Stephanie Nelson, PhD, ABPP, ABPdN, is a board-certified pediatric neuropsychologist and founder of The Peer Consult, where she provides consultation and mentorship to assessment psychologists. Her interests include clinical reasoning, psychological assessment, and understanding human behavior through developmental, contextual, and systems-oriented perspectives.
This guest post is the second of a pair of articles by Stephanie Nelson in response to recent posts on Psychiatry at the Margins. It is a comment on:
In his essay, “Adiposity by Self-Report,” Dr. Aftab asks us to imagine a world in which body fat is invisible and impossible to measure directly. How would we diagnose obesity from subjective experiences, self-reports, and associated behaviors? How would we disentangle medically necessary treatment from much-wanted enhancement, or helpful intervention from potential harm? Wouldn’t we face the same epistemic problems we encounter when diagnosing ADHD, which cannot be determined by brain scans or neuropsychological scores?
It’s a thought-provoking analogy. But I think it understates something important about clinical diagnosis without a biomarker. In the absence of a definitive test for ADHD, we triangulate toward a latent construct using multiple imperfect indicators.
We certainly start with subjective experiences. As Dr. Aftab’s essays poignantly explore, our patients share heartbreaking stories of stress and distress. They may have spent years trying to understand their struggles. They long for compassion, clarity, and concrete support, and we long to provide all three. We begin with careful, and care-full, attention to their lived experiences.
But the bright beam of empathic attention can narrow our focus, so clinicians guard against attending only to subjective experience. We must determine whether patients’ descriptions are symptoms at all, and if so, where they belong within the diagnostic space. Descriptions that are unique, universal, incongruent, inconsistent, or improbable can all signal very real distress without being symptoms of the disorder in question. An experience can be clinically important while having little diagnostic value. Symptoms are also shapeshifters. What points in one direction may signal something else, so each symptom must expand curiosity rather than collapse into certainty. Clinicians also attend to symptomatic negative space, cautiously questioning whether expected core symptoms are missing.
Subjective symptoms are also, well, subjective. In our wish to help, and our patients’ wish to be helped, we can miss, mistake, misattribute, or reshape symptoms. Self-inferences can block or blur the diagnostic picture. The temptation to subtly mold symptoms to fit our preconceptions is enshrined in many of our clinical pearls: everything looks like a nail when all we have is a hammer, and we can miss a whole forest by focusing on the trees.
Since symptoms are slippery, we seek other imperfect indicators. While Dr. Aftab mentions self-report and behavioral inferences, and talks about self-perception and other-perception, I don’t think he fully explores the implications of triangulating from multiple lines of indirect evidence. We look for signs, or the behaviors observable to clinicians and people close to the patient. We consider continuity, examining the conditions under which signs and symptoms appear, persist, or diminish. We trace the course to determine developmental timing and trajectories. We evaluate evidence for functional impairment. We weigh relevant family and medical history and the presence of known correlates. We also carefully consider alternative explanations.
If we do neuropsychological assessment, we also bring the concerns into the consulting room. Scores on tests of attention or processing speed do not diagnose, yet the person’s approach can clarify. During testing, did we see signs of fatigue, frustration, fear of failure, forgetfulness, flagging focus, or freezing? Can we get curious about the patient’s experiences together? Research can contribute knowledge about group-level differences, such as distinct patterns of brain activity or discrepant outcomes over time. No one indicator is decisive, and conflicting evidence can reveal as much as agreement. Together, signs and symptoms suggest a clinical pattern, but the other indicators help us judge whether this pattern, in this person, warrants this diagnosis.
So let us reconsider how we would diagnose adiposity if it were literally invisible. We wouldn’t merely ask whether patients felt overweight or engaged in behaviors they associate with obesity. Instead, we would build a syndrome from converging evidence. We would start with subjective symptoms around hunger, eating, physical comfort, or exertion. But we would explore beyond self-report to ensure we were addressing the right problem and proceeding with appropriate caution. We would look for signs, such as specific patterns around eating and exertion, and consider the conditions under which those patterns occur. If we were assessors, we might devise tasks to observe some of those patterns ourselves. We would consider whether difficulty climbing stairs or finding clothes that fit constitutes functional impairment. We would trace the course to look for predisposing and precipitating factors. We would consider family and medical history and the presence of known correlates like physiological markers of inflammation. We would carefully exclude other causes; for example, hunger caused by severe caloric restriction, medication effects, or depression.
