Hiding in the open: Consideration of nonsuicidal self‐injury by proxy as a clinically meaningful construct

Abstract

Nonsuicidal self‐injury (NSSI) is defined as the deliberate destruction of one's own body tissue without suicidal intent and for purposes not socially sanctioned. However, this definition limits the understanding and assessment of NSSI by excluding a clinically relevant form of NSSI that is both self‐driven and associated with self‐injurious intentions: NSSI by proxy. Specifically, we propose that NSSI by proxy be defined as the intentional destruction of one's own body tissue through the elicitation of another being's (e.g., human, animal) actions, wherein the agency of the person being injured is a critical facet of the behavior. We review the literature supporting the clinical relevance of this behavior, as well as its similarities to traditional NSSI. Next, we propose four behaviors that may be conceptualized as NSSI by proxy, and identify two other behaviors that warrant further investigation. Finally, we identify future directions for research in this area and implications for the assessment and treatment of NSSI.

INTRODUCTION

Nonsuicidal self‐injury (NSSI) occurs in approximately 3%–6% of the adult general population (Klonsky, 2011; Liu, 2021; Plener et al., 2016; Swannell et al., 2014), 17%–27% of adolescents (Monto et al., 2018; Swannell et al., 2014; Zetterqvist et al., 2021), 13%–18% of young adults (Kiekens et al., 2021; Swannell et al., 2014), and up to 80% of clinical populations (Auerbach et al., 2014; Clarkin et al., 1983; Gunderson & Ridolfi, 2001). In addition to being associated with functional impairment and numerous psychiatric disorders (e.g., depression, posttraumatic stress disorder [PTSD], borderline personality disorder [BPD]; Bentley et al., 2015; Gratz et al., 2015; Patel et al., 2021), NSSI is a robust predictor of suicidal thoughts and behaviors (Asarnow et al., 2011; Franklin et al., 2017). Further, recurrent NSSI was included as a “condition for further study” within the 5th edition of the Diagnostic and Statistical Manual of Mental Disorders (DSM‐5; American Psychiatric Association, 2013), with the diagnostic criteria for NSSI disorder (NSSID) including engagement in NSSI on at least five days in the past year (Criterion A); the expectation that NSSI will regulate emotions and/or resolve interpersonal difficulties (Criterion B); the experience of negative feelings or thoughts immediately prior to NSSI, and/or preoccupation with or frequent thoughts related to NSSI (Criterion C); and the presence of significant distress or impairment (Criterion E), among others.

Although the inclusion of this new diagnosis offers great promise and hope for those who struggle with NSSI (Wilkinson & Goodyer, 2011), it will have less clinical utility if clinically‐relevant forms of NSSI are missed when considering this diagnosis due to overly restrictive definitions of NSSI that exclude important NSSI behaviors. Indeed, although NSSI is commonly defined as the deliberate destruction of one's body tissue without suicidal intent and for purposes not socially sanctioned (Chapman et al., 2006; Gratz, 2001; ISSS, 2018; Nock, 2009), there remain multiple inconsistencies and disagreements in the precise operational definitions of NSSI used across studies and the specific NSSI behaviors included in these definitions (see Lengel et al., 2022). More specifically, researchers have documented considerable variability in the restrictiveness of current operational definitions of NSSI, noting the negative implications of this for both research on and clinical assessment of NSSI (Lengel et al., 2022).

Notably, previous concerns about the operational definitions of NSSI have focused on issues such as gender bias in the behaviors identified as NSSI (Green & Jakupcak, 2016; Kimbrel et al., 2017), with researchers identifying relevant NSSI behaviors more common among men that have often been overlooked in the literature and excluded from assessment instruments (e.g., wall punching; Green & Jakupcak, 2016; Kimbrel et al., 2018; Whitlock et al., 2011). However, one limitation of extant definitions of NSSI that has not been widely addressed is the requirement that the behavior be exclusively self‐inflicted. Although the reasons behind this are understandable (given that self‐injurious intentions are a defining characteristic of NSSI), theoretical, clinical, and empirical literature suggest that the singular focus on self‐inflicted behaviors per se may be too restrictive and may limit the understanding and assessment of NSSI. In particular, literature suggests that NSSI by proxy (i.e., the deliberate destruction of one's own body tissue through the elicitation of another being's actions) is a clinically relevant form of NSSI that is both self‐driven and associated with self‐injurious intentions. Indeed, a recent survey of NSSI researchers and clinicians found that approximately half the respondents thought that NSSI by proxy should be considered a form of NSSI (Lengel et al., 2022).

NSSI BY PROXY

Although NSSI by proxy (e.g., initiating injury via sexual activity, an animal, tattooing, or a physical fight) has received limited attention within the literature to date, it has been recognized as clinically‐relevant and in need of further exploration (see Møhl, 2019). However, past work in this area has not put forth a clear conceptual definition of this form of NSSI, or identified the range of relevant behaviors that may qualify as NSSI by proxy. We propose that NSSI by proxy be defined as the deliberate destruction of one's own body tissue through the elicitation of another being's (e.g., human, animal) actions that is absent of suicidal intent and for purposes not socially sanctioned. Notably, and speaking to the clinical relevance of NSSI by proxy, these behaviors may also meet criteria for NSSID as outlined above. Specifically, like traditionally defined NSSI, NSSI by proxy behaviors may be used to regulate emotion or resolve interpersonal difficulties (Criterion B) and may be preceded by interpersonal difficulties or negative thoughts or emotions and/or associated with preoccupation with or frequent thoughts about the self‐injury (Criterion C) – all of which could, with repeated engagement, result in clinically significant impairment and/or distress (Criterion E). Thus, from a functional analytic perspective, NSSI by proxy may be considered a specific type of NSSI as traditionally‐defined.

