by Michael D. Anestis, M.S.
The other day, a reader named Ernest Miller asked an interesting question in the comment section of our article on a suicide prevention effort in China. The basic point of his question was whether any research had been done that differentiated those who attempt suicide in public versus those who do so privately and, if so, whether data exist on whether these groups differ from one another on variables such as pain tolerance. Both Joye and I were fascinated by the question and neither of us know of studies looking specifically at this idea. We can certainly make guesses, but as clinical scientists who value empirical data over intuition, we would rather examine what has already been discovered before sharing our own thoughts on the matter. Over the past few days, I have unfortunately not been in my office at the same time as colleagues who might have some thoughts on this, but I will continue to seek out their insights. In the meantime, I have done my own research into the matter and, today, I came across a relevant article recently published in Personality and Individual Differences by Catherine Glenn and David Klonsky of the University of British Columbia.
The article, while relevant, does not answer the exact same question. Instead, Glenn and Klonsky (2009) were interested in examining the differences between those who engage in non-suicidal self-injury (NSSI) exclusively in private and those who occasionally or regularly engage in NSSI in the presence of others. Specifically, the authors were curious about whether the two groups might differ in terms of suicide risk. As you might recall from our discussion of Joiner's interpersonal-psychological theory of suicidal behavior, research has indicated that, in order for an individual to make a serious suicide attempt, they must not only desire suicide, but develop the acquired capability to do so. This capability, which is marked by a diminished fear of death and a heightened tolerance to physiological pain, can be developed through exposure to others' pain, but is most directly impacted by self-inflicted injury. Although NSSI itself does not involve suicidal intent and is a distinct phenomenon, research has demonstrated that individuals who self-injure regularly are less likely to experience pain during self-injury than are individuals with less experience with the behavior (Nock & Prinstein, 2005), a clear indicator that this particular behavior can be a path to the acquired capability for suicide.
Along these lines and perhaps not surprisingly, NSSI has been demonstrated to be a strong predictor of suicidal behavior (Nock, Joiner, Gordon, Lloyd-Richardson, & Prinstein, 2006). In the study we are looking at today, Glenn and Klonsky (2009) wanted to clarify this relationship by examining whether the context within which NSSI occurs can help predict who is at the greatest risk of eventually attempting and potentially dying by suicide.
NSSI - defined as the intentional infliction of harm upon one's own body without suicidal intent - is generally a private affair, performed in secrecy and concealed from others. That being said, as our articles on the functional model of NSSI have demonstrated, this is not always the case. Similarly, although the most commonly reported reason for engaging in NSSI is to reduce one's own negative emotions (Nock & Prinstein, 2004), there are some who engage in NSSI for interpersonal reasons as well.
Importantly, the reason an individual engages in NSSI has been linked to the risk for suicide. Specifically, individuals who self-injure to reduce negative emotions have been found to be more likely than individuals who self-injure for other reasons to have made a recent suicide attempt (Nock & Prinstein, 2005). In an "in press" article (e.g., already accepted for publication but not yet released in the journal), Klonsky and Glenn reported that individuals who engage in NSSI in order to reduce negative emotions are most likely to self-injure while alone. As such, the authors have established a link between a particularly harmful motive for self-injury and a particular context. In this study, they sought to follow-up on that finding and determine the degree to which the context itself contributed to suicide risk. In other words, is a pattern of self-injuring exclusively while alone in and of itself a risk factor for suicidal behavior?
In a sample of 801 college students in an introductory psychology course, the authors reported that 26% indicated they had a prior history of NSSI. The resulting sample of individuals with a history of NSSI consisted of 205 students (57% female, average age = 18.5). Each participant was given self-report questionnaires that assessed the frequency and severity of many different forms of NSSI (e.g., cutting, burning), the functions of NSSI (e.g., to reduce negative emotions), history of suicidal thoughts and behaviors, depression symptoms, anxiety symptoms, and borderline personality disorder (BPD) symptoms. Individuals without a history of self-injury were given all of the same questionnaires except those assessing form and function of NSSI. As such, comparisons could be made between groups on a number of variables.
The most commonly reported forms of NSSI were banging/hitting self (61%), pulling hair (47%), pinching (42%), and cutting (40%). 82% of the respondents reported using multiple methods of self-injury and 62% of the sample had self-injured within the past year. The mean age of onset for self-injury in this sample was 13. Individuals with a history of self-injury reported higher mean levels of depression, anxiety, BPD, and suicidality relative to individuals with no history of self-injury. Although the study was looking at college students, who typically do not represent the most severe psychopathology, there was a significant amount and variety of NSSI represented in this sample.
