Association Between the New York SAFE Act and Firearm Suicide and Homicide: An Analysis of Synthetic Controls, New York State, 1999‒2019

Abstract

Objectives. To assess the association between the New York Secure Ammunition and Firearms Enforcement Act (NY SAFE Act) and firearm suicide and homicide rates.

Methods. We employed a synthetic controls approach to investigate the impact of the NY SAFE Act on firearm suicide and firearm homicide rates. We collected state-level data on firearm mortality from the Centers for Disease Control and Prevention’s Wide-ranging Online Data for Epidemiologic Research (WONDER) database for the period 1999–2019. We derived statistical inference by using a permutation-based in-place placebo test.

Results. The implementation of the NY SAFE Act was associated with a significant reduction in firearm homicide rates, demonstrating a decrease of 63%. This decrease corresponds to an estimated prevention of 1697 deaths between 2013 and 2019. However, there was no association between the NY SAFE Act and firearm suicide rates.

Conclusions. As the responsibility for enacting firearm policies increasingly falls on states instead of the federal government, this study provides valuable information that can assist states in making evidence-based decisions regarding the development and implementation of firearm policies that prioritize public safety and aim to prevent firearm-related fatalities. (Am J Public Health. 2023;113(12):1309–1317. https://doi.org/10.2105/AJPH.2023.307400)

In 2022, the Supreme Court of the United States blocked a 111-year-old gun law that prohibited New York State residents from carrying concealed guns in public.¹ The Court ruled that the law was unconstitutional and violated residents’ rights in the Second and Fourteenth Amendments.¹ Loosening gun laws has been found to be linked to an increase in homicides from firearms.^2–4 Firearm violence is a preventable public health crisis and a leading cause of premature death in New York and the United States.^5,6 Therefore, preventing firearm-related deaths, disabilities, and injuries warrants an empirical investigation of policies that are effective in protecting the public’s health.

Firearm injuries account for 5.3 deaths per 100 000 New York population, which equates to 1052 deaths in 2020 alone.^7,8 On average, 1991 persons are injured by guns every year in New York, ranking the state 42nd in firearm injuries nationwide.⁹ Although New York is the fourth-lowest in gun deaths nationally, the mortality rate from guns increased by 7% from 2011 to 2020, which is equivalent to an absolute increase of 55 more deaths during the past decade.⁹ Firearm violence is also a cause of economic burden to New York, with $11.4 billion expended annually, including $301.2 million that is being subsidized by taxpayers.⁹

To limit gun-associated harm, several states, including New York, have enacted laws to regulate firearms.¹⁰ Examples include mandating licensing requirements beyond the standard set by the federal government, requiring background checks before the purchase of firearms and ammunition, prohibiting individuals that demonstrate signs of harm—to themselves or others—from purchasing a firearm (red flag laws), and mandating gun owners to make a reasonable effort at keeping firearms out of reach of children and other prohibited persons (safe storage laws).^4,10 State legislative efforts have been heterogeneous and have resulted in varied firearm-related outcomes.¹⁰ But, overall, states with a comprehensive set of legislative actions have recorded reduced firearm-related injuries and deaths.^10–13 Legislative actions are also more likely to result in reduced morbidity and mortality if they restrict child access to firearms, implement systematic measures to limit the process of firearm acquisition, regulate the use of assault weapons and large-capacity magazines, and prohibit individuals who demonstrate a risk of harm—to self or others—from purchasing a firearm.^11,13–16 Conversely, stand-your-ground and right-to-carry laws have been associated with increased firearm morbidity and mortality.^17,18

The New York Secure Ammunition and Firearms Enforcement Act (NY SAFE Act) was enacted in New York in January 2013.¹⁹ Under this Act, owning a magazine that can hold more than 7 rounds of ammunition became illegal.¹⁹ NY SAFE Act also mandates national background checks for private firearm purchases, expands the definition of illegal assault weapons to include rifles and shotguns with a thumbhole stock, requires a list of every firearm owned by an individual whenever a permit is renewed, establishes stronger penalties for illegal firearm use, requires safe storage of firearms from convicts of a felony or domestic violence, requires owners to report lost or stolen firearms or ammunitions within 24 hours, requires the recertification of handgun permits every 5 years, and amends the New York Mental Hygiene Law to require the reporting of patients at risk for inflicting significant harm to themselves or others, among other provisions.¹⁹

