Beyond DSM-5 Categories: A Clinician's Guide to Dimensional (HiTOP) Assessment
Why categorical diagnosis holds the field back — and how a dimensional model (HiTOP) predicts outcomes better.
Mental health spending is increasing (slowly). More people are accessing treatment than ever before. Research funding is holding steady. And yet—the prevalence of mental disorders is rising, treatment non-response rates remain stubbornly unchanged, and breakthroughs in new treatments are conspicuously absent.
So why aren’t we making more progress? The webinar hosted by NovoPsych and presented by Associate Professor Miri Forbes argues that a significant part of the answer lies in how we classify mental disorders—and that the DSM-5, the dominant diagnostic and classification system used globally, has fundamental structural limitations that constrain the research we produce, the clinical assessments we conduct, and ultimately the treatments we develop.
The DSM-5 lists 1,419 symptoms — but only 628 are unique, and a full third of its diagnoses have no unique symptoms at all (Forbes, 2023).
The DSM-5: What It Gets Right
The DSM-5 has genuine strengths. It standardized diagnosis globally, substantially improved the reliability of research by giving scientists a common set of constructs to study, and provided a shared clinical language across disciplines and countries (American Psychiatric Association, 2022). These are not trivial achievements. But classification does more than facilitate communication—it drives research by defining the constructs we study, shapes clinical assessment and treatment planning, and affects funding priorities. Which is precisely why its limitations carry such significant downstream consequences.
Dimensional HiTOP scores predicted functional impairment 186% better than DSM diagnoses.
The webinar Shortcomings of the DSM-5 and an Alternative Approach to Assessment and Classification, presented by Associate Professor Miri Forbes, walks through six evidence-based limitations of categorical diagnosis—from rampant comorbidity and symptom repetition to arbitrary thresholds—and introduces HiTOP, a dimensional alternative that predicts real-world outcomes substantially better. Watch the recording for the full evidence base and its practical implications for clinical assessment.
Limitations of the DSM-5
The DSM-5 uses a categorical classification system: you either meet criteria for a disorder, or you don’t. The evidence suggests this approach has at least six significant limitations.
- Comorbidity is the rule, not the exception. The DSM treats co-occurring disorders as distinct entities that happen to coincide. But people with Major Depressive Disorder are ten times more likely to also meet criteria for Generalized Anxiety Disorder than the general population—more likely reflecting shared underlying dimensions of psychopathology than genuinely independent conditions (Kotov et al., 2017).
- Symptom repetition inflates the diagnostic count. The DSM-5 lists 1,419 symptoms in total—but only 628 are unique. A full third of DSM-5 diagnoses have no unique symptoms whatsoever (Forbes, 2023).
- Diagnostic heterogeneity makes the same label clinically meaningless. MDD permits over 100,000 distinct symptom profiles; in one large study, 90% of profiles were observed in a single person only (Forbes, 2023). PTSD, with 46,778,256 possible profiles, makes MDD look tidy by comparison.
- Low interrater reliability. The DSM-5 field trials found that 40% of diagnoses had unacceptable levels of interrater reliability—clinicians applying the same criteria to the same patient frequently arrived at different conclusions (Regier et al., 2013).
- Arbitrary diagnostic thresholds. A patient with four depressive symptoms is categorically “well”; one with five is “disordered.” The underlying continuous gradient of severity is artificially bifurcated at a line with no particular biological or psychological justification.
- Diagnostic instability over time. Because thresholds are categorical, small fluctuations in symptom counts flip a person’s diagnostic status entirely—not because their psychological functioning changed, but because the scoring system poorly captures dimensional variation.
A Data-Driven Dimensional Alternative: HiTOP
The Hierarchical Taxonomy of Psychopathology, or HiTOP, is an empirically derived, dimensional alternative to categorical diagnosis (Kotov et al., 2017). Rather than asking “does this person meet criteria for Disorder X?”, HiTOP asks “where does this person sit along the relevant dimensions of psychopathology—and at what level of severity?”
HiTOP rests on three core features. It is data-driven, derived from factor analyses of psychopathology data rather than committee consensus. It is dimensional, measuring severity along continuous spectra rather than imposing arbitrary cutoffs. And it is hierarchical, organized into nested levels from broad superspectra down to individual symptom components.
The Six HiTOP Spectra
The HiTOP model organizes psychopathology into six main spectra: Internalizing (negative emotions including depression, worry, and panic), Externalizing (antisocial behavior and substance use), Thought Disorder (delusions, hallucinations, paranoia), Detachment (social withdrawal and low social drive), Antagonism (aggression and callousness), and Somatoform (unexplained physical symptoms and health anxiety).
Critically, this structure does not just look more elegant on paper—it predicts outcomes better. Compared to DSM diagnoses, dimensional HiTOP scores have shown a 186% improvement in predicting functional impairment, a 135% improvement in predicting medication prescription needs, and a 135% improvement in predicting the course of illness prospectively over up to 20 years (Kotov et al., 2017). A meta-analysis of 35 studies and over 120,000 participants confirmed that the hierarchical dimensional structure replicates robustly across samples (Ringwald, Forbes, & Wright, 2021).
Limitations of the HiTOP Model
Critics have noted that the original HiTOP model was built on DSM diagnoses (arguably repackaging categorical thinking at a higher level of abstraction), that early iterations relied on narrative reviews rather than systematic meta-analyses, and that HiTOP initially provided insufficient coverage of the full DSM. The consortium has addressed these critiques methodically: a meta-analytic review confirmed the structure empirically, and a symptom-level rebuild—working directly from individual symptoms across 14,762 participants—largely replicated the framework, now incorporating 167 disorders instead of 71 and adding neurodevelopmental and cognitive dimensions (Forbes et al., 2025). The model is evolving in response to evidence.
The Impact of HiTOP on Your Clinical Practice
The DSM-5 is not going away anytime soon, and there is no suggestion that clinicians should abandon it wholesale—diagnostic labels remain necessary for billing, communication, and access to services. But understanding their limitations changes how you use them. A dimensionally-informed perspective means you stop treating a diagnosis as a complete description of a person and start treating it as a rough coordinate: being genuinely curious about which specific symptom dimensions are elevated, how severely, and at what level in the hierarchy they are best conceptualized. Two clients with the same diagnosis may need quite different formulations and treatment targets—because they probably do.
NovoPsych’s validated psychological assessment library supports exactly this kind of dimensional, symptom-level assessment, allowing clinicians to assess across the spectra that HiTOP identifies without being constrained by siloed categorical thinking. NovoPsych also offers the Brief Hierarchical Taxonomy of Psychopathology (B-HiTOP), a broad screening assessment of dimensions that span the hierarchical structure of psychopathology in adults.
Clinical Takeaway: Why Classification Matters More Than You Think
Classification is not an abstract academic exercise. It defines the constructs we research, the questions we ask in assessment, the treatment targets we prioritize, and the policies that determine what gets funded. If the classification system has structural flaws, those flaws propagate through the entire field.
The NovoPsych webinar Shortcomings of the DSM-5 and an Alternative Approach to Assessment and Classification, presented by Associate Professor Miri Forbes of Macquarie University, walks through the evidence for each of these limitations in detail—and makes a structured, data-grounded case for why a dimensional approach like HiTOP is not just theoretically appealing, but empirically compelling.
Because it turns out the problem isn’t that mental illness is necessarily resistant to treatment. It may be that we have been measuring it with the wrong ruler.
View over 150 of the current psychometric assessment tools available on NovoPsych.
Warm regards,
Dr Ben Buchanan
Psychologist
NovoPsych Co-founder
[email protected]
LinkedIn