Richard L. Frierson, MD

Nonsuicidal self-injury (NSSI) is the direct and deliberate destruction of body tissue without intent to die.¹ Common forms of NSSI include cutting, burning, scraping/scratching skin, biting, hitting, and interfering with wound healing.² Functional theories suggest that NSSI temporarily alleviates overwhelming negative emotions and can produce feelings of relief, resulting in a reinforcing effect.³

NSSI has been shown to be a risk factor for future suicide attempts.⁴ A 2018 study found that NSSI is associated with an increased risk of subsequent suicidal ideation (odds ratio [OR] 2.8), suicide plan (OR 3.0), and suicide attempt (OR 5.5).⁵ NSSI is also associated with individuals who had suicidal ideation and formed a suicide plan, and individuals who had a suicide plan and attempted suicide (ORs 1.7 to 2.1).⁵ Another study found that 70% of adolescents who engage in NSSI have attempted suicide during their lifetime, and 55% have multiple attempts.⁶

Given the overlap between suicide attempts and NSSI, performing a thorough suicide risk assessment (which is beyond the scope of this article) is crucial. This article describes the static and dynamic risk factors for NSSI in adolescents and adults, which can help us perform a suicide risk assessment and allow us to formulate an appropriate treatment plan that includes safety-based interventions.

NSSI risk factors for adolescents

From developing sexual identity and undergoing puberty to achieving increased independence from their parents and developing a sense of autonomy, adolescents undergo many biological, psychological, and social changes before reaching adulthood.⁷ Data suggest that NSSI often begins in adolescence, with a typical onset at age 13 or 14.³ Community studies show that one-third to one-half of adolescents in the United States have engaged in NSSI.^8,9 Previously, NSSI during adolescence was associated with 3 major diagnostic categories: eating disorders, developmental disabilities, and borderline personality disorder (BPD).¹⁰ However, recent data suggest that NSSI is also common outside of these categories. Here we describe static and dynamic risk factors for NSSI in adolescents (Table 1^11-42). Table 2^11-42 summarizes the studies of NSSI in adolescents that we reviewed.

Static risk factors

Female adolescents and adults engage in NSSI at higher rates than males. The difference is larger in clinical populations compared to the general population.¹¹

A large portion of research about NSSI has been conducted in studies in which the majority of participants were White.¹² Most studies report a higher prevalence of NSSI among non-Hispanic White youth,¹³ but some suggest other ethnic groups may also experience high rates of self-harm and NSSI.^13-15 Several studies have demonstrated high rates of self-harm among South Asian adult females compared with White adult females, but this difference may be less pronounced in adolescents.¹⁴ One study in the United Kingdom found that White females age 10 to 14 had higher rates of self-harm compared to South Asian females,¹⁴ while another found that risk and rates of self-harm in young South Asian people varied by city and country of origin.¹⁵ Young Black females¹⁵ and young Black males¹³ also may be at an increased risk of self-harm. One review found that Black females were more likely to self-harm than Asian or White groups.¹⁵

Several studies suggest that sexual minority adolescents (SMA) (eg, lesbian, gay, bisexual, transgender, queer) are at greater risk for NSSI than heterosexual adolescents.¹⁶ SMA have been shown to engage in a significantly greater frequency of NSSI and more types of NSSI than heterosexual adolescents.¹⁶ Furthermore, on the Inventory of Statements about Self-Injury, SMA self-reported using NSSI for intrapersonal functions (eg, for affect regulation, antisuicide, self-punishment) significantly greater than their heterosexual peers; however, there were no significant differences between the 2 groups on interpersonal functions (eg, autonomy, interpersonal boundaries, peer bonding, sensation-seeking).¹⁶

Continue to: Transgender and gender nonconfirming...

Transgender and gender nonconfirming (GNC) youth are at a particularly high risk for NSSI; 30% to 45.5% of transgender adolescents report self-injury.¹⁷ Factors shown to distinguish transgender/GNC youth who engage in NSSI from those who do not include having a mental health problem, depression, running away from home, substance use, lower self-esteem/greater self-criticism, experiencing transphobia victimization, and having more interpersonal problems.^18,19 Among transgender/GNC youth, those whose biological sex is female are more likely to report NSSI than those whose biological sex is male (ie, transgendered adolescent males are more likely to report NSSI than transgendered adolescent females).^18,19

Most forms of childhood maltreatment have been associated with NSSI. In a recently published review, Liu et al²⁰ found that childhood maltreatment (including sexual abuse, physical abuse, emotional abuse, and physical neglect) was associated with an increased risk for NSSI. However, conflicting evidence suggests that when confounders are removed, only childhood emotional abuse was directly associated with NSSI.²¹ Current evidence is modest for childhood emotional neglect as a risk factor for NSSI.²⁰

Increasing research is investigating the biological processes that may be implicated in NSSI. Some studies suggest that endogenous opioids,²² monoamine neurotransmitters,²² and the hypothalamic-pituitary-adrenal (HPA) axis²³ may play a role in NSSI. Compared to healthy controls, adolescents engaging in NSSI have been shown to have lower pain intensity (P = .036), higher pain thresholds (P = .040), and lower beta-endorphins (endogenous opioid hormones involved in mediating stress and pain) (P = .002).²⁴ There may be alterations in the HPA axis among adolescents who engage in NSSI, more specifically stronger cortisol awakening responses.²³ Both functional and standard MRI have been used to study the neurobiology of NSSI. One study demonstrated differences in functional connectivity between brain areas linked to neuroregulation of emotions in adolescents who engage in NSSI,²⁵ while another found volume reduction in the insula of these adolescents, which suggests a possible neurobiological reason for impulsivity and the increased risk of suicidal behavior.²⁶

Dynamic risk factors

Research has repeatedly shown bullying is a risk factor for NSSI.²⁷ One study found that younger children who were victimized reported significantly more NSSI than older children.²⁸ New data suggest that perpetrators of bullying are also at risk for deliberate self-harm behavior (SHB), which this study defined as a behavior that is intended to cause self-harm but without suicidal intent and having a nonfatal outcome.²⁹ Victims of cyberbullying also are at a greater risk for self-harm, suicidal behaviors, and suicide attempt.³⁰ To a lesser extent, cyberbullying perpetrators are at greater risk for suicidal behaviors and suicidal ideation.³⁰ Bullying is a risk factor for NSSI not only in adolescence, but also in adulthood. Lereya et al³¹ found that victims of bullying in childhood and early adolescence were more likely to have mental health problems (including anxiety and depression) and more likely to engage in SHB—which this study defined as hurting oneself on purpose in any way—as adults.

The effects of internet use on adolescents’ mental health also has been investigated. A recent review that explored the relationship between all types of internet use (general use, internet addiction, social media, self-harm websites, forums, etc) and SHB/suicidal behavior found that young people with internet addiction, high levels of internet use, and a tendency to view websites with self-harm or suicidal content were at higher risk of engaging in SHB/suicidal behavior.³² This study did not use a specific definition for SHB or suicidal behavior.³²

Continue to: Membership in certain youth...

Membership in certain youth subcultures (eg, emo or goth) has been evaluated as potential risk factors for depression and deliberate self-harm. Bowes et al³³ found that for each unit increase in goth affiliation (not at all, not very much, somewhat, more than somewhat, very much), youth were 1.52 times more likely to engage in SHB; these researchers also reported a dose-response association between goth identification and future SHB. This study asked participants if they have ever tried to harm or hurt themselves in any manner, but did not distinguish between individuals who had harmed themselves with and without suicidal intent.³³

Personality traits such as impulsiveness and loneliness have been linked to NSSI among adolescents.^34,35 A recent study found that adolescents who met the proposed DSM-5 diagnostic criteria for NSSI scored higher on the Barratt Impulsiveness Scale, specifically in measures of:

• motor impulsiveness (ie, acting without thinking)
• attentional impulsiveness (ie, making decisions quickly)
• impulsiveness due to lack of planning (ie, failure to plan for the future).³⁴

This study also found that adolescents who identified as being lonely based on scores on the Brazilian Loneliness Scale were at a higher risk for NSSI.³⁴

A recent systematic review (32 studies) and meta-analysis (9 studies) found that school absenteeism was associated with a risk of self-harm (pooled aOR 1.37, P = .01) and suicidal ideation (pooled aOR 1.20, P = .03).³⁶ This study suggested that school absenteeism, an important marker of social exclusion, was associated with both SHB and suicidal ideation in young people.³⁶ It defined SHB as any act of self-injury or self-poisoning, regardless of intent.³⁶

Finally, family-related factors have been associated with an increased risk of NSSI. One study of 11,814 children age 9 and 10 revealed that high family conflict (OR 1.09; 95% CI, 1.05 to 1.14) and low parental monitoring (OR 0.95; 95% CI, 0.93 to 0.98) were associated with NSSI.³⁷ A smaller, community-based study found that adolescents with NSSI reported significantly less maternal support and warmth than nonclinical controls, but a cause-and-effect relationship has not yet been determined.³⁸ Parental history alone may influence adolescents’ risk of NSSI. A study that included nearly 76,000 youth found that adolescents with perceived parental alcohol problems had higher odds of self-injury, suicidal ideation, and suicide attempts.³⁹ Adolescents exposed to maternal or paternal adversities were also at a higher risk of self-harm (hazard ratio 1.5 to 5.4 among males, 1.7 to 3.9 among females).⁴⁰

Continue to: NSSI risk factors for adults

NSSI risk factors for adults

Although data regarding the prevalence of NSSI in adults are lacking, available studies report a 12-month prevalence of 0.9%² and a lifetime prevalence of 5.5% to 5.9%.⁴³ There is a significant overlap in risk factors for NSSI in adolescent and adult populations, but there are also many important differences. The static and dynamic risk factors for NSSI in adults are described in Table 3.^44-66 Table 4^44-66 summarizes the studies of NSSI in adults that we reviewed.