Despite the passage of a decade since its enactment, the impact of the NY SAFE Act on firearm deaths remains largely unexplored in the existing literature. Previous studies have primarily focused on specific aspects of the Act, including its implications for individuals with mental health issues or its constitutionality.^20–24 This study aims to fill this significant research gap by employing a novel quasi-experimental design, the synthetic control method (SCM), to examine the association between the NY SAFE Act and firearm mortality. SCM has garnered widespread recognition as a groundbreaking innovation in the evaluation literature and has been hailed as “the most important innovation in the policy evaluation literature in the last fifteen years.”^25(p9) Unlike conventional approaches, SCM utilizes a data-driven approach to construct synthetic controls that closely resemble the treatment state, thereby enhancing the internal validity of the findings.²⁵ The aim of the current study was to assess the relationship between the implementation of the NY SAFE Act and firearm suicide and homicide rates.

METHODS

We used a quasi-experimental design to examine the relationship between the implementation of the NY SAFE Act and counterfactual firearm mortality rates, specifically firearm suicide and homicide. The analysis involved estimating the counterfactual rates using data from weighted control states. The study period spanned from 1999 to 2019, with the preintervention period covering 1999 to 2012 and the postintervention period from 2013 to 2019. We excluded the year 2020 from our analysis because of the higher number of recorded firearm deaths in the United States during the COVID-19 pandemic.²⁶

Outcome Measures

Our primary outcome measures included age-adjusted rates of overall firearm mortality, firearm suicide, and firearm homicide per 100 000 persons. Data on firearm deaths were obtained from the Centers for Disease Control and Prevention’s Wide-ranging Online Data for Epidemiologic Research (WONDER) database. We abstracted the following variables using the following International Classification of Diseases, 10th Revision (Geneva, Switzerland: World Health Organization; 1992) codes: W2‒W34, X72‒X74, X93‒X95, Y22‒Y24, Y35.0, and U01.4. All rates were age-adjusted to the 2000 US standard population using the direct method.

Mortality rates were occasionally suppressed in WONDER for confidential reasons or because of low counts deemed unreliable.²⁷ To address missing values, multiple imputation was employed, replacing them with the average mortality rate for the respective state. States with complete missing values were excluded from the analysis.

Covariates

We derived state-level characteristics that could potentially confound the association between the NY SAFE Act and firearm mortality from national registries and included as covariates in the analysis. From the US Census Bureau,²⁸ we obtained the state-level proportion of residents who were aged younger than 18 years or aged 65 years or older; the proportion of females; proportion who were non-Hispanic White, non-Hispanic Black, Hispanic, American Indian/Alaska Native, and Asian/Pacific Islander; the proportion of residents that were high-school graduates or higher, and those with bachelor’s degree or higher; and for the entire population, the median household income, the proportion living in poverty, and the proportion without health insurance aged younger than 65 years. From the US Bureau of Labor Statistics,²⁹ we obtained the unemployment rate for every state included in this study.

Statistical Analysis

We used SCM to estimate the counterfactual rates of overall firearm mortality, firearm suicide, and firearm homicide for the state of New York. SCM offers a valuable alternative to conventional evaluation approaches and effectively addresses significant limitations in the field of firearm policy research, including the challenges of finding comparable intervention and control groups and the sensitivity of findings to specific modeling choices.³⁰

This innovative approach involved leveraging preintervention data on outcomes and relevant predictors to assign weights to control states in the donor pool.³⁰ The objective was to create a weighted combination of control states that closely resembled the intervention state during the preintervention period.^25,30 By carefully selecting the weights, certain states were excluded from the synthetic control analysis. The resulting weighted average of states provided an approximation of what the outcomes would have been in the posttreatment period had there been no intervention.^31,32

To evaluate the impact of the NY SAFE Act on firearm mortality rates, we conducted a comparative analysis between the postintervention rates of New York and its synthetic control group from 2013 to 2019. This approach allowed us to assess the effectiveness of the intervention by examining the differences between the observed rates in New York and the counterfactual rates predicted by the synthetic control group.

