Abstract
Extreme risk protection orders (ERPOs) are civil restraining orders that aim to prevent firearm-related harm by temporarily separating firearms from people at high and imminent risk of harming themselves or others. We examined arrests for any crime, violent crimes, and firearm-related crimes among ERPO respondents over time in relation to the ERPO. This was a retrospective cohort study of all individuals issued an ERPO in Washington state from 12/8/2016 to 10/2/2020, with follow-up on arrests through 7/1/2023. Exposures were ERPO respondent and case characteristics and time relative to ERPO filing: (1) 1 year prior to ERPO; (2) precipitating event window (4 days before-4 days after ERPO filing); (3) during the remainder of ERPO; and (4) 1 year after ERPO expiration. Outcomes were arrests for any crime, violent crime, and firearm-related crime. There were 488 ERPO respondents. Approximately 18.0% of respondents (n = 88) were arrested during the year prior to ERPO (0.72 arrests per 1000 person-days, 95% CI = 0.60-0.85), 18.0% (n = 88) during the 9-day precipitating event window (20.81 arrests per 1000 person-days, 95% CI = 16.80-25.49), 19.5% (n = 95) during the remainder of ERPO (1.07 arrests per 1000 person-days, 95% CI = 0.92-1.23), and 10.3% (n = 48) in the year after ERPO (0.46 arrests per 1000 person-days, 95% CI = 0.37-0.58). Results were similar for violent and firearm-related arrests. Results suggest that ERPOs are being used when respondents are behaving dangerously and that more can be done to leverage ERPO as an intervention point to address underlying and ongoing causes of criminalized or other high-risk behavior.
Highlights
● Arrest rates among ERPO respondents were highest during the precipitating event surrounding ERPO filing and lowest after ERPO.
● Findings suggest ERPOs in Washington are being applied to a population at risk of using violence during heightened times of risk, as intended.
● Higher rates of arrest during vs. prior to ERPO suggest opportunities to address underlying and ongoing causes of criminalized or other high-risk behavior.
Introduction
Extreme risk protection orders (ERPOs) are civil restraining orders that temporarily separate firearms from people at high and imminent risk of harming themselves or others.1,2 They are a relatively new legal tool designed to prevent interpersonal and self-directed firearm-related harm based on indicators of dangerous behavior, filling gaps in existing firearm policies and risk reduction strategies.3,4 A small but growing body of evidence suggests ERPOs may prevent firearm violence, particularly suicide, 3 and are frequently used in cases of potential mass shootings.5,6 As of May 2025, ERPO laws and similar risk-based firearm removal laws exist in 21 states, the District of Columbia, and the US Virgin Islands, with most enacted since 2018. 7 Washington state, the focus of the current study, implemented its ERPO law on 12/8/2016.
Washington’s ERPO process has been previously described. 8 Briefly, a law enforcement officer, intimate partner, family, or household member petitions the court for an ERPO by providing written evidence as to why the respondent (the person subject to the order) poses substantial risk of imminent harm to themselves or others by having a firearm. 9 Individuals need not be involved in the criminal legal system for law enforcement to petition for an ERPO for them; instead, law enforcement may decide to petition after responding to a crisis call from family members, friends, intimate partners, or others. 10 After reviewing the sworn petition in support of the ERPO and hearing any additional evidence (testimony, if necessary), a judge decides whether to grant an ex parte (temporary) ERPO, which lasts 14 days. The ex-parte hearing must occur the day of or after petition filing. If granted, law enforcement serve the ex parte order and, if applicable, ensure the respondent is safely dispossessed of firearms and any concealed pistol license for the duration of the order. A full ERPO hearing must be set within 14 days of the ex parte order being granted, and the respondent must be given notice and served with the ERPO petition and ex-parte order at least 5 court days before the scheduled full ERPO hearing. At the full hearing, the judicial officer verifies that the respondent had notice of the hearing and was properly served. A judicial officer may also hear testimony from the petitioner and respondent and any witnesses and review any written and sworn declarations or additional evidence. At the conclusion of the hearing, the court determines if the petitioner has met its burden and can issue a 1-year ERPO.