Static risk factors

Research findings regarding the prevalence of NSSI based on gender are varied. For years, it has been believed that women are more likely to engage in NSSI than men. Recent meta-analyses that have examined this relationship closely found that the gender difference is larger for clinical samples compared to community samples and more pronounced in younger individuals.¹¹

As is the case with adolescents, there may be ethnic variations in rates of self-harm and NSSI among adults. A 2013 study by Chesin et al⁴⁴ found that Asian and White young adults experience higher rates of NSSI than their Hispanic and Black counterparts. Evidence suggests that relative rates of self-harm for older South Asian adults are lower than in older White adults.¹⁵

Compared to heterosexual or cisgender individuals, members of sexual and gender minorities have a higher past-year and lifetime prevalence of NSSI.⁴⁵ One study found that the weighted effect size between sexual orientation and NSSI had an OR of 3 (95% CI, 2.46 to 3.66), indicating a medium-to-large effect.⁴⁶ Bisexual and transgender individuals appear to be at the highest risk for NSSI when compared to members of other sexual and gender minority groups.⁴⁵ One review that included mostly cross-sectional studies found that individuals identifying as bisexual had up to 6 times the odds of engaging in NSSI when compared to those of other sexual orientations.⁴⁷

Incarceration is a risk factor for NSSI. The rates of NSSI in criminal justice settings are higher (up to 61%) than in the general adult population (approximately 4%).⁴⁸ Recent research found that NSSI serves similar functions in correctional and non-correctional settings, primarily to regulate emotions.⁴⁸ However, there is also evidence of higher rates of NSSI being motivated by an attempt to influence the environment (ie, engaging in NSSI in order to be transferred to another prison unit) compared to NSSI in community settings.⁴⁸

Continue to: Though less robust than data...

Though less robust than data published regarding adolescents, the role of biological processes in adults engaging in NSSI has also been studied. A 2021 study by Störkel et al⁴⁹ found that levels of salivary beta-endorphins were significantly lower in adults immediately before engaging in NSSI compared to after NSSI. Furthermore, adults who engage in NSSI have lower levels of met-enkephalin (P < .01), an opioid growth factor, compared to adults who have never engaged in NSSI.²²

Dynamic risk factors

Individuals who engage in NSSI often report substance use, but there is little data on whether substance use is an independent risk factor for NSSI. Although limited, recent evidence suggests illicit substance use in both adolescents⁴¹ and adults⁵⁰ increases risk for NSSI. Richardson et al⁵⁰ found that the use of barbiturates, opiates, and sedatives significantly increased the frequency of NSSI, whereas use of marijuana, phencyclidine, and medications used to treat anxiety significantly increased the severity of NSSI. A smaller study conducted in South Africa found that individuals who engage in substance use and NSSI were more likely to be male (P < .001).⁵¹

Eating disorders and NSSI are highly comorbid.⁵² The lifetime prevalence of NSSI among individuals with eating disorders ranges from 20.6%to 37.1%.^52,53 Results are inconsistent regarding which eating disorders (if any) are greater risk factors for NSSI. One study found that the prevalence of NSSI in patients with bulimia nervosa was 32.7% (95% CI, 26.9% to 39.1%) vs 21.8% in patients with anorexia nervosa (95% CI, 18.5% to 25.6%).⁵⁴ Another study found that individuals with binge eating/purging–type eating disorders reported engaging in NSSI more frequently than those with other types of eating disorders.⁵⁵ Among patients with eating disorders who reported NSSI, risk factors included younger age of onset, more negative self-evaluation, more impulsive behavior, concomitant substance use, history of suicide attempts, childhood abuse, and peer aggression.^53,55 Body image dissatisfaction and self-criticism, even in individuals not formally diagnosed with an eating disorder, are small but significant predictors of NSSI.^56,57

Mood disorders have also been linked to NSSI.^58,59 Anxiety disorders (including generalized anxiety disorder, social phobia, panic disorder, and agoraphobia) as well as anxiety-related disorders such as obsessive-compulsive disorder have been significantly associated with NSSI (P < .001), but this relationship decreased in strength when mood instability was removed as a confounder.⁵⁸ Among patients with anxiety and anxiety-related disorders, panic disorder and posttraumatic stress disorder (PTSD) have shown the strongest association with NSSI, with pooled aORs of 2.67 and 2.06, respectively.⁵⁹

Recent studies have examined the association of other mental health disorders and symptoms with NSSI, including psychosis⁶⁰ and dissociative symptoms.⁶¹ One study found that paranoia, thought control, and auditory hallucinations were significantly associated with NSSI⁶⁰; however, after controlling for concomitant BPD, only paranoia was significantly associated with NSSI.⁶⁰ Individuals diagnosed with dissociative disorders were more likely than patients without such disorders to endorse NSSI and suicide attempts.⁶¹

Continue to: Emotional dysregulation...

Emotional dysregulation (EDR)—defined as difficulty understanding, recognizing, and managing one’s emotions—has been researched extensively in relation to NSSI.⁶² A recent review that included studies of both adolescents and adults reported a significant association between EDR and NSSI, with an OR of 2.40 (95% CI, 2.01 to 2.86).⁶² A larger effect size was observed between EDR and lifetime NSSI (OR 3.21; 95% CI, 2.63 to 3.91) compared to past-year NSSI (OR 2.32; 95% CI, 1.84 to 2.92).⁶² Patient age, sex, and sample type (clinical vs community) were not significant moderators of strength between the reported associations.⁶²

Studies examining intimate partner violence (IPV) and NSSI have found that young adults who engage in IPV (both as victims and as perpetrators) are more likely to report NSSI.^63-65 Researchers have proposed that anxiety over abandonment may explain this relationship.⁶⁴ A recent study found that individuals with bidirectional IPV (ie, both victimization and perpetration) engaged in NSSI at a higher prevalence than those engaging in unidirectional IPV or no IPV.⁶⁵ This suggests that relationship violence in general (rather than just being a victim of IPV) may be a risk factor for NSSI.⁶⁵

Finally, studies suggest that adolescents and adults who have sleep problems (insomnia, short sleep duration, long sleep onset latency, waking after sleep onset, and poor quality sleep) are more likely to report self-harm or NSSI than those without sleep problems.^42,66 In adults, this relationship is partially mediated by depressive symptoms, EDR, and PTSD.⁶⁶ In adolescents, depressive symptoms are a mediator for this relationship.⁴²

Bottom Line

Nonsuicidal self-injury (NSSI) is a significant health concern due to its association with suicide attempts. Although there are similarities in NSSI risk factors between adolescents and adults, there are also important differences. Understanding these differences is necessary to develop appropriate treatment plans.

Related Resources

• American Foundation for Suicide Prevention. https://afsp.org/
• Cipriano A, Cella S, Cotrufo P. Nonsuicidal self-injury: a systematic review. Front Psych. 2017;8:1946. doi:10.3389/ fpsyg.2017.01946
• Gold LH, Frierson RL, eds. Textbook of Suicide Risk Assessment and Management. 3rd ed. American Psychiatric Association Publishing; 2020.

Nonsuicidal self-injury (NSSI) is the direct and deliberate destruction of body tissue without intent to die.¹ Common forms of NSSI include cutting, burning, scraping/scratching skin, biting, hitting, and interfering with wound healing.² Functional theories suggest that NSSI temporarily alleviates overwhelming negative emotions and can produce feelings of relief, resulting in a reinforcing effect.³

NSSI has been shown to be a risk factor for future suicide attempts.⁴ A 2018 study found that NSSI is associated with an increased risk of subsequent suicidal ideation (odds ratio [OR] 2.8), suicide plan (OR 3.0), and suicide attempt (OR 5.5).⁵ NSSI is also associated with individuals who had suicidal ideation and formed a suicide plan, and individuals who had a suicide plan and attempted suicide (ORs 1.7 to 2.1).⁵ Another study found that 70% of adolescents who engage in NSSI have attempted suicide during their lifetime, and 55% have multiple attempts.⁶

Given the overlap between suicide attempts and NSSI, performing a thorough suicide risk assessment (which is beyond the scope of this article) is crucial. This article describes the static and dynamic risk factors for NSSI in adolescents and adults, which can help us perform a suicide risk assessment and allow us to formulate an appropriate treatment plan that includes safety-based interventions.

NSSI risk factors for adolescents

From developing sexual identity and undergoing puberty to achieving increased independence from their parents and developing a sense of autonomy, adolescents undergo many biological, psychological, and social changes before reaching adulthood.⁷ Data suggest that NSSI often begins in adolescence, with a typical onset at age 13 or 14.³ Community studies show that one-third to one-half of adolescents in the United States have engaged in NSSI.^8,9 Previously, NSSI during adolescence was associated with 3 major diagnostic categories: eating disorders, developmental disabilities, and borderline personality disorder (BPD).¹⁰ However, recent data suggest that NSSI is also common outside of these categories. Here we describe static and dynamic risk factors for NSSI in adolescents (Table 1^11-42). Table 2^11-42 summarizes the studies of NSSI in adolescents that we reviewed.

