📄 medicine

Parent–adolescent agreement in suicide-related symptom reporting among psychiatric inpatients with depressive disorders: a multi-informant study

This study of psychiatric inpatients with depressive disorders reveals that while parent and adolescent reports of suicide-related symptoms show modest agreement and both contribute to clinical risk assessment, they are not interchangeable, suggesting that discrepancies between informants warrant clinical clarification rather than being treated as equivalent indicators of future risk.

Original authors: Kexin Zhou¹, Jinquan Zhang¹, Qiyuan Cau¹, Jiaqi Xu¹, Wenjing Li¹, Yan Zhang¹, Xuehua Huang¹, Lijun Jiang

Published 2026-07-31

📖 6 min read🧠 Deep dive

CC BY 4.0

Original authors: Kexin Zhou¹, Jinquan Zhang¹, Qiyuan Cau¹, Jiaqi Xu¹, Wenjing Li¹, Yan Zhang¹, Xuehua Huang¹, Lijun Jiang

Original paper licensed under CC BY 4.0 (https://creativecommons.org/licenses/by/4.0/). ✨ This is an AI-generated explanation of the paper below. It is not written or endorsed by the authors. For technical accuracy, refer to the original paper. Read full disclaimer

The Two-Headed Detective: Why One Story Isn't Enough

Imagine you are trying to solve a mystery about how a person is feeling inside their head. In the world of child and adolescent psychiatry, scientists have long known that you can't just ask the person one question and expect the whole truth. It's a bit like trying to understand a movie by only watching the audio track; you miss the visual cues, the body language, and the context. This is where "multi-informant assessment" comes in. It's the idea that to get the full picture of a young person's mental health, you need to listen to at least two different voices: the teenager themselves and the person who cares for them, usually a parent or guardian.

Think of a teenager's mind as a house with many rooms. The teenager holds the keys to the private rooms where their deepest thoughts, fears, and secret plans live. They know exactly what is happening in the dark corners. The parent, however, is like a security guard walking the halls outside. They can't see inside the locked rooms, but they can notice if the lights are flickering, if someone is pacing nervously, or if the front door is being barricaded. Both perspectives are real, but they are different. The big question for doctors and researchers is: Do these two stories match up? If the teenager says, "I'm fine," but the parent says, "They seem really sad," who should we believe? And more importantly, can we swap one story for the other, or do we need both to keep everyone safe?

The Study: When the Teen and Parent Tell Different Stories

This paper dives right into that messy, complicated middle ground. The researchers went to a hospital in China where teenagers aged 11 to 18 were staying because they were struggling with depression. They asked a simple but tricky question: When these families are in the hospital, do the parents and the teenagers agree on whether the teen is thinking about suicide or has tried to hurt themselves?

To find out, the team acted like detectives gathering clues. They had 229 pairs of parents and teens fill out separate questionnaires right when they arrived. They also looked at notes written by nurses and doctors who were rating the risk of suicide at that exact same moment. The goal wasn't to predict the future, but to see how well the two stories lined up right now.

The Big Reveal: They Don't Match Perfectly
The results were a bit like two people describing the same weather report but using different words. The agreement between parents and teens was "modest." In the language of statistics, the numbers were around 0.32. To put that in plain English: if you asked a hundred pairs, they would only agree on the answer about one-third of the time. They weren't totally random, but they certainly weren't saying the same thing.

The study found that you absolutely cannot swap one report for the other. If a parent says "no" to suicide thoughts, it doesn't mean the teen is safe. In fact, in about 21% of the cases where they could be sure of the answer, the teen said "yes" (they were thinking about it) while the parent said "no" (they had no idea). This is a huge clue that parents often miss what's happening inside their child's head. Conversely, in about 11% of cases, the parent saw signs of trouble that the teen didn't report.

The "Unknown" Factor
One of the clever parts of this study was how they handled the "I don't know" answers. Sometimes, a parent would check a box that said "unknown" because they genuinely didn't know if their child had ever tried to hurt themselves. The researchers were careful not to count these "unknowns" as "yes." If they had, it would have looked like parents knew more than they actually did. When they looked only at the clear "yes" and "no" answers, they found that the most common pattern was when both the teen and the parent agreed that there were suicide-related thoughts or attempts (52.7% of the time). But the next most common group was the one where only the teen spoke up (21.0%).

What the Numbers Mean for Safety
The study also checked how these different stories matched up with the doctors' and nurses' ratings of risk. Here is what they found:

When the teen said "yes" to having thoughts, the nurses' risk scores went up significantly.
When the parent said "yes," the risk scores also went up, but not quite as high as when the teen said it.
When both said "yes," the risk scores were the highest of all.

This suggests that both voices matter. The teen's voice is the loudest regarding their own internal thoughts, but the parent's voice adds important extra information that the teen might be hiding or the parent might have noticed in behavior.

What This Study Does NOT Say
It is very important to understand what this paper doesn't claim. The researchers were very clear: they did not prove that a parent's report can predict if a teen will try to hurt themselves in the future. They also didn't prove that a teen's report is a crystal ball for the future. They only looked at what was happening in the hospital at that moment. The study shows that the two reports overlap but are not interchangeable. It's like having two different maps of the same city; they show the same main streets, but one might show a hidden alley the other misses.