While ERPOs are civil orders, ERPO respondents may interact with the criminal legal system, including prior to the ERPO, while the ERPO is in place, and after ERPO expiration. One measure of criminal legal system interaction is arrest, which could be an indication of dangerous, criminalized behavior (though potentially differentially across the population given well-established inequities in social conditions and policing and criminal legal system outcomes) 11 ; a strategy intended to ameliorate risk of harm if there is probable cause a crime was committed (eg, assault); and/or a consequence of violating ERPO, which prohibits firearm purchase and possession. ERPOs may also be a strategy for firearm restriction during pending criminal cases. Prior research has shown high levels of criminal legal system interaction, including arrest, among ERPO respondents prior to and around the time of ERPO filing. For example, while approximately 2% of people are arrested annually in Washington, 12 ERPO petitions in Washington noted histories of arrest, conviction, or incarceration among respondents in 37% of cases. 8 In Indiana and Connecticut, 3% and 4% of respondents, respectively, were convicted of a crime occurring around the precipitating event that prompted gun removal (a similar court-sanctioned action to modern ERPO laws).13,14 In California, 33% of respondents were arrested at initial police contact or ERPO service, and 21% had prior arrests/charges. 15
Understanding patterns of arrest among ERPO respondents can help characterize those subject to ERPOs and inform issues of equity, uptake, and impact of the policy. For example, people, including family members/loved ones, policy makers, and others, may be less supportive of or willing to use ERPOs if they worry the policy will create new pathways to criminalization (ie, arrests for violations of the order), raising equity concerns. 16 Additionally, better understanding how often ERPO respondents are arrested around the time of ERPO filing can shed light on arrest as a coincident intervention with ERPO, which may influence inferential analyses of policy effects. Finally, characterizing arrests among individuals before they became an ERPO respondent as well as subsequent arrests during or after the order may inform our understanding of the potential consequences of the policy and ongoing needs of ERPO respondents to further reduce their risks.
In this state-wide study, we examined arrests for any crime, violent crimes, and firearm-related crimes among ERPO respondents prior to, during, and after ERPO, stratified by characteristics of respondents and ERPO cases which may relate to arrest patterns (eg, target of harm). No prior studies, to our knowledge, have longitudinally assessed arrests among ERPO respondents in Washington. While similar longitudinal research has been conducted in Indiana 13 and Connecticut 14 and is underway in California, 17 given variation in populations, ERPO processes, and policy provisions, it is important to examine arrests among ERPO respondents in different jurisdictions. For example, in Connecticut, medical professionals may petition for an ERPO (in addition to law enforcement and family and household members), whereas, in Indiana, only law enforcement may petition. 18 Further, compared with California, ERPOs in Washington are more often filed for concern for harm to self than others alone.8,15 Findings from this study will inform implementation and effectiveness research, public awareness of the tool, and potential refinement of ERPO policy and processes.
Methods
Study Design, Setting, and Population
This was a retrospective cohort study of individuals issued an ex-parte ERPO (some of whom ultimately received a final ERPO) in Washington from 12/8/2016 to 10/2/2020, with arrest data from 12/8/2015 to 7/1/2023. We excluded one individual for whom we could not determine whether the ex-parte order was granted.
Exposure
Exposure was time relative to ERPO: (1) 1 year prior to ERPO; (2) precipitating event (4 days before-4 days following ERPO filing); (3) during ERPO (after the precipitating event and before order expiration); and 1 year after ERPO expiration. Few respondents had gaps in time between multiple ERPOs, so we did not present results for these periods. To construct exposure periods, we used the date the ex-parte ERPO petition was filed; the date the ex-parte order expired (including any continuances); the date the final ERPO was granted (if granted); and the date the final order expired (including termination and renewals). Dates were abstracted from court records. We focused on the 9-day precipitating event window for consistency with prior research 13 and because (1) the events that lead to ERPO are not always discrete and contained, (2) it may take time to conduct an investigation and make an arrest, and (3) risk may not immediately dissipate with ERPO filing. In fact, the domestic violence literature suggests legal intervention during crises has the potential to increase risk of retaliation. 19 However, we conducted an additional analysis separating the precipitating event window by ERPO filing date.
