Risk markers for suicidality in autistic adults

Seventy-two percent is the share of autistic adults in this study who scored above the psychiatric threshold for suicide risk. Hold that number for a moment. Now, compare it to the general population figure of thirty-three percent. The gap here is significant. Cassidy and colleagues set out to answer not just how large that gap is, but why it exists and whether the forces driving it are specific to autism in ways that standard suicide prevention has overlooked entirely. The problem, as Cassidy and colleagues explain, is that many known risk factors for suicide—such as depression, unemployment, and non-suicidal self-injury—are significantly more common in autistic adults than in the general population. You might assume that these shared markers explain the gap. However, that assumption is incomplete. After statistically controlling for factors like depression, anxiety, employment, and living arrangements, there were still independent predictors of suicidality that appeared unique to autism. The study aimed to identify those. What makes the research design noteworthy is how the survey was constructed. A steering group of eight autistic adults took part in six focus groups, shaping the questionnaire through three successive drafts to ensure it was comprehensive, relevant, and clear. This co-design is not just a procedural footnote; it allowed the team to incorporate measures that standard psychiatric surveys often overlook, including a unique measure of camouflaging. More on that shortly. The final sample comprised one hundred sixty-four autistic adults and one hundred sixty-nine general population adults, carefully matched on age and sex ratio. Suicidality was assessed using the Suicidal Behaviours Questionnaire-Revised, or SBQ-R, which is a four-item validated instrument evaluating lifetime suicidal behavior, ideation over the past year, threat of attempt, and likelihood of future behavior. The SBQ-R has two clinically recommended cut-offs: a score of seven for the general population and eight for psychiatric populations. Those thresholds correspond to the seventy-two percent and thirty-three percent figures. Autistic adults were more than five times as likely to exceed the psychiatric cut-off compared to general population adults—an odds ratio of five point zero four. Lifetime suicide attempts were reported by thirty-eight percent of autistic participants, compared to eight percent in the general population. These are substantial numbers that were consistent across various statistical tests. Non-suicidal self-injury—deliberate self-harm without suicidal intent—also showed a stark difference. Sixty-five percent of autistic adults reported lifetime non-suicidal self-injury, in contrast to just under thirty percent in the general population. Within the autistic group, it was more prevalent among females: seventy-four percent of autistic women reported it, compared to fifty-four percent of autistic men. Furthermore, in the regression models, lifetime non-suicidal self-injury predicted suicidality independently, accounting for an additional four percent of variance in SBQ-R scores after controlling for demographics, depression, and anxiety. While non-suicidal self-injury is a shared risk marker—existing in the general population as well—its prevalence in autism is roughly double, amplifying its impact. Now, here's where the findings become more specific. The team ran hierarchical regression models, a method that allows you to add predictors in blocks to assess how much additional explanatory power each contributes. After including all the shared risk factors, two variables unique to autism still had a significant effect. The first was camouflaging. In this study, camouflaging is defined as the practice of masking or suppressing autistic behaviors to fit into social situations—mimicking others, suppressing stimming, forcing eye contact, and scripting conversations. The measure captured how many social contexts a person camouflages in, how frequently they do it, and how much of their waking time it occupies, resulting in a maximum composite score of twenty. Nearly ninety percent of the autistic adults in the sample—both male and female—reported ever camouflaging. Females averaged a score of around fourteen point seven out of twenty, compared to twelve point nine for males. The total score for camouflaging predicted suicidality in the autistic group even after controlling for depression and anxiety, accounting for an additional three point five percent of variance. While that increase may seem small, it is statistically significant, suggesting that camouflaging contributes something beyond what depression alone captures. Think about what camouflaging involves. It requires sustained, effortful concealment of identity across most social interactions, often for much of the day and frequently for years before anyone even receives a diagnosis. The mean age at which autism spectrum condition diagnosis occurred in this sample was thirty-four, indicating decades of masking before individuals even understand what they are masking. The second autism-specific predictor was unmet support needs. The team operationalized this directly by asking participants how many areas of life they would ideally like support in—such as housing, employment, healthcare, mental health, finances, social activities, and others—and subtracted the number of areas where they actually received support. Autistic participants reported an average of around three point one to three point four unmet needs, while the general population reported roughly one point six. This mismatch itself predicted suicidality independently, contributing an additional three point one percent of variance in SBQ-R scores. This is a structural finding. It suggests that autistic adults are not just feeling bad; the systems around them are failing to meet the needs they have clearly articulated. One additional finding merits attention because it expands the risk landscape beyond those with a formal diagnosis. Self-reported autistic traits in the general population also correlated significantly with suicidality, showing a correlation of zero point thirty-three. In the combined regression models, an autism spectrum condition diagnosis explained an additional four point five percent of variance in SBQ-R scores beyond demographics and mental health diagnoses. Risk scales with trait load, meaning it is not solely tied to having a clinical label. Experiences related to autism appear to pose an inherent risk, regardless of whether a person has ever been assessed. The practical implications of these findings are direct. Standard suicide risk screening tools were constructed based on data from the general population. They are well-designed to flag depression and prior attempts, and these factors are important in autism as well. However, they do not inquire about camouflaging or measure unmet support needs. Cassidy's study reveals that these gaps are not minor; they represent where a significant portion of autism-specific risk resides. If clinicians rely solely on standard screening when evaluating an autistic adult, they may end up systematically underestimating what truly drives that individual's risk. Cassidy and colleagues assert that new, tailored suicide prevention strategies for autistic individuals are essential. The study's methodological strengths make this argument difficult to dismiss. The survey was co-designed with autistic adults, utilized validated instruments with reported internal consistencies, included a matched comparison group, and employed bootstrapping—a resampling technique that produces more reliable estimates when data distributions are non-normal, which is often the case with suicidality data. The authors are also candid about a limitation: the additional variance explained by camouflaging and unmet support needs was small, the design was cross-sectional, and the sample had unusually high rates of lifetime depression—eighty percent in the autistic group. While the findings identify risk markers, they do not yet tell us if interventions that reduce camouflaging or meet support needs will actually lower suicidality. That work still lies ahead. However, this research prompts a reframe. The conventional narrative surrounding autism and suicide typically focuses on comorbidity: autism leads to depression, which in turn leads to suicidal risk. While that pathway is valid, Cassidy and colleagues present a second narrative that runs parallel to it—one in which the constant effort to conceal one's identity and the persistent failure of services to provide for needed support are factors contributing to suicidal risk, partly independent of diagnosed mental illness. Camouflaging is not merely incidental suffering; it is reflected in the regression. It has coefficients and predicts outcomes. Rather than concluding the discussion, this study opens a specific and actionable avenue: prevention strategies that lessen the conditions requiring camouflage and urgently identify and address the unmet needs that autistic individuals have already articulated. Those at highest risk are telling us what they require. The data validate this. The current task is to create systems that genuinely respond. This lecture was created by ennepō. Go to https://ennepo.ai to Discover, Create and Follow the latest research in your field. Read when you can. Listen when you want to.