Article 5 of 6

Stimulants can substantially reduce ADHD symptoms. They can also suppress appetite, disturb sleep, raise pulse and blood pressure, and make some people feel unlike themselves. Good prescribing starts by accepting both statements and asking whether treatment helps this person live better.

Pemberton is right about one thing that can become strangely difficult to say in arguments about ADHD: the medicines are not trivial.

Methylphenidate and amphetamine-based medicines alter catecholamine signalling. Lisdexamfetamine is converted in the body to dexamfetamine. These medicines can affect appetite, sleep, pulse, blood pressure and mood. They are controlled drugs, can be diverted or misused, and require careful selection, titration and review. A child who feels emotionally flattened should not be told that this is the unavoidable price of behaving properly in class.

That is the strongest form of Pemberton’s concern. It deserves a serious answer.

The answer is not that medication is harmless. Nor is it that adverse effects reveal a project to drug ordinary difference. The clinical question is whether a particular treatment produces worthwhile benefit for a particular person at an acceptable cost, compared with realistic alternatives — including the risks of leaving significant ADHD unsupported.

Do the medicines work?

For short-term reduction of core symptoms, the answer is yes, on average.

A large 2018 network meta-analysis found several ADHD medicines more effective than placebo over about 12 weeks. Among the drugs compared, methylphenidate had the most favourable balance of efficacy and tolerability in children and adolescents, while amphetamines did so in adults.¹ A 2024 systematic review of treatments for children and adolescents similarly found that stimulant and non-stimulant medicines improve ADHD symptoms.²

Those findings should not be turned into a claim of certainty about lifelong outcomes.

The 2023 Cochrane review of methylphenidate in children and adolescents found probable improvement in teacher-rated symptoms and general behaviour, but rated confidence in the evidence very low because trials were often short and at risk of bias.³ Non-serious adverse effects were more common, while evidence about serious harm and quality of life remained limited.

Stimulant trials are difficult to blind because appetite, sleep and subjective effects can reveal treatment allocation. Many trials have industry involvement. Rating scales may improve without telling us enough about learning, friendships, family life, autonomy or how somebody feels about themselves. Randomised evidence extending beyond a year is much thinner than the language of indefinite treatment can imply.

These limitations do not erase a consistent short-term symptom effect. They narrow the claim we should make: medication has comparatively strong evidence for reducing core symptoms in the short term, while evidence about the size and durability of many wider long-term benefits is less complete.

A symptom score is not a life

Reduced distractibility or impulsivity may allow a child to learn, maintain friendships and finish a school day without repeated failure. An adult may drive more safely, complete essential administration or have more time between impulse and action. Those are meaningful outcomes.

Equally, quieter behaviour can conceal a poor trade. If a child is easier to manage but loses appetite, pleasure, spontaneity or a tolerable sense of self, a lower rating-scale score does not decide whether treatment is worthwhile.

A 2025 meta-analysis found that ADHD medication improved quality of life compared with placebo, but the effects were smaller than those usually found for core symptoms.⁴ That is unsurprising. Medication cannot undo years of shame, create organisational systems, repair a hostile workplace or provide a suitable education. It may make some of those changes easier; it cannot substitute for them.

Treatment goals should therefore be concrete and personal. What would improvement look like: safer driving, fewer exclusions, completing invoices, reading independently, less conflict at home, enough appetite to grow, or still feeling recognisably oneself? Review should return to those goals rather than continue a prescription merely because it was once started.

Personality and the authority of experience

Pemberton describes lisdexamfetamine as turning him “down from 10 to about 7.5”.⁵ In a later account, he reported that he could remain at his desk but felt dulled, less curious and less funny; he stopped after three weeks.⁶

That experience should be believed as an account of how the drug felt to him. It does not tell us whether the diagnosis was correct, whether another dose or drug would have been different, or what other people experience. Medication response is not a diagnostic test.

Some patients report emotional blunting, irritability, loss of spontaneity or feeling like a “zombie”. Those are reasons to reassess dose, timing, formulation and treatment — and sometimes to stop. They are not intended therapeutic effects. Other patients describe less mental noise and more ability to choose where attention goes. They feel more, not less, themselves.

Neither testimony cancels the other. The person taking the medicine has authority over whether their resulting state is acceptable; no individual testimony establishes the population effect.

NICE guidance recommends discussing preferences and concerns, including fear of personality change, recording symptoms and adverse effects during titration, and adjusting the dose until symptoms and impairment improve with tolerable harms.⁷ Shared decision-making is not decorative. A person may reasonably decline medication, stop one that feels wrong or accept an adverse effect for a benefit they value.

Mason: treatment for whose benefit?

The documentary’s advance account describes Mason, a ten-year-old who had taken lisdexamfetamine for two years. During a supervised six-week period without it, with several lifestyle changes at the same time, he was reportedly happier but harder to manage at home and struggled more at school.⁸

Article 2 explains why this cannot establish the general effectiveness of a package of screens, diet and activity. The treatment question is different: whose outcomes count, and how should competing outcomes be weighed?

Mason’s own experience matters. So do learning, friendships, family relationships, appetite, sleep and longer-term opportunities. “More himself” cannot simply be overruled by classroom order. Nor can serious educational impairment be dismissed as only the school’s problem. A child may value both wellbeing now and capacities that expand later choices.

The ethical mistake is to make either convenience or a romantic idea of the unmedicated child decisive. Children should be involved at a level appropriate to their development. Families and clinicians should name trade-offs openly, consider whether the dose, drug or school can change, and review the decision over time.