Outcomes
Outcomes were arrests for any crime, violent crime, and firearm-related crime. As done previously,20,21 violent crimes were defined using: (1) the Federal Bureau of Investigation Uniform Crime Reporting (UCR) program definition, including murder and nonnegligent manslaughter, forcible rape, robbery, and aggravated assault (“UCR violent crime”); and (2) a more expansive definition 22 (“any violent crime”). Arrests were categorized per the Revised Code of Washington. We obtained arrest data from Washington State Patrol, which maintains all fingerprinted arrests submitted by state and municipal authorities, and probabilistically linked ERPO respondents to arrest records by name, gender, birth date and, when available, race and ethnicity using the Link King (a SAS/AF application for record linkage and unduplication) in SAS version 9.4 (SAS Institute, Cary NC).
Analysis
We calculated frequencies, percentages, and incidence rates per 1000 person-days and associated 95% confidence intervals for each period and arrest type, stratifying by the following variables abstracted from ERPO court records: respondent race and ethnicity (social constructs reflecting exposure to systemic [dis]advantage 23 ), age, gender, target of harm in the events leading to ERPO (self only, others only, both), specific targets of other-directed harm (intimate partner, law enforcement, mass shooting threat, other family, other targets), petitioner (law enforcement, family/household member or intimate partner), firearm removal pursuant to ERPO, and whether the final order was granted. As previously described, 5 records were abstracted using a standardized protocol. Cases were double-coded until staff reached inter-rater reliability of 0.80.
The denominator of rates accounts for both the number of respondents and the amount of time respondents were observed. This allows for a valid comparison of arrest risk between periods with different lengths, and the possibility that respondents had multiple arrests per period. Rates per 1000 person-days are calculated as [arrests/(number of respondents × days observed)] × 1000.
The analysis was conducted at the person level (12 individuals had multiple ERPO cases). Characteristics pertained to the first ERPO case, for those with multiple cases. We censored individuals at the end of the study or their death, whichever occurred first. Death data through 12/31/2021 (most recent available) were obtained from the Washington State Department of Health and linked using the procedures described above.
Analyses were conducted in R version 4.1.1 (R Foundation for Statistical Computing, Vienna, Austria). The University of Washington Institutional Review Board approved this study and waived consent (IRB ID: STUDY00009018; approved 1/6/2020). This study followed The Strengthening the Reporting of Observational Studies in Epidemiology guidelines for cohort studies. 24
Results
Description of Sample
There were 488 ERPO respondents (Table 1). Most were male (n = 413, 84.6%), mean age was 42.3 years (range 18-91 years), 81.8% were identified as White (n = 399), 7.0% as Black (n = 34), 5.3% as Asian (n = 26), 3.3% as having unknown race or ethnicity (n = 16), and 2% or less each as Latine, Native Hawaiian Pacific Islander, and American Indian/Alaska Native. The sex, age, and racial and ethnic distribution of ERPO respondents differed from the general population of Washington; this does not necessarily indicate disparities in appropriate use of ERPOs, however, since we do not have data on the general population for whom ERPOs would theoretically be indicated.
Description of Study Sample, ERPO Respondents in Washington State (WA).
Note. NA = data not applicable or accessible.
Measured with data from ERPO petitions and court records.
Missing for 18 respondents; there were no ERPO respondents under age 18.
Categories are not mutually exclusive. Data for ERPO respondents obtained for court records. Data for WA obtained from American Community Survey data, and reflect each race alone or in combination with other races.
Among those with a final hearing (no. = 467).
Among those with evidence of current firearm possession at the time of the ERPO (no. = 371).
From the 2016 to 2020 American Community Survey, obtained from IPUMS NHGIS. 25
Most cases were petitioned for by law enforcement (n = 439, 90.0%; Table 1). Approximately 43.0% of petitions noted respondents had a concern for harm to others only (n = 210), and 28.3% for harm to self only (n = 138). Most respondents (n = 381) were issued a final order, representing 78.1% of all respondents and 81.6% of those with a final order hearing (n = 467). There was evidence of firearm removal pursuant to ERPO for 73.0% of respondents (n = 271) with documented firearm possession (n = 371). Approximately 5.1% (n = 25) of respondents died during follow-up (up to 1 year after ERPO expiration).
