University students who deliberately injure themselves without intending to die rarely reach out for help on their own, and a new study suggests that the strongest push toward getting support may come not from the pattern of their psychological symptoms but from other people encouraging them to seek it. In a cross-sectional survey of 1,476 Swedish university students, researchers identified four distinct profiles of borderline personality symptoms and then asked whether help-seeking for non-suicidal self-injury differed across those profiles. It largely did not. What did differ was how often students were encouraged by others to seek help, and that encouragement emerged as the most consistent correlate of actually seeking it.
The study, published in the journal Borderline Personality Disorder and Emotion Dysregulation, was led by Rosie James and colleagues including Sofie Westling and Magnus Nilsson. All 32,001 enrolled students at a Swedish university were invited by email to complete a comprehensive survey on self-harming behaviors and attitudes. Responses were recorded from 1,500 individuals, and 24 were removed after a predefined attention check designed to verify that participants were reading the questions, leaving an analytic sample of 1,476 students with a mean age of 26.27 years, of whom 68.5 percent were female. Participation was voluntary and anonymous, no compensation was provided, and the study was approved by the Swedish Ethical Review Authority.
Non-suicidal self-injury, often abbreviated NSSI, is defined as the deliberate harm of one’s body tissue, through acts such as cutting, burning, or hitting, without the intention to die. The researchers note that roughly one in eight young adults worldwide engage in such behavior, yet rates of help-seeking among those who self-injure remain low. That gap matters clinically because NSSI is associated with anxiety and depression, with an increased risk of suicidal thoughts and behaviors, and with borderline personality disorder. Self-injurious behavior is encompassed within one of the nine diagnostic criteria for borderline personality disorder, is common among people with subthreshold borderline features, and may emerge early in trajectories associated with later borderline pathology. Help-seeking for NSSI may therefore offer an opportunity to recognize clinically relevant borderline symptoms and support needs that would otherwise go unnoticed.
Because borderline pathology is heterogeneous, the researchers used a person-centered approach rather than treating symptoms as a single dimension. They applied latent class analysis to nine indicators from the McLean Screening Instrument for Borderline Personality Disorder, combining two items that represent the paranoia and dissociation criterion. Competing solutions with three to six classes were compared using standard information criteria, likelihood ratio tests, and substantive interpretability, and a four-class solution showed the best balance of statistical fit and interpretability, with adequate classification quality reflected in an entropy value of 0.665.
The four profiles differed clearly in their symptom patterns. The Full Spectrum profile, comprising 234 students or 15.8 percent of the sample, showed high endorsement of most borderline symptoms. The Internalizing profile, with 426 students or 28.9 percent, showed high levels of emptiness, identity disturbance, and paranoia or dissociation. The Externalizing profile, with 263 students or 17.8 percent, showed elevated endorsement of mood instability and anger. The Asymptomatic profile, the largest group at 553 students or 37.5 percent, showed low endorsement of all symptoms. The researchers emphasize that these classes represent patterns of symptom endorsement within a community student sample and do not correspond to clinical diagnoses or diagnostic subtypes. Validation analyses confirmed meaningful differences: all clinical validators, including depression, anxiety, identity pathology, and emotion dysregulation, differed significantly across profiles, with the Full Spectrum profile showing the highest levels and the Asymptomatic profile the lowest.
The central help-seeking analyses focused on the 581 participants who reported lifetime non-suicidal self-injury. Help-seeking was measured by asking whether participants had ever sought help for their self-injury and whether they had ever been encouraged to do so, with responses recoded into binary variables. Attitudes were assessed with the Mental Help-Seeking Attitudes Scale. Using three-step auxiliary methods that correct for classification error, the researchers found that the omnibus test of profile differences in help-seeking behavior was not statistically significant. An exploratory pairwise contrast suggested higher help-seeking in the Full Spectrum compared with the Asymptomatic profile, but the overall test did not establish a reliable difference. In contrast, encouragement to seek help did differ significantly across profiles, with students in the Full Spectrum profile more often encouraged than those in the Internalizing and Asymptomatic profiles. Help-seeking attitudes also differed, with the Full Spectrum profile reporting the least positive attitudes, significantly lower than the Externalizing and Asymptomatic profiles.