Medication and accommodation are not rivals. A rigid classroom can magnify impairment; adapting it may help. Medication may still give the child more control within that environment. Schools should not offer a tablet instead of support, and children should not be denied an effective treatment until society becomes ideal.

Common harms and safe monitoring

Common adverse effects of stimulants include reduced appetite, weight loss, difficulty sleeping, headache, abdominal symptoms, dry mouth and increases in pulse or blood pressure. Mood changes may occur. Growth can be modestly affected in some children, which is why height and weight are monitored. Rare psychiatric or cardiovascular events require appropriate assessment and a clear route to review.

Diversion and misuse also matter. Long-acting formulations and lisdexamfetamine’s prodrug design may reduce some opportunities for misuse but do not remove them. Prescribers should discuss safe storage, sharing and the risks of taking a controlled drug other than as prescribed.

NICE recommends baseline review of physical health, current medication, height and weight, pulse and blood pressure, and a cardiovascular assessment before treatment. Routine electrocardiography is not required without specified clinical indications. During treatment, pulse and blood pressure should be checked before and after dose changes and every six months; height and weight monitoring varies by age.⁹

Safe prescribing also requires accessible follow-up. A normal baseline assessment is not enough if titration is rushed, adverse effects cannot be reported or responsibility becomes fragmented between a private provider and GP.

Interpreting the cardiovascular evidence

Pemberton has highlighted a Swedish study reporting a four per cent increase in cardiovascular-disease risk for each additional year of medication exposure.¹⁰ The figure is accurate within the authors’ model, but needs its context.

This was an observational nested case-control study, not a randomised trial; it shows association rather than proving causation. The relationship was nonlinear: adjusted odds rose most rapidly during the first three years and then stabilised. Compared with no use, more than five years of exposure was associated with 23 per cent higher adjusted odds of cardiovascular disease. Associations were statistically significant for hypertension and arterial disease, but not for the other cardiovascular categories examined.¹¹

The result is clinically relevant, especially for treatment lasting decades. A 2025 network meta-analysis also found small average changes in blood pressure and pulse with several stimulant and non-stimulant medicines, reinforcing the need for monitoring.¹²

Relative effects do not tell an individual their absolute risk. The implications differ between a healthy teenager and an older adult with hypertension. The appropriate response is baseline risk assessment, regular pulse and blood-pressure checks, attention to smoking and other modifiable risks, investigation of concerning symptoms and periodic reconsideration of whether benefit still justifies exposure.

It is equally important not to use an individual cardiac event to imply causation without adequate clinical evidence. Anecdotes can signal a possible harm. They cannot estimate its probability.

The comparator is not “no risk”

Untreated, impairing ADHD is not a neutral condition. At population level it is associated with injuries, substance misuse, educational and occupational difficulty, self-harm and premature death. Association does not make these outcomes inevitable, and it does not show that medication prevents all of them. They still belong in a balanced decision.

Observational studies suggest treatment may reduce some serious harms. In a 2024 Swedish target-trial emulation of 148,578 people diagnosed with ADHD, medication initiation was associated with lower two-year all-cause mortality: 39.1 rather than 48.1 deaths per 10,000 people, with the difference driven mainly by unnatural causes.¹³ Because treatment was not randomised, unmeasured confounding may contribute; the study does not prove that medication saved those lives.

The cardiovascular and mortality studies point in different directions because treatment can introduce one risk while reducing others through better attention, impulse control, engagement with care or daily functioning. Selectively presenting only one side is poor risk communication.

Disclosure without generalisation

I take lisdexamfetamine and have found it helpful. I experience more space between thought and action, not less personality. That explains why Pemberton’s account does not feel universal to me. It is not evidence that the medicine is right for anybody else.

I accept monitoring and a possible cardiovascular cost for a hoped-for functional benefit. That bargain should remain open to review. If benefit disappeared, harms became unacceptable or my health changed, the decision should change.

My experience and Pemberton’s are equal in one respect: each is an anecdote of one. Population evidence begins beyond both.

Better prescribing

The honest case for ADHD medication is not that every diagnosed person needs it, that it is harmless, or that questions about it are stigmatising.

It is that these medicines produce worthwhile benefits for many people; the evidence for short-term symptom reduction is substantial despite limitations; broader and longer-term outcomes are less certain; and harms require meaningful consent, monitoring and review.

Good prescribing therefore means:

• a careful diagnosis and a clear account of the impairment being treated;
• shared goals based on the patient’s life rather than other people’s convenience;
• baseline physical and cardiovascular assessment;
• cautious titration, taking subjective experience seriously;
• monitoring symptoms, function, appetite, sleep, growth, pulse and blood pressure;
• access to environmental adjustments and non-drug support;
• periodic review of whether medication is still wanted and needed; and
• no shame attached either to taking it or deciding its balance is wrong.

The useful questions are not treating a disorder or drugging difference? They are: what does this person want to change, what options are available, what benefit appeared, what harm appeared, and does the balance still favour continuing?

Stimulants are powerful. That is why they can help, why they can harm and why the person taking them must remain an active author of the decision.

---

Series disclosure: I am a doctor diagnosed with ADHD in adulthood and take lisdexamfetamine. This is context, not evidence; the clinical claims in this series should stand or fall on the sources cited.

Sources and further reading

Series navigation

← Previous: Article 4 — Can you buy an ADHD diagnosis? · Series index · Next: Article 6 — Are we diagnosing too much ADHD? →