Any Arrest
Swimmer plots showing arrests per person are in eFigure 1, Additional File 1. Approximately 18.03% of respondents (n = 88) were arrested during the year prior to ERPO (0.72 arrests per 1,000 person-days, 95% CI = 0.60-0.85), 18.0% (n = 88) during the 9-day precipitating event window (20.81 arrests per 1000 person days, 95% CI = 16.80-25.49), 19.5% (n = 95) during the remainder of ERPO (1.07 arrests per 1000 person days, 95% CI = 0.92-1.23), and 10.3% (n = 48) in the year after ERPO (0.46 arrests per 1000 person days, 95% CI = 0.37-0.58; Table 2). Arrest rates during the precipitating event window were higher prior to ERPO filing (29.18 arrests per 1000 person days, 95% CI = 22.15-37.72) than on/after filing (14.11 arrests per 1000 person days, 95% CI = 9.83-19.62, eTable 1, Additional File 1). While most respondents (82.0%) were not arrested during the precipitating event, respondents were more often arrested only during the pre-filing precipitating event window (11.3%) than the post (6.4%; eTable 2, Additional File 1).
Arrests Among ERPO Respondents in Washington State.
Per 1000 person-days.
Defined as the period of 4 days before to 4 days following ERPO petition filing.
Four arrests, 1 during the precipitating event, 2 during the ERPO, and 1 post-ERPO, were specifically for unlawful possession of a firearm which violated the terms of ERPO.
Stratified results showed some variation (Figure 1). For example, arrest rates during the precipitating event were lower among respondents with a family/household member petitioner (no arrests during the precipitating event) and concern for self-harm only. Asian, Black, and White ERPO respondents had lower arrest rates after versus before ERPO, and Black respondents were the only group to have lower arrest rates during versus before ERPO.
Arrest rate among ERPO respondents in Washington State by time period and respondent and case characteristics. Incidence rates are plotted on the log scale. Bars indicate 95% confidence intervals. Rates of zero are not plotted. The precipitating event is defined as the period of 4 days before to 4 days following ERPO petition filing. Target of harm refers to the precipitating event.
Violent and Firearm-Related Arrests
Patterns in arrests for any violent arrests, violent UCR arrests, and firearm-related arrests were largely similar to those for any arrest, though estimates for UCR violent crime and firearm-related crime were based on a small number of arrests (Table 2; eFigures 2-5, Additional File 1). Approximately 17.2% of arrests during the precipitating event were firearm-related, and 4 arrests were for ERPO violations (1.1% of all arrests occurring after ERPO filing).
Discussion
In this retrospective cohort study of individuals issued an ERPO in Washington state, 18.0% of respondents were arrested during the 9-day precipitating event window. This percentage was similar to periods prior to and during ERPO; however, given differences in period length, the arrest rate during the precipitating event was substantially higher. This is a meaningful finding, as the arrest rate prior to ERPO was still higher than the general population (18.2 ERPO respondents arrested per 100 in the year prior to ERPO, including 11.6 per 100 for violent crimes, vs 2.5 persons arrested per 100 in Washington in 2019), 26 but not an unexpected one, as arrest is another potential response for behaviors that may qualify someone for an ERPO. Furthermore, high rates of arrests for violent and firearm-related crimes specifically suggest respondents were exhibiting dangerous behaviors for which ERPOs are an intended response. Separating the precipitating event window by pre- and post-ERPO filing revealed that arrests were more common immediately prior to versus after ERPO filing. We also found arrest rates were lowest in the post-ERPO period and generally higher during versus prior to ERPO, although it is important to note that arrests during the ERPO may have been related to the original precipitating event.
In Indiana, 5.1% of respondents were charged with an offense during the 9-day period surrounding the event prompting gun removal, and criminal charges were more common in the year after (14.7%) than before removal (9.6%). 13 In Connecticut, 17% of respondents were arrested during the gun removal period (similar to our results), and 6% of respondents had an arrest resulting in conviction in the year before and after removal. 14 Estimates for years before and after gun removal in Indiana and Connecticut, and for the period surrounding removal in Indiana, should be lower than ours—all else equal—because we examined arrests, whereas the Connecticut and Indiana studies examined arrests leading to convictions and charges, respectively. However, all else is likely not equal across these contexts; for example, most orders in Indiana and Connecticut were for suicide risk, while most in Washington were for interpersonal violence risk (and those with self-harm concerns only had lower arrest rates). Yet, estimates for the post-order period in Indiana were higher than ours (10.3%). These differences may be explained by aspects of the laws, for example, Indiana’s law includes warrantless seizure and, under Connecticut law, a search warrant is included with granted ERPOs when there is probable cause to believe that respondents possess firearms, potentially putting respondents at greater risk of subsequent arrests. Differences may also be due to COVID-19 pandemic-related factors, as the post-ERPO period was during the pandemic for many respondents in our sample.
