Here’s how it typically works. A parent brings their child to their pediatrician, concerned about a potential mental health condition — maybe it’s anxiety, or maybe it’s depression. That pediatrician asks the child a set of questions, then assigns a numerical score based on the answers. If the score exceeds a certain threshold, the child is diagnosed with a condition and referred for treatment — maybe it’s pills, maybe it’s therapy. This process relies on classifications created in a tome called the Diagnostic and Statistical Manual of Mental Disorders, a book that has been the dominant paradigm in children’s mental health diagnosis since the 1970s and is now on its fifth edition (DSM-5). To James Li, this process is, to put it succinctly, “a crock. 

Portrait of James Li.
James Li: “If you want to do good science in mental health, the first thing you want to do is think about mental health in terms of degrees —  a spectrum, instead of ‘I have ADHD,’ or ‘I have autism,’”

“The diagnosis of mental health is in dire need of a paradigm shift, and the way that we’ve been assessing, diagnosing and providing help to people has been really inefficient and not helpful for the clinician nor the patient,” says Li, the A.A. Alexander Associate Professor of Psychology and Psychiatry in the Department of Psychology. “If you want to do good science in mental health, the first thing you want to do is move away from the DSM and think about mental health in terms of degrees — a spectrum, instead of ‘you have this’ or ‘you don’t have this.’”

For the past few years, Li has been trying to do just that. In 2020, he became part of a growing nationwide consortium of researchers and medical professionals looking to replace the DSM-5 with something that’s more data-driven, nuanced and patient-specific. The emerging new model is called the Hierarchical Taxonomy of Psychopathology (HiTOP), and it’s designed to reconceptualize mental health into something more than a system that efficiently places patients into pigeonholes. Li co-chairs the Diversity, Equity, and Inclusion Scientific Workgroup within the consortium, which is focusing on ensuring that HiTOP applies to all populations.

Most patients are coming in with some combination of a lot of different psychological and mental health traits and behaviors that don’t fit neatly into one treatment category or diagnosis. 

“Most patients don’t come in as garden variety, ‘I have ADHD,’ or ‘I have autism,’ and that’s all you’ve got,” explains Li. “Most patients are coming in with some combination of a lot of different psychological and mental health traits and behaviors that don’t fit neatly into one treatment category or diagnosis.” 

Children’s mental health remains a growing and serious issue nationwide. According to the American Psychiatric Association, one in five children struggles with a mental health disorder, 20% of children are receiving help from a mental health care provider and the number of visits to those providers has grown a whopping 300% since 2010. At the same time, the U.S. health care system often struggles to accommodate patients’ mental health needs, creating a situation in which efficiency sometimes ends up trumping nuance and accuracy. 

Li’s lab in the Waisman Center and the Department of Psychology focuses its research on neural mechanisms and genetics. It’s not lost on him that these two things are based on science and biological constructs. The DSM-5 — and by extension the physicians who use it — relies largely on physician observation.

“What ends up happening is that most of these clinicians are just giving a diagnosis for the sake of giving one and then figuring out through some sort of guesswork what the best combination of treatments is,” says Li. 

A Dimensional Approach to Care

HiTOP has the potential to provide a better model. Psychology researchers have long argued that there are five dimensions of an individual’s personality (“The Big Five,” a designation that includes things like openness, agreeableness and neuroticism). HiTOP is based on a similar principle — that there are also five dimensions of mental health: internalizing, detachment, thought disorder, antagonistic externalizing and disinhibited externalizing. [See “On the Spectrum,” below.] To most of us, those terms might seem confusing and scientific, but within the HiTOP model, each of the traits and symptoms of common mental health problems like depression can be mapped to one of those dimensions. Physicians can then use that framework to measure where a patient stands on each one — a dimensional approach that places the patient on a spectrum, rather than putting them into the box of a singular diagnosis.

“What this does is give researchers and clinicians a common language,” says Li. “Instead of studying an eating disorder as its own thing, you can study eating pathology as part of the internalizing dimension because that’s where it belongs. That’s where all the symptoms seem to cluster.”

Li notes that in this example, years of research and data have shown that eating disorders are often accompanied by depression and anxiety, a nuance the HiTOP model could capture and highlight. Mapping patients and their symptoms in this way also allows physicians to begin creating norms based on specific patient populations, on gender, age, race and ethnicity, to help hone future diagnoses. The DSM-5 model doesn’t include many of these factors.