Static risk factors

Female adolescents and adults engage in NSSI at higher rates than males. The difference is larger in clinical populations compared to the general population.¹¹

A large portion of research about NSSI has been conducted in studies in which the majority of participants were White.¹² Most studies report a higher prevalence of NSSI among non-Hispanic White youth,¹³ but some suggest other ethnic groups may also experience high rates of self-harm and NSSI.^13-15 Several studies have demonstrated high rates of self-harm among South Asian adult females compared with White adult females, but this difference may be less pronounced in adolescents.¹⁴ One study in the United Kingdom found that White females age 10 to 14 had higher rates of self-harm compared to South Asian females,¹⁴ while another found that risk and rates of self-harm in young South Asian people varied by city and country of origin.¹⁵ Young Black females¹⁵ and young Black males¹³ also may be at an increased risk of self-harm. One review found that Black females were more likely to self-harm than Asian or White groups.¹⁵

Several studies suggest that sexual minority adolescents (SMA) (eg, lesbian, gay, bisexual, transgender, queer) are at greater risk for NSSI than heterosexual adolescents.¹⁶ SMA have been shown to engage in a significantly greater frequency of NSSI and more types of NSSI than heterosexual adolescents.¹⁶ Furthermore, on the Inventory of Statements about Self-Injury, SMA self-reported using NSSI for intrapersonal functions (eg, for affect regulation, antisuicide, self-punishment) significantly greater than their heterosexual peers; however, there were no significant differences between the 2 groups on interpersonal functions (eg, autonomy, interpersonal boundaries, peer bonding, sensation-seeking).¹⁶

Continue to: Transgender and gender nonconfirming...

Transgender and gender nonconfirming (GNC) youth are at a particularly high risk for NSSI; 30% to 45.5% of transgender adolescents report self-injury.¹⁷ Factors shown to distinguish transgender/GNC youth who engage in NSSI from those who do not include having a mental health problem, depression, running away from home, substance use, lower self-esteem/greater self-criticism, experiencing transphobia victimization, and having more interpersonal problems.^18,19 Among transgender/GNC youth, those whose biological sex is female are more likely to report NSSI than those whose biological sex is male (ie, transgendered adolescent males are more likely to report NSSI than transgendered adolescent females).^18,19

Most forms of childhood maltreatment have been associated with NSSI. In a recently published review, Liu et al²⁰ found that childhood maltreatment (including sexual abuse, physical abuse, emotional abuse, and physical neglect) was associated with an increased risk for NSSI. However, conflicting evidence suggests that when confounders are removed, only childhood emotional abuse was directly associated with NSSI.²¹ Current evidence is modest for childhood emotional neglect as a risk factor for NSSI.²⁰

Increasing research is investigating the biological processes that may be implicated in NSSI. Some studies suggest that endogenous opioids,²² monoamine neurotransmitters,²² and the hypothalamic-pituitary-adrenal (HPA) axis²³ may play a role in NSSI. Compared to healthy controls, adolescents engaging in NSSI have been shown to have lower pain intensity (P = .036), higher pain thresholds (P = .040), and lower beta-endorphins (endogenous opioid hormones involved in mediating stress and pain) (P = .002).²⁴ There may be alterations in the HPA axis among adolescents who engage in NSSI, more specifically stronger cortisol awakening responses.²³ Both functional and standard MRI have been used to study the neurobiology of NSSI. One study demonstrated differences in functional connectivity between brain areas linked to neuroregulation of emotions in adolescents who engage in NSSI,²⁵ while another found volume reduction in the insula of these adolescents, which suggests a possible neurobiological reason for impulsivity and the increased risk of suicidal behavior.²⁶

Dynamic risk factors

Research has repeatedly shown bullying is a risk factor for NSSI.²⁷ One study found that younger children who were victimized reported significantly more NSSI than older children.²⁸ New data suggest that perpetrators of bullying are also at risk for deliberate self-harm behavior (SHB), which this study defined as a behavior that is intended to cause self-harm but without suicidal intent and having a nonfatal outcome.²⁹ Victims of cyberbullying also are at a greater risk for self-harm, suicidal behaviors, and suicide attempt.³⁰ To a lesser extent, cyberbullying perpetrators are at greater risk for suicidal behaviors and suicidal ideation.³⁰ Bullying is a risk factor for NSSI not only in adolescence, but also in adulthood. Lereya et al³¹ found that victims of bullying in childhood and early adolescence were more likely to have mental health problems (including anxiety and depression) and more likely to engage in SHB—which this study defined as hurting oneself on purpose in any way—as adults.

The effects of internet use on adolescents’ mental health also has been investigated. A recent review that explored the relationship between all types of internet use (general use, internet addiction, social media, self-harm websites, forums, etc) and SHB/suicidal behavior found that young people with internet addiction, high levels of internet use, and a tendency to view websites with self-harm or suicidal content were at higher risk of engaging in SHB/suicidal behavior.³² This study did not use a specific definition for SHB or suicidal behavior.³²

Continue to: Membership in certain youth...

Membership in certain youth subcultures (eg, emo or goth) has been evaluated as potential risk factors for depression and deliberate self-harm. Bowes et al³³ found that for each unit increase in goth affiliation (not at all, not very much, somewhat, more than somewhat, very much), youth were 1.52 times more likely to engage in SHB; these researchers also reported a dose-response association between goth identification and future SHB. This study asked participants if they have ever tried to harm or hurt themselves in any manner, but did not distinguish between individuals who had harmed themselves with and without suicidal intent.³³

Personality traits such as impulsiveness and loneliness have been linked to NSSI among adolescents.^34,35 A recent study found that adolescents who met the proposed DSM-5 diagnostic criteria for NSSI scored higher on the Barratt Impulsiveness Scale, specifically in measures of:

• motor impulsiveness (ie, acting without thinking)
• attentional impulsiveness (ie, making decisions quickly)
• impulsiveness due to lack of planning (ie, failure to plan for the future).³⁴

This study also found that adolescents who identified as being lonely based on scores on the Brazilian Loneliness Scale were at a higher risk for NSSI.³⁴

A recent systematic review (32 studies) and meta-analysis (9 studies) found that school absenteeism was associated with a risk of self-harm (pooled aOR 1.37, P = .01) and suicidal ideation (pooled aOR 1.20, P = .03).³⁶ This study suggested that school absenteeism, an important marker of social exclusion, was associated with both SHB and suicidal ideation in young people.³⁶ It defined SHB as any act of self-injury or self-poisoning, regardless of intent.³⁶

Finally, family-related factors have been associated with an increased risk of NSSI. One study of 11,814 children age 9 and 10 revealed that high family conflict (OR 1.09; 95% CI, 1.05 to 1.14) and low parental monitoring (OR 0.95; 95% CI, 0.93 to 0.98) were associated with NSSI.³⁷ A smaller, community-based study found that adolescents with NSSI reported significantly less maternal support and warmth than nonclinical controls, but a cause-and-effect relationship has not yet been determined.³⁸ Parental history alone may influence adolescents’ risk of NSSI. A study that included nearly 76,000 youth found that adolescents with perceived parental alcohol problems had higher odds of self-injury, suicidal ideation, and suicide attempts.³⁹ Adolescents exposed to maternal or paternal adversities were also at a higher risk of self-harm (hazard ratio 1.5 to 5.4 among males, 1.7 to 3.9 among females).⁴⁰

Continue to: NSSI risk factors for adults

NSSI risk factors for adults

Although data regarding the prevalence of NSSI in adults are lacking, available studies report a 12-month prevalence of 0.9%² and a lifetime prevalence of 5.5% to 5.9%.⁴³ There is a significant overlap in risk factors for NSSI in adolescent and adult populations, but there are also many important differences. The static and dynamic risk factors for NSSI in adults are described in Table 3.^44-66 Table 4^44-66 summarizes the studies of NSSI in adults that we reviewed.

Static risk factors

Research findings regarding the prevalence of NSSI based on gender are varied. For years, it has been believed that women are more likely to engage in NSSI than men. Recent meta-analyses that have examined this relationship closely found that the gender difference is larger for clinical samples compared to community samples and more pronounced in younger individuals.¹¹

As is the case with adolescents, there may be ethnic variations in rates of self-harm and NSSI among adults. A 2013 study by Chesin et al⁴⁴ found that Asian and White young adults experience higher rates of NSSI than their Hispanic and Black counterparts. Evidence suggests that relative rates of self-harm for older South Asian adults are lower than in older White adults.¹⁵

Compared to heterosexual or cisgender individuals, members of sexual and gender minorities have a higher past-year and lifetime prevalence of NSSI.⁴⁵ One study found that the weighted effect size between sexual orientation and NSSI had an OR of 3 (95% CI, 2.46 to 3.66), indicating a medium-to-large effect.⁴⁶ Bisexual and transgender individuals appear to be at the highest risk for NSSI when compared to members of other sexual and gender minority groups.⁴⁵ One review that included mostly cross-sectional studies found that individuals identifying as bisexual had up to 6 times the odds of engaging in NSSI when compared to those of other sexual orientations.⁴⁷

Incarceration is a risk factor for NSSI. The rates of NSSI in criminal justice settings are higher (up to 61%) than in the general adult population (approximately 4%).⁴⁸ Recent research found that NSSI serves similar functions in correctional and non-correctional settings, primarily to regulate emotions.⁴⁸ However, there is also evidence of higher rates of NSSI being motivated by an attempt to influence the environment (ie, engaging in NSSI in order to be transferred to another prison unit) compared to NSSI in community settings.⁴⁸

Continue to: Though less robust than data...