Notably, however, unique to NSSI by proxy is the added component of an “other” as the inflictor of the injury. This “other” allows for an external attribution of the injury for the recipient and viewer/witness – something that can obfuscate the self‐injurious intentions of the behavior and increase its perceived social acceptability. Indeed, it is the presence of an “other” and related external attribution of the injury that can, for some individuals, make NSSI by proxy a particularly appealing form of NSSI that may accompany or even replace more traditionally‐recognized NSSI behaviors.

Importantly, however, although the nature of NSSI by proxy allows for external attributions of the injury, a critical facet of this form of NSSI is the agency of the person being injured: specifically, these behaviors are initiated by the individual and can be maintained and/or terminated by the individual. Although specific NSSI by proxy behaviors may differ in the formality of the arrangement or interaction with the other, ranging from formalized transactions (e.g., tattooing, piercing) to informal interactions (e.g., initiating injury via a fight or dog bite), all behaviors remain initiated, maintained, and/or terminated by the individual. For example, regarding the formalized NSSI by proxy behaviors, the individual chooses when (e.g., sets an appointment), where (e.g., chooses a body location for tattoo, piercing, etc.), and for how long (e.g., chooses size/detail of tattoo, can leave the appointment willingly or rescind consent) to have the injury. Likewise, for the informal NSSI by proxy behaviors, the individual chooses to initiate or engage in behaviors that can result in injury if the person so chooses (e.g., initiates rough play with a dog, goads an opponent, disobeys a police order). As with traditional forms of NSSI, what is key to NSSI by proxy is the individual's choice to initiate, maintain, and terminate the injurious behavior, making the injury self‐directed even though inflicted via the actions of an “other.”

Notably, the inclusion of an “other” introduces some conceptual overlap between NSSI by proxy and indirect self‐injury, as some researchers have conceptualized NSSI by proxy behaviors as forms of indirect self‐injury (given that it is the proxy causing the injury, rather than the individual doing so directly; Fliege et al., 2002; Green et al., 2017). Yet, indirect self‐injury is a distinct construct that differs in important ways from NSSI by proxy, in that it refers to behaviors that, while ultimately destructive, are not immediately or deliberately damaging to body tissue (Møhl et al., 2014; St. Germain & Hooley, 2012). For example, prototypical forms of indirect self‐injury include alcohol and drug abuse, disordered eating, problematic over‐exercise, risky behaviors, and “continuous engagement in abusive relationships” (St. Germain & Hooley, 2012) – none of which result in immediate injury to the body. Rather, the injury associated with these behaviors often appears only after repeated engagement in the behavior over time (Green et al., 2017). Likewise, Green et al. (2017) suggest that the injury associated with indirect self‐injurious behaviors is not as deliberate as in NSSI. Thus, indirect self‐injury differs from NSSI by proxy in that the latter results in injury that is intentional, physical, and (more) immediate (consistent with contemporary definitions of NSSI). For example, with behaviors such as tattooing and piercing, the presence of a formalized transactional relationship produces an obligation that ensures the physical injury (e.g., puncturing or laceration of skin) immediately upon an agreed upon time. Likewise, for NSSI by proxy behaviors involving informal interactions (e.g., initiating a fight or sexual encounter), the individual can choose to engage in these behaviors in such a way as to cause immediate tissue damage (e.g., broken bones, lacerations, bruising, etc.). Conversely, indirect self‐injury is characterized by injury that occurs only after prolonged exposure to or engagement in repeated behaviors (e.g., deterioration of esophagus from repeated purging, liver disease from regular binge drinking, etc.). Furthermore, whereas the self‐injurious intention of indirect self‐injurious behaviors in the moment is either absent or unclear, this intent is a defining characteristic of NSSI by proxy. Therefore, just as intent differentiates NSSI from suicide attempts, so too does intent distinguish NSSI by proxy from indirect self‐injury.

Although there have been no systematic investigations of NSSI by proxy to date, key NSSI by proxy behaviors have received preliminary attention from researchers in the context of traditional NSSI research, with studies examining behaviors such as tattooing and piercing (Solís‐Bravo et al., 2019; Wessel & Kasten, 2014), initiating a pet to inflict injury (Mann et al., 2020), initiating sex as self‐injury (Fredlund et al., 2020; Jonsson et al., 2019; Zetterqvist et al., 2018), and initiating a fight (Green et al., 2017). However, current investigations have not delineated the critical facets of these behaviors, most notably that they are other‐caused yet self‐driven. Given the unique characteristics of NSSI by proxy as previously described, NSSI by proxy may differ from both other forms of NSSI and topographically similar behaviors that do not involve self‐injurious intentions (e.g., aesthetic body modification) in important ways. Therefore, NSSI by proxy behaviors require critical examination as a distinct and clinically‐relevant construct. Based on existing empirical and clinical literature, as well as clinical experience, we propose four behaviors that should be considered examples of NSSI by proxy, followed by two other behaviors that warrant further consideration as potential forms of NSSI by proxy (see Table 1). Discussion of each behavior follows.

TABLE 1.