This process would be messy. Many indicators would be neither necessary nor sufficient; some would be consequences rather than causes; others would prove misleading. Our constructed syndrome would only imperfectly approximate the underlying inferred phenomenon. Still, triangulating to adiposity from multiple vantage points would give us something to investigate, validate, and revise, in a way that relying on subjective experience alone would not.
This discussion may sound abstract, or like I’m congratulating clinicians on the comprehensiveness of our (idealized) assessments. It may even sound unnecessary or unachievable in real practice: we have limited time, so we do often center our decisions around self-reported symptoms and behaviors. But reminding ourselves, our patients, and the public about how we infer diagnoses from multiple indicators is important and has real consequences.
Consider adult ADHD. Many more adults are reporting experiences and behavior that may align with a missed diagnosis, and we are having a public conversation about what that means. When the discussion focuses primarily on self-report, it can sound like we’re debating whether someone’s subjective experiences are true or worthy of care. This is a profoundly unhelpful frame, ungenerous to patients and clinicians alike. It’s no wonder people feel dismissed, defensive, and distrustful, and the debate quickly devolves. Looking beyond self-report gives us a way out of this impasse. Bringing the other indicators into the frame makes room for more expansive, and potentially more productive, conversations.
Because what worries me is not that more adults are reporting potential symptoms of ADHD. It’s that when we consider their experiences alongside the other indicators, we may be looking at a different pattern altogether. Perhaps the more productive question is whether we’re missing something about the pattern itself.
In clinical and consultation work, and conversations with colleagues, I’ve noticed some striking shifts. Some adults are reporting breathtakingly high levels of executive dysfunction, even compared with those I’ve historically encountered in patients with ADHD and extensive documented impairment. Nonspecific and transdiagnostic experiences such as rejection sensitivity, and seemingly paradoxical experiences such as hyperfocus, are increasingly folded into the ADHD phenotype, despite limited value as discriminating indicators. These experiences are not necessarily new, but their place within the narrative has changed; they sometimes receive more emphasis than the canonical symptoms of inattention, hyperactivity, and impulsivity. Self-report profiles can also be extraordinarily broad, with elevations that cross multiple diagnostic categories, and frequent endorsement of low-incidence symptoms and concerns not traditionally associated with ADHD.
Sometimes these symptoms occur alongside relatively few observable signs, even when we consider observations from people who have known the patient well for decades. Because signs are imperfect (people compensate, observers are biased, memories fail), an increasingly common clinical narrative posits that the absence of signs can never be useful or revealing. We also increasingly see behavioral signs recast as subjective symptoms: not “I routinely fail to complete tasks,” but “it feels much harder for me to complete tasks than it seems to be for other people”; not “I repeatedly lose things,” but “I feel as though I would lose things if I didn’t compensate so intensely.” These descriptions introduce additional layers of inference. What was once relatively visible must now be inferred, sometimes primarily from what the patient imagines might happen if they weren’t compensating so intensely.
These are clinical observations, rather than established population-level findings. But surely they deserve serious investigation, without the questions themselves being treated as evidence of inadequate empathy, disregard for lived experience, or attempts to gatekeep care. Have you noticed these shifts too? Are the distribution and content of complaints changing? Do they converge with the signs, developmental histories, and functional consequences we expect? Are we seeing new expressions of an established syndrome, new ways of describing old difficulties, overdue recognition of overlooked experiences, or something genuinely different?
Broadening a phenotype can be scientifically legitimate, of course, but we would need evidence that its new manifestations belong to the same construct. If this emerging pattern turns out to be associated with less observable impairment, weaker cross-setting corroboration, or different developmental histories, familial patterns, or outcomes, shouldn’t we ask: Are we still talking about the same syndrome? Is this really a different picture of the same diagnosis, or might we be trying to cram a star-shaped peg into a round hole?
