One pattern across many diagnostic labels
Psychiatric classification divides distress into named disorders. Real people often cross those borders. Anxiety, depression, trauma-related symptoms, impulsivity, sleep disruption, substance use, and interpersonal difficulty overlap in ways that diagnostic categories only partly capture.
Emotion regulation has long been proposed as one process that may cut across diagnoses. A 2026 systematic review and meta-analysis in Translational Psychiatry tested that proposition at substantial scale by comparing people with psychiatric disorders and healthy controls on established self-report measures of emotion regulation.
The result supports a transdiagnostic pattern of impairment. The meaning of that pattern requires restraint.
Source card
Title: Emotion regulation impairment across psychiatric disorders: a systematic review and meta-analysis
Authors: Matthew E. Sloan and colleagues
Journal: Translational Psychiatry
Volume/article: 16, article 386 (2026)
Published online: 29 May 2026
DOI: https://doi.org/10.1038/s41398-026-04119-x
Design: systematic review and meta-analysis of case-control comparisons
Scope: 188 studies, 563 case-control comparisons, 15 psychiatric diagnoses, 11,201 cases and 9,609 controls, according to the published paper
Search coverage: databases searched from inception through September 2023
The research question
The authors asked whether impaired emotion regulation is evident across major classes of psychiatric disorder and therefore qualifies, at least descriptively, as a transdiagnostic feature.
Included studies compared adults with a psychiatric disorder against healthy controls and used one of four self-report instruments: the Difficulties in Emotion Regulation Scale, Emotion Regulation Questionnaire, Cognitive Emotion Regulation Questionnaire, or Emotion Regulation Skills Questionnaire.
This measure-based inclusion rule provides comparability. It also defines what “emotion regulation impairment” means in the analysis. The meta-analysis is not a synthesis of every behavioral, physiological, clinical, or ecological measure of regulation.
What the study found
Across diagnoses, people in psychiatric groups generally reported greater difficulties on the DERS and less favorable patterns on relevant ERQ dimensions than healthy controls.
For DERS comparisons, effect sizes were large across the disorders with sufficient data, ranging in the paper from approximately Hedges' g 0.94 for bipolar disorder to 2.55 for borderline personality disorder. On the ERQ, cognitive reappraisal was lower in most psychiatric groups, with panic disorder an exception in the available data, while expressive suppression was generally higher, with exceptions depending on diagnosis and sensitivity analysis.
The overall pattern is consistent with the proposition that emotion-regulation difficulty is not confined to one diagnosis.
It is not evidence that all psychiatric disorders are the same, or that regulation difficulty explains the entirety of any one disorder.
What “transdiagnostic” means here
Transdiagnostic can be used in several ways. In this paper, the strongest supported meaning is descriptive: similar categories of self-reported regulation difficulty appear across multiple diagnoses.
A stronger meaning would be causal: one shared regulatory mechanism causes or maintains all those disorders. The meta-analysis cannot establish that. Most included comparisons are cross-sectional. They show group differences at a point in time, not which process came first.
A still stronger clinical claim would be that one regulation-focused treatment should work across all diagnoses. That requires intervention evidence.
Keeping these meanings separate prevents an important finding from becoming an oversized theory.
The DERS findings
The DERS contributed the largest body of comparisons in the meta-analysis. It measures several perceived difficulties involving acceptance, goals, impulse control, awareness, clarity, and access to strategies, depending on version and scoring.
Large case-control differences can arise because these difficulties are genuinely common across psychiatric disorders. They may also partly reflect shared distress, negative self-evaluation, symptom severity, or overlapping questionnaire content.
That is not a trivial objection. When a predictor and an outcome both involve self-reported difficulty, common-method variance can inflate relationships. The study's case-control design reduces some ambiguities but does not remove them.
The practical conclusion is that DERS differences are robust descriptive evidence of reported difficulty, not a direct assay of a single neural regulatory deficit.
The ERQ findings
The ERQ measures habitual cognitive reappraisal and expressive suppression. Across many diagnostic groups in the review, reappraisal tended to be lower and suppression higher than in healthy controls.
That pattern fits a large literature associating reappraisal with more favorable mental-health outcomes and suppression with less favorable ones on average. It should still be interpreted contextually. Habitual strategy frequency is not equivalent to strategy effectiveness, and culture moderates some strategy-outcome associations.
Sensitivity analyses in the 2026 paper also matter. When analyses were restricted to higher-quality studies, some suppression differences changed significance for particular diagnoses. This is exactly why a meta-analysis should be read as a pattern with uncertainty rather than a list of immutable disorder profiles.
Sample and design boundaries
The review focused on adults. Mean participant age across the evidence base was in the mid-thirties, and the sample was approximately two-thirds female. Studies were limited by the available English-language full-text literature and by minimum evidence requirements for disorder-measure combinations.
The search ran through September 2023, even though the synthesis was published in 2026. That lag is normal in systematic review work but important for update planning.
Case-control designs also select a particular comparison: diagnosed groups versus healthy controls. They do not show how regulation varies dimensionally across the whole population, nor how it changes before, during, and after illness.