Encouragement also stood out as a predictor within profiles. In one-step mixture logistic regressions predicting whether students had ever sought help, adjusted for age, sex, NSSI repetition and recency, encouragement, and attitudes, encouragement was associated with higher odds of help-seeking in all four profiles, reaching statistical reliability in the Full Spectrum profile, where the odds ratio was 17.87, and in the Asymptomatic profile, where it was 3.31. Estimates in the remaining profiles were positive but imprecise. Repetitive self-injury occurring more than six months ago predicted help-seeking in the Internalizing and Asymptomatic profiles, while recent repetition, attitudes, age, and sex showed no reliable associations in any profile. A joint Wald test of all class-specific thresholds and slopes was non-significant, indicating no evidence that the predictors of help-seeking differed across profiles.
The pathway into care also appeared broadly similar once help-seeking had begun. Of the 581 students with lifetime self-injury, 304, or 52.3 percent, reported any help-seeking, and help-source data were available for 297 of them. Students could indicate sources including primary healthcare, psychiatry, student health, private healthcare, family and friends, internet or social media, community services, and a helpline, which the researchers collapsed into formal and informal categories. Help-source categories did not differ significantly by profile. The most common response was the use of both formal and informal support rather than a clear preference for one source, which the authors note contrasts with earlier findings suggesting a preference for informal sources and instead indicates that help-seeking for self-injury commonly involves engagement with formal care, often alongside informal support.
The authors interpret these results through two competing accounts that they set out to test: a profile-specific account, in which the configuration of symptoms shapes help-seeking, and a profile-general account, in which help-seeking is driven by processes operating across profiles, such as encouragement from others. The findings favor the second account. One plausible mechanism involves the visibility of distress. Presentations marked by outwardly expressed dysregulation, such as interpersonal conflict and impulsive actions, may make distress more observable to others, whereas internalizing presentations may involve greater concealment, leaving fewer opportunities to receive encouragement. Notably, the Internalizing profile reported rates of repetitive self-injury comparable to the Full Spectrum profile, so the lower rate of encouragement in that group is unlikely to be explained by self-harm frequency alone. The authors suggest this may represent one route to delayed access to support for students whose distress is less visible.
The study has limitations that shape how its findings should be read. The response rate was low and participation was voluntary, so the sample may be affected by selection bias of unknown direction and should not be considered representative of a defined at-risk population; prevalence estimates of the profiles should be interpreted with caution. Information on alcohol and substance use was not collected, limiting the characterization of externalizing features. The repetition indicator was a frequency-based severity proxy rather than a standardized measure, and the study did not assess self-injury methods, medical severity, age at onset, or duration. Because the design was cross-sectional, the researchers cannot establish the temporal ordering between encouragement and help-seeking, cannot determine whether previous help or treatment contributed to lower current symptom levels, and did not assess the source, form, intensity, or content of encouragement. Even so, the authors conclude that interpersonal processes are particularly relevant to the initiation of help-seeking for self-injury, and that improving recognition of self-injury-related distress and actively encouraging help-seeking may create opportunities for timely identification of support needs, especially for individuals with more internalizing presentations whose distress could otherwise go unnoticed.
Subject of Research: Help-seeking for non-suicidal self-injury across borderline symptom profiles in university students
Article Title: Help-seeking patterns for non-suicidal self-injury across borderline symptom profiles in a large sample of university students
Article References: James, R., Overgaard, R., Westling, S., & Nilsson, M. (2026). Help-seeking patterns for non-suicidal self-injury across borderline symptom profiles in a large sample of university students. Borderline Personality Disorder and Emotion Dysregulation, 13(1), Article 30. https://doi.org/10.1186/s40479-026-00375-x
Image Credits: AI Generated
DOI: 10.1186/s40479-026-00375-x
Keywords: non-suicidal self-injury, borderline personality disorder, help-seeking, latent class analysis, university students, encouragement, internalizing symptoms, externalizing symptoms, cross-sectional study, mental health attitudes, Swedish students, patterns
News Source: Glenn Wilkins. (October 28, 2026). Research report: Help-seeking patterns for non-suicidal self-injury across borderline symptom profiles in a large sample of university students. Scienmag.