Proposed NSSI by proxy behaviors

Behavior	Description
Initiating injury via tattooing or body piercing	Getting a tattoo or piercing for the explicit purpose of injuring oneself physically (vs. socially sanctioned aesthetic or artistic expression purposes)
Initiating injury via an animal	Intentionally provoking an animal to scratch, bite, or otherwise cause direct physical injury to oneself
Initiating injury via sexual activity with a partner	Initiating sexual activity with a partner for the purpose of causing direct injury to oneself
Initiating injury by instigating a physical fight	Instigating a physical fight with the explicit intention to cause immediate and direct physical injury to oneself
Initiating injury via sports	Participating in sports for the explicit purpose of causing immediate and direct physical injury to oneself by deliberately putting oneself in harm's way or circumventing regulations in place to reduce risk of injury
Initiating injury by engaging in elective or cosmetic medical procedures	Engaging in cosmetic or elective medical procedures for the explicit purpose of causing immediate physical injury to oneself

NSSI BY PROXY BEHAVIORS

Initiating injury via tattooing or body piercing

Based on the definition above, tattooing and body piercing may, for some individuals and in some contexts, be conceptualized as a form of NSSI by proxy. Historically, consideration of tattooing and body piercing as forms of NSSI has been a controversial topic within the NSSI literature, both because these behaviors generally involve another person inflicting the bodily tissue damage and because tattooing and body piercing are often engaged in for aesthetic or artistic expression purposes (Aizenman & Jensen, 2007; Claes & Vandereycken, 2007; Wessel & Kasten, 2014; Wohlrab et al., 2007). Yet, despite the fact that these behaviors may commonly serve socially sanctioned purposes, there are notable examples of individuals seeking out tattooing and body piercing for the explicit purpose of self‐injury (Anderson & Sansone, 2003; Mann et al., 2020; Solís‐Bravo et al., 2019; Wessel & Kasten, 2014). These latter examples complicate the established distinction between conventional tattooing and body piercing and NSSI (Aizenman & Jensen, 2007; Claes & Vandereycken, 2007) and suggest that there may be times when these behaviors are better considered as a form of NSSI by proxy.

Indeed, although tattooing and body piercing are most frequently performed by an “other” (Bone et al., 2008), past NSSI research has examined these behaviors as NSSI when they are performed by the individual themselves (see Lloyd‐Richardson et al., 2007). For example, the Self‐Injurious Thoughts and Behaviors Interview (SITBI; Nock et al., 2007) and the Functional Assessment of Self‐Mutiliation (FASM; Lloyd et al., 1997) include tattooing oneself as a form of NSSI, the Screen for Nonsuicidal Self‐Injury (Halverson et al., 2022) includes the self‐insertion of needles into one's body as a form of NSSI (without excluding body piercing), and the Alexian Brothers Assessment of Self‐Injury (Washburn et al., 2015) includes both tattooing and piercing oneself as forms of NSSI. Thus, given that the only difference between self‐tattooing/self‐piercing and conventional tattooing/body piercing by a professional/“other” is the involvement of another person as the inflictor of the bodily tissue damage, considering tattooing and piercing by an “other” as a potential form of NSSI by proxy is consistent with extant operational definitions of NSSI that consider self‐tattooing/self‐piercing as a form of NSSI. Rather, the key to distinguishing conventional tattooing and body piercing from the forms of these behaviors that are better conceptualized as NSSI by proxy is the purpose of the behavior and the function it is serving (with socially sanctioned purposes excluded from the proposed definition of NSSI by proxy).

Specifically, literature suggests that tattooing and body piercing may be motivated by self‐injurious intentions, functioning in the same way as prototypical NSSI. Although this is less common than aesthetic‐ or expression‐related purposes, when it does occur, it warrants clinical consideration as a form of NSSI by proxy. Indeed, failure to recognize tattooing and body‐piercing for the purpose of self‐injury as a form of NSSI by proxy may result in the under‐identification of clinically relevant forms of NSSI and obscure clinically important treatment targets. For example, research suggests that tattooing and body‐piercing may serve as a behavioral substitution for traditional NSSI, serving the same function as NSSI but in such a way that the self‐injurious intentions of the behavior are obscured. Specifically, studies have found that a majority of individuals with a history of NSSI reported cessation of or decreases in NSSI following initiation of tattooing and body‐piercing (Stirn & Hinz, 2008; Wessel & Kasten, 2014). Likewise, Solís‐Bravo et al. (2019) found that individuals with NSSI and tattoos reported greater severity of NSSI and had greater endorsement of all NSSID criteria (including past‐year NSSI frequency, negative emotional antecedents, NSSI‐related cognitions, and resultant distress/impairment) when compared to individuals with NSSI but no tattoos. Although preliminary, these findings suggest that tattooing and piercing may function in the same way as NSSI and be associated with increased NSSI severity and related impairment among individuals with a history of this behavior, supporting the conceptualization of a subset of these behaviors as NSSI by proxy.

Published clinical case studies also support the conceptualization of some tattooing behaviors as a form of NSSI by proxy. For example, Anderson and Sansone (2003) describe the case of Mr. B, a 19‐year‐old man who dealt with NSSI urges and emotional distress through tattooing. In discussing his decisional process of choosing to engage in either prototypical NSSI (i.e., cutting) or NSSI by proxy (i.e., tattooing), Mr. B reported that cutting would result in more embarrassment if noticed by others, whereas tattooing would not have the same negative social and emotional consequences (and could even have positive social consequences of the tattoo being judged as “cool”). Consequently, although engaging in tattooing would have the same positive consequences of relieving emotional distress, it would avoid the negative social and emotional consequences of cutting. Notably, Mr. B also reported that he was able to dictate the level of pain he received from the tattooing behavior by directing placement of the tattoo to either a less sensitive or more sensitive area on his body (Anderson & Sansone, 2003).