Though less robust than data published regarding adolescents, the role of biological processes in adults engaging in NSSI has also been studied. A 2021 study by Störkel et al⁴⁹ found that levels of salivary beta-endorphins were significantly lower in adults immediately before engaging in NSSI compared to after NSSI. Furthermore, adults who engage in NSSI have lower levels of met-enkephalin (P < .01), an opioid growth factor, compared to adults who have never engaged in NSSI.²²

Dynamic risk factors

Individuals who engage in NSSI often report substance use, but there is little data on whether substance use is an independent risk factor for NSSI. Although limited, recent evidence suggests illicit substance use in both adolescents⁴¹ and adults⁵⁰ increases risk for NSSI. Richardson et al⁵⁰ found that the use of barbiturates, opiates, and sedatives significantly increased the frequency of NSSI, whereas use of marijuana, phencyclidine, and medications used to treat anxiety significantly increased the severity of NSSI. A smaller study conducted in South Africa found that individuals who engage in substance use and NSSI were more likely to be male (P < .001).⁵¹

Eating disorders and NSSI are highly comorbid.⁵² The lifetime prevalence of NSSI among individuals with eating disorders ranges from 20.6%to 37.1%.^52,53 Results are inconsistent regarding which eating disorders (if any) are greater risk factors for NSSI. One study found that the prevalence of NSSI in patients with bulimia nervosa was 32.7% (95% CI, 26.9% to 39.1%) vs 21.8% in patients with anorexia nervosa (95% CI, 18.5% to 25.6%).⁵⁴ Another study found that individuals with binge eating/purging–type eating disorders reported engaging in NSSI more frequently than those with other types of eating disorders.⁵⁵ Among patients with eating disorders who reported NSSI, risk factors included younger age of onset, more negative self-evaluation, more impulsive behavior, concomitant substance use, history of suicide attempts, childhood abuse, and peer aggression.^53,55 Body image dissatisfaction and self-criticism, even in individuals not formally diagnosed with an eating disorder, are small but significant predictors of NSSI.^56,57

Mood disorders have also been linked to NSSI.^58,59 Anxiety disorders (including generalized anxiety disorder, social phobia, panic disorder, and agoraphobia) as well as anxiety-related disorders such as obsessive-compulsive disorder have been significantly associated with NSSI (P < .001), but this relationship decreased in strength when mood instability was removed as a confounder.⁵⁸ Among patients with anxiety and anxiety-related disorders, panic disorder and posttraumatic stress disorder (PTSD) have shown the strongest association with NSSI, with pooled aORs of 2.67 and 2.06, respectively.⁵⁹

Recent studies have examined the association of other mental health disorders and symptoms with NSSI, including psychosis⁶⁰ and dissociative symptoms.⁶¹ One study found that paranoia, thought control, and auditory hallucinations were significantly associated with NSSI⁶⁰; however, after controlling for concomitant BPD, only paranoia was significantly associated with NSSI.⁶⁰ Individuals diagnosed with dissociative disorders were more likely than patients without such disorders to endorse NSSI and suicide attempts.⁶¹

Continue to: Emotional dysregulation...

Emotional dysregulation (EDR)—defined as difficulty understanding, recognizing, and managing one’s emotions—has been researched extensively in relation to NSSI.⁶² A recent review that included studies of both adolescents and adults reported a significant association between EDR and NSSI, with an OR of 2.40 (95% CI, 2.01 to 2.86).⁶² A larger effect size was observed between EDR and lifetime NSSI (OR 3.21; 95% CI, 2.63 to 3.91) compared to past-year NSSI (OR 2.32; 95% CI, 1.84 to 2.92).⁶² Patient age, sex, and sample type (clinical vs community) were not significant moderators of strength between the reported associations.⁶²

Studies examining intimate partner violence (IPV) and NSSI have found that young adults who engage in IPV (both as victims and as perpetrators) are more likely to report NSSI.^63-65 Researchers have proposed that anxiety over abandonment may explain this relationship.⁶⁴ A recent study found that individuals with bidirectional IPV (ie, both victimization and perpetration) engaged in NSSI at a higher prevalence than those engaging in unidirectional IPV or no IPV.⁶⁵ This suggests that relationship violence in general (rather than just being a victim of IPV) may be a risk factor for NSSI.⁶⁵

Finally, studies suggest that adolescents and adults who have sleep problems (insomnia, short sleep duration, long sleep onset latency, waking after sleep onset, and poor quality sleep) are more likely to report self-harm or NSSI than those without sleep problems.^42,66 In adults, this relationship is partially mediated by depressive symptoms, EDR, and PTSD.⁶⁶ In adolescents, depressive symptoms are a mediator for this relationship.⁴²

Bottom Line

Nonsuicidal self-injury (NSSI) is a significant health concern due to its association with suicide attempts. Although there are similarities in NSSI risk factors between adolescents and adults, there are also important differences. Understanding these differences is necessary to develop appropriate treatment plans.

Related Resources

• American Foundation for Suicide Prevention. https://afsp.org/
• Cipriano A, Cella S, Cotrufo P. Nonsuicidal self-injury: a systematic review. Front Psych. 2017;8:1946. doi:10.3389/ fpsyg.2017.01946
• Gold LH, Frierson RL, eds. Textbook of Suicide Risk Assessment and Management. 3rd ed. American Psychiatric Association Publishing; 2020.

Nonsuicidal self-injury (NSSI) is the direct and deliberate destruction of body tissue without intent to die.¹ Common forms of NSSI include cutting, burning, scraping/scratching skin, biting, hitting, and interfering with wound healing.² Functional theories suggest that NSSI temporarily alleviates overwhelming negative emotions and can produce feelings of relief, resulting in a reinforcing effect.³

NSSI has been shown to be a risk factor for future suicide attempts.⁴ A 2018 study found that NSSI is associated with an increased risk of subsequent suicidal ideation (odds ratio [OR] 2.8), suicide plan (OR 3.0), and suicide attempt (OR 5.5).⁵ NSSI is also associated with individuals who had suicidal ideation and formed a suicide plan, and individuals who had a suicide plan and attempted suicide (ORs 1.7 to 2.1).⁵ Another study found that 70% of adolescents who engage in NSSI have attempted suicide during their lifetime, and 55% have multiple attempts.⁶

Given the overlap between suicide attempts and NSSI, performing a thorough suicide risk assessment (which is beyond the scope of this article) is crucial. This article describes the static and dynamic risk factors for NSSI in adolescents and adults, which can help us perform a suicide risk assessment and allow us to formulate an appropriate treatment plan that includes safety-based interventions.

NSSI risk factors for adolescents

From developing sexual identity and undergoing puberty to achieving increased independence from their parents and developing a sense of autonomy, adolescents undergo many biological, psychological, and social changes before reaching adulthood.⁷ Data suggest that NSSI often begins in adolescence, with a typical onset at age 13 or 14.³ Community studies show that one-third to one-half of adolescents in the United States have engaged in NSSI.^8,9 Previously, NSSI during adolescence was associated with 3 major diagnostic categories: eating disorders, developmental disabilities, and borderline personality disorder (BPD).¹⁰ However, recent data suggest that NSSI is also common outside of these categories. Here we describe static and dynamic risk factors for NSSI in adolescents (Table 1^11-42). Table 2^11-42 summarizes the studies of NSSI in adolescents that we reviewed.

Static risk factors

Female adolescents and adults engage in NSSI at higher rates than males. The difference is larger in clinical populations compared to the general population.¹¹

A large portion of research about NSSI has been conducted in studies in which the majority of participants were White.¹² Most studies report a higher prevalence of NSSI among non-Hispanic White youth,¹³ but some suggest other ethnic groups may also experience high rates of self-harm and NSSI.^13-15 Several studies have demonstrated high rates of self-harm among South Asian adult females compared with White adult females, but this difference may be less pronounced in adolescents.¹⁴ One study in the United Kingdom found that White females age 10 to 14 had higher rates of self-harm compared to South Asian females,¹⁴ while another found that risk and rates of self-harm in young South Asian people varied by city and country of origin.¹⁵ Young Black females¹⁵ and young Black males¹³ also may be at an increased risk of self-harm. One review found that Black females were more likely to self-harm than Asian or White groups.¹⁵

Several studies suggest that sexual minority adolescents (SMA) (eg, lesbian, gay, bisexual, transgender, queer) are at greater risk for NSSI than heterosexual adolescents.¹⁶ SMA have been shown to engage in a significantly greater frequency of NSSI and more types of NSSI than heterosexual adolescents.¹⁶ Furthermore, on the Inventory of Statements about Self-Injury, SMA self-reported using NSSI for intrapersonal functions (eg, for affect regulation, antisuicide, self-punishment) significantly greater than their heterosexual peers; however, there were no significant differences between the 2 groups on interpersonal functions (eg, autonomy, interpersonal boundaries, peer bonding, sensation-seeking).¹⁶

Continue to: Transgender and gender nonconfirming...

