Why ADHD Assessment Keeps Missing Women
ADHD is underdiagnosed in women because the diagnosis depends on a childhood record of visible disruption, and girls are socialized to produce a quiet one. The symptoms don't disappear; they turn inward, and the chart calls them anxiety.
Shafi Lodhi, MD · September 2026
ADHD assessment misses women for a structural reason rather than a careless one. The diagnosis requires evidence of impairment beginning in childhood, and the evidence most instruments are built to find is visible disruption — the child who couldn't stay seated, the report card with conduct comments, the parent called in twice a semester. Girls with ADHD present more often with predominantly inattentive symptoms, and they face earlier and stronger expectations of organization and compliance, so by nine or ten many of them have already built scaffolding — lists, rituals, late nights — that holds the outward behavior steady while the internal experience stays punishing. The disruption goes inward. What reaches the chart instead is anxiety, then depression, then a decade of treatment that never quite works.
The presentation gap is really a visibility gap
The hyperactivity in adult women with ADHD is usually there. It simply isn't in the legs. It shows up as a mind that will not stop narrating, a physical restlessness that gets discharged into cleaning at eleven at night, an inability to sit through a film without a phone in hand. None of that interrupts a classroom or a meeting. None of it generates a complaint from anyone but the patient.
Inattention is similarly quiet. A woman who loses the thread of a conversation and reconstructs it from context, who reads a page four times and then reads it aloud in her head to make it stick, who has never once in her adult life opened her mail on the day it arrived — she produces no incident. She produces a feeling, which she reports to her primary care physician as being overwhelmed. Overwhelm is not a diagnostic category, so it gets translated into the nearest one that is.
Emotional dysregulation is where the most diagnostic damage happens. Rejection sensitivity, rapid mood shifts that resolve within hours, irritability that arrives with fatigue and cognitive load, the crash after a long day of holding it together — these are read, reasonably enough, as mood pathology. A clinician hearing about mood lability and chronic low self-worth in a thirty-four-year-old woman has a short list of things to consider, and ADHD is rarely at the top of it. So the mood symptoms get treated as the primary condition instead of what they often are in these presentations: the downstream cost of running an attentional system at maximum effort all day, every day, for twenty years.
The referral history is itself a diagnostic finding
There is a chart pattern I have come to treat as a soft signal in its own right. Anxiety diagnosed in the late teens or early twenties, often around a transition — leaving home, starting college, the first job with no external structure. An SSRI, which helps some. A second SSRI when the first plateaus. A depression diagnosis a few years later, usually after a period of genuine functional collapse. Therapy that the patient describes as useful for insight and useless for follow-through. Sometimes a bipolar II consideration, prompted by the mood lability, that never quite fit and never quite got withdrawn. And then, twelve or fifteen years in, a sister's diagnosis or a child's evaluation, and the first question anyone has ever asked her about attention.
This history is not evidence that the earlier diagnoses were wrong. Anxiety and depression are genuinely common alongside ADHD, and treating them was not a mistake. What the history tells you is something narrower and more useful: that the sequence of care has been reactive to distress and has never once gone looking for a developmental cause. Nobody asked about childhood because the presenting complaint was mood, and mood questionnaires do not have a childhood section.
The chart doesn't say ADHD because nobody asked. It says anxiety because somebody asked about anxiety.
So I ask about the treatment response with some care. Not did the medication help, which invites a polite yes, but what specifically changed and what specifically didn't. The answer is often precise and clinically loud: the dread lifted, the crying stopped, and she still could not start the report, still could not find her keys, still lay awake at one in the morning writing tomorrow's list because she did not trust herself to remember it. A partial response that spares the executive complaints entirely is worth more than any screening score.
Socialization corrupts the childhood evidence the diagnosis depends on
Here is the part that makes this population genuinely hard rather than merely overlooked. DSM-5 requires several symptoms present before age twelve. That requirement is sound — ADHD is a neurodevelopmental condition and the childhood history is what separates it from acquired attentional problems. But the requirement assumes the childhood record is a reasonably honest transcript of the child's neurology, and for girls it often isn't.
Girls are held to expectations of tidiness, preparation, and social attunement earlier and more consistently than boys are. A girl who forgets her homework is corrected sooner, corrected more often, and corrected in a register that makes the forgetting a moral failure rather than a logistical one. The predictable result is that she starts compensating young — while the symptoms are still developing, before she has any language for them, and long before anyone would think to evaluate her. She copies the assignment book twice. She befriends the organized girl and shadows her. She gets the grades by staying up until two.