Likewise, Mann et al. (2020) describe a similar case of a veteran whose tattooing was specifically done for the principal purpose of intentional injury rather than artistic expression. Notably, this case would have met all criteria for NSSID if it were not for the exclusion of behaviors involving an “other.” Specifically, the purpose of the tattooing behavior for this veteran was to cause immediate injury to himself in order to regulate his emotions (in the same way that his other, more prototypical NSSI behaviors functioned), rather than to express himself artistically or for other aesthetic purposes. Furthermore, the tattooing behavior of this veteran was preceded by negative emotions (e.g., tension, sadness, loneliness, etc.), and associated with significant distress and impairment (Mann et al., 2020). Moreover, this veteran explicitly described the tattooing as a form of NSSI, engaged in specifically to cause pain and harm himself physically in a way that was more socially acceptable than traditionally examined forms of NSSI. In fact, he reported that he preferred tattooing over more prototypical NSSI behaviors (e.g., cutting, burning, wall punching) precisely because tattooing was a more socially acceptable form of NSSI, which appears consistent with Mr. B's decision‐making process (Anderson & Sansone, 2003). Nonetheless, despite being a more socially accepted form of NSSI, this behavior was a cause for serious clinical concern, as the veteran experienced both subjective distress (e.g., regret) and impairment (e.g., financial difficulties, relationship problems with his fiancée; Mann et al., 2020) as a result of the tattooing. Finally, consistent with the conceptualization of NSSI by proxy described above, the veteran emphasized his agency and control over this behavior, noting that he initiated the business contract with the tattoo artist, had control of where the pain occurred and to what extent (similar to Mr. B; Anderson & Sansone, 2003), and had the ability to end the tattoo session. For these reasons, Mann et al. (2020) suggested that excluding this type of behavior from definitions of NSSI may detract from the clinical utility of the NSSID diagnosis and interfere with the identification of a clinically relevant treatment target and source of functional impairment.

Initiating injury via an animal

In the same case study series describing clinically important but overlooked forms of NSSI (including NSSI by proxy) noted above, Mann et al. (2020) described another case that involved the NSSI by proxy behavior of initiating injury by an animal. Specifically, this individual reported seeking to purposefully harm himself by getting his puppy to bite him through play, describing it as a “painful massage” that “blocked out everything else” (Mann et al., 2020), consistent with the conceptualization of NSSI by proxy presented here. Likewise, and in further support of the clinical relevance of this form of NSSI, this veteran reported that this behavior was motivated by a desire to regulate his emotions (e.g., decrease negative emotions, relieve tension, etc.) and preceded by negative emotions (e.g., sadness, upset, worthlessness, etc.). Indeed, this was his primary form of NSSI, and he explicitly noted that this behavior functioned to relieve tension and emptiness, punish himself, distract himself, and stop negative thoughts to a greater degree than his more traditional NSSI behaviors of head‐banging and self‐biting.

Likewise, the last author of this article had a patient who reported intentionally getting her cat to scratch her in order to harm herself. As in the case of the veteran described above, this patient explicitly described this behavior of initiating injury via her cat as a form of NSSI comparable to the cutting and burning behaviors she engaged in regularly. She described initiating injury via interactions with her cat as a way of injuring herself when other implements were unavailable, someone else was present (thus precluding more private NSSI behaviors), or she wanted to hide the self‐injurious intentions of her behaviors and resulting tissue damage. Thus, as in the case of certain forms of tattooing and body piercing noted above, the explicit and intentional elicitation of injury from an animal for the purpose of causing injury to oneself may be considered a clinically meaningful form of NSSI by proxy in need of further examination.

Initiating injury via sexual activity with a partner

Although sex is a context in which consensual violence may be considered socially acceptable, the purposeful initiation of and reception to injury from one's sexual partner may be motivated by factors beyond the often socially sanctioned motivation of sexual gratification. In these instances, the purposeful initiation of injury via sex may be considered a form of NSSI by proxy if it meets the other criteria outlined above. Notably, although the DSM‐5 (American Psychiatric Association, 2013) includes a disorder that overlaps with this form of sexual activity (i.e., sexual masochism disorder [SMD], in which sexual violence is used as a means to alter an emotional state), SMD differs from the sexual behavior of NSSI by proxy in that SMD does not require injury or even the manifestation of behavior. Rather, SMD may manifest via fantasies or urges. Moreover, if a behavior does occur, it may include other forms of suffering that do not result in injury (e.g., humiliation). Therefore, behaviors resulting in purposeful physical injury from a sexual partner may be more precisely conceptualized as NSSI by proxy, especially if the motivation is to cause immediate physical injury for the purpose of regulating emotions or resolving interpersonal difficulties.

Likewise, although researchers have recently proposed the construct of sex as self‐injury (SASI; Fredlund, 2018; Fredlund et al., 2020; Jonsson et al., 2019; Zetterqvist et al., 2018), this construct overlaps with but is distinct from sexual activity as a form of NSSI by proxy. Specifically, because SASI encompasses both psychological and physical harm in sexual situations, as well as self‐inflicted sexual injury (Fredlund et al., 2020), SASI does not necessarily have to involve the direct elicitation of physical injury from one's sexual partner during sexual activity, as it may involve self‐inflicted harm or indirect or psychological harm. Indeed, researchers have conceptualized SASI as indirect self‐injury (Zetterqvist et al., 2018), direct self‐injury comparable to traditional forms of NSSI (Fredlund et al., 2020), and a continuum of behavior ranging from indirect to direct (Jonsson et al., 2019). Notably, as research on SASI has progressed, there is greater support for conceptualizing some forms of SASI as NSSI by proxy (i.e., when direct and immediate injury elicited from an “other” is a defining characteristic). However, as a whole, SASI that does not involve another individual or that does not result in direct and immediate physical harm remains unique to the construct of SASI and different from NSSI by proxy.

