For a field growing this quickly, the research base behind ADHD coaching has stayed remarkably thin. In January, the first systematic survey of the American ADHD coaching workforce appeared in JAMA Network Open. Its authors reached 481 coaches and found a profession that had expanded rapidly and recently, with nearly two thirds of respondents having begun practicing during or after the pandemic. Eighty-nine percent had no professional mental health background. Roughly 63 percent had completed a curriculum endorsed by the ADHD Coaches Organization. The median session cost about 150 dollars, comparable to a session with a licensed therapist, and nearly two thirds of coaches reported receiving referrals from health providers. What that survey could not answer, and did not claim to, was the obvious next question. Does it work?

Two studies published this year in the American Journal of Lifestyle Medicine take the most serious run at that question yet attempted with adults. Both come from Elizabeth Ahmann and Micah Saviet of the Springer Institute, working with the independent statistician Mark Otto. Ahmann is based in the health and wellness coaching program at Notre Dame of Maryland University’s School of Integrative Health, and she and Saviet have spent roughly a decade building the methodological groundwork this research required, including a published manual describing exactly what a 12-session adult ADHD coaching engagement should contain.

The first paper went online in March. The second appeared on August 20, and it is the one that makes the pair genuinely interesting.

What the Studies Actually Did

The design was a single-arm prospective mixed-methods study. Coaches who met defined criteria for training and credentialing recruited clients from their own private practices and screened them for eligibility. Those clients then completed a 12-session individual coaching engagement that followed a published manual rather than each coach’s personal improvisation. That last detail is not a technicality. Manualization is what makes a coaching intervention a thing that can be studied at all, because it lets a researcher say what was delivered, check whether it was delivered as intended, and give another investigator a chance to repeat it.

Outcomes were assessed with validated instruments covering three domains: ADHD symptoms, executive functioning, and functional impairment in daily life. The papers cite the standard measures this literature relies on, including the Barkley Deficits in Executive Functioning Scale and the Weiss Functional Impairment Rating Scale, along with a short form of the Working Alliance Inventory to gauge the coach and client relationship. Both papers sit behind a paywall, so the published abstracts do not confirm which instrument was mapped to which outcome.

From pre-test to post-test, improvements in all three domains were statistically significant, with effect sizes the authors describe as medium to large. Goal attainment results were positive. The working alliance between coaches and clients scored high, which matters given that alliance quality is one of the better established predictors of client outcome across the coaching literature generally.

The Part Nobody Should Have Assumed

Behavior change interventions have a well earned reputation for producing gains that evaporate once the support ends. The support itself is often doing more of the work than the skills it was supposed to install, and when the weekly appointment disappears, so does the structure. This is why follow-up data is worth more than post-test data, and why so few coaching studies bother to collect it.

The August paper collected it. Six weeks after their coaching engagements ended, 42 adults were reassessed. Symptoms, executive functioning, and functional impairment had all improved slightly further rather than sliding back. For executive functioning and functional impairment, that additional change reached statistical significance, though at a small effect size. The authors are careful about this, describing the additional movement as small and framing the real finding as durability rather than continued growth. Their own conclusion is that coaches, clients, and referring clinicians should find the persistence of gains encouraging, and that longer follow-up is warranted.

Six weeks is a short window. It is not a year, and nobody in this literature is claiming otherwise. But it is long enough to catch the fastest and most common form of decay, and the decay did not appear.

What “Executive Function Improved” Actually Means

Here is where readers who work in executive function should slow down, because the phrase is doing more work than it looks like it is doing.

Executive function is measured two very different ways. One is a performance-based task, administered under controlled conditions, that records how quickly and accurately a person inhibits a response or holds information in working memory. The other is a rating scale, on which a person reports how often executive lapses disrupt their actual life: missed deadlines, abandoned projects, the second half of a conversation lost because the first half triggered a tangent.

