What helps with ADHD? You might immediately think of specific medications or therapies. But as you read this umbrella review by Gosling and colleagues in the BMJ, it quickly becomes clear that this is not a study that ends with one clear answer. Quite the opposite. Its main contribution is to bring structure to a field where much has been said, but a clear overview is often missing.

The design is robust. The authors analyse more than 100 meta-analyses of randomised controlled trials and recalculate them using a consistent method. That matters. It allows them to compare results across very different types of interventions, from medication to behavioural approaches.

The first conclusion is straightforward. Medication works, at least in the short term. It reduces core ADHD symptoms in a clear and consistent way. But—and this matters—those effects come with trade-offs. Patients experience side effects, and we still know far less about the long-term impact.

For non-pharmacological interventions, the picture quickly becomes more complex. Some approaches show promise, including behavioural interventions and certain forms of training. But the evidence behind them is often weaker. Not necessarily because they do not work, but because researchers rely on smaller, more heterogeneous, or methodologically weaker studies.

This is where the review becomes particularly interesting. It draws a sharp distinction between effect and evidence. An intervention may produce a positive effect in studies, but when the underlying evidence base is limited or inconsistent, we should interpret that effect with caution.

That point is not trivial. In discussions about ADHD, people often jump too quickly to “what works,” as if it were a simple yes-or-no question. This review shows it is not. What works depends on how you define “works,” for whom, in which context, and based on what kind of evidence.

Another important nuance: many outcomes extend beyond core symptoms. Some interventions may not strongly affect attention or impulsivity, yet still improve well-being, functioning, or academic performance. Researchers measure these outcomes less consistently, which makes them harder to compare.

So what does this mean in practice?

First, medication remains one of the best-supported options, especially for short-term symptom reduction. But it is no silver bullet. Any decision to use it should weigh the side effects and long-term uncertainties.

Second, we should not dismiss non-pharmacological interventions. Some show real promise, but the evidence base still needs to grow. That calls for cautious interpretation and, better, high-quality research.

And finally, the key takeaway: we need more precision when we talk about “what works.” Not every observed effect rests on equally strong evidence. And weaker evidence does not automatically mean an intervention is ineffective.

In a field as complex as ADHD, that distinction matters.