Updated Estimate of the Number of Extreme Risk Protection Orders Needed to Prevent 1 Suicide

Key Points Question How effective are extreme risk protection orders (ERPOs) in preventing suicide for the people subject to these emergency measures? Findings This cohort study, which combines information about the suicide methods used by handgun-owning suicide decedents in California, the lethality of different suicide methods, and the distribution of suicide deaths among ERPO respondents in Connecticut, found that 1 suicide death was prevented for every 22 ERPOs issued. Meaning These findings suggest that ERPOs can play an important role in averting deaths among high-risk individuals.


Introduction
Extreme risk protection orders (ERPOs) are civil court orders designed to temporarily prevent a person from purchasing or possessing firearms and ammunition. 1 These laws-also known as red flag, risk warrant, and gun violence restraining orders-authorize law enforcement, family members, and sometimes others (eg, health care professionals or coworkers) to petition a court to remove firearms from and prevent the acquisition of new firearms by a person judged to pose an immediate danger to themselves or others.By March 2024, 21 states and the District of Columbia had enacted such laws. 1 As with many firearm laws, most US adults do not know whether their state has ERPOs or what they entail. 2However, once informed, a majority of the public express support for ERPO laws. 35][6] The most recent of these population-level studies 5 called for additional investigation to estimate individual-level benefits of ERPOs, noting specifically that strong effects on those directly affected by these orders could be present even if the population-level effect is minimal.
To our knowledge, no study to date has directly compared the suicide rate among ERPO respondents (ie, persons who received an ERPO) to the rate among a comparable population who are not ERPO respondents.However, 2 prominent studies, both by Swanson and colleagues, 7,8 used indirect methods to estimate the number of ERPOs issued to prevent 1 suicide death (ie, a number needed to treat [NNT]).Those studies focused on ERPO respondents who died by suicide (21 in   Connecticut, 1999-2013; 14 in Indiana, 2006-2013).The overlapping conclusion of both studies was that 10 ERPOs need to be issued to prevent 1 suicide death, or, as the abstract of the later study states, "one life was saved for every 10 gun-removal actions, similar to results of a previous study in Connecticut." 8In the Connecticut study, 7 Swanson et al also present what they considered to be an overly conservative estimate of approximately 20 suicides averted for every ERPO issued to Connecticut residents.
The approach Swanson and colleagues 7 used involved 3 steps, which we describe here to illustrate how data from the Connecticut study produced an NNT of 10.First, using information about suicide deaths by method among all ERPO respondents, Swanson et al 7 derived an estimate of the number of suicidal acts (ie, suicides plus nonfatal suicide attempts) made by all ERPO respondents by summing across all methods the number of suicidal acts by each method (the latter was derived by dividing the number of method-specific suicide deaths by the corresponding method-specific case fatality ratio [CFR] for male individuals).Male CFRs were used because 86% of suicides were among male individuals, as were 92% of suicides among ERPO respondents.Second, Swanson et al 7 assumed that the same number of suicidal acts would have occurred had no ERPOs been issued.Third, to calculate the counterfactual number of suicide deaths, method-specific CFRs were applied to a conjectured counterfactual method distribution among the estimated total number of suicidal acts.In arriving at an estimated NNT of 10, Swanson et al 7 assumed that in the absence of ERPOs, 70% of suicidal acts by ERPO respondents would have involved firearms.For the remaining 30%, they assumed a distribution of nonfirearm methods that mirrored the (CFR-derived) nonfirearm method distribution produced in the first step.Given the method-specific CFRs for each method, this distribution corresponded to an aggregate counterfactual suicide death in the Connecticut study 7 of 93, which was 72 suicides in excess of the 21 suicides that had, in fact, occurred among the 762 ERPO respondents.The difference between the counterfactual and the observed suicide death counts is the basis for the widely cited estimate that 1 suicide death was prevented for approximately every 10 ERPOs served-that is, NNT = 1 / [(93 − 21) / 762] = 10.6-which is equivalent to an estimated causal risk difference of 9.4 percentage points (ie, 72 / 762). 1,9e NNT that Swanson et al chose to report as their only estimate in the Indiana study 8 and to exemplify in the Connecticut study 7 was an NNT of 10.In the Connecticut study, 7 the NNT of 10 depended critically on the conjecture that 70% of suicidal acts by ERPO respondents would have involved firearms in the absence of ERPOs.Had the Indiana study 8 used the 70% statistic, rather than 39% as they did in that study, their NNT would have been approximately 5. The current study updates the NNT estimates from Swanson and colleagues 7,8 by incorporating individual-level data about handgun ownership among suicide decedents in California to inform the counterfactual distribution of methods that would have been used in suicide acts by ERPO respondents in Connecticut under the counterfactual scenario that the intervention had not occurred and their firearms had not been removed.Specifically, we used individual-level data about lawful handgun ownership in California linked to cause-specific mortality for a cohort of more than 25 million adults tracked from October 18, 2004, through December 31, 2015, to identify the method-specific distribution of suicide deaths for handgun owners.Next, we derived an estimate of the aggregate number and distribution of suicidal acts by handgun owners in our California cohort (based on the observed method distribution among our suicide decedents and published sex-specific and methodspecific CFRs).Finally, following the approach used by Swanson et al, 7,8 but substituting our method-specific distributional estimate of suicidal acts for the investigators' conjectured counterfactual distribution, we provide an empirically updated estimate of the outcomes of ERPOs for the ERPO respondents in the Connecticut study, 7 expressed as the estimated number of ERPOs issued for every suicide death prevented.

