Extreme Risk Protection Orders in Older Adults in Six U.S. States: A Descriptive Study

Abstract

Objectives:

Extreme Risk Protection Orders (ERPOs) allow a court to restrict firearm access for individuals (“respondents”) at imminent risk of harm to self/others. Little is known about ERPOs use for older adults, a population with higher rates of suicide and dementia.

Methods:

We abstracted ERPO cases through June 30, 2020, from California, Colorado, Connecticut, Florida, Maryland, and Washington. We restricted our analysis to petitions for older (≥65 years) respondents, stratified by documented cognitive impairment.

Results:

Among 6,699 ERPO petitions, 672 (10.0%) were for older adults; 13.7% (n = 92) of these noted cognitive impairment. Most were white (75.7%) men (90.2%). Cognitively impaired (vs. non-impaired) respondents were older (mean age 78.2 vs 72.7 years) and more likely to have documented irrational/erratic behavior (30.4% vs 15.7%), but less likely to have documented suicidality (33.7% vs 55.0%). At the time of the petition, 56.2% of older adult respondents had documented firearm access (median accessible firearms = 3, range 1–160).

Conclusions:

Approximately 14% of ERPO petitions for older adults involved cognitive impairment; one-third of these noted suicide risk. Studies examining ERPO implementation across states may inform usage and awareness.

Clinical Implications:

ERPOs may reduce firearm access among older adults with cognitive impairment, suicidality, or risk of violence.

Introduction

Firearm-related death among older adults is predominantly due to suicide, as suicide rates rise with age and a higher proportion of suicides among older adults are by firearm (CDC, 2023; Price & Khubchandani, 2021). Major neurocognitive disorders, commonly referred to as dementia, are a risk factor for suicide in certain populations (e.g., those with psychiatric comorbidity) and in the months after dementia diagnosis (Alothman et al., 2022; Choi et al., 2021; Conejero et al., 2018). In a multi-country study in Europe, suicide ideation among individuals with home-dwelling individuals with dementia ranged from 6% to 24%, particularly in individuals with depression or depressive symptoms (Holmstrand et al., 2021). Dementia can also increase the risk of committing interpersonal firearm violence, particularly when there are dementia-related behavior changes (e.g., aggression or paranoia) (Polzer et al., 2022; Rosen et al., 2019).

In a large survey of U.S. adults aged 50–80 years, approximately one-quarter (26.8%) owned at least one firearm, and most (80.6%) supported removing firearms from older adults with confusion or dementia (Carter et al., 2022). Older firearm owners prefer to make their own decisions about firearm access (Betz et al., 2022), yet in a prior survey, only one-fifth of older firearm owners had made a plan for securing, removing, or transferring firearms if they became unsafe to handle them (Betz et al., 2021). The aging of the U.S. population and the growing number of individuals living with dementia have highlighted the importance of addressing firearm access for older adults with a risk of suicide, homicide, or unintentional injury (Doucette et al., 2020; Prater et al., 2022; Simonetti et al., 2022).

In cases where counseling or voluntary methods of reducing firearm access for persons at elevated risk of firearm-related harm fail, an Extreme Risk Protection Order (ERPO) may be another option. ERPOs, or “red flag laws,” are civil court orders that temporarily prohibit firearm purchase and possession by people deemed at imminent risk of harm to themselves or others (Swanson, 2021). As of summer 2023, ERPO-type laws had been passed in 21 states and the District of Columbia. Under these laws, a “petitioner” (generally, law enforcement or a family member, depending on state statute) can submit evidence that a “respondent” is at imminent risk. If a court grants the petition, the respondent (the at-risk individual named in the ERPO) must relinquish their firearms and is not allowed to purchase or otherwise obtain firearms while the ERPO is in effect. Generally, states allow an emergency hearing for a temporary ERPO (which is put in place before a full hearing and lasts for 7 days to 1 month, depending on the state), followed by a full hearing for an ERPO lasting up to one to 5 years. ERPO details, including allowable petitioners and order lengths, vary across state (Extreme Risk Protection Order: A Tool to Save Lives, 2022).