Transgender and gender nonconfirming (GNC) youth are at a particularly high risk for NSSI; 30% to 45.5% of transgender adolescents report self-injury.¹⁷ Factors shown to distinguish transgender/GNC youth who engage in NSSI from those who do not include having a mental health problem, depression, running away from home, substance use, lower self-esteem/greater self-criticism, experiencing transphobia victimization, and having more interpersonal problems.^18,19 Among transgender/GNC youth, those whose biological sex is female are more likely to report NSSI than those whose biological sex is male (ie, transgendered adolescent males are more likely to report NSSI than transgendered adolescent females).^18,19

Most forms of childhood maltreatment have been associated with NSSI. In a recently published review, Liu et al²⁰ found that childhood maltreatment (including sexual abuse, physical abuse, emotional abuse, and physical neglect) was associated with an increased risk for NSSI. However, conflicting evidence suggests that when confounders are removed, only childhood emotional abuse was directly associated with NSSI.²¹ Current evidence is modest for childhood emotional neglect as a risk factor for NSSI.²⁰

Increasing research is investigating the biological processes that may be implicated in NSSI. Some studies suggest that endogenous opioids,²² monoamine neurotransmitters,²² and the hypothalamic-pituitary-adrenal (HPA) axis²³ may play a role in NSSI. Compared to healthy controls, adolescents engaging in NSSI have been shown to have lower pain intensity (P = .036), higher pain thresholds (P = .040), and lower beta-endorphins (endogenous opioid hormones involved in mediating stress and pain) (P = .002).²⁴ There may be alterations in the HPA axis among adolescents who engage in NSSI, more specifically stronger cortisol awakening responses.²³ Both functional and standard MRI have been used to study the neurobiology of NSSI. One study demonstrated differences in functional connectivity between brain areas linked to neuroregulation of emotions in adolescents who engage in NSSI,²⁵ while another found volume reduction in the insula of these adolescents, which suggests a possible neurobiological reason for impulsivity and the increased risk of suicidal behavior.²⁶

Dynamic risk factors

Research has repeatedly shown bullying is a risk factor for NSSI.²⁷ One study found that younger children who were victimized reported significantly more NSSI than older children.²⁸ New data suggest that perpetrators of bullying are also at risk for deliberate self-harm behavior (SHB), which this study defined as a behavior that is intended to cause self-harm but without suicidal intent and having a nonfatal outcome.²⁹ Victims of cyberbullying also are at a greater risk for self-harm, suicidal behaviors, and suicide attempt.³⁰ To a lesser extent, cyberbullying perpetrators are at greater risk for suicidal behaviors and suicidal ideation.³⁰ Bullying is a risk factor for NSSI not only in adolescence, but also in adulthood. Lereya et al³¹ found that victims of bullying in childhood and early adolescence were more likely to have mental health problems (including anxiety and depression) and more likely to engage in SHB—which this study defined as hurting oneself on purpose in any way—as adults.

The effects of internet use on adolescents’ mental health also has been investigated. A recent review that explored the relationship between all types of internet use (general use, internet addiction, social media, self-harm websites, forums, etc) and SHB/suicidal behavior found that young people with internet addiction, high levels of internet use, and a tendency to view websites with self-harm or suicidal content were at higher risk of engaging in SHB/suicidal behavior.³² This study did not use a specific definition for SHB or suicidal behavior.³²

Continue to: Membership in certain youth...

Membership in certain youth subcultures (eg, emo or goth) has been evaluated as potential risk factors for depression and deliberate self-harm. Bowes et al³³ found that for each unit increase in goth affiliation (not at all, not very much, somewhat, more than somewhat, very much), youth were 1.52 times more likely to engage in SHB; these researchers also reported a dose-response association between goth identification and future SHB. This study asked participants if they have ever tried to harm or hurt themselves in any manner, but did not distinguish between individuals who had harmed themselves with and without suicidal intent.³³

Personality traits such as impulsiveness and loneliness have been linked to NSSI among adolescents.^34,35 A recent study found that adolescents who met the proposed DSM-5 diagnostic criteria for NSSI scored higher on the Barratt Impulsiveness Scale, specifically in measures of:

• motor impulsiveness (ie, acting without thinking)
• attentional impulsiveness (ie, making decisions quickly)
• impulsiveness due to lack of planning (ie, failure to plan for the future).³⁴

This study also found that adolescents who identified as being lonely based on scores on the Brazilian Loneliness Scale were at a higher risk for NSSI.³⁴

A recent systematic review (32 studies) and meta-analysis (9 studies) found that school absenteeism was associated with a risk of self-harm (pooled aOR 1.37, P = .01) and suicidal ideation (pooled aOR 1.20, P = .03).³⁶ This study suggested that school absenteeism, an important marker of social exclusion, was associated with both SHB and suicidal ideation in young people.³⁶ It defined SHB as any act of self-injury or self-poisoning, regardless of intent.³⁶

Finally, family-related factors have been associated with an increased risk of NSSI. One study of 11,814 children age 9 and 10 revealed that high family conflict (OR 1.09; 95% CI, 1.05 to 1.14) and low parental monitoring (OR 0.95; 95% CI, 0.93 to 0.98) were associated with NSSI.³⁷ A smaller, community-based study found that adolescents with NSSI reported significantly less maternal support and warmth than nonclinical controls, but a cause-and-effect relationship has not yet been determined.³⁸ Parental history alone may influence adolescents’ risk of NSSI. A study that included nearly 76,000 youth found that adolescents with perceived parental alcohol problems had higher odds of self-injury, suicidal ideation, and suicide attempts.³⁹ Adolescents exposed to maternal or paternal adversities were also at a higher risk of self-harm (hazard ratio 1.5 to 5.4 among males, 1.7 to 3.9 among females).⁴⁰

Continue to: NSSI risk factors for adults

NSSI risk factors for adults

Although data regarding the prevalence of NSSI in adults are lacking, available studies report a 12-month prevalence of 0.9%² and a lifetime prevalence of 5.5% to 5.9%.⁴³ There is a significant overlap in risk factors for NSSI in adolescent and adult populations, but there are also many important differences. The static and dynamic risk factors for NSSI in adults are described in Table 3.^44-66 Table 4^44-66 summarizes the studies of NSSI in adults that we reviewed.

Static risk factors

Research findings regarding the prevalence of NSSI based on gender are varied. For years, it has been believed that women are more likely to engage in NSSI than men. Recent meta-analyses that have examined this relationship closely found that the gender difference is larger for clinical samples compared to community samples and more pronounced in younger individuals.¹¹

As is the case with adolescents, there may be ethnic variations in rates of self-harm and NSSI among adults. A 2013 study by Chesin et al⁴⁴ found that Asian and White young adults experience higher rates of NSSI than their Hispanic and Black counterparts. Evidence suggests that relative rates of self-harm for older South Asian adults are lower than in older White adults.¹⁵

Compared to heterosexual or cisgender individuals, members of sexual and gender minorities have a higher past-year and lifetime prevalence of NSSI.⁴⁵ One study found that the weighted effect size between sexual orientation and NSSI had an OR of 3 (95% CI, 2.46 to 3.66), indicating a medium-to-large effect.⁴⁶ Bisexual and transgender individuals appear to be at the highest risk for NSSI when compared to members of other sexual and gender minority groups.⁴⁵ One review that included mostly cross-sectional studies found that individuals identifying as bisexual had up to 6 times the odds of engaging in NSSI when compared to those of other sexual orientations.⁴⁷

Incarceration is a risk factor for NSSI. The rates of NSSI in criminal justice settings are higher (up to 61%) than in the general adult population (approximately 4%).⁴⁸ Recent research found that NSSI serves similar functions in correctional and non-correctional settings, primarily to regulate emotions.⁴⁸ However, there is also evidence of higher rates of NSSI being motivated by an attempt to influence the environment (ie, engaging in NSSI in order to be transferred to another prison unit) compared to NSSI in community settings.⁴⁸

Continue to: Though less robust than data...

