Introduction

Violence, both interpersonal and self-directed, is an enduring public health problem in the USA. In 2019 alone, 66,652 people died from violence, incurring approximately $672 billion in costs to society [1]. However, violence does not affect all communities equally, and a considerable amount of research reveals disproportionate rates of violence affecting people who identify as lesbian, gay, bisexual, or transgender (LGBT). In recent years, the LGBT abbreviation has broadened to sexual and gender minority (SGM) to encompass the heterogeneity of identities and experience, such as people who identify as gender non-binary or gender non-conforming or identify as queer or pansexual [2]. For the purposes of this review, SGM will be used unless referring to studies that focused on specific sub-populations.

Interpersonal violence against SGM people, driven by bias, is a well-known phenomenon [3], but most research has been limited to convenience-based sampling. From Miller and Humphries’ initial attempt in 1980 to stoke empirical study in gay men’s victimization [4], nearly four decades would pass before sexual orientation data were gathered in the National Crime Victimization Survey (NCVS) for the first time in 2017. The results from the NCVS [5], utilizing robust nationally representative sampling, corroborated findings from numerous studies about interpersonal violence among sexual minorities gathered through convenience samples.

Similarly for self-directed violence, in 1999, Remafedi questioned whether the scientific community could end equivocal questions about disparities in suicide risk for sexual minorities, with the evidence at that time seemingly compelling and concordant [6]. Here again, it would take 16 years for sexual orientation data collection to be added to the National Survey of Drug Use and Health [7], the only ongoing population-based survey in the USA that includes surveillance of suicidal ideation and attempt. Concomitantly, additions of sexual orientation items to the Centers for Disease Control and Prevention’s (CDC) Youth Risk Behavior Survey in 2015 finally equipped researchers to examine both suicidal ideation, suicide attempt, and peer victimization among sexual minority adolescents [8].

Thus, disparities in violence have been clarified through the gold standard of probability-based sampling, but largely only within the last 5 years in the USA and after decades of research by scientists forced to use convenience samples due to lack of available data from national surveys [9]. Consequently, framing violence research among SGM populations remains unclear despite a flurry of individual studies and seemingly numerous reviews of them.

One framework for health disparities research created by Kilbourne and colleagues outlines a three-generation approach [10]. In the first generation, disparities are detected, i.e., evidence that a disparity exists. Subsequent second-generation studies aim to understand the factors driving disparities, which then informs third-generation studies that target those driving forces through interventions to reduce the disparities. By placing research studies along this continuum, one can observe both where progress occurs and where research seemingly has stalled.

This scoping review was guided with the question “What is the breadth of research reviews about violence among SGM populations?” There were two main reasons for conducting a scoping review rather than a systematic review of reviews. First, the intent of the review was not to answer specific questions about prevalence or incidence of violence or effectiveness of interventions. Second, within the two main categories of interpersonal and self-directed violence, there are further categories of violence, (e.g., within interpersonal violence, there is intimate partner violence, peer victimization, childhood abuse). Thus, a scoping review aligned best with an endeavor “to provide an overview or map of the evidence.”[11]

Methods

The author conducted an initial search on December 1, 2021 to review titles and abstracts and repeated the search on February 1, 2022 to assure no new reviews had been published in the time during the manuscript development. January 1, 1990 was selected as the starting point because it was unlikely that the literature on SGM individuals was populated or developed enough by that time point to lend itself for reviews. A simultaneous search of several databases was conducted, including Scopus, IngentaConnect, Medline, ProQuest, SAGE Premier, Web of Science, JSTOR, and LGBTQ + Source.

Based on the overarching research question, the literature search consisted of three main terms for: population (“sexual minority” OR “sexual minorities” OR “gender minority” OR “gender minorities” OR transgender OR nonconform* OR lesbian* OR gay* OR bisexual* OR lgb* OR “men who have sex with men” OR “women who have sex with women” OR MSM OR WSW OR “same-sex” OR “sexual orientation” OR “gender identity”), type of study (review OR meta-analysis), and topical focus (violen* OR abuse OR victim* OR suic* OR harm OR injury OR assault OR crime OR injury OR homicide).

Inclusion criteria were as follows: (a) must be a scientific review (e.g., systematic, meta-analysis, scoping); (b) explained search criteria (e.g., databases, search terms, time period searched); (c) written in English; (d) published in a peer-reviewed journal. Despite limiting inclusion to studies published in English, there was no exclusion based on country or locale. The references of included articles were scanned for any studies potentially missed in the initial search.

In addition to key characteristics of each review to assess the breadth of research (e.g., years of search, number of studies included, countries included in the review), each review was coded regarding whether its scope aligned with first-generation (i.e., documenting), second-generation (i.e., understanding), or third-generation (i.e., reducing) health disparities research [10]. Lastly, key findings of each review are summarized based on data supplied in each original study: for first-generation studies, summaries of prevalence were extracted; for second-generation studies, examples of risk factors identified by each review were extracted; for third-generation studies, narrative summaries of findings were extracted.