Importantly, support for the conceptualization of initiating injury via sexual activity as a form of NSSI by proxy has been provided by research on SASI and NSSI. Utilizing a sample of 1027 Swedish high school students, Jonsson et al. (2019) investigated functional differences and similarities between NSSI‐only, SASI‐only, and combined NSSI and SASI groups. They found that functional similarities exist between NSSI and SASI, particularly regarding emotion regulation motives, although SASI was associated with more interpersonal motives. Similarly, in another study of 199 Swedish adults (M age = 27.9; SD = 9.3), 94% of individuals with SASI endorsed emotion regulation motives for the behavior (Fredlund et al., 2020). Moreover, participants reported that SASI operated similarly to NSSI and that they alternated between SASI and NSSI behaviors, with 21.6% of the sample reportedly replacing NSSI with SASI (Fredlund et al., 2020). This preference for SASI as an alternative for NSSI was explained through SASI's invisibility and the convenience of finding willing men to participate in physically violent sex (Fredlund et al., 2020). These findings support the conceptualization of some forms of SASI (i.e., those involving direct other‐elicited physical injury) as NSSI by proxy rather than indirect self‐injury. Further supporting the conceptualization of some forms of SASI as NSSI by proxy, the first author of this article interviewed a female veteran in the context of administering the Clinician Administered Nonsuicidal Self‐injury Disorder Index (CANDI; Gratz et al., 2015) as part of a research study (see Patel et al., 2021). During the interview, the veteran reported initiating sexual activity for the purpose of causing direct injury to her body. Specifically, this veteran reported that she had a series of men to whom she would reach out when she was feeling especially distressed. On her self‐described hierarchy of NSSI behaviors, she placed initiating injury via sex as the primary behavior (above both prototypical forms of NSSI and other NSSI by proxy behaviors) she would use for the most intense self‐injurious urges and emotional distress. She reported using these sexual interactions to elicit bruises and other injuries that she could hide from others (or explain away) and also continue to exacerbate throughout the following week in order to elicit further pain and relieve emotional distress (e.g., by pressing on the bruises when feeling distressed in order to replace her emotional pain with physical pain). Importantly, this veteran also reported significant distress and impairment as a result of this behavior, spending numerous hours at a time trying to secure a partner for this behavior, noting difficulties concentrating on other tasks in the hours leading up to this behavior, and noting instances when the behavior escalated beyond her desires/intentions to put her life at risk or result in more severe injury than she intended.

Initiating injury by instigating a physical fight

Although initiating a physical fight is considered by some researchers to be a risky or reckless behavior (i.e., a form of indirect self‐injury; Green et al., 2017; St. Germain & Hooley, 2012), this behavior can also be done with the explicit intention to cause immediate and direct physical injury and tissue damage to oneself. Moreover, and consistent with the conceptualization of NSSI by proxy proposed here, this type of behavior can be initiated, maintained, and ended by the individual receiving the injury. Thus, consistent with Møhl (2019), we suggest that initiating a physical fight in order to injure oneself may be best considered a form of NSSI by proxy. An example of this form of NSSI by proxy is provided by Møhl (2019), who documents the case of a man who would intentionally provoke police into beating him during political demonstrations and taunt opposing supporters of sports teams into fistfights in order to cause immediate injury to himself. Given that the injury in question was intentional, self‐driven, and other caused, then categorizing this type of behavior as NSSI by proxy is warranted.

As another example of this form of NSSI by proxy, a patient seen by the second author reported several instances of instigating physical fights at bars for the purpose of injuring himself. This patient engaged in other forms of NSSI (i.e., burning himself with cigarettes, banging his head, punching walls) to punish himself when experiencing shame and self‐disgust. The patient reported that instigating physical fights served the same function, while also having the added benefit of ensuring that the punishment would be sufficient (as some of his more minor NSSI behaviors did not always result in the level of tissue damage he preferred). This behavior was conceptualized as NSSI by proxy because the patient was specifically choosing to initiate physical fights in order to bring about immediate physical injury to himself. Furthermore, the patient reported significant distress associated with this behavior, noting that he experienced self‐disgust shortly after the fights, as well as guilt for bringing others into his NSSI. Likewise, because there were times when he was not able to hide the injuries incurred via this behavior, he reported that this behavior had interfered with personal and work relationships on multiple occasions.

POTENTIAL NSSI BY PROXY BEHAVIORS REQUIRING FURTHER CONSIDERATION

Although there is clinical and empirical literature suggesting that there are instances in which the aforementioned behaviors may be best conceptualized as NSSI by proxy, there are additional behaviors that overlap with the previously described behaviors conceptually and may also meet our definition of NSSI by proxy. However, further research is needed to determine if these behaviors can indeed be considered NSSI by proxy. Thus, although we discuss the available literature and relevant clinical case examples supporting the conceptualization of these behaviors as forms of NSSI by proxy, we acknowledge the need for further examination of the extent to which these behaviors are better conceptualized as NSSI by proxy, indirect self‐injury, or another type of behavior.

Initiating injury via sports

In the context of contact sports and martial arts (e.g., rugby, football, boxing, taekwondo, etc.), injury is both accepted and expected (accidental or intentional). Consequently, for some individuals, sports may provide an ideal context for initiating injury to the self by deliberately putting themselves in harm's way or circumventing regulations in place to reduce risk of injury. For example, in order to increase the likelihood of bodily injury, individuals could choose to wear less protective gear, sabotage their own safety equipment, or, in the case of boxing, temporarily increase their weight to achieve a higher weight class before returning to their usual, lighter weight. ¹ In all of these instances, the intention of the behavior is to initiate injury to the self via an opponent or opposing player in the context of a sport (where one's self‐injurious intentions may be obscured).