“This could be used to guide treatment planning at a more nuanced level,” says Li. “Because instead of saying, ‘Well, I’m just going to treat the anxiety disorder,’ you can also say part of the intervention and the treatment planning should be more tailored to the individual, according to the data that we’re getting.”

On the Spectrum

One of the most significant advantages of the HiTOP model is its ability to chart patients and their traits and symptoms to a spectrum that is divided into five mental health dimensions. The key is that patients are rarely mapped to a single dimension but may be scored on all five.  

  1. Somatoform dimension encompasses physical symptoms like malaise, head pain, gastrointestinal symptoms and cognitive symptoms. Anxiety and somatic disorders are reflected in this dimension. 
  2. Internalizing dimension comprises maladaptive traits of emotionality, liability, anxiousness, separation anxiety and anhedonia, as well as symptoms characteristic of distress, fear and eating problems. Disorders like major depression, generalized anxiety, obsessive-compulsive disorder and social phobias may fit here.   
  3. Thought Disorder dimension comprises maladaptive traits of peculiarity, unusual beliefs, unusual experiences and fantasy proneness, as well as symptom dimensions of disorganization and reality distortion. It reflects conditions such as schizophrenia and bipolar disorder. 
  4. Detachment dimension comprises maladaptive traits of emotional detachment, anhedonia, social withdrawal and romantic disinterest, as well as symptom dimensions of avolition and blunted affect. It reflects many personality disorders (e.g., avoidant personality) and schizophrenia-related disorders.   
  5. Externalizing dimension is broken into two related (but still distinct) categories: Disinhibited and Antagonizing. The former reflects traits like impulsivity, disorganization and risk-taking, and the latter reflects traits like manipulativeness, deceitfulness and aggression. Collectively, these dimensions reflect disorders like ADHD, conduct disorder and substance use disorders. 

Changing Hearts and Minds

Shifting mindsets — of physicians, insurance companies and patients — may be the hardest part of championing the new system. Li points to a confluence of factors that have kept each new edition of the DSM in place for more than five decades. The first is the need for physicians to provide patients with a diagnosis so that the care can be reimbursed by insurance companies, and the ease with which the DSM-5 model can often provide it. Li also argues that physicians aren’t always sufficiently trained in mental health care.

“Sometimes there is not a strong appreciation for the complexities of mental health beyond what they learned in textbooks,” says Li.

The third barrier? Until now, there’s never been an acceptable alternative.

“Without an alternative, then what else is there but the status quo?” asks Li.

Li worries that clinicians might struggle without being able to rely on the strict numeric thresholds of the DSM-5 to provide a single diagnosis. After all, most treatment plans are the result of clinical trials and studies designed to test a medication’s effectiveness using patients who meet the DSM-5’s criteria for a mental health disorder. For instance, under the DSM-5’s criteria, a child meets the clinical definition of ADHD if they meet six of a possible nine symptoms — a very yea-or-nay approach. A child who has only five of those symptoms might not receive an ADHD diagnosis or be included in trials to test a potential treatment.

“If you’re a pediatrician who isn’t trained in mental health — and these are the people that are on the frontline seeing kids and figuring out whether or not they should treat for something like ADHD — they’re not really relying on the science anymore,” says Li. “Because if they want to help that kid, they might just decide, let’s just put them on a drug like methylphenidate, even if they’re very young. And is that the right approach?”

For as long as I’ve been in the game, there’s never been an alternative to something that we all knew was a problem. I’m really excited to get this into the ethos. Let’s try to make sure people know something else exists. 

Li joined the HiTOP consortium three years after it first formed in 2017. Today, there are more than 200 research members worldwide, in addition to a clinical network of physicians who are trying the HiTOP model with their patients. Meanwhile, the research surrounding HiTOP continues to explode, with hundreds of studies around the world (including in Li’s own lab) exploring the idea of using a dimensional approach to diagnosing mental health problems. Meanwhile, the American Psychiatric Association, the organization that created and published the DSM-5, is beginning to pay closer attention, as are clinicians in Wisconsin. Li recently gave a presentation on HiTOP at the Medical College of Wisconsin in Milwaukee and noted a lot of heads nodding in agreement when he finished speaking. That’s the kind of reaction that leaves him energized and hopeful.

“For as long as I’ve been in the game, there’s never been an alternative to something that we all knew was a problem,” says Li. “I’m really excited to get this into the ethos. Let’s try to make sure people know something else exists.”

More From Fall 2023


A roundup of the latest happenings on campus in UW–Madison's College of Letters & Science

Just One Byte

L&S researchers team up to bring an ethical approach to the thorniest issues in AI and computing.