The synthetic control method does not rely on traditional measures of uncertainty or statistical significance.^31–34 Therefore, to assess the likelihood of observing firearm suicide and homicide rate differences equal to or greater than those observed in New York attributable to chance alone, we employed permutation-based in-place placebo tests. These tests involved conducting the same analysis multiple times by randomly shuffling (permuting) the assignment of the intervention among control states.^31–34 The underlying premise is that in the absence of an actual intervention, no effect is expected, although random variation in the postintervention period could occur.^31–34 We calculated the root mean square prediction error (RMSPE) ratio (RMSPE posttreatment/pretreatment) and examined the P value below .10 as described by Abadie et al., Chrisinger, and Galiani et al.^32–34 We excluded states with a poor fit before the intervention, indicated by an RMSPE exceeding twice that of the intervention state, from the analyses.

To calculate the estimated number of firearm deaths prevented by the NY SAFE Act from 2013 to 2019, we multiplied the relative difference in rate between New York and its synthetic control by New York’s population size each year, summing the values across those years.

We conducted all statistical analyses with Stata version 17.0 (StataCorp LP, College Station, TX).

Sensitivity Analysis

We assessed whether the implementation of the NY SAFE Act was associated with any potential substitution effect leading to increased deaths from nonfirearms. The results of this analysis, including the corresponding tables and charts, are provided in Appendix B, Tables A‒C and Figures F‒K, available as a supplement to the online version of this article at https://ajph.org.

RESULTS

We constructed a synthetic control for New York’s overall firearm mortality, firearm suicide, and firearm homicide rates by utilizing predictive covariates and mortality rates from the period preceding the implementation of the NY SAFE Act. The inclusion of states as synthetic controls was based on their nonzero weights, as summarized in Table 1. Specifically, for the overall firearm mortality, the synthetic controls for New York included California, Hawaii, Maryland, Massachusetts, Nebraska, New Hampshire, Rhode Island, and District of Columbia.

TABLE 1—

Weight Allocation for Synthetic Control Analysis of New York State: 1999‒2019

State	Weight
	Firearm Mortality	Firearm Suicide	Firearm Homicide
Alabama	0.000	0.000	0.000
Alaska	0.000	0.000	0.158
Arizona	0.000	0.000	0.058
Arkansas	0.000	0.000	0.000
California	0.192	0.000	0.000
Connecticut	0.000	0.037	0.000
Delaware	0.000	0.000	0.000
Florida	0.000	0.000	0.000
Georgia	0.000	0.000	0.000
Hawaii	0.305	0.082	0.000
Idaho	0.000	0.000	0.000
Illinois	0.000	0.098	0.000
Kansas	0.000	0.000	0.156
Kentucky	0.000	0.000	0.000
Louisiana	0.000	0.000	0.000
Maine	0.000	0.000	0.000
Maryland	0.031	0.000	0.000
Massachusetts	0.376	0.649	0.017
Minnesota	0.000	0.000	0.123
Mississippi	0.000	0.000	0.000
Missouri	0.000	0.000	0.010
Montana	0.000	0.000	0.000
Nebraska	0.041	0.000	0.000
Nevada	0.000	0.000	0.060
New Hampshire	0.026	0.000	0.000
New Jersey	0.000	0.063	0.000
New Mexico	0.000	0.000	0.000
North Dakota	0.000	0.000	0.000
Ohio	0.000	0.000	0.000
Oklahoma	0.000	0.000	0.000
Oregon	0.000	0.000	0.106
Pennsylvania	0.000	0.071	0.000
Rhode Island	0.012	0.000	0.000
South Carolina	0.000	0.000	0.000
South Dakota	0.000	0.000	0.000
Texas	0.000	0.000	0.000
Utah	0.000	0.000	0.241
Virginia	0.000	0.000	0.064
West Virginia	0.000	0.000	0.000
Wisconsin	0.000	0.000	0.000
Wyoming	0.000	0.000	0.000
Washington, DC	0.017	0.000	0.009
RMSPE synthetic control/all control states	0.159/0.618	0.102/0.424	0.104/0.533

Note. RMSPE = root mean square prediction error. The RMSPE provides a measure of the fit, with lower values indicating a better alignment with New York during the preintervention period.

We used RMSPE to evaluate the accuracy of the synthetic controls in approximating New York’s firearm mortality rate. Lower RMSPE values signify a better fit.^31,32 As indicated in the last row of Table 1, the synthetic control outperformed the simple average of all states in the donor pool in terms of fit. For instance, in the case of overall firearm mortality, the synthetic control achieved an RMSPE of 0.159, which is significantly lower than the RMSPE that would have been obtained if a simple average of all control states (i.e., the remaining 49 US states and the District of Columbia) had been used (0.618; Table 1).