Indeed, the likelihood or inevitability of injury in some sports may complicate determination of an individual's specific motives for engaging in these types of sports (e.g., self‐injurious vs. other) and differentiation of engagement in sports for socially sanctioned versus clinically relevant purposes (see Green & Jakupcak, 2016), especially for boys and men (for whom such injury may be viewed as a means of becoming a ‘man’; see Gard & Meyenn, 2000). For example, in their qualitative study of Australian boys aged 12–15 years, Gard and Meyenn (2000) describe how pain (both the ability to give and receive) acts like a ‘currency’ with which boys develop their masculine identities. One of the boys in this study described enjoying “smashing people” as well as “getting smashed myself” (p. 26, Gard & Meyenn, 2000). Although this boy was not able to articulate the reasons he enjoyed “getting smashed” and, ultimately, upon repeated questioning in front of his peers, recanted that getting hurt was the intended outcome (Gard & Meyenn, 2000), the transcribed interview of this participant underscores the ambiguity of motives for engaging in sports where injury may be expected or anticipated, as well as the fact that, for some individuals, getting injured by others in the context of sport is the desired outcome.

As another example of sports serving as the context for NSSI by proxy, the same female veteran who reported initiating sexual encounters for the purpose of injuring herself also reported participating in rugby and jujitsu for the purposes of self‐injury, as doing so provided a controlled way in which she could injure herself in a socially acceptable way. Specifically, she noted that the opposing players were already willing to engage in violent behavior, and any conflict between teams and players was context‐ and time‐limited to the field and duration of play. Although this case example demonstrates a clearer purpose behind initiating injury via sports than the interview with the Australian youth, further investigation is necessary to determine whether initiating injury via sports should be considered a form of NSSI by proxy.

Notably, although this behavior has some similarities to the NSSI by proxy behavior of initiating injury by instigating a physical fight (e.g., the obfuscation of self‐injurious intentions, both in general and from the other inflicting the injury in particular), there are some important distinctions between the two that support their conceptualization as different forms of NSSI by proxy. In particular, although both behaviors (consistent with all of the proposed NSSI by proxy behaviors reviewed here) involve the initiation of injury to the self via another being, they differ in their social acceptability and, thus, expected interpersonal consequences (with initiation of injury via sports versus a fight considered more socially acceptable). Indeed, the female veteran described above specifically noted a preference for initiating injury in the context of sports versus interactions on the street or in a bar because the former is viewed as far more socially acceptable and was expected to have far fewer negative interpersonal consequences. Moreover, these differences in the social acceptability of the context and resultant injury may also have implications for the extent to which the behavior may be recognized as a form of NSSI by proxy, with initiating injury via sport being less clearly recognizable as a clinically concerning behavior with self‐injurious intentions due to the expected nature of injury in this context.

Initiating injury by engaging in elective or cosmetic medical procedures

Many cosmetic or elective medical procedures are characterized by bodily change (e.g., altered nose, removal of blood for donation) that is the result of tissue damage. These behaviors are commonly motivated by a desire to improve one's appearance (e.g., rhinoplasty, micro‐needling), improve one's health (e.g., requesting intravenous fluids, tooth extraction, wet cupping), reduce potential risk (e.g., breast cancer prophylaxis, biopsies, blood tests), or provide life‐saving care for others (e.g., donating blood or bone marrow). However, there are times when these behaviors are not engaged in for these more common health‐ and aesthetic‐related purposes, but for the purpose of causing immediate injury to the self. In those instances, engagement in cosmetic or elective medical procedures may be viewed as a behavior of clinical concern that warrants intervention.

For example, when done repeatedly, some of these behaviors may be considered a symptom of body dysmorphic disorder. However, repeated elective surgery related to body dysmorphic disorder differs from NSSI by proxy in that the intention behind NSSI by proxy is injuring the self rather than enduring an injuring process to correct or modify a negatively perceived body part or feature as in body dysmorphic disorder. Likewise, repeated (elective) medical procedures may, for some individuals, be a form of indirect delegated self‐harm related to factitious disorder (Fliege et al., 2002). Indirect delegated self‐harm in the context of factitious disorder involves an individual claiming to have a disease or symptoms that they do not have and/or inducing relevant symptoms in an effort to obtain medical treatment from professionals, including operations or amputations (Fliege et al., 2002). For example, patients may feign heart pain and require cardiac catheters, or they may report localized problems on their skin and require skin biopsies (Gieler et al., 2013). Although these behaviors are also considered to be a clinically concerning variant of engaging in elective medical procedures, they are distinct from NSSI by proxy due to the absence of direct self‐injurious intentions. Specifically, whereas the injury associated with NSSI by proxy is, by definition, intentional, physical, and immediate, this is not a necessary feature of indirect delegated self‐harm, which does not require the individual to experience direct and immediate tissue damage (as in the case of repeated x‐rays); rather, any medical procedure provoked by the patient that may increase risk to health over the long‐term or indirectly would qualify (similar to other forms of indirect self‐injury; Fliege et al., 2002).