Though less robust than data published regarding adolescents, the role of biological processes in adults engaging in NSSI has also been studied. A 2021 study by Störkel et al⁴⁹ found that levels of salivary beta-endorphins were significantly lower in adults immediately before engaging in NSSI compared to after NSSI. Furthermore, adults who engage in NSSI have lower levels of met-enkephalin (P < .01), an opioid growth factor, compared to adults who have never engaged in NSSI.²²

Dynamic risk factors

Individuals who engage in NSSI often report substance use, but there is little data on whether substance use is an independent risk factor for NSSI. Although limited, recent evidence suggests illicit substance use in both adolescents⁴¹ and adults⁵⁰ increases risk for NSSI. Richardson et al⁵⁰ found that the use of barbiturates, opiates, and sedatives significantly increased the frequency of NSSI, whereas use of marijuana, phencyclidine, and medications used to treat anxiety significantly increased the severity of NSSI. A smaller study conducted in South Africa found that individuals who engage in substance use and NSSI were more likely to be male (P < .001).⁵¹

Eating disorders and NSSI are highly comorbid.⁵² The lifetime prevalence of NSSI among individuals with eating disorders ranges from 20.6%to 37.1%.^52,53 Results are inconsistent regarding which eating disorders (if any) are greater risk factors for NSSI. One study found that the prevalence of NSSI in patients with bulimia nervosa was 32.7% (95% CI, 26.9% to 39.1%) vs 21.8% in patients with anorexia nervosa (95% CI, 18.5% to 25.6%).⁵⁴ Another study found that individuals with binge eating/purging–type eating disorders reported engaging in NSSI more frequently than those with other types of eating disorders.⁵⁵ Among patients with eating disorders who reported NSSI, risk factors included younger age of onset, more negative self-evaluation, more impulsive behavior, concomitant substance use, history of suicide attempts, childhood abuse, and peer aggression.^53,55 Body image dissatisfaction and self-criticism, even in individuals not formally diagnosed with an eating disorder, are small but significant predictors of NSSI.^56,57

Mood disorders have also been linked to NSSI.^58,59 Anxiety disorders (including generalized anxiety disorder, social phobia, panic disorder, and agoraphobia) as well as anxiety-related disorders such as obsessive-compulsive disorder have been significantly associated with NSSI (P < .001), but this relationship decreased in strength when mood instability was removed as a confounder.⁵⁸ Among patients with anxiety and anxiety-related disorders, panic disorder and posttraumatic stress disorder (PTSD) have shown the strongest association with NSSI, with pooled aORs of 2.67 and 2.06, respectively.⁵⁹

Recent studies have examined the association of other mental health disorders and symptoms with NSSI, including psychosis⁶⁰ and dissociative symptoms.⁶¹ One study found that paranoia, thought control, and auditory hallucinations were significantly associated with NSSI⁶⁰; however, after controlling for concomitant BPD, only paranoia was significantly associated with NSSI.⁶⁰ Individuals diagnosed with dissociative disorders were more likely than patients without such disorders to endorse NSSI and suicide attempts.⁶¹

Continue to: Emotional dysregulation...

Emotional dysregulation (EDR)—defined as difficulty understanding, recognizing, and managing one’s emotions—has been researched extensively in relation to NSSI.⁶² A recent review that included studies of both adolescents and adults reported a significant association between EDR and NSSI, with an OR of 2.40 (95% CI, 2.01 to 2.86).⁶² A larger effect size was observed between EDR and lifetime NSSI (OR 3.21; 95% CI, 2.63 to 3.91) compared to past-year NSSI (OR 2.32; 95% CI, 1.84 to 2.92).⁶² Patient age, sex, and sample type (clinical vs community) were not significant moderators of strength between the reported associations.⁶²

Studies examining intimate partner violence (IPV) and NSSI have found that young adults who engage in IPV (both as victims and as perpetrators) are more likely to report NSSI.^63-65 Researchers have proposed that anxiety over abandonment may explain this relationship.⁶⁴ A recent study found that individuals with bidirectional IPV (ie, both victimization and perpetration) engaged in NSSI at a higher prevalence than those engaging in unidirectional IPV or no IPV.⁶⁵ This suggests that relationship violence in general (rather than just being a victim of IPV) may be a risk factor for NSSI.⁶⁵

Finally, studies suggest that adolescents and adults who have sleep problems (insomnia, short sleep duration, long sleep onset latency, waking after sleep onset, and poor quality sleep) are more likely to report self-harm or NSSI than those without sleep problems.^42,66 In adults, this relationship is partially mediated by depressive symptoms, EDR, and PTSD.⁶⁶ In adolescents, depressive symptoms are a mediator for this relationship.⁴²

Bottom Line

Nonsuicidal self-injury (NSSI) is a significant health concern due to its association with suicide attempts. Although there are similarities in NSSI risk factors between adolescents and adults, there are also important differences. Understanding these differences is necessary to develop appropriate treatment plans.

Related Resources

• American Foundation for Suicide Prevention. https://afsp.org/
• Cipriano A, Cella S, Cotrufo P. Nonsuicidal self-injury: a systematic review. Front Psych. 2017;8:1946. doi:10.3389/ fpsyg.2017.01946
• Gold LH, Frierson RL, eds. Textbook of Suicide Risk Assessment and Management. 3rd ed. American Psychiatric Association Publishing; 2020.

The FDA defines “off-label” prescribing as prescribing an FDA-approved medication for an unapproved use, such as for an unapproved clinical indication, for a higher-than-approved dose, or for a patient who is not part of the FDA-approved population (eg, children or geriatric patients).¹ Off-label prescribing is common in psychiatry; approximately 13% of psychiatry patients are prescribed off-label psychotropic medications.² The American Psychiatric Association strongly supports “the autonomous clinical decision-making authority of a physician” and “a physician’s lawful use of an FDA-approved drug product or medical device for an off-label indication when such use is based upon sound scientific evidence in conjunction with sound medical judgment.”³ Because many psychiatric diagnoses have no FDA-approved medications, off-label prescribing often may be a psychiatrist’s only pharmacologic option.

Unfortunately, off-label prescribing can increase a psychiatrist’s risk for liability when treatment falls short of patients’ expectations, or when patients allege that they were injured by the use of an off-label medication. Off-label prescribing does not automatically lead to losing a malpractice suit because the FDA states that physicians can prescribe approved medications for any scientifically supported use, including off-label.¹ Medical malpractice lawsuits alleging negligence in prescribing practices, such as off-label prescribing, typically include allegations against the psychiatrist for failure to⁴:

• adequately assess the patient
• consult the patient’s medical records
• obtain informed consent from the patient
• appropriately prescribe a medication for the clinical indication, dosage, patient’s age, etc.
• monitor for adverse effects and therapeutic effectiveness.

Steps to minimize your risk

When prescribing a medication off-label, the following approaches can help reduce your liability risk:

Conduct a comprehensive clinical assess­ment. This should include requesting and reviewing your patient’s medical records.

Explain your motivation. Explain to your patient how prescribing an off-label medication can directly benefit him/her. Make it clear that you are not conducting experimental research by prescribing off-label because some patients might perceive this as a covert form of research.

Know the medications you prescribe. Although this sounds obvious, psychiatrists should thoroughly understand how each medication they prescribe is likely to clinically affect their patient. This information is available from many sources, including the FDA’s medication information sheets and the manufacturer’s medication package inserts. If possible, make sure that your off-label prescribing is supported by reputable, peer-reviewed literature.

Obtain informed consent. Tell your patient that the medication you are recommending is being prescribed off-label. Discuss the medication’s risks, benefits, adverse effects, associated “black-box” warnings, off-label uses, and alternatives to the off-label medication.⁴ Allow time for the patient to ask questions about these treatments.

Continue to: Document all steps

Document all steps. There is an adage in medicine that “If it’s not written, it wasn’t done.” To help reduce your liability risk when prescribing off-label, be sure to document the following⁴:

• your clinical assessment
• information you gleaned from the patient’s medical records
• your review of information regarding both therapeutic and adverse effects of the medication you want to prescribe
• your discussion of informed consent, including documentation that the patient is aware that the medication is being prescribed off-label
• your clinical rationale for why the off-label medication is in the patient’s best interest.

Also, document the steps you take to monitor for adverse events and therapeutic effectiveness.⁴ Overall, the goal of documentation should be to support the adequate continuing care of our patients.

The FDA defines “off-label” prescribing as prescribing an FDA-approved medication for an unapproved use, such as for an unapproved clinical indication, for a higher-than-approved dose, or for a patient who is not part of the FDA-approved population (eg, children or geriatric patients).¹ Off-label prescribing is common in psychiatry; approximately 13% of psychiatry patients are prescribed off-label psychotropic medications.² The American Psychiatric Association strongly supports “the autonomous clinical decision-making authority of a physician” and “a physician’s lawful use of an FDA-approved drug product or medical device for an off-label indication when such use is based upon sound scientific evidence in conjunction with sound medical judgment.”³ Because many psychiatric diagnoses have no FDA-approved medications, off-label prescribing often may be a psychiatrist’s only pharmacologic option.

Unfortunately, off-label prescribing can increase a psychiatrist’s risk for liability when treatment falls short of patients’ expectations, or when patients allege that they were injured by the use of an off-label medication. Off-label prescribing does not automatically lead to losing a malpractice suit because the FDA states that physicians can prescribe approved medications for any scientifically supported use, including off-label.¹ Medical malpractice lawsuits alleging negligence in prescribing practices, such as off-label prescribing, typically include allegations against the psychiatrist for failure to⁴:

• adequately assess the patient
• consult the patient’s medical records
• obtain informed consent from the patient
• appropriately prescribe a medication for the clinical indication, dosage, patient’s age, etc.
• monitor for adverse effects and therapeutic effectiveness.

Steps to minimize your risk

When prescribing a medication off-label, the following approaches can help reduce your liability risk:

Conduct a comprehensive clinical assess­ment. This should include requesting and reviewing your patient’s medical records.

Explain your motivation. Explain to your patient how prescribing an off-label medication can directly benefit him/her. Make it clear that you are not conducting experimental research by prescribing off-label because some patients might perceive this as a covert form of research.

Know the medications you prescribe. Although this sounds obvious, psychiatrists should thoroughly understand how each medication they prescribe is likely to clinically affect their patient. This information is available from many sources, including the FDA’s medication information sheets and the manufacturer’s medication package inserts. If possible, make sure that your off-label prescribing is supported by reputable, peer-reviewed literature.