The central aim of the study was not simply to look at how often individuals self-injured. Instead, the authors were interested in determining whether individuals who self-injure exclusively while alone differed from those who self-injure in the presence of others both in their motive for self-injury and in their risk for suicide. 23 individuals reported "never" engaging in NSSI while alone and 67 reported engaging in NSSI while alone only "sometimes." These two groups, representing a total of 90 individuals, were collapsed into a single group. 108 individuals reported that they are "always" alone when they self-injure.
Looking at their data, the authors found that individuals who self-injure while alone have significantly higher levels of suicidality, even when controlling for the effects of biological sex, than did individuals who never or only sometimes self-injure while alone. Additionally, individuals who self-injure exclusively while alone were most likely to report engaging in NSSI in an effort to reduce negative emotions. As a follow-up, the authors ran a mediational analysis in order to see if the motive for self-injury explained the relationship between self-injuring while alone and suicidality. As it turns out, it did not. Although the relationship between context of self-injury and suicidality weakened when controlling for motives, it remained significant, meaning that self-injuring exclusively while alone is related to suicide risk for more reasons that the fact that doing so likely reflects a tendency to self-injure in an effort to regulate emotions.
So, having established that the motive for self-injury did not fully explain the link between self-injuring alone and suicide risk, the authors decided to examine whether other important variables might serve as mediators. Specifically, they considered depression symptoms, BPD symptoms, frequency of cutting behavior, and frequency of carving behavior. When these variables were entered into the equation along with context of self-injury, the relationship between self-injuring exclusively while alone and suicidality again remained significant. In other words, individuals who self-injure only while they are alone do not appear to be at a greater risk for suicide simply because of their reason for self-injury, their level of depression or BPD, or the frequency with which they engage in cutting or carving of their own skin. Those variables made significant contributions to suicide risk, but self-injuring while alone significantly predicted suicidality above and beyond those effects.
Looking over all of these findings, Glenn and Klonsky's data indicate that individuals who self-injure alone are at a greater risk for suicide than are individuals who self-injure in the presence of others and that those who self-injure alone tend to do so in an effort to reduce negative emotions. Given that none of the variables they examined mediated the relationship between self-injuring alone and suicidality, the authors were left to explain why self-injuring alone might, in and of itself, reflect a risk for suicide. They provided several hypotheses on this front. The one that, to me, seemed the most compelling was the idea that those who self-injure might either lack social support or at least believe that others could not relate to their experiences. This idea is highly consistent with Joiner's (2005) theory and the idea that an individual must not only be capable of lethal self-harm, but must desire suicide in order to engage in a serious attempt. Prior research (e.g, Van Orden et al., 2008) has demonstrated that a thwarted sense of belongingness is a strong contributing factor to suicidal desire. In this sense, the mere presence of other people during an NSSI episode is not necessarily the important factor. Simply placing another person in the room while somebody who would otherwise only self-injure in private is not likely to directly impact suicide risk. Instead, it is the belief that others can relate to their experience that might make individuals who self-injure in the presence of others at lower risk.
Importantly, other interpretations could prove true as well. In fact, it might simply be that those who self-injure in the presence of others do so in an effort to fit in with peers and that their behavior represents a less clinically severe stage in development rather than the presence of severe psychopathology. Regardless, further data are needed in order to reach stronger conclusions and, ideally, future research will involve more complex methods for data collection rather than relying on self-report and a single time point. In the meantime, Glenn and Klonsky (2009) provided an interesting view of the importance of context in self-injurious behaviors.
If you would like to learn more about NSSI or suicide, we recommend the following resources, all of which are available through our online store. Additionally, if you are experiencing thoughts of suicide, please dial 1-800-273-TALK for quality care that is anonymous and available 24 hours, seven days per week.
- Why People Die by Suicide by Thomas Joiner
- Treating Suicidal Behavior: An Effective, Time-Limited Approach by David Rudd, Thomas Joiner, and Hasan Rajab
- The Interpersonal Theory of Suicide: Guidance for Working With Suicidal Clients by Thomas Joiner, Kim Van Orden, Tracy Witte, and David Rudd