Table 2 summarizes the predictive covariates associated with firearm mortality. These variables are presented for both New York and the synthetic control group designed to optimize firearm mortality overall, firearm suicide, and firearm homicide. The similarities in the average balance of these covariates serve as further evidence of the close resemblance between the synthetic controls (synthetic New York) and the actual state of New York during the pre‒NY SAFE Act period.

TABLE 2—

Covariate Balance for New York and Synthetic New York State: 1999‒2012

Covariate	Firearm Mortality		Firearm Suicide		Firearm Homicide
	New York	Synthetic New York	New York	Synthetic New York	New York	Synthetic New York
Age, y, %
< 18 y	20.3	20.4	20.3	19.8	20.3	23.7
≥ 65 y	18.1	18.3	18.1	18.3	18.1	15.6
Female, %	51.1	50.3	51.1	50.8	51.1	49.5
Race/ethnicity, %
Non-Hispanic White	54.2	47.8	54.2	63.7	54.2	68.7
Non-Hispanic Black	17.7	7.6	17.7	10.1	17.7	6.0
American Indian/Alaska Native	1.0	0.76	1.0	0.5	1.0	3.9
Asian	9.6	17.9	9.6	9.8	9.6	4.9
Pacific Islander	0.1	3.3	0.1	0.9	0.1	0.7
Hispanic	19.7	17.4	19.7	13.9	19.7	13.9
Education level, %
High school education or lower	87.4	90.4	87.4	91.2	87.4	92.1
Bachelor’s degree or higher	38.1	39.3	38.1	40.3	38.1	34.7
Persons in poverty, %	13.9	11.1	13.9	10.7	13.9	10.7
Household income, median, $	75 157.0	86 683.0	75 157.0	85 648.0	75 157.0	74 698.0
Persons without health insurance, %	6.1	5.0	6.1	4.3	6.1	9.9
Unemployment rate, %	3.9	3.3	3.9	3.1	3.9	3.1
Mortality rate,a %
2012	4.8	4.9	2.5	2.6	2.2	2.3
2011	5.0	5.0	2.5	2.4	2.4	2.5
2010	5.1	5.1	2.3	2.5	2.7	2.5
2009	4.8	4.9	2.0	2.0	2.6	2.6
2008	4.9	5.1	2.1	2.2	2.6	2.7
2007	5.1	5.1	2.2	2.2	2.7	2.7
2006	5.2	5.0	2.1	2.0	3.0	2.9
2005	5.3	5.0	2.3	2.1	2.7	2.8
2004	4.9	5.1	2.0	2.0	2.7	2.8
2003	5.3	5.3	2.1	2.2	3.1	3.1
2002	5.1	5.3	2.1	2.1	2.9	2.9
2001	5.5	5.4	2.3	2.3	3.1	3.0
2000	5.7	5.5	2.3	2.3	3.2	3.1
1999	5.3	5.3	2.3	2.4	2.9	3.0

Note. The preintervention period includes 1999–2012, which is the period before the enactment of the NY SAFE Act. Variable balance is assessed by comparing the values in New York (the treatment state) and synthetic New York (its synthetic control). A difference closer to zero indicates better balance between the 2.

^a Firearm mortality rates are reported per 100 000 population and age-adjusted to the US population in 2000.

We conducted a graphical comparison of the firearm mortality rate over time between New York and synthetic New York. During the preintervention period, before the implementation of the NY SAFE Act, both groups exhibited a closely tracked trajectory, as also supported by the RMSPE of 0.159 and the covariate balance data presented in Table 2. However, starting from 2013, a divergence in rates occurred. While New York’s firearm mortality rate continued to decrease through 2019, the rate for synthetic New York increased notably throughout the study period. We compared the rate differences between New York and synthetic New York from 2013 to 2019 and found that the implementation of the NY SAFE Act was associated with a 22.4% reduction in firearm mortality deaths compared with the counterfactual scenario (4.10 vs 5.02; Figure 1; Table 3).

FIGURE 1—

Age-Adjusted Firearm Mortality Rate (per 100 000) for New York and Synthetic New York State: 1999‒2019

Note. The vertical line indicates the enactment of the New York Secure Ammunition and Firearms Enforcement Act in January 2013.