Obtain informed consent. Tell your patient that the medication you are recommending is being prescribed off-label. Discuss the medication’s risks, benefits, adverse effects, associated “black-box” warnings, off-label uses, and alternatives to the off-label medication.⁴ Allow time for the patient to ask questions about these treatments.

Continue to: Document all steps

Document all steps. There is an adage in medicine that “If it’s not written, it wasn’t done.” To help reduce your liability risk when prescribing off-label, be sure to document the following⁴:

• your clinical assessment
• information you gleaned from the patient’s medical records
• your review of information regarding both therapeutic and adverse effects of the medication you want to prescribe
• your discussion of informed consent, including documentation that the patient is aware that the medication is being prescribed off-label
• your clinical rationale for why the off-label medication is in the patient’s best interest.

Also, document the steps you take to monitor for adverse events and therapeutic effectiveness.⁴ Overall, the goal of documentation should be to support the adequate continuing care of our patients.

The FDA defines “off-label” prescribing as prescribing an FDA-approved medication for an unapproved use, such as for an unapproved clinical indication, for a higher-than-approved dose, or for a patient who is not part of the FDA-approved population (eg, children or geriatric patients).¹ Off-label prescribing is common in psychiatry; approximately 13% of psychiatry patients are prescribed off-label psychotropic medications.² The American Psychiatric Association strongly supports “the autonomous clinical decision-making authority of a physician” and “a physician’s lawful use of an FDA-approved drug product or medical device for an off-label indication when such use is based upon sound scientific evidence in conjunction with sound medical judgment.”³ Because many psychiatric diagnoses have no FDA-approved medications, off-label prescribing often may be a psychiatrist’s only pharmacologic option.

Unfortunately, off-label prescribing can increase a psychiatrist’s risk for liability when treatment falls short of patients’ expectations, or when patients allege that they were injured by the use of an off-label medication. Off-label prescribing does not automatically lead to losing a malpractice suit because the FDA states that physicians can prescribe approved medications for any scientifically supported use, including off-label.¹ Medical malpractice lawsuits alleging negligence in prescribing practices, such as off-label prescribing, typically include allegations against the psychiatrist for failure to⁴:

• adequately assess the patient
• consult the patient’s medical records
• obtain informed consent from the patient
• appropriately prescribe a medication for the clinical indication, dosage, patient’s age, etc.
• monitor for adverse effects and therapeutic effectiveness.

Steps to minimize your risk

When prescribing a medication off-label, the following approaches can help reduce your liability risk:

Conduct a comprehensive clinical assess­ment. This should include requesting and reviewing your patient’s medical records.

Explain your motivation. Explain to your patient how prescribing an off-label medication can directly benefit him/her. Make it clear that you are not conducting experimental research by prescribing off-label because some patients might perceive this as a covert form of research.

Know the medications you prescribe. Although this sounds obvious, psychiatrists should thoroughly understand how each medication they prescribe is likely to clinically affect their patient. This information is available from many sources, including the FDA’s medication information sheets and the manufacturer’s medication package inserts. If possible, make sure that your off-label prescribing is supported by reputable, peer-reviewed literature.

Obtain informed consent. Tell your patient that the medication you are recommending is being prescribed off-label. Discuss the medication’s risks, benefits, adverse effects, associated “black-box” warnings, off-label uses, and alternatives to the off-label medication.⁴ Allow time for the patient to ask questions about these treatments.

Continue to: Document all steps

Document all steps. There is an adage in medicine that “If it’s not written, it wasn’t done.” To help reduce your liability risk when prescribing off-label, be sure to document the following⁴:

• your clinical assessment
• information you gleaned from the patient’s medical records
• your review of information regarding both therapeutic and adverse effects of the medication you want to prescribe
• your discussion of informed consent, including documentation that the patient is aware that the medication is being prescribed off-label
• your clinical rationale for why the off-label medication is in the patient’s best interest.

Also, document the steps you take to monitor for adverse events and therapeutic effectiveness.⁴ Overall, the goal of documentation should be to support the adequate continuing care of our patients.

As a psychiatrist, you could be called to court to testify as a fact witness in a hearing or trial. Your role as a fact witness would differ from that of an expert witness in that you would likely testify about the information that you have gathered through direct observation of patients or others. Fact witnesses are generally not asked to give expert opinions regarding forensic issues, and treating psychiatrists should not do so about their patients. As a fact witness, depending on the form of litigation, you might be in one of the following 4 roles¹:

• Observer. As the term implies, you have observed an event. For example, you are asked to testify about a fight that you witnessed between another clinician’s patient and a nurse while you were making your rounds on an inpatient unit.
• Non-defendant treater. You are the treating psychiatrist for a patient who is involved in litigation to recover damages for injuries sustained from a third party. For example, you are asked to testify about your patient’s premorbid functioning before a claimed injury that spurred the lawsuit.
• Plaintiff. You are suing someone else and may be claiming your own damages. For example, in your attempt to claim damages as a plaintiff, you use your clinical knowledge to testify about your own mental health symptoms and the adverse impact these have had on you.
• Defendant treater. You are being sued by one of your patients. For example, a patient brings a malpractice case against you for allegations of not meeting the standard of care. You testify about your direct observations of the patient, the diagnoses you provided, and your rationale for the implemented treatment plan.

Preparing yourself as a fact witness

For many psychiatrists, testifying can be an intimidating process. Although there are similarities between testifying in a courtroom and giving a deposition, there are also significant differences. For guidelines on providing depositions, see Knoll and Resnick’s “Deposition dos and don’ts: How to answer 8 tricky questions” (Current Psychiatry. March 2008, p. 25-28,36,39-40).² Although not an exhaustive list, we offer the following practical tips for testifying as a fact witness.

Don’t panic. Although your first reaction may be to panic upon receiving a subpoena or court order, you should “keep your cool” and remember that the observations you made or treatment provided have already taken place.¹ Your role as a fact witness is to inform the judge and jury about what you saw and did.¹

Continue to: Refresh your memory and practice

Refresh your memory and practice. Gather all required information (eg, medical records, your notes, etc.) and review it before testifying. This will help you to recall the facts more accurately when you are asked a question. Consider practicing your testimony with the attorney who requested you to get feedback on how you present yourself.¹ However, do not try to memorize what you are going to say because this could make your testimony sound rehearsed and unconvincing.

Plan ahead, and have a pretrial conference. Because court proceedings are unpredictable, you should clear your schedule to allow enough time to appear in court. Before your court appearance, meet with the attorney who requested you to discuss any new facts or issues as well as learn what the attorney aims to accomplish with your testimony.¹

Speak clearly in your own words, and avoid jargon. Courtroom officials are unlikely to understand psychiatric jargon. Therefore, you should explain psychiatric terms in language that laypeople would comprehend. Because the court stenographer will require you to use actual words for the court transcripts, you should answer clearly and verbally or respond with a definitive “yes” or “no” (and not by nodding or shaking your head).

Testimony is also not a time for guessing. If you don’t know the answer, you should say “I don’t know.”

As a psychiatrist, you could be called to court to testify as a fact witness in a hearing or trial. Your role as a fact witness would differ from that of an expert witness in that you would likely testify about the information that you have gathered through direct observation of patients or others. Fact witnesses are generally not asked to give expert opinions regarding forensic issues, and treating psychiatrists should not do so about their patients. As a fact witness, depending on the form of litigation, you might be in one of the following 4 roles¹:

• Observer. As the term implies, you have observed an event. For example, you are asked to testify about a fight that you witnessed between another clinician’s patient and a nurse while you were making your rounds on an inpatient unit.
• Non-defendant treater. You are the treating psychiatrist for a patient who is involved in litigation to recover damages for injuries sustained from a third party. For example, you are asked to testify about your patient’s premorbid functioning before a claimed injury that spurred the lawsuit.
• Plaintiff. You are suing someone else and may be claiming your own damages. For example, in your attempt to claim damages as a plaintiff, you use your clinical knowledge to testify about your own mental health symptoms and the adverse impact these have had on you.
• Defendant treater. You are being sued by one of your patients. For example, a patient brings a malpractice case against you for allegations of not meeting the standard of care. You testify about your direct observations of the patient, the diagnoses you provided, and your rationale for the implemented treatment plan.

Preparing yourself as a fact witness

For many psychiatrists, testifying can be an intimidating process. Although there are similarities between testifying in a courtroom and giving a deposition, there are also significant differences. For guidelines on providing depositions, see Knoll and Resnick’s “Deposition dos and don’ts: How to answer 8 tricky questions” (Current Psychiatry. March 2008, p. 25-28,36,39-40).² Although not an exhaustive list, we offer the following practical tips for testifying as a fact witness.

Don’t panic. Although your first reaction may be to panic upon receiving a subpoena or court order, you should “keep your cool” and remember that the observations you made or treatment provided have already taken place.¹ Your role as a fact witness is to inform the judge and jury about what you saw and did.¹

Continue to: Refresh your memory and practice

Refresh your memory and practice. Gather all required information (eg, medical records, your notes, etc.) and review it before testifying. This will help you to recall the facts more accurately when you are asked a question. Consider practicing your testimony with the attorney who requested you to get feedback on how you present yourself.¹ However, do not try to memorize what you are going to say because this could make your testimony sound rehearsed and unconvincing.