Causality remains open
The most consequential limitation is temporal. Emotion-regulation difficulties may increase vulnerability to psychopathology. Psychiatric symptoms may impair regulation. Both may be influenced by a third process such as chronic stress, trauma exposure, sleep disruption, cognitive load, social adversity, or neurodevelopment.
Longitudinal designs can clarify temporal relationships. Intervention studies can test whether deliberately changing a regulatory process changes symptoms. Experimental paradigms can test specific mechanisms. None of those causal questions are settled by cross-sectional group differences alone.
The paper acknowledges this limitation, and any downstream summary should preserve it.
Measurement heterogeneity
Four questionnaire families were included, and they do not operationalize the same construct. DERS measures broad perceived difficulties. ERQ focuses on two strategies. CERQ emphasizes cognitive strategies. ERSQ captures regulation skills.
Calling all of them “emotion regulation” is defensible at a high level but risks conceptual blending. A transdiagnostic pattern that appears across several measures is stronger than one found on a single instrument, yet differences among instruments remain scientifically meaningful.
Future work would benefit from multimethod assessment and from models that distinguish general distress from regulation-specific variance.
What the evidence does not show
This meta-analysis does not show that emotion-regulation impairment causes psychiatric disorders. It does not show that a high DERS score diagnoses psychopathology. It does not show that every person with a psychiatric diagnosis has poor regulation, or that healthy controls have uniformly strong regulation.
It does not show that one transdiagnostic treatment will be equally effective across 15 diagnoses. It does not justify telling a person that their disorder is “really” an emotion-regulation problem.
The analysis also does not cover children and adolescents in the same way, nor does it establish equivalence across all cultures and languages represented by the underlying measures.
Relation to the 2026 clinical-intervention review
Johan Bjureberg's 2026 Annual Review examines the role of emotion regulation in clinical interventions and the possibility that regulatory change mediates treatment outcomes. The Sloan meta-analysis and Bjureberg review answer complementary questions.
Sloan and colleagues show that self-reported regulation impairment is broadly distributed across diagnoses. Bjureberg examines how interventions may engage regulatory abilities and processes. The first strengthens the rationale for transdiagnostic attention. The second makes clear that mechanism and treatment claims require additional evidence.
Together they support research investment without collapsing descriptive, causal, and intervention evidence into one category.
Why this matters in Northern Europe
The paper is internationally authored rather than specifically Nordic, but its implications are relevant to Nordic mental-health research because Swedish groups have made substantial contributions to DERS measurement and digital emotion-regulation interventions.
A Nordic research agenda could use strong health registries, longitudinal cohorts, digital treatment platforms, and multimethod assessment to test whether regulation difficulties predict onset, course, relapse, or treatment response beyond general distress.
The regional opportunity is methodological. It is not a reason to invent a geographically unique form of emotion regulation.
Institute synthesis: shared difficulty does not mean shared destiny
The following is Institute synthesis.
For Inner Technology, transdiagnostic evidence highlights capacities that matter across conventional categories. The ability to notice emotion, tolerate it, maintain goal-directed action, choose responses, and recover from activation can be relevant to many people regardless of diagnosis.
But the Institute does not treat clinical impairment as the template for all human development. A capacity framework must distinguish prevention, education, ordinary variation, and clinical care. It must also avoid turning universal human difficulty into pathology.
The most useful translation is therefore structural: if a capacity appears relevant across multiple forms of distress, it deserves careful measurement and serious research. That is not permission to promise treatment outcomes outside healthcare.
At an institutional level, the paper also warns against siloed programmes that assume each diagnostic label requires an entirely separate account of regulation. Shared mechanisms can support integrated research while clinical decisions remain individualized.
Open questions
Does emotion-regulation impairment predict onset before symptoms emerge? Which dimensions are most transdiagnostic after controlling for general distress? Which are disorder-specific? How much do culture and age alter the pattern? Can intensive longitudinal measures distinguish capacity from retrospective self-evaluation? Which regulatory changes actually mediate treatment outcomes?
The 2026 meta-analysis establishes a strong descriptive platform. The next work is causal, temporal, and mechanistic.
Large effects do not identify one underlying mechanism
Large standardized differences can feel mechanistically decisive. If a disorder group differs from healthy controls by more than a standard deviation on a DERS total score, it is tempting to imagine a common regulatory deficit underneath the diagnosis.
Several different processes can create the same statistical separation. Participants may have more intense or frequent emotions, making regulation harder even if regulatory skill is similar. They may evaluate themselves more negatively. Medication, sleep, social adversity, trauma history, cognitive symptoms, or current episode severity may affect questionnaire responses. Diagnostic groups may also differ in age, education, comorbidity, or treatment exposure.
Meta-analysis can average these differences and test some moderators, but it cannot reconstruct every unmeasured pathway. “Large” describes separation on the measure, not the singularity of the cause.
Healthy-control comparison has a hidden denominator
Healthy controls are not a neutral zero point. Studies define and screen them differently. Some exclude any psychiatric history; others rely on current status. Recruitment may come from universities, communities, online panels, or local advertisements.
