The New York University Institutional Review Board deemed that the study was not considered human subjects research because the social media post data were publicly available (ie, from the public social media accounts of elected public officials). Therefore, further ethics approval was not required.

This study used Quorum, a public affairs database, to identify all X posts by US state legislators that mentioned suicide between December 1, 2023, and November 30, 2024. Quorum is a web-based public affairs software program that centralizes, in near real-time, policymakers’ social media posts and other textual artifacts of policymaking processes [24]. We limited our analysis to a 1-year period given the exploratory, cross-sectional nature of the study and used the most recent data that were available when analyses began.

Search terms were informed by extant literature [25] and developed by the study team to identify X posts that mentioned suicide and related terms. The search query used was: (“suicide” OR “suicidal” OR “suicide” OR “suicide prevention” OR “suicidal prevention” OR “suicidal ideation” OR “suicide ideation” OR “suicidality” OR “unalive” OR “unaliving” OR “killing oneself” OR “killing themselves” OR “offing” OR “sucide” OR “suicde” OR “unalive” OR “unaliving” OR “offing”). Reposts were excluded from the analysis to prevent the analytic sample from being skewed by the content of single posts that were reposted with high frequency.

Preliminary coding categories and definitions were developed based on studies related to the epidemiology of suicide policy, approaches to addressing suicide [4,26], and prior content analyses of textual news and social media data [27]. Following a review of a random sample of 50 posts by 3 study team members, the codebook was revised to update definitions and add emergent categories. This process was repeated 2 more times, with the 3 team members coding 25 posts each time, followed by codebook revisions.

The codebook included the following primary domains: characteristics of the legislator (ie, political party and state), the population that suicide was being mentioned in reference to (eg, youth and veterans), causes of suicide (eg, bullying and loneliness), solutions to prevent suicide (eg, restricting access to lethal means and increasing access to mental health services), and collateral consequences of suicide (eg, suicide contagion and bereavement). The study team also examined whether each X post was about an individual person, included statistics about suicide, cited a specific state bill about suicide, or encouraged self-help (eg, use of the 988 Lifeline). Posts were further coded according to whether they used language that was concordant with guidelines for reporting on suicide by Suicide Awareness Voices of Education (SAVE), which were collaboratively developed with input from organizations such as the American Foundation for Suicide Prevention and the US Substance Abuse and Mental Health Services Administration [28]. Examples of posts that were nonconcordant with these guidelines include those that mentioned “kill themselves” or “commit suicide,” or that made light of suicide or used the word suicide or related terms in a joking manner. The complete codebook is included in the Multimedia Appendix 1.

Three study team members coded posts using the finalized coding guide. Of 1,049 posts, a random sample of 200 (19.1%) were double coded. The mean interrater agreement was 91% (SD 11.2%) for binary (yes/no) coding categories and 88% (SD 7.6%) when including codes with multiple response options. The same 3 coders then independently coded the remaining 80.9% (n=849) of posts. This coding team met weekly to discuss any ambiguities or issues in the coding process and came to a consensus. Duplicate or near-duplicate posts identified throughout the coding process were removed from the sample to maintain consistency with the exclusion of reposts. Near-duplicate posts included instances in which a X post was uploaded a second time to correct grammar or spelling, add a hashtag, or include similar minor edits. Posts were coded using Qualtrics (Qualtrics, Inc), a web-based survey tool.

The final dataset of coded X posts was imported into SPSS (version 29.0.2.0; IBM Corp) for analysis. Univariate descriptive statistics were generated to characterize the prevalence of themes in the sample. The study team then stratified the sample according to the political party affiliation of the legislators, calculated the prevalence of themes among legislators in each party, and used chi-square tests to assess the statistical significance of differences in themes between Democrats and Republicans. Furthermore, to account for differences in the number of Democrat and Republican legislators and variability in the extent to which each party is active on X, Quorum was used to estimate the total number of posts by legislators of each party every month. Party-specific rates of suicide-related posts per 10,000 X posts were calculated. t tests were then used to assess statistical significance of differences between Democrat and Republican legislators, an approach that has been used in prior work [23].

Posts that primarily mentioned suicide in reference to war (eg, “suicide bombing”; 18/1049, 1.7% of posts) and physician-assisted suicide (88/1049, 9.3% of posts) were excluded from the analytic sample as we focused on self-directed harm only. More specifically, suicidal acts of terrorism were excluded because of the primary intent of harming others, and physician-assisted suicide was excluded because the death is not the result of self-directed harm but instead includes a lengthy process, approval, and a physician.

