setting. Although the study lacked the power to prove validity of the SaFETy score in PC, it did demonstrate 45% of participants had a SaFETy score >0, compared with 93% observed among a high-risk population in the original SaFETy research, where 60% were presenting to the ED for an assault injury. Any score >0 likely indicates an increased risk of future firearm violence and opportunity for intervention to decrease future risk.14 Further, a strong correlation was found between the SaFETy score and gun violence exposure, suggesting the SaFETy score is a strong proxy for firearm exposure and future risk. Study data also indicated a strong correlation between the SaFETy score and ACEs, suggesting adverse experiences in childhood can be related to increased chances of firearm risk later in life.17,18
SaFETy as a Primary Care Screening Tool
This study was among the first to assess and demonstrate the SaFETy score is a straightforward way to assess gun violence risk in a nonurban PC setting. The majority of patients agreeing to participate in this study completed the entire 126-item questionnaire. This suggests patients may be willing to answer the 4 questions of the SaFETy questionnaire in PC settings. Items are worded in a nonthreatening manner, creating an opportunity to obtain valuable information about high-risk behaviors without asking adolescents or young adults confrontational questions such as “do you have a gun?” or “do you use guns for protection?” Although further studies are needed to validate the SaFETy score in a PC setting, this scoring can be easily added to discussions during well child visits or annual exams.
Significance of Findings
There were several notable findings in the study. Most significantly, at baseline and at 6 months, 45% of participants had a SaFETy score greater than zero, suggesting potentially elevated violence risk among nearly half of the study population. Many reported concerning levels of gun violence exposure and gun access and lifetime exposure to adverse events. Many also admitted to carrying a gun at night, having seen someone shoot a gun in a public place, knowing someone close to them being murdered, being “really scared” due to threats, and reported having easy access to a gun. It is important for PCPs to know their patients are being exposed to gun violence which threatens their well-being. This rate of exposure is lower than the 93% rate found in the original SaFETy study, but that study population was higher risk, and over half were at the ED for a violent injury. The percentage of participants who reported having a close friend or family member being murdered in the previous 6 months was higher than the averages of ‘homicide survivorship’; the closest proxy, found in the limited and often outdated literature, which ranged from 8 to 15%, but notably, was not always from directly compatible samples.19–21 Despite this, results require further exploration to ensure correct interpretation from the respondents and direct correlation to murders by firearms. Finally, the majority of participants believed retaliation to transgressions was appropriate compared with only one-quarter preferring to “forgive and forget.”
These findings demonstrate that teens and young adults in PC are a high-need population and a potential high leverage point for gun violence prevention. In particular, this population has high rates of gun violence exposure, easy access to firearms, elevated rates of ACEs, and beliefs that retaliation is appropriate. The information learned here, combined with the increasing rate of death and injury caused by gun violence in the US, support a compelling case for PCPs’ active involvement in assessing their patients’ gun violence risk.
PCPs screen and counsel patients on a variety of highly complex topics. To date no evidence is available to cite PCP’s perception of their patient’s gun violence exposure. While it is possible nonurban PCPs assume their adolescent/young adult patients do not have significant gun violence exposure, this research reveals the prevalence of gun violence exposure and retaliatory attitudes necessitates gun violence exposure screening of this population. Parents overwhelming support firearm injury prevention counseling by physicians (80%) but report this is rarely received (9%).22 A study of pediatric residents revealed 98% believed they have a responsibility to counsel patients and families about the risks posed by firearms.23 Once risk is assessed, physicians can use focused counseling and evidence-based recommendations, such as those outlined in BulletPoints Project.24 PCPs have the potential to provide lifesaving counseling to address risk of gun violence if a validated tool revealed an increased level of gun violence risk. Studies have found both patients and clinicians value discussions about firearm safety in clinical settings as a means for injury prevention, especially for high-risk patients,11,12 yet clinicians lack a single tool to screen for the number 1 cause of death for individuals aged 1 to 19.