A stringent healthy-control group can increase case-control contrast relative to the general population, where subclinical symptoms and mixed histories are common. This is one reason transdiagnostic findings should eventually be tested dimensionally across population cohorts rather than only by diagnostic category.
Comorbidity complicates diagnostic bins
Many psychiatric disorders co-occur. A participant assigned to a primary diagnosis may also meet criteria for another condition or carry substantial symptoms outside that category. Emotion-regulation impairment attributed to one diagnosis can therefore reflect shared comorbidity.
The transdiagnostic interpretation is partly compatible with this reality: shared regulation difficulty may help explain overlap. But comorbidity can also make disorder-specific effect sizes difficult to interpret. Studies with careful characterization, dimensional symptom measures, and multivariate models are needed to separate shared from specific variance.
From association to prevention
One of the most attractive implications of a transdiagnostic factor is prevention. If regulation difficulty precedes multiple forms of psychopathology, strengthening relevant capacities might reduce risk across diagnoses. The 2026 meta-analysis itself cannot establish that sequence because most source studies are cross-sectional.
Prospective cohorts are needed to test whether regulation measures predict new onset after accounting for baseline symptoms and confounders. Prevention trials can then test whether changing a specific process reduces incidence or severity. Without those steps, prevention remains a plausible research direction rather than a demonstrated outcome.
This distinction is especially important for schools and workplaces, where clinical language can travel quickly. A transdiagnostic research signal does not justify population screening or treatment claims outside healthcare.
A better future evidence stack
The strongest next generation of work would combine self-report measures such as DERS and ERQ with experience sampling, behavioral tasks, clinician assessment, and longitudinal outcomes. It would examine general distress explicitly, test measurement invariance, include more diverse age groups and cultures, and preregister primary hypotheses.
Such multimethod work could reveal whether there is a broad general factor of regulation difficulty, several partially independent dimensions, or different pathways across disorders. It could also show whether the most clinically useful level is diagnosis, symptom dimension, regulatory process, or some combination.
Heterogeneity inside the meta-analytic average
A pooled transdiagnostic pattern can coexist with substantial variation among studies. Different recruitment settings, diagnostic procedures, illness severity, medication use, cultures, and instrument versions can all shift effect sizes. Heterogeneity is not statistical noise to be wished away; it can point toward moderators that matter clinically.
The paper's sensitivity analyses are therefore part of the result, not a technical appendix. When significance changes under stricter study-quality criteria for some disorder-strategy comparisons, confidence should become more specific rather than collapsing into a yes/no conclusion. The robust claim is that regulation difficulties are widespread across diagnoses. The precise profile for each diagnosis remains a more conditional question.
Disorder categories and dimensional science
Modern psychopathology research increasingly examines dimensions that cut across diagnoses, including negative affect, compulsivity, threat sensitivity, cognitive control, and social processes. Emotion regulation fits naturally into that movement because it can be measured continuously.
The 2026 meta-analysis uses conventional diagnoses because that is how much of the source literature is organized. Its transdiagnostic conclusion can therefore be read as a bridge: categorical groups differ from healthy controls on a process that itself is dimensional. Future studies can test the same process without requiring a hard diagnostic boundary.
Dimensional models also help avoid an unfortunate interpretation in which clinical groups are imagined as lacking a capacity that healthy people possess. Regulation varies across everyone. Psychiatric disorders may shift the distribution, intensity, context, or consequences of difficulty rather than create an entirely separate kind of human functioning.
Treatment status is another source of variation
Many participants in psychiatric research are receiving medication, psychotherapy, or other support. Treatment can influence both symptoms and regulation scores. Studies differ in whether they recruit newly diagnosed, currently treated, remitted, inpatient, outpatient, or community participants.
Meta-analytic group labels compress those differences. Where source studies report treatment status, future syntheses can examine it as a moderator. Until then, a case-control effect should not be treated as the untouched signature of a disorder.
Clinical communication matters
A transdiagnostic result can be communicated in two very different ways. “People across many diagnoses report regulation difficulties” is evidence-based and potentially destigmatizing. “All mental illness is emotion dysregulation” is reductive and unsupported.
The difference is not cosmetic. Explanatory language shapes how patients understand responsibility, prognosis, and treatment. A shared process can create useful treatment targets without becoming a total explanation of suffering.
Research references
Sloan, M. E., et al. (2026). Emotion regulation impairment across psychiatric disorders: a systematic review and meta-analysis. Translational Psychiatry, 16, 386. https://doi.org/10.1038/s41398-026-04119-x
Bjureberg, J. (2026). The Role of Emotion Regulation in Clinical Interventions. Annual Review of Clinical Psychology, 22, 105–133. https://doi.org/10.1146/annurev-clinpsy-061324-072853
Gratz, K. L., & Roemer, L. (2004). Multidimensional assessment of emotion regulation and dysregulation. Journal of Psychopathology and Behavioral Assessment, 26, 41–54. https://doi.org/10.1023/B:JOBA.0000007455.08539.94
Gross, J. J., & John, O. P. (2003). Individual differences in two emotion regulation processes. Journal of Personality and Social Psychology, 85, 348–362. https://doi.org/10.1037/0022-3514.85.2.348