Despite the risk of suicide or unintentional firearm injury among older adults, little is known about how ERPOs have been used with this population, including in the context of cognitive impairment. One prior study examined older adult suicide rates (firearm, non-firearm, and total) from 2012 to 2016 in association with total firearm-related laws and with ERPO-type laws; for the period of analysis, four states (California, Connecticut, Indiana, and Washington) had ERPO laws (Saadi et al., 2020). In that analysis, there was a negative association between the total number of firearm-related laws (including ERPO-type laws) and suicide rates among adults 55–64 and >65 years. ERPO-type laws were associated with lower total suicide rates (Saadi et al., 2020). However, that study did not examine the reasons for initiating ERPO petitions for older adults, or how often cognitive impairment was noted. A single state (Washington) study examined ERPO petitions (2016–19) that cited cognitive impairment (n = 9) in describing the respondent and identified common themes (Prater et al., 2021). Findings included that firearms were often stored unsafely (e.g., on the respondents’ person or unlocked), and the case files often noted psychotic disturbances (e.g., paranoia, delusions, hallucinations), substance abuse, domestic violence, or suicidal ideation (Prater et al., 2021).

Here, we sought to expand the knowledge base concerning how ERPOs are used in cases of older adult respondents, including in the presence of cognitive impairment. An understanding of current patterns of use across multiple states can inform future policy, implementation, and education efforts.

Methods

As described previously (Zeoli et al., 2022), we abstracted ERPO petitions filed through June 30, 2020, beginning from the time that laws were first implemented in: California (January 1, 2016); Colorado (January 1, 2020); Connecticut (October 1, 2013); Florida (March 3, 2018); Maryland (October 1, 2018); and Washington (December 8, 2016). These states were chosen for geographic diversity and varying policy and cultural contexts.

For each state, study staff obtained ERPO court case files (e.g., ERPO petition form, attached affidavits, firearm ownership documents, and other court documents) via state-specific processes. Staff reviewed all available case information and abstracted data into a centralized, secure database. Staff were trained on abstraction, and each staff member double-coded with a lead coder until there was an inter-rater reliability score of 0.80. We coded all cases through June 30, 2020, except for Florida, where the volume of cases exceeded our available personnel resources for abstraction. For Florida, we randomly sampled 50% of cases to code in counties with more than 10 ERPO cases and coded 100% of cases where the number was less than 10.

We abstracted the following data elements: respondent demographics (age, gender, race, ethnicity); petitioner type (e.g., law enforcement officer [LEO], intimate partner, other family member); reason for the petition (threats or acts against self, others or both); documented mental health concerns or cognitive impairment; documentation of firearm possession and/or dispossession; and outcome of the petition (e.g., ex parte hearing requested, outcome of ex parte and full hearings). For mental health or cognitive concerns, staff completed separate variables for various concerns, marking “yes” if there was any mention in available documentation. ERPO court forms varied by state and sometimes within states; most of the available information came from the court documents for the initial ERPO petition and had been included to demonstrate risk. Petitions did not include comprehensive medical or mental health history, and we did not have access to respondent medical records. Elements were missing in some cases across states, most commonly because the information had not been included in the original petition; in Colorado, some counties redacted data so elements were not visible.

We restricted this analysis to petitions for older (≥65 years) respondents, based on the documented age at the time of petition filing. We stratified the sample on documented presence or absence of cognitive impairment, which included specific documentation within the petition of cognitive decline, age-related cognitive decline, or impaired cognitive abilities from unstated cause.

For analysis, data were combined across the six states. Variables were summarized for all data and stratified by cognitive impairment status. Categorical variables were reported with frequencies and percentages, and continuous variables were reported with means (standard deviations) and medians (minimum, maximum) to comprehensively summarize distributions. Differences between the two cognitive groups were tested with Fisher’s exact tests for categorical variables due to some small cell sizes and two-sample t-tests for differences in the means of continuous variables. Data cleaning and analysis were performed in R version 4.2.1 (Vienna, Austria). This study was approved by the Michigan State University Institutional Review Board.

Results

Among 6,699 ERPO petitions filed in six states through June 30, 2020, 672 (10.0%) of petitions were for older adult respondents aged ≥65 years. All subsequent analyses were restricted to these 672 petitions.

Most respondents were aged 65–74 years (61.0%), white (75.7%), and men (90.2%); 10.9% were veterans (Table 1). Nearly one-fifth (18.9%) of petitions noted alcohol use by the respondent that was part of the event precipitating the ERPO petition. Most (82.3%) petitions noted a mental health or cognitive concern, including (≥1 allowed): a general reference to mental illness (57.1%); past psychiatric hospitalizations, actual mental illness diagnosis, or evidence of being prescribed psychiatric medication (25.3%), or irrational or erratic behavior (21.5%; Table 2).