Methods
This cohort study follows the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) reporting guidelines for observational studies and was approved by the institutional review board at Stanford University. 10Informed consent was not obtained because all individuals were deceased.

Firearm Ownership Data in the LongSHOT Cohort
Virtually all lawful transfers of firearms in California, including transfers between private parties, gifts, and loans, must be transacted through a licensed firearms dealer.Dealers relay details of the transfers and transferees electronically to the California Department of Justice, where the information is archived in the Dealer Record of Sale (DROS) database.People who move to California with firearms are required to report or transfer their weapons within 60 days after arrival, and these reports are also entered into the database.Although this legal regimen has governed handgun transfers for decades, transfers of long guns (rifles and shotguns) were not routinely archived until January 1, 2014. 112][13] Briefly, the LongSHOT cohort was formed by linking information on handgun transactions and all-cause mortality among adults in California to a series of historical extracts of the California Statewide Voter Registration Database (SVRD).The SVRD enumerates all registered voters in the state.Because the SVRD must be kept up to date-with new registrations, deregistrations, and changes of names and addresses-extracts present a sample of adults known to be alive and residing in California on the extract date.Using records of the 9.1 million handgun and long gun transfers recorded in the DROS database over a 32-year period (January 1, 1985, through December 31, 2016), we constructed 3 additional variables.
First, to identify cohort members who already owned a handgun at the beginning of the study period, we linked data on handgun transfers in the 19.8 years leading up to the study period (2004-2016).
Second, we created a time-varying variable that indicated the cumulative number of handguns owned (based on acquisitions and deacquisitions) and used it to identify divestments-that is, transfers of the last known handgun a cohort member owned.The age and sex of cohort members were derived from the SVRD.Data on race and ethnicity were derived from the SVRD and are categorized as Asian, Black, Hispanic, White, and other (any other US Census category not otherwise specified).Data on race and ethnicity were included in this study to provide context about gun ownership in California.Missing values for sex and race were imputed with use of validated methods

JAMA Network Open | Public Health
Updated Estimate of the Number of ERPOs Needed to Prevent 1 Suicide described in detail elsewhere. 11,14,15Because California enacted its first ERPO in 2016, we restrict our analytic sample to data from the LongSHOT cohort through December 31, 2015.

California Mortality Data
Causes of death were coded according to the International Statistical Classification of Diseases and Related Health Problems, Tenth Revision, in which suicides are specified according to method (codes X60-X84).Mortality data were linked to handgun exposure status using DROS data that indicated which cohort members acquired handguns and the dates of acquisition.

Estimation of Suicidal Acts by Handgun Ownership Status and Sex in LongSHOT
The number of suicidal acts (fatal plus nonfatal) by method was estimated by dividing the number of method-specific suicide fatalities by their corresponding method-specific CFRs. 7,16The CFR we used was based on all self-inflicted deliberate self-harm episodes that resulted in an emergency department visit (Swanson et al, 7 by contrast, used an inpatient database). 17

Respondents in Connecticut
Borrowing from the study by Swanson et al, 7 we estimated the number of suicidal acts made by the 762 ERPO respondents in Connecticut by dividing the number of method-specific suicide deaths that occurred by the corresponding method-specific CFR for male individuals and aggregating over all methods.We assumed that the same number of suicidal acts would have occurred had no ERPOs been issued and distributed these acts by method to match the method-specific distribution of suicidal acts derived above for male handgun owners in LongSHOT.Finally, we derived the counterfactual number of suicide deaths among ERPO respondents by multiplying the methodspecific CFR for male individuals by the counterfactual method-specific number of suicidal acts, and

Characteristics of Handgun Owners and Nonowners Who Died by Suicide in LongSHOT
A total of 1216 handgun owners and 15 164 nonowners died by suicide during the study period.