As with the other behaviors described previously in this review and consistent with the definition of NSSI by proxy proposed here, initiating injury via engagement in elective cosmetic or medical procedures may be best conceptualized as a form of NSSI by proxy. Specifically, engagement in elective medical procedures may meet criteria for NSSI by proxy if immediate injury is the intended purpose of these behaviors. An example of this form of NSSI by proxy was reported by a patient of the authors of this article, who described excessively donating blood as a form of self‐injury. In this case, the patient utilized a socially accepted (and esteemed) behavior that resulted in immediate injury through the actions of another. This patient initiated and participated in this behavior for the expressed purpose of incurring injury in the same way as her other NSSI behavior of cutting. Indeed, this particular behavior served the same functions as her cutting behavior (i.e., to relieve emotional distress) and was reportedly used interchangeably with the cutting, with the choice being driven by other contextual factors (e.g., this patient was a nurse and so could give blood at work but did not feel comfortable cutting herself at work). Furthermore, this behavior was associated with significant distress (similar to her cutting behavior), as well as functional impairment (e.g., interfering with work; interfering with social engagements). Similar to initiating injury via sports, initiating injury through elective or cosmetic medical procedures requires further examination as a form of NSSI by proxy. In particular, and in comparison to the other behaviors described previously, the presence of more numerous diagnostic differentials (i.e., factitious disorder and body‐dysmorphic disorder) requires greater and clearer justification for this behavior as a possible form of NSSI by proxy.

FUTURE RESEARCH DIRECTIONS

NSSI by proxy is a clinically relevant behavior in need of further empirical attention. In particular, research is needed to clarify the prevalence of the various NSSI by proxy behaviors reviewed above, as well as the immediate and long‐term emotional and interpersonal consequences of these behaviors and the extent to which these overlap with and differ from those associated with more traditional NSSI behaviors. The first step in facilitating the systematic progression of research in this area is the development of empirically‐supported and psychometrically sound measures of NSSI by proxy and its characteristics. One such measure based on the CANDI (Gratz et al., 2015) is currently being evaluated by the authors of this article, and holds promise for faciliating future research in this area. Specifically, this NSSI by proxy questionnaire was developed to assess not only the presence and frequency of the previously reviewed NSSI by proxy behaviors, but characteristics of these behaviors relevant to an NSSID diagnosis, including the emotional and cognitive antecedents of and expectancies/motivations for these behaviors. This questionnaire is currently being simultaneously validated in a general population sample and a clinically‐relevant community sample of individuals with current NSSI behaviors.

Importantly, however, future research is also needed to identify other relevant NSSI by proxy behaviors that were not reviewed here. Indeed, although the behaviors described here were identified through a thorough review of the literature and the authors' clinical experiences, this is likely not an exhaustive list and additional research is needed to identify other NSSI by proxy behaviors that warrant further consideration. For example, by definition, the elicitation of another person to injure oneself via more traditional NSSI methods (e.g., cutting, burning, etc.) would qualify as NSSI by proxy. Although we could not find any examples of this type of NSSI by proxy in the literature and have not encountered any instances of this in our clinical practice, further research is needed to systematically assess for the presence of this behavior among a subset of individuals with NSSI and/or NSSI by proxy. In addition to clarifying the relevance of this behavior as a form of NSSI by proxy, such research may help elucidate the extent to which the ability to obfuscate the self‐injurious intentions of the behavior is a defining versus commonly associated characteristic of NSSI by proxy.

Once psychometrically sound measures of NSSI by proxy are available, research can examine the prevalence of NSSI by proxy within different populations (e.g., community, clinical samples) and the extent to which NSSI by proxy co‐occurs with traditional NSSI behaviors. Moreover, research is needed to examine the developmental trajectory of NSSI by proxy, both in general and relative to traditional NSSI. For example, whether NSSI by proxy tends to precede or follow more traditional NSSI behaviors remains unknown. Although the literature reviewed here suggests that it may be more common for individuals to transition from NSSI to NSSI by proxy due to the stigma associated with traditional NSSI behaviors and the greater ability to hide self‐injurious intentions in the context of NSSI by proxy, it is also possible that learned associations of physical injury with emotional relief in the context of socially sanctioned activities (e.g., sports, sexual activity) may lead to the development of NSSI by proxy and, for some individuals, the later escalation of these behaviors to NSSI.

Research is also needed to examine individual difference factors associated with the initiation of and preference for NSSI by proxy versus more traditional forms of NSSI. For example, given that NSSI by proxy behaviors may be more easily hidden in the context of socially sanctioned activities, individuals who engage in NSSI by proxy versus traditional NSSI may be higher in shame proneness. Future research should also evaluate whether NSSI by proxy is associated with a similar increased risk for suicide as traditional forms of NSSI (see Hamza et al., 2012). Finally, research is needed to examine if differences in the social acceptability of specific NSSI by proxy behaviors and expectations of the likelihood and acceptability of injury in these contexts translate into differences in the interpersonal and emotional consequences of engagement in these behaviors, as well as recognition of these behaviors as forms of NSSI by proxy.

CLINICAL IMPLICATIONS

Further consideration of NSSI by proxy as a distinct construct has a number of important clinical implications. First, given that most extant assessments of NSSI fail to recognize or assess NSSI by proxy, individuals with NSSI by proxy behaviors may not be identified by currently available instruments, interfering with the identification of clinically relevant treatment targets and clinically indicated treatments, as well as an important source of functional impairment. Furthermore, inattention to NSSI by proxy behaviors along with traditional NSSI behaviors may result in erroneous conclusions regarding the extent and scope of an individual's NSSI (as clinically relevant NSSI by proxy behaviors are overlooked). Moreover, individuals who are receiving treatment for prototypical NSSI may, over time, replace their prototypical NSSI behaviors with a more socially acceptable and non‐assessed NSSI by proxy behavior (as 21.6% of the Swedish adults reported doing with NSSI by proxy sexual activity [Fredlund et al., 2020]). Critically, this latter scenario may erroneously suggest clinical improvement when, in actuality, behavioral substitution is occurring and resulting in the maintenance or even exacerbation of functional impairment. Indeed, although behavioral substitution of NSSI with NSSI by proxy may protect against some negative interpersonal or emotional consequences of NSSI (e.g., judgment by others, embarrassment; Anderson & Sansone, 2003), this is not always the case, as some of the patients described in the aforementioned clinical case examples noted experiencing guilt, self‐disgust, and interpersonal difficulties as a result of their NSSI by proxy. Moreover, NSSI by proxy shares many of the same negative intrapersonal consequences as traditional NSSI (e.g., paradoxical increase in emotional distress in the long term, distorted relationship with one's body, increase in the acquired capability for suicide). Furthermore, even though NSSI by proxy could be viewed as a form of harm reduction (i.e., obtaining a tattoo could be viewed as a safer alternative to cutting due to the injury occurring through sterile methods and being administered by a professional), tattooing behaviors associated with the same motives and serving the same functions as more traditionally‐defined forms of NSSI (e.g., experiential avoidance) would still have a number of very serious negative consequences for emotional and general functioning and, thus, remain a clinical concern. Therefore, it becomes paramount to effectively assess for NSSI by proxy behaviors within clinical practice.