The Takeaway
So, what's the lesson for parents, doctors, and anyone caring for a struggling teen? The study suggests that disagreement is not a mistake to be ignored; it's a signal. If a teen says they are fine but a parent is worried, or vice versa, that gap is a red flag that needs investigation. You can't just pick the "winner" and ignore the other. The best approach is to listen to both, realize they are telling different parts of the same story, and use that disagreement to ask better, more careful questions. The data confirms that while parents and teens often see eye-to-eye, they are not seeing the exact same thing, and that difference is exactly where the most important clues for safety might be hiding.

Technical Summary: Parent–Adolescent Agreement in Suicide-Related Symptom Reporting

Problem Statement
Multi-informant assessment is a cornerstone of child and adolescent psychiatry, yet the interchangeability of adolescent and caregiver reports regarding suicide-related symptoms remains uncertain. In inpatient settings, where acute risk assessment is critical, discrepancies between informants are common but often lack clear clinical interpretation. Key measurement challenges include the conflation of distinct constructs (e.g., recent suicidal ideation vs. lifetime suicide attempts) within combined domains, the handling of caregiver "unknown" responses which may distort prevalence estimates, and the difficulty in distinguishing between measurement noise and clinically meaningful information. Furthermore, while clinical risk instruments (such as the NGASR and C-SSRS) are often used as anchors, associations between informant reports and these same-window clinical scores may reflect shared content rather than independent predictive validity.

Methodology
This cross-sectional observational study utilized routinely collected baseline data from 229 adolescents (aged 11–18 years) hospitalized with depressive disorders at the West China Hospital of Sichuan University (May 2025–February 2026).

Data Collection: Adolescents and caregivers completed separate electronic questionnaires based on DSM-5 Level 1 cross-cutting symptom items. Clinical anchors included the Nurses' Global Assessment of Suicide Risk (NGASR) and the Columbia-Suicide Severity Rating Scale (C-SSRS), completed within the same admission window.
Analytic Cohort: The primary cohort consisted of 229 pairs with complete baseline data. A secondary classifiable cohort of 205 participants was derived by excluding cases where caregiver "unknown" responses precluded confirmed yes/no classification for suicide items.
Statistical Analysis: Agreement was quantified using Spearman correlation, absolute-agreement intraclass correlation coefficients (ICC), and quadratic weighted kappa. Item-level analyses calculated Cohen's kappa, exact agreement, and positive/negative agreement. Linear and proportional-odds logistic regression models assessed the association between confirmed informant endorsements and clinical anchors (NGASR and C-SSRS), controlling for age and sex. Sensitivity analyses examined completion timing (≤24h, ≤72h), age range (11–17 years), and alternative scoring thresholds.

Key Results

Agreement Metrics: Agreement for the platform-exported suicide domain was modest across all metrics: Spearman $\rho$ = 0.321, ICC = 0.318, and weighted $\kappa$ = 0.317. These estimates remained consistent across sensitivity analyses regarding timing and age.
Item-Level Discrepancies: After excluding "unknown" responses, exact agreement for recent suicidal ideation/planning was 66.0% ( $\kappa$ = 0.321), and for lifetime suicide attempt was 66.2% ( $\kappa$ = 0.303). Positive agreement was higher than negative agreement for both items.
Report Patterns: Among the 205 classifiable participants, endorsement patterns were distributed as follows: neither informant (15.1%), parent only (11.2%), adolescent only (21.0%), and both informants (52.7%).
Clinical Convergence: In joint models, confirmed endorsement by both adolescents and parents was independently associated with higher NGASR scores (Adolescent: $B$ = 2.53; Parent: $B$ = 1.77) and higher C-SSRS risk levels. However, the authors note that these associations likely reflect overlapping content between the questionnaires and the clinical anchors rather than independent predictive power.
NSSI Findings: While baseline informant endorsements were associated with past-month non-suicidal self-injury (NSSI), exploratory follow-up analyses (1 and 3 months) did not show clear independent associations between baseline endorsements and subsequent NSSI, though limited by incomplete follow-up data.

Key Contributions

Quantification of Modest Agreement: The study provides robust, multi-metric evidence that parent and adolescent reports of suicide-related symptoms are not interchangeable, even in a controlled inpatient setting.
Methodological Refinement: The paper highlights the impact of caregiver "unknown" responses on data classification, advocating for the use of confirmed raw yes/no items over platform-exported domain counts to avoid inflating apparent parent positivity.
Differentiation of Constructs: By analyzing recent ideation and lifetime attempts separately, the study demonstrates that combined domains obscure clinically distinct information, suggesting future assessments should specify reference periods for each construct.
Interpretation of Discrepancies: The findings support a clinical framework where discrepancies are treated as meaningful data requiring clarification rather than measurement error. Specifically, "adolescent-only" endorsement was associated with higher clinical risk scores than "neither" endorsement, suggesting potential concealment or limited caregiver awareness.

Significance and Claims
The paper concludes that parent and adolescent reports overlap but cannot substitute for one another. The primary significance lies in the clinical recommendation that disagreement between informants should prompt structured, developmentally sensitive clarification rather than the dismissal of one report in favor of the other.

Crucially, the authors maintain a modest stance regarding predictive validity. They explicitly state that the observed associations between informant reports and clinical anchors are cross-sectional and confounded by content overlap; therefore, the data do not support claims that parent reports improve the prediction of future suicidal behavior. The study emphasizes that while multi-informant discrepancies contain clinically relevant information, establishing their prognostic value requires prospective studies with independent outcome assessment and blinded raters.

The Two-Headed Detective: Why One Story Isn't Enough

The Study: When the Teen and Parent Tell Different Stories

Technical Summary: Parent–Adolescent Agreement in Suicide-Related Symptom Reporting

More like this