This is not the same as asking whether patients suffer or are impaired. Something is happening that is both real and really distressing. We’re asking whether ADHD best explains the pattern and guides our response, rather than assuming a familiar label settles those questions. This matters especially once treatment enters the picture. Interventions can feel profoundly helpful while reinforcing dysfunctional processes or introducing new harms. For example, someone might experience enormous relief from an appetite-suppressing medication without having excess adiposity and without that medication being an appropriate treatment.
Our obligation to ask these questions comes from our desire to provide good care to those seeking our help. But I also feel a duty of care toward the people who are not in my office at this moment. Those include people with more traditional, sometimes severely impairing ADHD, whose difficulties may be less easily articulated or less readily translated into today’s language for accessing care. Expanding a diagnosis can change who qualifies for services, how limited resources are distributed, and whose struggles are recognized. We should be curious about whether the people who have historically experienced the most severe manifestations of a disorder are being helped or displaced as its boundaries change.
Diagnostic categories also ripple outward, shaping self-understanding, family and school responses, clinical attention, research priorities, and expectations of ordinary human functioning. A diagnosis can alter the expectations we place on a child, what we ask of a spouse, or what we attribute to an employee. And as certain experiences become linked to specific diagnoses, they may become less visible as possible responses to mismatched environments, strained relationships, impossible expectations, or the ordinary demands of being human. Meanwhile, the diagnostic explanations we offer influence which of their own experiences people attend to and consider worth exploring. We risk changing not just the boundaries of a diagnosis, but the very meaning of the experiences inside them.
I want us to be able to ask what we’re uncovering, what we might be overlooking, and what follows from the explanations we choose. I’m grateful to Dr. Aftab for providing a space for serious, thoughtful, pluralistic conversations like this one.
See also:
I scanned this thoughtful post. I would like to add two points that weren't mentioned.
1) "Adult ADHD" is not in the DSM. I don't mean to stand on principle or play games with words. Check the DSM, the ADHD diagnosis exclusively describes children, if you apply it to adults you have to use your own imagination to translate it to adult problems. Furthermore, ,I only have the DSM on my phone (in a an app). I don't have time to hunt around for "proposed" adult symptoms or the consensus of some advocacy organization I don't know. Translating the ADHD criteria to adult experience is as, or more challenging than differentiating mood disorders from psychotic disorders in the DSM famework, for example.
2) The author made no mention that ADHD is a specific type of cognitive impairment that appears in childhood. I don't mean to fight the battle of "if it wasn't there before...". I mean to point out that there are numerous etiologies of this sort of impairment besides the disorder we have called "ADHD". Sleep deprivation for example; a promotion to a job for which you are unprepared; young children; multiple jobs that exceed a manageable amount of waking time. Adults with reported ADHD haven't really been systematically studied to explore these etiologies. Patients in menopause report this; patients after chemotherapy; patients with multiple schlerosis. We have to acknowledge this then decide what to do about it. We all know the question: who gets stimulants? We can't discuss ADHD diagnosiss without mentioning stimulants.
The list of indicators stops one indicator short of the clinic. Signs, continuity, course, impairment, history, alternatives: all six are gathered before anyone prescribes. The one gathered after is what happens when the drug is taken away, and her own appetite-suppressant example is where that shows. In the STEP 1 extension, adults with obesity who stopped semaglutide had regained about two-thirds of the lost weight within a year, which says the drug managed a mechanism it never changed. Stimulants behave the same way in the randomized withdrawal trials collected in the 2021 Frontiers in Psychiatry review: most stable responders relapse within two weeks of switching to placebo, and roughly 30 percent do not. Nobody, as far as I can tell, has asked which presentation that 30 percent came from. If the newer profile she describes, high self-reported dysfunction with few corroborated signs, clusters among the people who stop and stay well, that is a different syndrome by a test the clinic already runs. If it clusters among the relapsers, the phenotype widened and the construct held. The withdrawal trials on file were run on the older presentation, so they cannot settle this by themselves, and the non-relapsers may just be people whose demands dropped or who built the compensations she mentions. I don't know. What strikes me is that NICE already requires an annual review and permits a trial of discontinuation, so the data point is being generated and then filed as a dosing decision. The record that would move me is a withdrawal trial run on the new presentation.