What if ADHD medication doesn’t work?

What “not working” usually turns out to mean

The phrase collapses at least five different problems into one sentence, and telling them apart is most of the work.

The dose. NICE, the National Institute for Health and Care Excellence, which sets clinical guidance for the NHS in England and Wales, asks prescribers to be familiar with the pharmacokinetic profiles of all the short- and long-acting preparations, to tailor treatment to the individual, and to take account of variations in bioavailability between preparations so as to avoid reduced effect. A starting dose suits some people and not others. If a trial stayed at a starting dose, it may not have reached that person’s effective dose, and whether it did is something to check with the prescriber.

The timing. The same guidance suggests thinking about modified-release once-daily preparations for convenience, adherence and reduced stigma. Cover that runs out mid-afternoon can read as “it doesn’t work” when the issue is when it works rather than whether.

Tolerability rather than efficacy. These are separate problems that arrive wearing the same words. Stopping because something felt unbearable is not the same as stopping because nothing happened, and they lead to different next steps.

Adherence. ADHD symptoms make treatment plans hard to keep, which NICE names directly, down to remembering to order and collect a prescription. It also asks clinicians to check that adherence problems aren’t down to misconceptions about the medication.

The target. Some of what people want from medication was never what it was measured on. That one gets its own section below.

What the guidelines say to do next

Formal guidance is jurisdictional, and it’s worth knowing how unevenly it’s distributed before reading the sequence below.

If you’re in the US, there isn’t one to read. APSARD, the professional society writing the first American adult ADHD guidelines, says on its own site that there are currently no guidelines in the United States. It announced on 27 August 2026 that its guidelines had reached a final review stage, with publication anticipated in the autumn. Until then, what happens after a first medication disappoints depends on your prescriber’s judgment and your insurer, not on a document you can look up.

The sequence below comes from NICE guideline NG87, which applies in England and Wales, published in March 2018 and last updated in September 2019. It’s quoted because it’s specific and public, not because it governs you. Drug names and their order differ between countries. The underlying logic travels better than the specifics: confirm the dose and the duration, change drug family, try a non-stimulant, then escalate to someone more specialised.

NICE recommends lisdexamfetamine or methylphenidate as first-line pharmacological treatment for adults with ADHD. From there:

Switch families before giving up on stimulants. NICE recommends considering a switch to lisdexamfetamine for adults who’ve had a six-week trial of methylphenidate at an adequate dose without enough benefit in terms of reduced symptoms and associated impairment, and considering a switch to methylphenidate for adults in the same position on lisdexamfetamine. The two aren’t interchangeable, and the guidance treats them as separate attempts.

A different effect profile, not a different drug. For adults whose symptoms respond to lisdexamfetamine but who can’t tolerate the longer effect profile, NICE recommends considering dexamfetamine. That’s a case where the medication is working and the shape of the day around it isn’t.

Then a non-stimulant. NICE recommends offering atomoxetine to adults who can’t tolerate lisdexamfetamine or methylphenidate, or whose symptoms haven’t responded to separate six-week trials of both, having considered alternative preparations and adequate doses.

Then a second opinion. NICE recommends obtaining a second opinion or referring to a tertiary service when ADHD symptoms are unresponsive to one or more stimulants and one non-stimulant.

That’s several distinct next steps before the guidance reaches specialist review. A first prescription that disappoints is near the start of them.

How often the first prescription isn’t the last

Two studies give a sense of the scale, and both need reading carefully, because neither measures what the question asks.

Biederman and colleagues followed 86 consecutively referred, previously unmedicated adults meeting ADHD criteria in the DSM-5, the American Psychiatric Association’s diagnostic manual, starting stimulant treatment at a US clinic. Twenty-four percent, 21 of the 86, needed to switch from their initially prescribed stimulant family within 60 days. The reason recorded was poor tolerability, not lack of effect. So this figure describes how often the first stimulant proved hard to take, not how often it failed to work. Switching was more common among those started on methylphenidate, while the need for a change of formulation or an added antianxiety or antidepressant medication was more common among those on amphetamine.

Brancati and colleagues followed 150 adults with ADHD naturalistically in Italy for at least four months. Of those, 58 (38.7%) were lost to follow-up and 75 completed, on average after about 26 weeks; 40 were treated with methylphenidate and 35 with atomoxetine. Treatments were moderately effective overall, and 37 patients, 49.3% of the completers, were responders on a threshold of at least a 30% reduction in observer-rated ADHD severity. That 49.3% is among people who stayed, in a study where nearly two in five did not, so it isn’t a response rate for everyone who starts.

Read together, and with those limits in mind, they describe a process with a lot of adjustment in it rather than a single pass-or-fail test. Neither is large, and neither followed anyone for long.