Plan ahead, and have a pretrial conference. Because court proceedings are unpredictable, you should clear your schedule to allow enough time to appear in court. Before your court appearance, meet with the attorney who requested you to discuss any new facts or issues as well as learn what the attorney aims to accomplish with your testimony.¹

Speak clearly in your own words, and avoid jargon. Courtroom officials are unlikely to understand psychiatric jargon. Therefore, you should explain psychiatric terms in language that laypeople would comprehend. Because the court stenographer will require you to use actual words for the court transcripts, you should answer clearly and verbally or respond with a definitive “yes” or “no” (and not by nodding or shaking your head).

Testimony is also not a time for guessing. If you don’t know the answer, you should say “I don’t know.”

As a psychiatrist, you could be called to court to testify as a fact witness in a hearing or trial. Your role as a fact witness would differ from that of an expert witness in that you would likely testify about the information that you have gathered through direct observation of patients or others. Fact witnesses are generally not asked to give expert opinions regarding forensic issues, and treating psychiatrists should not do so about their patients. As a fact witness, depending on the form of litigation, you might be in one of the following 4 roles¹:

• Observer. As the term implies, you have observed an event. For example, you are asked to testify about a fight that you witnessed between another clinician’s patient and a nurse while you were making your rounds on an inpatient unit.
• Non-defendant treater. You are the treating psychiatrist for a patient who is involved in litigation to recover damages for injuries sustained from a third party. For example, you are asked to testify about your patient’s premorbid functioning before a claimed injury that spurred the lawsuit.
• Plaintiff. You are suing someone else and may be claiming your own damages. For example, in your attempt to claim damages as a plaintiff, you use your clinical knowledge to testify about your own mental health symptoms and the adverse impact these have had on you.
• Defendant treater. You are being sued by one of your patients. For example, a patient brings a malpractice case against you for allegations of not meeting the standard of care. You testify about your direct observations of the patient, the diagnoses you provided, and your rationale for the implemented treatment plan.

Preparing yourself as a fact witness

For many psychiatrists, testifying can be an intimidating process. Although there are similarities between testifying in a courtroom and giving a deposition, there are also significant differences. For guidelines on providing depositions, see Knoll and Resnick’s “Deposition dos and don’ts: How to answer 8 tricky questions” (Current Psychiatry. March 2008, p. 25-28,36,39-40).² Although not an exhaustive list, we offer the following practical tips for testifying as a fact witness.

Don’t panic. Although your first reaction may be to panic upon receiving a subpoena or court order, you should “keep your cool” and remember that the observations you made or treatment provided have already taken place.¹ Your role as a fact witness is to inform the judge and jury about what you saw and did.¹

Continue to: Refresh your memory and practice

Refresh your memory and practice. Gather all required information (eg, medical records, your notes, etc.) and review it before testifying. This will help you to recall the facts more accurately when you are asked a question. Consider practicing your testimony with the attorney who requested you to get feedback on how you present yourself.¹ However, do not try to memorize what you are going to say because this could make your testimony sound rehearsed and unconvincing.

Plan ahead, and have a pretrial conference. Because court proceedings are unpredictable, you should clear your schedule to allow enough time to appear in court. Before your court appearance, meet with the attorney who requested you to discuss any new facts or issues as well as learn what the attorney aims to accomplish with your testimony.¹

Speak clearly in your own words, and avoid jargon. Courtroom officials are unlikely to understand psychiatric jargon. Therefore, you should explain psychiatric terms in language that laypeople would comprehend. Because the court stenographer will require you to use actual words for the court transcripts, you should answer clearly and verbally or respond with a definitive “yes” or “no” (and not by nodding or shaking your head).

Testimony is also not a time for guessing. If you don’t know the answer, you should say “I don’t know.”

The ubiquitous use of e-mail has opened the proverbial “Pandora’s box” of access to psychiatrists. Our e-mail addresses are readily available online via search engines or on hospital Web sites. E-mail has become a convenient method of communicating with patients; however, it also has resulted in a proliferation of unsolicited e-mails sent to physicians from people they don’t know seeking professional advice.¹ If you publish medical literature or make media appearances, you may be contacted by such individuals requesting your expertise.

Unsolicited e-mails present psychiatrists with ethical and legal quandaries that force them to consider how they can balance the human reflex to offer assistance against the potential ramifications of replying. These conundrums include:

• whether the sender is an actual person, and whether he or she is asking for advice
• the risks of replying vs not replying
• the possibility that there is a plausible crisis or danger to the sender or others
• the potential for establishing a doctor–patient relationship by replying
• the legal liability that might be incurred by replying.²

Take preemptive measures

There is guidance on how to e-mail your patients and respond to solicited e-mails, but there is a dearth of literature on how to respond to unsolicited e-mails. Anecdotal reports and limited literature suggest several possible measures you could take for managing unsolicited e-mails:

• Establish a policy of never opening unsolicited e-mails
• Create a strict junk-mail filter to prevent unsolicited e-mails from being delivered to your inbox
• Set up an automatic reply stating that unwanted or unsolicited e-mails will not be read and/or that no reply will be provided
• Read unsolicited e-mails, but immediately delete them without replying
• Acknowledge the sender in a reply, but state that you are unable to assist and decline further contact
• Send a generic reply clarifying that you are unable to provide medical assistance, and encourage the sender to seek help locally.²

Despite the urge to help, consider the consequences

In addition to taking up valuable time, unsolicited e-mails create legal and ethical predicaments that could subject you to legal liability if you choose to reply. Even though your intentions may be altruistic and you want to be helpful, you may unknowingly create problems for yourself. Clinicians should carefully weigh the consequences of replying before clicking “send.”

The ubiquitous use of e-mail has opened the proverbial “Pandora’s box” of access to psychiatrists. Our e-mail addresses are readily available online via search engines or on hospital Web sites. E-mail has become a convenient method of communicating with patients; however, it also has resulted in a proliferation of unsolicited e-mails sent to physicians from people they don’t know seeking professional advice.¹ If you publish medical literature or make media appearances, you may be contacted by such individuals requesting your expertise.

Unsolicited e-mails present psychiatrists with ethical and legal quandaries that force them to consider how they can balance the human reflex to offer assistance against the potential ramifications of replying. These conundrums include:

• whether the sender is an actual person, and whether he or she is asking for advice
• the risks of replying vs not replying
• the possibility that there is a plausible crisis or danger to the sender or others
• the potential for establishing a doctor–patient relationship by replying
• the legal liability that might be incurred by replying.²

Take preemptive measures

There is guidance on how to e-mail your patients and respond to solicited e-mails, but there is a dearth of literature on how to respond to unsolicited e-mails. Anecdotal reports and limited literature suggest several possible measures you could take for managing unsolicited e-mails:

• Establish a policy of never opening unsolicited e-mails
• Create a strict junk-mail filter to prevent unsolicited e-mails from being delivered to your inbox
• Set up an automatic reply stating that unwanted or unsolicited e-mails will not be read and/or that no reply will be provided
• Read unsolicited e-mails, but immediately delete them without replying
• Acknowledge the sender in a reply, but state that you are unable to assist and decline further contact
• Send a generic reply clarifying that you are unable to provide medical assistance, and encourage the sender to seek help locally.²

Despite the urge to help, consider the consequences

In addition to taking up valuable time, unsolicited e-mails create legal and ethical predicaments that could subject you to legal liability if you choose to reply. Even though your intentions may be altruistic and you want to be helpful, you may unknowingly create problems for yourself. Clinicians should carefully weigh the consequences of replying before clicking “send.”

The ubiquitous use of e-mail has opened the proverbial “Pandora’s box” of access to psychiatrists. Our e-mail addresses are readily available online via search engines or on hospital Web sites. E-mail has become a convenient method of communicating with patients; however, it also has resulted in a proliferation of unsolicited e-mails sent to physicians from people they don’t know seeking professional advice.¹ If you publish medical literature or make media appearances, you may be contacted by such individuals requesting your expertise.

Unsolicited e-mails present psychiatrists with ethical and legal quandaries that force them to consider how they can balance the human reflex to offer assistance against the potential ramifications of replying. These conundrums include:

• whether the sender is an actual person, and whether he or she is asking for advice
• the risks of replying vs not replying
• the possibility that there is a plausible crisis or danger to the sender or others
• the potential for establishing a doctor–patient relationship by replying
• the legal liability that might be incurred by replying.²

Take preemptive measures

There is guidance on how to e-mail your patients and respond to solicited e-mails, but there is a dearth of literature on how to respond to unsolicited e-mails. Anecdotal reports and limited literature suggest several possible measures you could take for managing unsolicited e-mails:

• Establish a policy of never opening unsolicited e-mails
• Create a strict junk-mail filter to prevent unsolicited e-mails from being delivered to your inbox
• Set up an automatic reply stating that unwanted or unsolicited e-mails will not be read and/or that no reply will be provided
• Read unsolicited e-mails, but immediately delete them without replying
• Acknowledge the sender in a reply, but state that you are unable to assist and decline further contact
• Send a generic reply clarifying that you are unable to provide medical assistance, and encourage the sender to seek help locally.²

Despite the urge to help, consider the consequences

In addition to taking up valuable time, unsolicited e-mails create legal and ethical predicaments that could subject you to legal liability if you choose to reply. Even though your intentions may be altruistic and you want to be helpful, you may unknowingly create problems for yourself. Clinicians should carefully weigh the consequences of replying before clicking “send.”