Results

The search produced 431 results, and after de-duplication, there were 375 unique citations to review. After reviewing the titles and abstracts, 293 were not scientific reviews (e.g., book reviews), leaving 82 papers for full-text review of which 29 did not meet the inclusion criteria. Fifty-three reviews met the inclusion criteria, and 18 additional reviews were recovered from those papers’ reference lists and met the inclusion criteria, producing a total of 73 review studies (Fig. 1). One review included outcomes for both interpersonal and self-directed violence [12], so that study was included within both of the two major categories of violence. In total, there were 32 reviews related to self-directed violence [12,13,14,15,16,17,18,19,20,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36,37,38,39,40,41,42,43] and 42 reviews related to interpersonal violence [12, 44,45,46,47,48,49,50,51,52,53,54,55,56,57,58,59,60,61,62,63,64,65,66,67,68,69,70,71,72,73,74,75,76,77,78,79,80,81,82,83,84].

Fig. 1
figure 1

Search and screening process

Full size image

In terms of the type of disparities research, the majority of reviews focused on summarizing first generation research (Tables 1 and 2). For self-directed violence, 69% were first generation, 53% were second generation, and 6% were third generation. Interpersonal violence reviews followed a similar cascade, with 67% first generation, 43% second generation, and 7% third generation. Within the interpersonal violence reviews, 6 reviews (14%) could not be categorized within the generations of disparities research framework because their foci were either summaries of methodologies (rather than prevalence, risk factors/correlates, or interventions) or theoretical synthesis of reviews [53, 65, 71, 75, 82, 83].

Table 1 Reviews of self-directed violence among sexual and gender minority populations, 1990–2022
Full size table
Table 2 Reviews of interpersonal violence among sexual and gender minority populations, 1990–2022
Full size table

For specific topics within each of the two major types of violence, most reviews in self-directed violence combined suicidal ideation and attempt (n = 18; 56%), and most reviews of interpersonal violence focused on intimate partner violence (n = 20; 48%). The majority of reviews across both major types of violence included studies between 2000 and 2020. To better depict the breadth of current reviews, Supplemental Figs. 2 and 3 illustrate reviews according to type of violence and timespans of the review by specific population.

Tables 1 and 2 also depict that although there was a varied landscape of risk factors identified across reviews, some risk factors were applicable to all populations (e.g., substance use, depression, history of victimization). Other reviews highlighted risk factors that were more unique to SGM populations, which most centered around minority stressors (e.g., family rejection, internalized homophobia or transphobia, discrimination). The scant reviews of intervention studies were concordant in emphasizing the overall lack of research for addressing violence-related health disparities for SGM populations.

Discussion

This scoping review of reviews of both interpersonal and self-directed violence illustrates many key points about the breadth of research reviews on violence among SGM populations. First, the reviews included in this scoping review contained a total of 1148 articles on self-directed violence and 1895 articles on interpersonal violence, suggesting a substantial amount of research, most of which being first-generation disparity research. Some of these studies are likely repeated because of reviews’ overlapping topics and time spans, but it was beyond the scope of the present review to critically analyze all of the reference lists across the 73 reviews for duplication. Still, the concordance across studies, which substantiates disparities across multiple forms of violence, from microaggressions to intimate partner violence to suicide attempt, echoes a simple question posed by Fish in a recent commentary: what now? [85] There is ample epidemiologic evidence of violence disparities — 72 reviews’ worth of hundreds of studies — so how does the field of health equity sail beyond the eddies of documenting prevalence and risk factors and into the uncharted waters of reducing disparity?

Future Directions

The reviews by Coulter et al. (12) and Russon et al. (38) are the rare examples that summarized the literature about intervention studies to reduce or prevent violence for SGM individuals, both of which found sparse results. There are three main challenges that may be scientific barriers to developing and testing violence intervention and prevention efforts for SGM populations.

First, despite minority stress being a major theoretical underpinning for the production of SGM-related health disparities [86], specific measurement of SGM minority stress to operationalize it in research has been a relatively recent development [87, 88]. Thus, with the proliferation of more specific measurement of key intervenable risk factors, researchers can identify salient prevention points.

Second, in terms of interpersonal violence, the majority of research focuses on victims or survivors, and there is a clear paucity of research about perpetrators and primary prevention efforts [89]. Moreover, extant programs and efforts to combat intimate partner violence are too frequently limited by not understanding dynamics of or adequately serving individuals who are in same-sex relationships or in relationships that are not characterized with socially constructed binary gender identities [90,91,92]. For self-directed violence, the scope of inquiry has historically relied on individual-level psychopathology (e.g., depression, bipolar disorder) [93], with considerably less focus on the role of life disruption and other social environmental factors germane to distress in SGM populations (i.e., family rejection, discrimination).

Third, the roots of violence often trace to “wicked problems,” such as intergenerational poverty, historical abuse and trauma, and institutionalized racism, homophobia, sexism, and heterosexism. Thus, attacking the roots of violence require broader application of monetary, social, and intellectual resources to foster the interdisciplinary capacity to meet such lofty challenges [94, 95]. However, large-scale public health interventions for violence are few in comparison to individual-level interventions [94]. By bolstering efforts in these arenas, new avenues of intervention and prevention — at both individual and structural levels — may eventually build a critical mass to answer the disparities in violence experienced by SGM communities.