Over the 1-year study period, state legislators posted 955 unique X posts mentioning suicide, and 849 (88.9%) posts were included in the analytic sample after removing posts primarily mentioning terrorism-related suicide and physician-assisted suicide. A total of 573 (67.5%) posts were from Democrats, 268 (31.6%) were from Republicans, and 8 (0.9%) were from Independents. The annual suicide tweet rate per 10,000 X posts was 13.2 (0.1% of all posts) among Democrats and 7.4 (0.1% of all posts) among Republicans (P=.09). Legislators from both parties posted the most about suicide in September, which is Suicide Prevention Awareness Month (Figure 1). When September was excluded from the analysis, there was a significant difference in the rate of suicide posts between Democrats and Republicans during the rest of the year (405/394,700, 0.001% vs 219/341,100, 0.001%; P=.003).

Of 849 posts, 443 (52.2%) mentioned a specific population in reference to suicide. The most frequently mentioned populations were youth (n=186, 21.9%), veterans (n=106, 12.5%), firearms owners or people with easy access to firearms (n=103, 12.1%), and LGBTQ+ (lesbian, gay, bisexual, transgender, queer, and more) individuals (n=82, 9.7%; Table 1). Democrats were significantly more likely than Republicans to mention youth (138/573, 24.1% vs 48/268, 17.9%; P=.04) and firearm owners or people with easy access to firearms (92/573, 16.1% vs 11/268, 4.1%; P<.001), while Republicans were significantly more likely to mention LGBTQ+ individuals (36/268, 13.4% vs 45/573, 7.9%; P=.04). First responders (n=13, 1.5%) were infrequently mentioned while older adults were never mentioned (n=0, 0%).

Of 849 posts, 315 (37.1%) identified at least one factor as a cause of suicide, with no significant difference in the prevalence of a cause identified between Democrats and Republicans. The most frequent causes for suicide identified were lethal means (n=135, 15.9%), exposure to trauma (n=41, 4.8%), and public policies or other government actions (eg, anti-LGBTQ+ laws and ending no-fault divorce laws; n=41, 4.8%; Table 2). Among the 135 posts that identified lethal means as a cause, firearms were the most frequently identified method (n=111, 82.1%), followed by poisoning (n=13, 9.3%), and jumping (n=10, 7.1%). Firearms were identified as a cause of suicide significantly more frequently among Democrats (115/573, 20.1%) than Republicans (20/268, 7.5%; P<.001). Democrats were also significantly more likely than Republicans to identify causes related to barriers to accessing mental health services (17/573, 3% vs 2/268, 0.7%; P=.047), public policies or other government actions (34/573, 5.9% vs 7/268, 2.6%; P=.04), and any lethal means (115/573, 20.1% vs 20/268, 7.5%; P<.001).

In contrast, Republicans were significantly more likely than Democrats to identify social media (9/268, 3.4% vs 5/573, 0.9%; P=.009) and substance use or addiction (7/268, 2.6% vs 4/573, 0.7%; P=.04) as a cause of suicide. Republicans were also significantly more likely identify gender-affirming care as a cause of suicide than Democrats (30/268, 12.6% vs 1/573, 0.2%; P<.001).

Of 849 posts, 558 (65.7%) identified at least one solution to prevent suicide, with Democrats being significantly more likely to identify a solution than Republicans (443/573, 77.3% vs 114/268, 42.5%; P<.001). The most frequently identified solutions were education or public awareness campaigns (eg, recognizing World Mental Health Day; n=299, 35.2%), policy changes (n=198, 23.3%), the 988 Suicide and Crisis Lifeline (n=175, 20.6%), help seeking not explicitly related to 988 (n=169, 19.9%), and restricting access to lethal means (n=105, 12.4%; Table 3).

Democrats were significantly more likely than Republicans to identify the 988 Suicide and Crisis Lifeline (140/573, 24.4% vs 34/268, 12.7%; P<.001) as well as other crisis lines (34/573, 5.9% vs 2/268, 0.7%; P<.001). Democrats were also significantly more likely to identify policy changes as a solution (149/573, 26% vs 49/268, 18.3%; P=.01) as well as restricting access to lethal means (93/573, 16.2% vs 12/268, 4.5%; P<.001).