The elevated rate of violent and firearm-related arrest pre-ERPO and the findings that arrest rates were highest during the precipitating event and lowest in the year following ERPO could indicate that ERPOs in Washington are being applied to a population at risk of using violence during heightened times of risk, as intended. Together, the high pre-ERPO arrest rate, the fact that several respondents had multiple arrests prior to ERPO, and the finding that over 80% of respondents were not arrested at all prior to ERPO further indicate that ERPOs are filling gaps in firearm-related harm reduction strategies (eg, ERPOs are reaching people or being used to address risk in ways the criminal legal system, as measured by arrests, is not). The higher likelihood of arrest in the 4 days preceding versus following ERPO petition filing could additionally indicate that ERPOs are deflecting some from the criminal legal system toward the civil system, or that arrests immediately preceding ERPO reduced opportunity for subsequent arrest via incarceration. However, that arrest rates were higher during versus prior to ERPO (except for UCR violent crime) may indicate that additional support is needed for some ERPO respondents.
ERPOs are designed to reduce firearm-related risks by removing or restricting firearm access during dangerous situations. ERPO petitions and associated court records noted that 73% of respondents with evidence of firearm possession had firearms removed, suggesting opportunities to increase court-ordered firearm removal pursuant to ERPO. Failure to do so may contribute to elevated rates of arrest during ERPO. However, estimates of firearm removal in our data are likely a lower bound (reflecting actions recorded at the time of court adjudication), and results may be confounded if certain individuals were both more likely to have firearms removed and be arrested following ERPO.
While ERPOs are a narrow intervention focused on temporary firearm restriction, individuals subject to ERPOs often have multifaceted social, economic, and health needs (eg, substance use, housing instability) shaped by structural forces and social-ecological context. 10 These needs can themselves be criminalized 27 or they can manifest in criminalized behaviors, for example, threats or uses of violence, including firearm violence. Although ERPOs were not designed to meet such underlying needs, our findings suggest potential to better leverage ERPO as an intervention point to attend to otherwise unaddressed needs that may continue to manifest in criminal legal system involvement. This could be done via the courts (eg, judges may order behavioral health evaluations and courts can work to ensure respondents have timely and affordable access to care) or complementary supports. For example, mental health providers in police departments or crisis intervention teams can assist in connecting respondents to needed resources. Further, in King County, Washington, the Regional Domestic Violence Firearms Enforcement Unit includes an ERPO advocate who, among other things, helps family petitioners with referrals to social services to support their loved one. 28
Involving family members in coordinated care plans can help support such underlying needs, 29 and our results suggest that respondents with family/household petitioners also had lower risk of arrest during the precipitating event. This is promising but could be due to chance owing to the small number of family/household member petitioners or differences in characteristics of respondents or precipitating events (rather than petitioners). For example, precipitating events that involve law enforcement petitioners may be more likely to be “emergencies,” including cases of violence against others, in which family and household members call police.
One potential concern about ERPOs is that, while they are not criminalizing themselves, they can create new paths toward criminal legal system involvement because it is a crime to purchase or possess firearms when subject to an ERPO, and that, especially considering the racialized nature of policing in the US, arrests for such violations could deepen racial inequities. 16 We found that arrests for ERPO violations were relatively rare (n = 4, with all among White respondents), suggesting either that few individuals unlawfully possessed firearms during ERPO or police did not often learn of and arrest people for such violations. Because police often respond to crises, petition for ERPOs, and recover firearms pursuant to the ERPO, it is also conceivable that family members or others (particularly those belonging to structurally marginalized groups) seeking non-criminal intervention may worry that involving police will cause their loved one to be arrested (either for the crisis behavior leading to the precipitating event or other unlawful behavior observed around that time).16,30,31 How family members of ERPO respondents perceive arrest is an area for future research and a topic our analysis cannot address. However, we found that arrests were less common immediately after versus prior to ERPO filing, and, while sample sizes were small for many racialized groups, there was little difference in arrest rates during the precipitating event window. We further found that Black respondents—the racialized group with one of the highest pre-ERPO arrest rates—were the only group to experience lower arrest rates both during and after ERPO than before. This could indicate that ERPOs helped deflect individuals at highest risk of criminal legal system involvement toward the civil legal system, but future research is needed.