Table 1.

Characteristics of ERPO respondents aged ≥65 years (n = 672), by documented presence (N = 92) or absence (n = 580) of cognitive impairment.

	Total n (%)	Cognitive impairment n (%)	No impairment n (%)	P valuea
Age group (years)				<0.001
65–74	410 (61.0)	29 (31.5)	381 (65.7)
75–84	201 (29.9)	47 (51.1)	154 (26.6)
85+	61 (9.1)	16 (17.4)	45 (7.8)
Female	64 (9.5)	8 (8.7)	56 (9.7)	0.348
Race/ethnicityb
White	509 (75.7)	68 (73.9)	441 (76.0)	0.695
Black/African American	46 (6.8)	6 (6.5)	40 (6.9)	1
American Indian/Alaska Native	0 (0.0)	0 (0.0)	0 (0.0)	-
Asian	4 (0.6)	0 (0.0)	4 (0.7)	1
Native Hawaiian/Pacific Islander	1 (0.1)	1 (1.1)	0 (0.0)	0.137
Latino/Latina/Latinx/Hispanic	9 (1.3)	1 (1.1)	8 (1.4)	1
Other/Unknown	105 (15.6)	16 (17.4)	89 (15.3)	0.643
Veteran	73 (10.9)	11 (12.0)	62 (10.7)	0.719
Documented mental health/cognitive concernb	553 (82.3)	92 (100.0)	461 (79.5)	<0.001
History of psychiatric diagnosis and/or treatment	140 (25.3)	24 (26.1)	116 (25.2)	0.896
Other concern about mental health	316 (57.1)	12 (13.0)	304 (65.9)	<0.001
Irrational or erratic behavior	119 (21.5)	28 (30.4)	91 (19.7)	0.026
Impaired from medical illness other than dementia	15 (2.7)	2 (2.2)	13 (2.8)	1
State of petition				0.004
California	95 (14.1)	13 (14.1)	82 (14.1)
Colorado	2 (0.3)	0 (0.0)	2 (0.3)
Connecticut	203 (30.2)	34 (37.0)	169 (29.1)
Florida	234 (34.8)	22 (23.9)	212 (36.6)
Maryland	80 (11.9)	6 (6.5)	74 (12.8)
Washington	58 (8.6)	17 (18.5)	41 (7.1)

Missing data not shown if < 5%.

^aP values from two-sample t tests for continuous variables and Fisher’s exact tests for categorical variables due to small cell sizes.

^b≥1 allowed.

Table 2.

Details of ERPO petition among respondents aged ≥65 years (n = 672), by documented presence (N = 92) or absence (n = 580) of cognitive impairment.

	Total n (%)	Cognitive impairment n (%)	No impairment n (%)	P valuea
Precipitating eventc
ERPO petitioner (excluding FL and CT; n=235)b				0.001
Intimate partner	16 (6.8)	4 (11.1)	12 (6.0)
Other immediate family	16 (6.8)	8 (22.2)	8 (4.0)
Health professional	0 (0.0)	0 (0.0)	0 (0.0)
Other/unknown	7 (3.0)	1 (2.8)	6 (3.0)
Law enforcement	196 (83.4)	23 (63.9)	173 (86.9)
If law enforcement petition, alerted by Intimate partner	113 (17.9)	19 (24.1)	94 (17.0)	0.156
Other immediate family (non-intimate partner)	65 (10.3)	9 (11.4)	56 (10.1)	0.693
Health professional	57 (9.0)	6 (7.6)	51 (9.2)	0.834
Friend	13 (2.1)	0 (0.0)	13 (2.3)	0.387
Self	83 (13.1)	14 (17.7)	69 (12.5)	0.212
Employer/Employee/Coworker	14 (2.2)	1 (1.3)	13 (2.3)	1
Other	124 (19.6)	13 (16.5)	111 (20.0)	0.545
Unknown	167 (26.4)	17 (21.5)	150 (27.1)	0.341
Suicide threat/act	350 (52.1)	31 (33.7)	319 (55.0)	<0.001
Use of interpersonal violence	132 (19.6)	24 (26.1)	108 (18.6)	0.119
Threats of interpersonal violence	345 (51.3)	56 (60.9)	289 (49.8)	0.056
Threat of mass shooting	44 (6.5)	3 (3.3)	41 (7.1)	0.254
Unlawful/reckless firearm use	223 (33.2)	32 (34.8)	191 (32.9)	0.722
State proceedings for mental health	414 (61.6)	56 (60.9)	358 (61.7)	0.908
Firearms at time of petitionb
Evidence of possession	378 (56.2)	57 (62.0)	321 (55.3)	0.259
Recently dispossessed as part of ERPO petition process	301 (44.8)	37 (40.2)	264 (45.5)	0.368
Dispossessed by law enforcement before precipitating event	6 (0.9)	2 (2.2)	4 (0.7)	0.193
Recently gave their firearms to someone else	16 (2.4)	3 (3.3)	13 (2.2)	0.471
Does not possess or have access to a firearm	24 (3.6)	2 (2.2)	22 (3.8)	0.760
Has access to, but does not possess, a firearm	7 (1.0)	1 (1.1)	6 (1.0)	1
Possession unknown	22 (3.3)	1 (1.1)	21 (3.6)	0.342
Petition indicated how many firearms respondent possessed/could access before petition filed	590 (87.8)	81 (88.0)	509 (87.8)	1
Median firearms (min, max)	3 (1, 160)	3 (1, 38)	3 (1, 160)
Mean firearms (SD)	5.9 (11.9)	5.6 (7.4)	5.9 (12.4)	0.717