Limitations
The study had several limitations. First, the study involved a small sample size over a short time frame due to stipulations from our funding source. To confirm the findings and validate the SaFETy score in PC, the study needs to be repeated on a substantially larger sample size over a longer period. Second, participating sites were nonurban PC clinics located in North Carolina, where firearm laws are relatively relaxed compared with more restrictive states and open carry is allowed without a permit.25 Future studies need to include patients from urban and nonurban settings, and from states with differing firearm laws, to assess feasibility and validity of PC firearm screening across different geographical and political environments. Third, the study was during the COVID-19 pandemic, in particular the Omicron variant wave, which may have negatively affected patient enrollment as the clinics participating in the study were inundated with COVID-19 patients. Fourth, the enrollment rate of patients may have also been impacted by the utilization of MedServe fellows in the recruitment process, whom most patients did not have a prior relationship. Enrollment rates may have been higher if the patient’s PCP had been the primary individual recruiting them to participate in the study. Lastly, participants’ answers during phone in-takes may not have been completely truthful as to not seem overly embedded in gun violence or not. However, as the respondents generally did not know the interviewer, and the number of respondents completing the survey over the phone was relatively low, we do not believe that this limits our results.
Conclusions
In the US, the issue of firearms is deeply complex and rooted in ideological beliefs, but as firearm-related deaths continue to rise and medical organizations increasingly recognize firearm violence as a public health crisis. PCPs should consider their role in addressing this crisis. As the primary clinicians of health care in the US, PCPs are uniquely positioned to identify future firearm violence risk in their patients through routine screening. Yet, even if a PCP identifies firearm risk in their patient, there are limited evidence-based recommendations or actions proven to reduce risk outside of counseling.26–32 Although this further complicates what role PCPs can play in addressing firearm violence in their patient population, the consequences of firearm violence cannot be ignored, namely patient death, and the trauma associated with surviving or being affected by firearm violence. Viewing firearm violence as a health care problem and screening patients for violence exposure or risk, is a critical first step in addressing firearm violence in the US.
This was one of the first studies attempting to assess the viability of screening for firearm violence and other associated risk factors in a nonurban PC setting.33 Although only a pilot study, our findings suggest adolescent and young adult PC patients are willing to disclose in a survey topics related to violence and firearms to their PCPs. Using tools such as the SaFETy score or ACEs questionnaire can provide PCPs with important insight into their patients’ past exposures and future violence risks. If validated in the PC setting, the SaFETy score could become a crucial tool for PCPs given the ease with which it can be implemented into a routine visit and the nonthreatening, nonconfrontational question design. Rigorous validation of the SaFETy score, and other tools for use in PC, are critically needed to provide PCPs with a means to identify firearm risk in their patients and guide at-risk individuals to resources for prevention and intervention.
Acknowledgments
This study would not have been possible without the participation of the following practices from North Carolina: Batish Family Medicine, Cabarrus Rowan Community Health Center, Piedmont Adult and Pediatric Medicine Associates, and Roanoke Chowan Community Health Center. The study authors also acknowledge Divia Batish, Darren Sanders, Godgive Umozurike, Isaiah Hamilton, Lakshmi Meenakshi Immaneni, Margaret Lister, and Sarah Kautz, for their contributions in participant recruitment and retention. Neil Khot is also thanked for his contributions to the final draft of this manuscript. Funding for this research was made possible through an award from the American Academy of Family Physicians Foundation and the support of the American Academy of Family Physicians National Research Network through the Family Medicine Discovers Rapid Cycle Scientific Discovery and Innovation (FMD RapSDI) program. SB, AG, CL, and EC coled the writing of the original and final drafts. CH and JG provided critical review, editing, and guidance on final drafts. No financial disclosures were reported by the authors of this article.
Appendix 1
Appendix 2. Serious fighting, Friend weapon carrying, community Environment, and firearm Threats (SaFETy) Score distribution.
Appendix 3. Consort diagram.
Notes
This article was externally peer reviewed.
Funding: This work was supported by the American Academy of Family Physicians Foundation (AAFPF), Rapid Cycle Scientific Discovery and Innovation (RapSDI) program.
Conflict of interest: Dr. Batish received funding from the American Academy of Family Physicians Foundation to conduct the study. Mr. Lutgen and Drs. Hester and Callen advised and supported the design and implementation of the study. Dr. Callen also provided statistical analysis and support. Dr. Goldstick served as a paid consultant to the study, providing design, analysis, and dissemination support.
- Received for publication October 1, 2024.
- Revision received May 1, 2025.
- Accepted for publication May 19, 2025.
References
Jump to section
Related Articles
Cited By...
- No citing articles found.