Limitations
Our study should be considered in light of its limitations. Arrests during the ERPO period may overestimate risk in that period because some arrests may have been related to the incident that originally gave rise to the ERPO, as cases may require further investigation to determine if there is a crime and if so, what laws were violated based on the evidence. Part of our study was during the COVID-19 pandemic, which may affect generalizability. We lacked mortality data for the entire study period, so mortality during and after ERPO may be underestimated. We also did not have data on incarceration or state of residence, which may affect exposure time. Incidence rates during the pre-ERPO period were calculated among those who subsequently received an ERPO, not the hypothetical cohort of people who could have received one at the start of the study. We defined the precipitating event window as the 9 days surrounding ERPO filing. Despite several reasons for this decision, the cutoff for this window is somewhat arbitrary; nevertheless, we do not expect results to material change with different precipitating event windows. Finally, we focused on arrests, which is only one of several criminal legal system interactions respondents might have. Future research should examine calls for police service (eg, related to crises or disturbances), police reports (which may have more context about incidents), and criminal charges and convictions.
Conclusion
Results of this analysis offer several valuable insights. First, ERPOs were used among people with no prior arrests and among others with multiple prior arrests, suggesting ERPOs are providing a novel risk reduction tool. Second, arrest rates among ERPO respondents were highest during the precipitating event surrounding ERPO filing and lowest after ERPO, suggesting ERPOs are being used when respondents are behaving dangerously. Third, arrest rates were highest in the days immediately preceding ERPO filing, which could indicate that ERPOs deflected some from the criminal legal systems toward the civil legal system. Fourth, elevated arrest rates during the duration of the ERPO suggest opportunities to better leverage the ERPO as an intervention point and further support respondents and prevent criminal legal system involvement. Finally, our finding that arrests for ERPO violations were rare informs conversations about potential adverse consequences of ERPOs.
Supplemental Material
sj-docx-1-inq-10.1177_00469580251349717 – Supplemental material for Arrests Among Extreme Risk Protection Order Respondents in Washington State: A Statewide Retrospective Cohort Study
Supplemental material, sj-docx-1-inq-10.1177_00469580251349717 for Arrests Among Extreme Risk Protection Order Respondents in Washington State: A Statewide Retrospective Cohort Study by Julia P. Schleimer, Nicole Asa, April Zeoli, Shannon Frattaroli and Ali Rowhani-Rahbar in INQUIRY: The Journal of Health Care Organization, Provision, and Financing
Footnotes
Acknowledgements
The authors thank the team of research associates who obtained and abstracted court records.
Ethical Considerations
The University of Washington Institutional Review Board approved this study and waived consent (IRB ID: STUDY00009018; approved 1/6/2020).
Consent to Participate
The University of Washington Institutional Review Board approved this study and waived consent (IRB ID: STUDY00009018; approved 1/6/2020).
Consent for Publication
Not applicable.
Author Contributions
JPS and ARR conceptualized and designed the study. JPS conducted data analyses and drafted the manuscript. NA contributed to data acquisition and management. All authors contributed to the interpretation of the data, critically reviewed the manuscript for important intellectual content, and gave final approval of the version to be published.
Funding
The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This project was supported by a grant from the National Collaborative on Gun Violence Research. The funder had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication. The views expressed in this manuscript are the authors’ and do not necessarily reflect the view of the National Collaborative on Gun Violence Research.
Declaration of Conflicting Interests
The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Data Availability Statement
The data that support the findings of this study are available from Washington State Patrol, but restrictions apply to the availability of these data, which were used under license for the current study, and so are not publicly available. Data are available from the authors with the permission of Washington State Patrol or directly from Washington State Patrol.
Supplemental Material
Supplemental material for this article is available online.
References
Supplementary Material
Please find the following supplemental material available below.
For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.
For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.