^aP values from two-sample t tests for continuous variables and Fisher’s exact tests for categorical variables due to small cell sizes.

^bIn Florida and Connecticut, only law enforcement could submit an ERPO petition during the study period.

^c≥1 allowed.

ERPO petitioners were usually law enforcement officers (83.4%), with intimate partners or other family members accounting for most other petitioners (Table 2). For petitions filed by law enforcement, intimate partners or immediate family members were often documented as the individuals who called for law enforcement assistance (Table 2). Petitions for older adults with cognitive impairment, versus those without, were more likely to come from intimate partners or other immediate family, highlighting likely involvement of dementia caregivers (Table 2).

Half of all ERPO petitions noted suicide risk (52.1%) or threats of violence to others (51.3%; Table 2) as the precipitating event; among these threats of violence to others, most were toward intimate partners (26.4%) or other adult family members (11.9%). One in five ERPO petitions (19.6%) documented a history of physical violence, again most often toward intimate partners (42.9%) or other adult family members (12.1%). One-third (33.2%) of older adult ERPO petitions noted unlawful or reckless use of firearms.

Half (56.2%) of ERPO petitions for older adults had documented evidence of firearm possession at the time of the petition; another 44.8% documented that the respondent had ≥1 firearm removed as part of the ERPO petition process (Table 2). Among those respondents with access to firearms, most (87.8%) petitions noted the number of firearms the respondent possessed or was able to access, and the number of firearms ranged from 1 to 160 (median 3, mean 5.9 [SD 11.9]).

Cognitive impairment was noted in 13.7% (n = 92) of petitions. Washington had the highest proportion of older adult ERPO petitions for cognitive impairment (17 out of 58 petitions in Washington; 29.3%), while Colorado had none (Table 1). Cognitively impaired (vs. non-impaired) respondents were older (mean age 78.2 vs 72.7 years, p < .001) and more likely to exhibit irrational or erratic behaviors according to the petition (30.4% vs 15.7%, p < .001). Their petitions were less likely to note alcohol use (9.8% vs 20.3%, p = .015) or suicidality (33.7% vs 55.0%, p < .001) compared with petitions for non-cognitively impaired respondents. Petitions for cognitively impaired (versus non-impaired) respondents were more likely to come from intimate partners (11.1% vs 6.0%) or other immediate family (22.2% vs 4.0%; p = .001). Among ERPO petitions that resulted in a final hearing, those involving cognitive impairment (vs. not) were not more likely to be granted (55.1% vs 57.6%, p = .631; Figure 1).

Figure 1.

Outcomes of ERPO petitions among respondents aged ≥65 years (n = 672), by documented presence (N = 92) or absence (n = 580) of cognitive impairment.