Collateral consequences of suicide were only identified in 1.4% (12/849) of posts, with bereavement being identified as the consequence in most (9/12, 75%) of these posts. There was no significant difference in frequency of collateral consequences being identified between Democrats (8/573, 1.4%) and Republicans (4/268, 1.5%; P=.91).

Of 849 posts, 78 (9.2%) used the word suicide in a way that was nonconcordant with the SAVE guidelines. Guideline nonconcordant posts were significantly less frequent in the posts of Democrats than Republicans (16/573, 2.8% vs 57/268, 21.3%; P<.001).

This study characterized the frequency and content of state legislators’ X posts mentioning suicide between December 1, 2023, and November 30, 2024. Consistent with research showing that suicide is the second leading cause of death among young people, youth were the most frequently identified population [29]. Other high-concern groups, such as veterans, firearm owners, and LGBTQ+ individuals, were also frequently identified. Populations with disproportionately high suicide rates, such as older adults and first responders [30-33], were either mentioned infrequently or not mentioned at all. This finding highlights a need to better educate state legislators about the epidemiology of suicide among these high-risk populations.

Legislators’ X posts mentioned solutions to address suicide more frequently than causes of suicide, which is consistent with the fact that legislators have the ability to address social problems via policymaking. Two-thirds of posts identified solutions to address suicide, with Democrats being significantly more likely to identify a solution than Republicans. This partisan difference was also observed for specific solutions, such as policy changes, implementation of the 988 Lifeline, and restricting access to lethal means. The most common solutions identified that shared bipartisan support were education or public awareness campaigns and help seeking in general.

For both Democrats and Republicans, the most frequent type of post was one recognizing an awareness campaign such as National Suicide Month or World Mental Health Day. While there are potential benefits to state legislators posting around a specific awareness day or month, such actions are more symbolic than substantive and do not reflect the power that state legislators have to help address risk factors for suicide [34-36]. Critically, state legislators’ ability to help prevent suicide lies in their ability to introduce and vote in favor of bills that address proximal risk factors for suicide and preventive interventions, such as limiting access to lethal means, improving access to mental health care, and increasing funding for the 988 Suicide and Crisis Lifeline [4]. Additionally, as was identified in a 2025 systematic review, there are many public policies that can help affect distal, structural determinants of suicide risk (eg, by reducing economic stress through increases in the minimum wage rate) [4]. These types of policies were rarely or never mentioned in legislators’ posts mentioning suicide. This suggests a need for improved dissemination of evidence to policymakers about the range of policies that have the potential to prevent suicide.

Democrats posted significantly more than Republicans about firearm owners or people with easy access to firearms, as well as lethal means restriction as a solution to prevent suicide. This is consistent with evidence indicating that Republicans are generally stronger supporters of the Second Amendment and have higher rates of firearm ownership than Democrats [37], and it follows that Republican legislators would be less focused on firearm ownership and access as they pertain to suicide. Suicide prevention dissemination campaigns may benefit from tailoring their language to Republican legislators and voters by focusing on evidence related to firearms and suicide. Democrats’ posts, however, are more consistent with research evidence, which suggests the potential importance of lethal means restriction and safe storage [4].

Nearly 10% (78/849) of posts used the word suicide in a way that was nonconcordant with the guidelines of SAVE. This finding suggests that policymakers would benefit from consultations with organizations such as the Suicide Prevention Resource Center or the American Foundation for Suicide Prevention to improve their use of appropriate and responsible suicide-related language.

This study is the first to examine state legislators’ social media posts mentioning suicide. However, the study has some limitations. All posts analyzed were from the period after Elon Musk purchased X, and it is possible that some legislators, especially Democrats, left the platform due to this change in leadership. Additionally, because not all state legislators are on social media, let alone X, the study is not a completely comprehensive assessment of how state legislators publicly discuss suicide. Furthermore, state legislators’ themselves may not be the ones posting social media content and may instead delegate posting responsibilities to social media management legislative aides who write and post the messages.

This is the first study to document the frequency and content of state legislators’ social media posts about suicide. When considered within the context of research about the epidemiology of suicide and evidence for policies that can prevent suicide, this study signals a need to improve communication pertaining to suicide to state legislators and encourage collaboration with suicide prevention organizations and researchers. Furthermore, given differences observed between Democrat and Republican legislators, the study findings suggest potential value in tailoring messages about suicide for legislators based on their political party affiliation.