Related to interventions, there remains a clear unmet need for ongoing population-base surveillance of violence for SGM individuals. For instance, the NCVS, YRBS, and NSDUH surveys only added sexual orientation and gender identity to their data collection relatively recently; thus, monitoring national prevalence of interpersonal and self-directed violence for SGM individuals — a population with known disparities in risk for violence — has scant data to estimate population-level trends over time. However, there is a more insidious consequence of historical exclusion of SOGI data from federal health surveillance. The lack of data to monitor trends of violence among SGM communities leaves prevention without a benchmark: even if the aforementioned need for interventions could be fulfilled, how would their effectiveness be evaluated without data to determine if rates of violence decrease?

In addition to violence as outcomes, epidemiologic data help to uncover novel risk and protective factors, necessary second-generation studies. One example to underscore the necessity of inclusion of SOGI information is Clark and colleagues’ analysis of the General Social Survey (GSS) [96], a robust dataset used to learn about Americans’ attitudes about firearms as well as their ownership of firearms [97, 98]. When the GSS added sexual orientation to the survey in 2008, it finally afforded an opportunity to examine potential sexual orientation–related differences in the presence of firearms in the home, which is of crucial importance for suicide prevention because access to firearms is a major moderator of suicide fatality [99]. The results of Clark et al.’s investigation revealed an interesting negative association of sexual orientation and firearms; sexual minorities were less likely to report having a firearm in the home [96]. These findings were recently replicated with BRFSS data from two US states (California and Texas), which both happened to gather SOGI and firearms ownership data in 2017 [100]. Together, the findings raise important future directions for violence prevention research. For example, does less access to firearms protect sexual minority populations from suicide? Would suicide prevention efforts focused on firearm safety [101] be less impactful for sexual minority populations?

As much as self-reported survey data play a role in population health surveillance, so too do administrative datasets, which largely lack SOGI data. For example, the CDC’s National Syndromic Surveillance Program gathers emergency department data to monitor national trends in suicide attempt injuries [102], but because SOGI data are largely missing in health care, these data cannot provide information about SGM communities. Thus, to fill gaps in intervention work to reduce violence among SGM communities, various sectors that generate administrative data — health care, social service agencies, and law enforcement — must begin to gather SOGI data alongside other demographic data they currently collect, such as age and race/ethnicity, which are typically used to monitor trends in indicators of population risk and health.

A specific form of administrative data that is paramount for monitoring violence outcomes for SGM communities is mortality surveillance. Because SOGI data are not identified in a standardized way at the time of a violent death [103], there is currently no way to determine if homicide and suicide rates are greater for SGM communities than their non-SGM peers, despite hundreds of articles suggesting SGM individuals have disproportionate rates of major predictors of violent deaths (e.g., rates of assault, rates of suicide attempt). Limited evidence from the National Violent Death Reporting System (NVDRS) suggests that, among youth, SGM people may die by suicide at higher rates [104], but importantly, NVDRS is missing SOGI data for nearly 80% of decedents. Can key questions about potential mortality disparities among SGM people be answered with only 20% of data? Efforts are underway to increase the likelihood for SOGI data to enter the mortality information pipeline by training death investigators to collect SOGI data, but this endeavor is still in its pilot phase [105]. Equal efforts will be needed across the aforementioned sectors (e.g., health care, social services, law enforcement) to discover ways to structurally change data systems, as well as the training and institutional culture around SOGI data collection.

Limitations

As with any scoping review, there are several limitations to note. Principally, relevant reviews may have been missed due to search criteria and parameters. For example, some highly cited review papers, such as Haas et al. [106] and Stotzer [107], were excluded due to a priori decisions for inclusion criteria requiring articles explain their search methodology. Additionally, because publication bias is a threat to review studies, this review of reviews may inherently have publication bias encoded within it due both to the original reviews’ methodologies and the inclusion criterion of reviews published in peer-reviewed journals. The restriction of studies to being published in English limited discovering the broader international scope of studies.

Conclusion

Violence is perhaps the most infuriating and puzzling threats to public health because rather than the culprit being a virus, bacterium, environmental toxin or disaster, cells that have turned against the body, or internal organs that fail, we only have ourselves and each other to hold to account. This first review of reviews about violence research on SGM communities revealed a surprising breadth of studies, albeit mostly focused on identifying disparities. There is some progress in second-generation studies to help understand disparities and identify potential targets for intervention, but the field clearly has quite far to go for generating evidence about efficacious and effective interventions to reduce violence. Researchers are quickly capitalizing on newly available population-based datasets that include SOGI data and violence-related outcomes [5, 7]. However, we must also focus attention to developing collection of SOGI data in administrative datasets, which are necessary to foster data infrastructures that facilitate evaluation for interventions at scale.