Discussion

In this multi-state study of ERPO use, older adults accounted for 10.0% of all respondents, which is lower than the population prevalence of older adults in those states in 2020 (ranging from 14.7% in California and Connecticut to 20.7% in Florida) (USAFacts, 2022). While firearm ownership rates among older adults in those states are unclear, older adults have higher rates of firearm suicide compared to younger adults (CDC, 2023), raising questions about whether ERPOs are being underused for older adults. It may be that older adults with suicide risk – and their families – pursue voluntary approaches rather than requesting ERPOs. It is also possible that ERPOs are requested less often in older adults because of ageist views around late-life depression or suicide (Schmutte et al., 2009) or because of misconceptions about the inevitability of suicide, which might lead healthcare providers, family, and other individuals to be less proactive in older adult suicide prevention when compared to younger adults and adolescents (Van Orden & Deming, 2018).

Approximately half of ERPO petitions for older adults noted threats of suicide and half noted threats of violence toward others; the latter was more common in the 13.7% of petitions noting cognitive impairment. This is consistent with the behavioral changes (including paranoia and aggression) that can occur with dementia (Volicer, 2018) and might prompt requests for an ERPO. Prior work has demonstrated that dementia caregivers are at risk of victimization, with approximately 20% of home-based caregivers experiencing violence or aggression (O’Leary et al., 2005). Within the general population, the annual prevalence of interpersonal violence perpetrated by an individual with dementia is 9.7 to 32.6 per 1,000 persons (Rosen et al., 2019). Attempting to reduce access to a firearm (e.g., hiding it or moving it out of the home) may prompt negative reactions from individuals with dementia (Polzer et al., 2022). This may make a legal option like an ERPO – with police assistance in firearm removal – particularly appealing to some caregivers. Indeed, in our multi-state study, petitions noting cognitive impairment were more likely to come from intimate partners or immediate family, rather than law enforcement, highlighting family involvement. Our findings are consistent with a prior survey in California, in which 70.5% of participants said they would be somewhat or very willing to request an ERPO in the case of severe dementia in a family member (Kravitz-Wirtz et al., 2021).

In this study, ERPO petitions that involved documented cognitive impairment were surprisingly not more likely to be extended after a hearing. Dementia-related cognitive impairment is generally irreversible and progressive and merits long-term reductions in firearm access (Betz et al., 2018). In other situations, risk might be abated through appropriate medical treatment (e.g., for mental illness or substance use disorder) and restoration of firearm access may be safe and appropriate. Available court records did not allow full exploration of the circumstances surrounding extension or not; it may be that, in cases of dementia, family had taken additional measures (e.g., movement into a facility) and did not require a full ERPO. Further exploration of ERPO usage, including continuation and outcomes over time, is warranted.

Moving forward, continued attention to implementation of ERPOs will be essential. Prior research has shown that ERPO uptake (e.g., awareness and use by the public and law enforcement) varies geographically, including by urbanicity and political ideology or declarations such as “Second Amendment Sanctuary” status of an area (Barnard et al., 2021; Pear et al., 2021). The education of the general public, law enforcement, judges, administrators, and other stakeholders is important and should include specific mention of how ERPOs can remove access to firearms from older adults at elevated risk of violence, whether from suicide risk, dementia, or violence against others. Coordination with specific organizations may facilitate dissemination of this information; examples include groups focused on dementia, caregiving, and older adult well-being. Future research should examine how implementation differences may affect ERPO use and violence outcomes.

Limitations of this study include that the ERPO laws were available for different lengths of time across states, making state-to-state comparisons difficult. Cognitive impairment may have been under-identified by the reliance on the petitions themselves; there may have been cases where an individual had cognitive impairment (diagnosed or undiagnosed) that was not known or noted by the petitioner. Additional details of interest, such as details on the cognitive concerns or whether the respondent lived with family, may also not have been captured by the ERPO court case files.

This is the first multi-state study examining ERPO use among older adults. It provides useful baseline information for understanding current ERPO patterns that can inform implementation efforts and generates questions for future studies.

Clinical Implications.

• Older adults have elevated rates of firearm suicide, and dementia can increase the risk of firearm-related injury or death in certain populations.

• Extreme Risk Protection Orders (ERPOs), available in many states, have been used for older adults with documented imminent risk of harm to themselves (1 in 2) or others (1 in 2). • To date, nearly 1 in 6 ERPOs for older adults involved cognitive impairment.

• In states with ERPOs, clinicians should consider educating families of at-risk patients about ERPO availability.

Data availability statement

A limited dataset, not including data from the state of Maryland, will be made available on Open Science Framework after publication of project-related publications.