TABLE 3—

Differences in Firearm Mortality Rates for New York State: 1999‒2019

	Pre-New York SAFE Act (1999–2012)				Post-New York SAFE Act (2013–2019)				Permutation-Based P
	New York	Synthetic New York	Relative Change, %	Absolute Difference (95% CI)	New York	Synthetic New York	Relative Change, %	Absolute Difference (95% CI)
Firearm mortality	5.14	5.15	0.2	−0.01 (−0.19, 0.17)	4.10	5.02	22.4	−0.92 (−1.29, −0.54)	0.08
Firearm suicide	2.21	2.23	0.9	−0.02 (−0.15, 0.12)	2.12	2.45	15.6	−0.33 (−0.49, 0.17)	0.22
Firearm homicide	2.77	2.78	0.4	−0.01 (−0.21, 0.19)	1.89	3.08	63.0	−1.20 (−1.69, −0.71)	0.026

Note. CI = confidence interval; SAFE = Secure Ammunition and Firearms Enforcement. The P value represents the proportion of control units’ effects that have posttreatment root mean square prediction error at least as large as the treated unit, while accounting for the quality of pretreatment matches.

The permutation test results were consistent with the graphical findings, indicating that there was a decrease in New York’s overall firearm mortality rate following the implementation of the NY SAFE Act. Furthermore, among the control states with an RMSPE no more than twice that of New York, none exhibited firearm mortality trends that diverged from their synthetic controls to the extent observed in New York (Table 3; Figure A, available as a supplement to the online version of this article at https://ajph.org).

Firearm Suicide

Firearm suicide rates in New York closely tracked those of synthetic New York during the preintervention period. However, a notable divergence between the 2 rates occurred after the implementation of the NY SAFE Act. By summing the differences in rates between New York and synthetic New York from 2013 to 2019, we determined that the NY SAFE Act was linked to 467 fewer deaths than anticipated. However, the permutation test indicated that the observed difference in firearm suicide rates between New York and synthetic New York was not statistically significant and could have occurred by chance (Tables 2 and 3; Appendix A, Figures B and C, available as supplements to the online version of this article at https://ajph.org).

Firearm Homicide

Firearm homicide rates in New York exhibited a close alignment with those of synthetic New York during the preintervention period from 1999 to 2012, as supported by the covariate balance, RMSPE of 0.104, and visual assessments. However, following the implementation of the NY SAFE Act, a divergence in firearm homicide rates occurred, with a continued decrease in New York through 2019 and an upward trend in synthetic New York until the end of the study period. The permutation test demonstrated that the difference in postintervention firearm homicide rates between New York and synthetic New York was significant and could not have arisen by chance. From 2013 to 2019, the implementation of the NY SAFE Act is estimated to have prevented a total of 1697 deaths from firearm homicides. This represents a decrease of 63% compared with the counterfactual, with an average rate of 1.89 deaths per 100 000 population as opposed to 3.08 deaths per 100 000 population (Tables 1–3; Appendix A, Figures D and E, available as supplements to the online version of this article at https://ajph.org).

Sensitivity Analysis for Nonfirearm Mortality

To investigate potential substitution effects, we examined the relationship between the NY SAFE Act and nonfirearm mortality overall, nonfirearm suicide, and nonfirearm homicide. The permutation tests conducted indicated that there was no significant association between any of these outcomes and the implementation of the NY SAFE Act (see Appendix B: Tables A‒C and Figures F‒K, for the detailed results).

DISCUSSION

In this quasi-experimental study, we identified a significant reduction in firearm homicide associated with the implementation of the NY SAFE Act. However, we did not observe an association between the NY SAFE Act and firearm suicide.

The NY SAFE Act, compared with other state firearm laws, incorporates a multitude of comprehensive provisions aimed at reducing gun-related harm.¹⁹ The Act stands out as a multifaceted measure with provisions, including the “1-feature test” for assault weapons, restrictions on large-capacity magazines, mandatory background checks for firearm purchases, and measures to enhance safe storage practices.¹⁹ These provisions collectively address various aspects of firearm safety and access, highlighting the Act’s comprehensive approach. While some states have implemented similar provisions individually, the NY SAFE Act goes beyond by encompassing a broader range of measures, making it one of the more comprehensive state laws enacted in recent years.²⁴ The inclusion of such provisions in the NY SAFE Act likely contributes to the observed reduction in firearm mortality found in this study, reinforcing the importance of comprehensive approaches to gun regulation.