These are widely treated as two windows onto the same thing. They are not. In a review that has shaped how careful clinicians read this literature, Maggie Toplak, Richard West and Keith Stanovich examined the relationship between the two measure types across 20 studies covering both children and adults. Of 286 relevant correlations, only 68, about a quarter, reached statistical significance. The median correlation was 0.19. Their conclusion was blunt: performance-based measures and rating measures of executive function assess different underlying constructs.

The distinction they drew is useful rather than merely deflating. Performance tasks index processing efficiency, how well the cognitive machinery runs when a person is instructed to run it as hard as they can. Rating scales index goal pursuit, what a person actually does across the loose, unconstrained situations that make up an ordinary week. Both are real. They are simply not the same question.

This is why the coaching finding is worth taking seriously and also why it should not be oversold. Nobody should expect twelve weeks of conversation to change processing efficiency, and no reasonable reading of these studies suggests it did.

That framing fits the dominant clinical account of adult ADHD, in which the central problem is not a deficit of knowledge but a failure to deploy what a person already knows at the moment it is needed. Adults with ADHD can usually describe exactly what they should do. The breakdown happens at the point of performance, where intention has to survive contact with a specific Tuesday afternoon. Interventions that work on this problem tend to work by changing the environment around the person rather than the machinery inside them: external cues, imposed structure, shortened feedback loops, someone expecting an answer on Thursday.

A rating scale is the right instrument to detect that kind of change, because that is precisely the domain rating scales measure. The improvement these studies found is therefore both more modest and more practically meaningful than the phrase “executive function improved” suggests on its own. Less went wrong in these adults’ weeks. That is the claim, and it is a claim worth having.

The Control Group That Was Not There

The single most important thing to understand about both papers is the one the authors themselves put in writing: there was no control group.

A single-arm study measures the same people before and after and reports the difference. It cannot separate the effect of the intervention from everything else that produces movement in a pre-post comparison. People usually seek help when things are at their worst, and things at their worst tend to improve somewhat on their own. Participants knew they were receiving coaching and knew they were being studied, and expectation alone moves self-report. Attention from a warm, attentive professional has measurable effects regardless of what that professional is doing. Some participants were likely receiving medication or therapy concurrently. Twelve weeks of ordinary life produces change for reasons no study is tracking.

None of these alternatives are exotic. Each is a routine explanation for exactly this pattern of results, and a single-arm design cannot rule out any of them. The correct summary is that adults who completed a structured coaching engagement improved, not that the coaching caused the improvement.

The authors add their own limits, which is to their credit. The coaches were selected for training, credentialing, and experience, so results may not generalize to a workforce where, as the JAMA Network Open survey documented, credentials vary enormously. The client sample was fairly homogeneous in gender, race, and education. And because coaches recruited from their own practices, the participants were people who had already chosen coaching, already found a coach, and could already afford one. That is not the general population of adults with ADHD.

What This Changes, and What It Does Not

What it changes is the honest description of the evidence. It is now accurate to say that emerging prospective research reports meaningful improvement in symptoms, executive function, and daily functioning following a structured, manualized adult ADHD coaching engagement, and that those improvements persisted at a six-week follow-up. That is a real step up from where this field stood two years ago, when the adult literature was mostly small pilots, dissertations, and college-student samples.

What it does not change is that ADHD coaching has not been shown effective in a controlled trial. Anyone who reads these two papers as proof of efficacy is reading something the authors did not write and would not endorse. Their own closing note asks for further longitudinal study, which is the request of researchers who know precisely what their design can and cannot support.

For practitioners, the more actionable finding may be the one hiding in the methods rather than the results. What was studied here was not coaching in general. It was a defined engagement of known length, following a published manual, delivered by coaches meeting stated criteria, measured with validated instruments, tracked for goal attainment, and checked for working alliance. Those are the conditions that made the question askable in the first place. A field that wants to be evaluated has to first become the kind of thing that can be evaluated, and that is a choice about how practice is structured long before it is a choice about study design.

The randomized trial will come or it will not. In the meantime, the most useful takeaway for coaches, clients, and the clinicians who refer between them is that structure appears to be doing real work, that its effects show up where rating scales look rather than where laboratory tasks look, and that six weeks after the last session those effects were still there.