By the time a teacher describes her, the description is accurate and diagnostically useless: quiet, dreamy, pleasant, disorganized but sweet, not living up to her potential. That last phrase appears in these histories with a regularity that has stopped surprising me. It is the sound of an adult noticing a gap between capability and output and assigning it to character. No one was disrupted, so no one investigated. The mother, interviewed twenty-five years later, remembers a good kid who cried a lot about school.
What this means practically is that the childhood inquiry cannot be a symptom checklist run against parental memory. It has to reconstruct effort. I ask what homework actually looked like from the outside — how long it took, who sat with her, how many times the fight happened. I ask about the tearful Sunday nights, the abandoned instruments and sports, the bedroom that was either immaculate or archaeological with nothing in between. I ask what the report cards said in the comments section, which is where the teachers who noticed something usually put it. The coping architecture these patients build is a large subject of its own — it deserves separate treatment, and the probes for high-functioning adults exist precisely because standard examples fail against it — but in women the architecture goes up early enough that it contaminates the developmental evidence, and that is a distinct problem from the one it poses in adulthood.
Hormonal context, stated carefully
Women in my practice describe symptom severity that moves with the cycle, and they describe it unprompted and consistently. The pattern they report most often is a worsening in the luteal phase — attention, emotional regulation, and word-finding all noticeably harder for several days, then a lifting. Many describe perimenopause as the period when their compensations stopped working, which is frequently what brings them in at forty-five having managed at forty.
I want to be careful about what that is and isn't. It is a clinical observation, made in an unblinded setting by a clinician who is listening for it, and the research literature on hormonal modulation of ADHD symptoms is thinner than the confidence with which it is sometimes discussed online. I do not use it to make a diagnosis. I use it as history: a cyclical pattern superimposed on a stable lifelong baseline is consistent with modulation of an underlying condition, whereas symptoms that exist only premenstrually and vanish otherwise point somewhere else entirely. Asking the question also tells the patient that her experience is a legitimate object of clinical attention, which matters in a population that has usually been told some version of everyone feels that way.
What the interview has to do differently
The adaptation is not a separate module. It is a change in what the questions are aimed at. Instead of probing for disruption, probe for effort and internal experience.
I ask about infrastructure, because the infrastructure is the symptom. What do the calendars look like, and how many are there? Which alarms are set, and for what? Is there a system, and what happened the last three times it collapsed? Where is the list — on paper, on the phone, on the hand? How late did she stay up to finish the thing she had two weeks to do, and how long has that been the only way things get finished? A woman who cannot describe a single unmanaged day in twenty years is not telling me she doesn't have ADHD. She is telling me the cost of not having it visible.
Then I ask what it costs. The recovery time after a functional week. Whether she drives home from work in silence because she cannot tolerate one more input. What the inside of a meeting feels like when she is nodding. Whether she has ever, once, done something the easy way. These questions surface impairment that no observer-based criterion will ever pick up, and they surface it in the patient's own words, which is the only form of it I trust.
A composite
A composite of many evaluations, no one patient: a thirty-nine-year-old attorney, referred after her son's ADHD diagnosis, with fifteen years of treated anxiety and two depressive episodes. She arrives early with a printed timeline of her own history, which she apologizes for. She meets every deadline at work and has never once started a brief before the night it is due; she bills the hours reading the same paragraph repeatedly. Her childhood, per her mother, was unremarkable — an A student, quiet, a crier. The fourth-grade comment was such a bright girl, if she could only focus. The SSRI stopped her panic and did nothing for the two a.m. lists. She was, in her own phrase, never a problem for anyone, which is exactly why it took thirty-nine years.
Nothing in that presentation matches the picture the field was built around, and nothing in it is subtle once you are looking for effort instead of disruption.
The LODHI-A was built around this problem, among others — the childhood section asks for narrative rather than recognition, and the functional inquiry asks what things cost rather than whether they fail. It has not undergone formal psychometric validation; studies are underway, and I would rather say that plainly than let a clinician assume otherwise. It is free to licensed clinicians. If the methodology is what interests you, the FAQ covers it directly; if the instrument is, you can request access and use it in your next evaluation.
Neuropsychiatrist · Bay Area Neuropsychiatry · Developer of the LODHI-A