Method-Specific Suicide Mortality Distribution in LongSHOT by Ownership and Sex
Of the 1216 suicides among handgun owners, 88% involved firearms among male and female owners, respectively (Table 2).Suicide by hanging and by drug poisoning comprised most of the remaining suicides among male and female handgun owners.One-third (33%) of all suicides by nonowners involved firearms (40% among male nonowners and 17% among female nonowners).

Estimated Number and Method Distribution of Suicidal Acts in LongSHOT by Ownership and Sex
Of the 4871 suicidal acts we estimated had occurred among all handgun owners, we estimate 3531 were made by male handgun owners, of which 28% involved firearms, 54% drug poisoning, 3% hanging or suffocation, 2% poisoning with solids or liquids, and 9% cutting or piercing (Table 3).
Among male suicide attempters who did not own handguns, we estimated that approximately 4% of suicidal acts involved firearms.

Estimated Number of Suicide Deaths Prevented by ERPOs Among ERPO Respondents in Connecticut
We estimate that the 21 suicide deaths among the 762 ERPO respondents in Connecticut resulted from a total of 187 suicidal acts, based on the method-specific distribution of suicide deaths reported by Swanson et al 7 and the corresponding method-specific CFRs for male individuals published by Conner et al 17 (Table 4).Applying the method distribution derived in Table 3 for handgun-owning Small rural town 23 (1.9) 296 (2) Missing 0 0 a Race and ethnicity for the cohort were imputed using the method developed by Imai et al 15 for imputing these variables in voter file records.Other refers to all categories reported in US Census data, except the 4 categories specified.male individuals in LongSHOT to these 187 suicidal acts produced a counterfactual estimate that 52 suicidal acts would have involved firearms had no ERPOS had been issued (ie, 28% of 187) (Table 4).
Of these 52 suicidal acts, we estimate that 47 would have been fatal.The counterfactual number of nonfirearm suicides was derived analogously for the remaining 135 suicidal acts involving nonfirearm methods, yielding 8 nonfirearm suicide deaths.Thus, our counterfactual suicide deaths consisted of a total of 55 suicides.Comparing this estimate with the 21 that had, in fact, occurred, yields an estimated 34 suicides that were prevented among the 762 ERPO respondents, or 1 suicide for every 22 ERPOs issued over the study period (NNT = 22.4).The estimated number of method-specific suicidal acts among the 21 total suicide deaths among ERPO respondents in Swanson et al 7 (column 4) are derived by dividing the number of suicide deaths (column 2) by the method-specific CFR in column 3.
Given that the estimated method-specific number of suicidal acts sums to 187 suicidal acts, applying the method-specific suicide attempt distribution taken from Table 3 (column 5), these 187 suicides attempts are distributed across rows of suicide methods (column 6).The counterfactual number of suicide deaths by method is calculated by dividing the number of method-specific suicidal acts (column 6) by the corresponding CFR (column 3).Finally, the counterfactual total number of suicides is calculated by summing across all methods.
d Data in this column are from Table 3.

1 (
summing across methods.Data analysis was performed in December 2023 using R statistical software version 4.1.1(R Project for Statistical Computing) and Stata statistical software version 14.StataCorp). c years) and were more likely to be male (1019

Table 1 .
Characteristics of Individuals Who Died by Suicide According to Handgun Ownership Status

Table 2 .
Method-Specific Suicide Mortality Distribution by Handgun Ownership and Sex in the LongSHOT Cohort

Table 3 .
Estimated Method Distribution of Suicidal Acts by Sex in the LongSHOT Cohort

Table 4 .
17ep-by-Step Derivation of the Estimated Number of ERPOs Issued for Every Suicide Death Prevented in Connecticut, 1999-2013 Data in this column are from Swanson et al. 7b Data in this column are from Conner et al.17 a