Promise for Remote Assessment of Self-Injury: Reliability and Validity of the SITBI-R in both Interview and Online Format
Interview and Online assessment of self-injurious thoughts and behaviors show equal validity and reliability. This is the bottom line of a recently published article in Psychological Assessment. Below is a summary of the research and findings as well as a translation of this research into practice.
Featured Article | Psychological Assessment | 2020, Vol. 32, No.7, 677-689
Katheryn, R. Fox, University of Denver
Julia A. Harris, University of Utah
Shirley B. Wang, Harvard University
Alexander J. Millner, Harvard University
Charlene A. Deming, Durham Veterans Affairs Health Care System
Matthew K. Nock, Harvard University
The Self-Injurious Thoughts and Behaviors Interview (SITBI) is a widely used measure of the presence, frequency, and characteristics of suicide and self-harming thoughts and behaviors. In response to advances in the conceptualization of these outcomes, and the potential for online data collection, we created a revised version of the SITBI (SITBI R) and tested its psychometric properties via in-person interview and online self-report formats. Across two studies, the SITBI-R demonstrated strong psychometric properties for both assessment formats. In Study 1, outcomes measured via the SITBI-R showed convergent validity with those assessed with the Columbia Suicide Severity Rating Scale, another interview assessing suicidal thoughts and behaviors. The SITBI-R also showed strong alternate-forms reliability across nearly all outcomes assessed via both assessment formats. In Study 2, the SITBI-R showed strong test–retest reliability via the online assessment format. Across both studies, reliability was strongest for more recent outcomes (e.g., past year vs. lifetime) and for more commonly assessed outcomes of suicidal thoughts, plans, and attempts than for other, less commonly assessed behaviors (e.g., suicide gestures, interrupted suicide attempts, and aborted suicide attempts). The results of these two studies suggest that the SITBI-R provides reliable and valid measurement of key self-injurious outcomes both in person and online.
assessment, self-injury, suicide, suicide attempts, nonsuicidal self-injury
Summary of the Research
“Despite decades of research on SITBs [self-injurious thoughts and behaviors], we have a limited understanding of why people engage in and how to best prevent these behaviors. Accurate and thorough measurement of SITBs is critical to increase our understanding of these outcomes. Working toward this goal, Nock, Holmberg, Photos, and Michel (2007) created the Self-Injurious Thoughts and Behaviors Interview (SITBI), a comprehensive and semistructured interview that provides standardized measurement of the presence of several different SITBs and characteristics of these behaviors (e.g., ages of onset, urge/intensity of thoughts, severity of behaviors).”p.677
“the original SITBI and research on this measure have several areas for improvement. First, the original SITBI assesses only a limited range of SITBs. Specifically, it assesses the presence and characteristics of five forms of SITBs: suicidal ideation, suicide plans, suicide gestures, suicide attempts, and nonsuicidal self-injury (NSSI). Other related behaviors not included in the original include aborted suicide attempts (i.e., getting close to killing oneself and at the last minute deciding not to), interrupted suicide attempts (i.e., getting close to killing oneself and at the last minute someone or something else intervening to prevent it from occurring), and dangerous behaviors enacted with ambivalent or no suicidal intent. Second, recent research suggests that some of the original items in the SITBI, in particular single items assessing suicide ideation, plan, and attempt, may lead to unreliable, or inconsistent, responses across people despite similar experiences. Third, data on the reliability and validity of the original SITBI are based on interviewer-assessed SITBs. The reliability and validity of an online format for the SITBI remains unknown. Similarly, data are needed on whether people prefer to answer questions about their SITB histories to someone face to face, or anonymously online. Fourth, as assumptions of suicidal intent as binary may be insufficient and inaccurate in some instances, assessing suicidal intentions outside the context of specific behaviors (e.g., suicide attempts) is necessary to provide a more comprehensive understanding of the full spectrum of SITBIs.”p.678