What medication was and wasn’t measured on

This is the part that changes how people read their own result.

A 2025 adult component network meta-analysis in The Lancet Psychiatry covered 113 randomised controlled trials and 14,887 adults. For reduction of core ADHD symptoms at around 12 weeks, stimulants and atomoxetine were the only interventions supported by both self-reported and clinician-reported ratings. Stimulants showed a standardised mean difference of -0.39 on self-report and -0.61 on clinician report against placebo; atomoxetine showed -0.38 and -0.51. Confidence in those estimates was rated between very low and moderate.

One finding from the same paper is uncontroversial: evidence in the longer term is underinvestigated, since the primary efficacy timepoint was closest to 12 weeks and what happens after that is much less well studied.

The quality-of-life evidence is less settled than the symptom evidence, and two syntheses from overlapping teams disagree.

Ostinelli and colleagues’ adult-only analysis found medications were not efficacious on additional relevant outcomes such as quality of life. A separate 2025 systematic review and meta-analysis by Bellato and colleagues, drawn from a different dataset, reached the opposite conclusion: across 17 randomised controlled trials and 5,388 participants, amphetamines (Hedges’ g 0.51), methylphenidate (0.38) and atomoxetine (0.30) were all significantly more efficacious than placebo at improving quality of life, with what the authors called a moderate effect size. Its trials covered people aged 6 and over, and the abstract reports the child-versus-adult comparison for atomoxetine only, so this is an all-age estimate whose applicability to adults is directly evidenced for the non-stimulant alone.

Different eligibility rules, populations and methods produced different answers, and Cortese is an author on both, so this isn’t rival camps talking past each other. The defensible conclusion is that quality-of-life effects are less consistent than symptom effects and, where detected, smaller. Not that medication can’t improve life.

So a medication can be doing what the symptom trials measured and still leave your week feeling much as it did. If what hasn’t changed is the load, the job, the sleep or the sense of being behind, that’s worth naming as its own problem rather than reading straight off as a failed trial.

There’s a second reason symptom gains may not become life gains, and it’s been measured directly. Jacobsson and colleagues followed 246 adults with ADHD (66% women, mean age 33.5) through routine stimulant treatment in Sweden, with repeated assessments and, by the authors’ own account, substantial attrition between waves. Higher personality dysfunction was associated with less improvement in functioning and more persistent ADHD symptoms, independent of stimulant dose and time in treatment, and the authors describe medication-related effects as modest by comparison. It’s a naturalistic study without a control group, so it identifies a source of variation rather than establishing a cause. What it suggests is that how much your functioning improves may depend on a good deal more than the drug.

NICE builds the same distinction into its guidance for adults, recommending non-pharmacological treatment in combination with medication for adults who’ve benefited from medication but whose symptoms still cause significant impairment in at least one domain.

When it’s a reason to revisit the diagnosis

The guideline puts this check before treatment rather than after, which is worth knowing if it was skipped. Before starting medication, it asks for a full assessment including a review to confirm the person continues to meet the criteria for ADHD and needs treatment, alongside a review of mental health and social circumstances and any coexisting conditions.

Coexisting conditions don’t change the medication choice on their own. NICE recommends offering the same medication choices to people with ADHD and anxiety disorder, tic disorder or autism as to anyone else. What they can change is what you’re measuring the medication against, since something else may be producing part of what you’re hoping it will lift.

A poor response is a reason to raise the question with a clinician, not a reason to conclude anything yourself. Plenty of people whose ADHD is correctly identified respond poorly to the first thing they’re given, which is why the guideline has a sequence in it at all.

What to take to the appointment

NICE asks for ADHD medication to be reviewed at least once a year, covering how well the treatment is working throughout the day, adverse effects, whether medication has been optimised, the impact on work or study, the effect of missed doses and planned reductions, and interactions with any other conditions. Those headings make a usable list to prepare against.

Worth writing down before you go: what specifically you hoped would change, what did change and at what times of day, what got worse, what dose you’re on and whether it’s ever been raised, how many doses you take in a typical week, and what else is going on that might be competing for the explanation.

The most useful thing you can bring is a description precise enough to be acted on. “It isn’t working” gives a prescriber almost nothing to move; “it lifts for about three hours and then I’m worse than baseline” points at a preparation, and “I can start things now but I still can’t finish them” points somewhere else entirely.

You’re also allowed to change your mind in either direction. NICE asks clinicians to reassure people they can revisit decisions about treatments, and to encourage people to raise any preference to stop or change medication and be involved in those decisions. A trial that disappointed you is information about one medication at one dose, not a verdict on the whole category.