The provision of the NY SAFE Act concerning assault weapons and large-capacity magazines may partially account for the observed decrease in firearm homicide in this study. Section 37 of the Act introduces the “1-feature test,” which effectively prohibits semiautomatic weapons with detachable magazines that possess 1 military-related feature.¹⁹ In addition, the Act bans the sale of new large-capacity magazines capable of holding more than 7 rounds of ammunition, with current owners required to transfer such magazines out of state or to firearms dealers.¹⁹ Given that ammunition and large-capacity magazines can contribute to mass shootings, which involve 4 or more victims, regulating the sale and use of these weapons logically has the potential to limit firearm homicides.¹⁶ Previous research examining mass shootings following the expiration of a federal ban on large-capacity magazines and military-style semiautomatic firearms found a significant increase in both the number of incidents and fatalities.³ A survey of 32 researchers conducted by the New York Times on best practices for reducing mass shootings revealed that an assault weapons ban was the most effective measure.¹⁶

We did not observe a significant association between the implementation of the NY SAFE Act and firearm suicide. Several potential explanations could account for this finding. A notable aspect of the NY SAFE Act is its provision that requires mental health professionals to report individuals they believe may pose a significant risk of harm to themselves or others.¹⁹ This provision has raised concerns regarding its constitutionality and its impact on the principles of the Hippocratic oath and the physician‒patient relationship.^20–24 One possible consequence of this provision is that individuals with mental health issues who are contemplating firearm suicide may be deterred from seeking medical care because of the fear of being flagged and reported.^22–24 This fear arises from the potential risk of having their firearm license revoked and their firearms confiscated.^22–24 The presence of this potential deterrent effect may contribute to the absence of a significant protective association between the NY SAFE Act and firearm suicide. These concerns underscore the importance of conducting further examination and fostering discussion regarding the ethical and legal considerations surrounding firearm policy and mental health.

Another potential explanation for the absence of a significant protective association between the NY SAFE Act and firearm suicide could be the presence of a time lag and long-term effects. The impact of the NY SAFE Act on firearm suicides may require a longer period to manifest.³⁵ Suicide prevention efforts typically involve comprehensive strategies, including improving access to mental health services, fostering community support, and reducing stigma surrounding mental health issues.³⁵ These multifaceted approaches may have a more delayed and cumulative effect, contrasting the immediate impact of measures targeting homicides.³⁵ Thus, the complex nature of suicide prevention efforts and the potential time lag involved may contribute to the lack of a detectable association between the NY SAFE Act and firearm suicide.

Further research is warranted to explore the potential effects of the NY SAFE Act on firearm mortality, including specific mechanisms that may explain the observed reduction in firearm homicides but not firearm suicides. Understanding these nuances can inform comprehensive firearm policies.

Limitations and Strengths

This study has some limitations. We used WONDER data, which were exclusively based on death certificate records of US residents.²⁷ Future studies may consider investigating the association between the NY SAFE Act and firearm injuries, rather than mortality. Second, our use of state-level data precludes individual-level inference—ecologic fallacy. To improve on this limitation, further studies may consider using individual-level data.

This study has several notable strengths that enhance the internal validity and robustness of our findings. First, we employed SCM, an innovative and novel approach to assess the impact of the NY SAFE Act.^25,31,32 Through SCM, we integrated data from multiple control states to construct a synthetic control that closely resembled New York, enabling a rigorous comparison of outcomes before and after implementation of the Act. Furthermore, our study addressed the issue of potential spillover effects from neighboring states by incorporating their data into the donor pool. The synthetic control analysis assigned weights based on similarities to the intervention state, considering factors such as geographic proximity and the potential for contamination.^31,32 Specifically, for overall firearm mortality, states adjacent to New York were assigned a weight of zero and were not included as synthetic controls. In the case of firearm suicide, Connecticut had a weight of 0.037 and New Jersey had a weight of 0.063, indicating their minimal contributions as synthetic controls. Finally, we explored the possibility of substitution effects by examining the association between the NY SAFE Act and nonfirearm suicide and homicide. This comprehensive analysis allowed us to assess whether any reductions in firearm mortality were counterbalanced by increases in non‒firearm-related deaths.

Public Health Implications

The implementation of the NY SAFE Act is associated with reduced firearm homicide rates. However, no significant association was observed with firearm suicide rates. As states assume a larger role in enacting firearm policies, our findings provide valuable insights for discussions and considerations on the potential impact of future firearm policies and state-level interventions in reducing firearm-related fatalities in the United States.