“Accordingly, the current series of studies aimed to develop and evaluate a revised version of the SITBI (SITBI-R) to addresses each of these limitations. Toward this goal, we first created the SITBI-R, including updated language and additional forms of SITBs. Second, we tested the validity of the SITBI-R compared to another widely used interview. Third, we adapted the SITBI-R to an online self-report format to assess reliability indices of this online format both across time and in comparison to the original interview format. Fourth, we assessed respondents’ preferences regarding self-report versus person-to-person interview. We provide further description of each aim below. These tests were conducted across two unique studies.”p.678
“ICC, kappa, and Spearman rho indices were primarily strong for outcomes assessed with the SITBI-R. This was true of both the in-person interview and the online, self-report versions of the SITBI-R. Furthermore, validity indices were good to excellent when comparing responses to the STIBI-R and C-SSRS, suggesting strong convergent validity.”p.685
“…interrater reliability of SITB engagement assessed within the SITBI-R was nearly perfect, and test–retest reliability for ever engaging in most SITBs was also quite strong. This is consistent with test–retest reliability reported for the original SITBI. Results suggest that the SITBI-R maintains strong agreement across raters…convergent validity between outcomes assessed via the SITBI-R and C-SSRS were also strong. This suggests that the SITBI-R provides valid assessment of SITB engagement.”p.686
“However, a few inconsistencies are worth noting. The SITBI-R and C-SSRS assess suicide planning in different ways. Whereas the SITBI-R defines a suicide plan as thinking of a suicide method, place, or time, and contains items to assess each of these components, the C-SSRS assesses suicide planning via items assessing thinking of a “specific” suicide method or via an item assessing whether participants have been thinking about how they might kill themselves. These discrepancies parallel the lack of clinical and research consensus regarding what constitutes a suicide plan”p.686
Translating Research into Practice
“…although assessment of suicide attempts showed strong reliability across the C-SSRS and SITBI-R, this should be interpreted with caution. In particular, across both assessments, the same suicide attempt definition was used: one requiring that the self-harming behavior was enacted with at least some, nonzero self-reported intent to die. However, the definition typically used in the C-SSRS is broader; in particular, self-harming behaviors are considered suicidal when a participant states suicidal intent or when suicidal intent can be inferred (i.e., from engaging in a highly lethal act even when suicidal intent is denied, from engaging in a behavior that the participant reports could have been lethal”p.686
“…results provide at least three lines of evidence that online, self-report versions of the SITBI-R can be administered reliably, at least for people willing and able to understand and answer questions about their SITB histories. In particular, Study 1 demonstrated strong equivalence across in-lab and online administrations of the SITBI-R…. results of Study 1 indicate that people respond similarly to an interviewer and to an online self-report questionnaire when answering questions about their SITB histories.”p.686
Other Interesting Tidbits for Researchers and Clinicians
“lower reliability indices may relate to construct definitions and/or validity. Regarding interrupted suicide attempts, participants may have experienced difficulty disentangling aborted from interrupted suicide attempts. For example, during in-person interviews, participants often tried to endorse a history of interrupted suicide attempt when an outside event (e.g., an unrelated phone call, text) changed their mindset and they decided after that event not to attempt suicide. However, within our coding schemes, such instances are coded as “aborted suicide attempts” unless someone or something specifically intervened on the suicide attempt (e.g., walked into the room as someone was attempting suicide, physically restrained the individual at risk, took away the method the individual was using, etc.). Given potential difficulty disentangling aborted from interrupted suicide attempts in online and in-person assessments, we recommend that researchers consider additional language to probe these distinctions. Additional research may also be useful to determine if such distinctions are meaningfully related to future SITB risk.”p.686
Join the Discussion
As always, please join the discussion below if you have thoughts or comments to add!
Looking for training? Here are a few suggestions:
Authored by Leila N. Wallach
Leila N. Wallach is a clinical psychology doctoral student at Palo Alto University. Her research and clinical interests focus on alternatives to incarceration, culture and trauma-informed care, policy in the juvenile justice system, and risk assessment for community offender management.