Consistent with other health risk behaviors that may or may not be cause for clinical concern (e.g., alcohol use, exercise), it is also imperative that clinicians differentiate between non‐pathological/normative behaviors and their maladaptive counterparts (i.e., NSSI by proxy). Specifically, for any behavior involving some form of physical injury or tissue damage, it is important that clinicians systematically assess the motives underlying and function of the behavior, the antecedents and consequences of the behavior, and the extent to which the behavior is associated with distress and/or impairment. Not only can a thorough evaluation of these behaviors identify clinically‐relevant treatment targets and instances of NSSI by proxy that may have otherwise gone unnoticed, such a thorough assessment can also prevent the stigmatization of normative or non‐pathological behaviors that are not intentional (e.g., accidental animal bite) or are associated with socially sanctioned motives (e.g., tattooing for self‐expression).

Given similarities between prototypical NSSI and NSSI by proxy, treatments shown to be efficacious for NSSI (e.g., dialectical behavioral therapy [DBT; Linehan, 1993; Stanley et al., 2007], emotion regulation group therapy [ERGT; Gratz et al., 2014], mentalization‐based treatment [MBT; Bateman & Fonagy, 2009; Calati & Courtet, 2016]) may be similarly useful in treating NSSI by proxy behaviors and warrant investigation. However, there may be important treatment considerations that are unique to NSSI by proxy behaviors. Notably, inherent to NSSI by proxy behaviors is their socially acceptable appearance, as well as the fact that clinically relevant self‐injurious intentions for these behaviors may be overlooked or obfuscated by the socially sanctioned motives typically attributed to these behaviors (e.g., body modification). Moreover, it is possible that the same behavior may serve multiple functions depending on the context, such as the emotional state of the individual. As an example, for an individual with NSSI by proxy who initiates injury via sexual activity, this sexual behavior provides not only the opportunity to self‐injure, but the simultaneous benefits of social connection (which may further reinforce the behavior and make it more difficult to stop). Additionally, the defining interpersonal nature of NSSI by proxy may introduce unique considerations relative to traditional NSSI. For example, in the case of initiating injury via sexual activity, consideration of a regular sexual partner's involvement and behavior may be critical in successfully reducing NSSI by proxy behavior. In these instances (i.e., when the proxy has a close and ongoing relationship with the individual), the patient may benefit from the inclusion of the “other” within the therapeutic process, including couples sessions.

Finally, although traditional NSSI can lead to relationship problems and negatively affect an individual's friends and loved ones (Turner et al., 2017), NSSI by proxy is unique in its involvement of another being as the inflictor of the injury (with or without their awareness). Thus, depending on the nature of their involvement with the individual engaging in NSSI by proxy, the duration of this relationship, and their awareness of their role as a proxy, the individual inflicting the injury may also benefit from clinical interventions focused on clarifying their emotional reactions to serving as the proxy and their goals within the relationship. DBT‐based interventions may also be helpful if the proxy is experiencing emotional distress about their role (e.g., shame, guilt, anger, or moral trauma) and/or wants to set limits with their partner around this behavior. For example, DBT distress tolerance and emotion regulation skills may be helpful in tolerating and modulating feelings of shame, guilt, and anger related to serving as the proxy, and DBT mindfulness skills may be helpful in approaching the situation and one's role with nonjudgmental awareness. Finally, DBT interpersonal skills may be useful for clarifying and asserting one's needs within the relationship and setting limits as needed.

CONCLUSION

NSSI by proxy is conceptualized as intentional, immediate, physical injury without suicidal intent and for purposes that are not socially sanctioned that is directed by the self yet resulting from the actions of another being. Similar to NSSI, NSSI by proxy may be used to regulate emotions and/or resolve interpersonal difficulties, is often preceded by negative thoughts or feelings and/or preoccupation with self‐injury, and has the potential to result in clinically significant distress and functional impairment. Although the use of an “other” as the inflictor of the injury, as well as the fact that many of the NSSI by proxy behaviors identified in this review are topographically (albeit not functionally) similar to socially acceptable behaviors, has heretofore interfered with the systematic assessment and investigation of these behaviors, emerging clinical, theoretical, and empirical literature highlights their clinical relevance as a target of treatment and research efforts. Consequently, NSSI by proxy is a unique construct that warrants further empirical and clinical consideration.

Mann, A. J. D. , Tull, M. T. , & Gratz, K. L. (2022). Hiding in the open: Consideration of nonsuicidal self‐injury by proxy as a clinically meaningful construct. Suicide and Life‐Threatening Behavior, 52, 1024–1036. 10.1111/sltb.12899