The UK will ban children under 16 from social media by early next year, replicating a policy that came into effect in Australia in 2025. The move is intended to put children’s well-being ahead of technology companies’ profits. But how will scientists study the effects of the measure and determine whether it is actually having a positive impact?

“We have no evidence either way,” says at Bath Spa University in the UK, who is working with the Australian government to analyse the effects of its own ban and is also advising the UK government. “It’s such new territory.”

Leading the way on research in the UK is the Wellcome Trust, which is already deep into the IRL Trial in Bradford. This has recruited about 4000 kids from 10 schools, aged 12 to 15, who installed an app on their phones to track their social media use. For half of those children, the app also limits access to social apps. The researchers expect to publish their first results in the middle of next year, after the newly announced ban has taken effect. However, at the Wellcome Trust says the findings should still help inform policy.

They will also improve upon existing research that tends to rely on self-reported measures, like asking children or their parents how much time they spend on social media, rather than using more objective metrics. What’s more, the few interventional studies that do exist . “If you’re talking about big changes here around mental health and those sorts of things, you’re not going to see changes in two weeks,” says Etchells.

Australia’s ban came into place in December, which is too recent to gauge its long-term effects. But once national bans like these are in place, it is impossible to do controlled studies, in which two groups of similar people are allocated access or not. What’s more, the results from larger studies on the wider population before and after the ban will be impossible to unpick from other social impacts.

For now, Sebastian is gearing up to run additional studies that hopefully generate at least some results soon after the ban. The UK government expects to bring legislation to Parliament before Christmas, with the policy coming into force in early 2027.

Free newsletter

Sign up to Health Check

Sign up to newsletter

The Wellcome Trust has invited 14 research teams to submit plans, some or all of which will be funded, into the effects of social media on young people’s well-being. These will take varying approaches to capture as much data as possible, with the goal of eventually synthesising the data into a conclusive result.

Some of these studies will follow existing cohorts and regularly interview them about their mental health and well-being over time, before and after the ban. Sebastian says these approaches can be insightful, but rigid. Other studies are proposing momentary assessments, where participants are sent text messages asking them to complete short surveys on the spur of the moment, capturing a different kind of data. Others may look to analyse data that is already being captured for some insight, such as the rate of hospital admissions or school absences.

With time against them, Sebastian hopes that some results could emerge relatively soon after the ban, but they are likely to be nuanced. For instance, a social media ban could have some positive impacts, but also disruptive ones in the short term, as online supportive networks are lost.

The effects of such bans are also likely to change over the years or decades, as today’s children and younger teenagers approach adulthood having never had access to social media. “It’s not that this is a done deal,” says Sebastian. “Policies could be continuing to change over the longer term, and it’s possible that the findings from our study and others will help to shape those policies iteratively.”

The truth about social media and screen time’s impact on young people

There are many scary claims about excess time on digital devices for children and teenagers. Here’s a guide to the real risks - and what to do about them

For now, some are wary that the UK government is taking a reactive stance in the complex problem of poor youth mental health, without the appropriate data. at the University of Oxford told the Science Media Centre that a blanket age ban is a “blunt tool” and a stronger step than current evidence can support, but adds that the Wellcome research is an opportunity to learn whether these measures will “help, harm or neither”.

One thing that could hinder research – and undermine government policy – is the ability of users to skirt the bans. Early reports suggest that facial-recognition technology designed to verify ages online can be , and VPNs make it trivial to appear to websites as a user from another country where age checks aren’t mandated.

Surveys in Australia by the Molly Rose Foundation, a suicide-prevention charity, found that 61 per cent of 12-to-15-year-olds who had accounts on restricted platforms before the ban came into force . The organisation said that given the findings, it would be a “high-stakes gamble” for the UK to follow suit at this stage.

Need a listening ear? UK Samaritans: 116123 (samaritans.org); US Suicide & Crisis Lifeline: 988 (988lifeline.org). Visit bit.ly/SuicideHelplines for services in other countries.

There is a saying in the halls of British political power that scientific advisers should be “on tap, not on top”. This pithy aphorism, often attributed to Winston Churchill, highlights that, in a democracy, science should guide – not dictate – policy-makers’ decisions.

Never was this truer than at the height of the covid-19 pandemic. Despite politicians in the UK claiming to “follow the science”, there were many decisions – from paying people to self-isolate to closing schools – that couldn’t be made on scientific advice alone. What’s more, some questions couldn’t immediately be answered by scientists. Policy-makers were forced to muddle through.

By contrast, the Trump administration is now turning off the scientific taps, as government health bodies overturn long-established guidelines on everything from vaccines to cell phone radiation, all in the name of the “Make America Healthy Again” movement.

By mid-2027, we should have much stronger evidence on social media harms

But what about situations where the science is still evolving, and we don’t face a global emergency? The question, then, centres on how long policy-makers should wait for scientific results to become clear.

One of the biggest debates in many countries at the moment is whether to ban under-16s from using social media, as Australia did at the end of last year. Proposed bans are extremely popular with the public, but the best available scientific evidence shows that, on a population level, the impact of social media on teenagers’ mental health is minimal. Should politicians ignore the evidence and go with the majority?

Doing so would be in keeping with Churchill’s maxim. But as we report here, by mid-2027, we should have much stronger evidence on social media harms, both from a randomised trial being conducted in the UK and the natural experiment afforded by the Australian ban. As such, the only sensible course is to wait for scientists to come up with the goods before charging ahead with policy. To coin a new phrase, science should be on tap, not on top – and given time.

Antidepressants vary greatly in the side effects they cause – from weight gain to weight loss and blood pressure changes. This may need to be taken into account when doctors decide which of the many drug options, which can be lifesaving, would be best for each individual.

About 1 in 10 people in and take antidepressants for conditions such as depression and anxiety. Withdrawal symptoms may be less common than we thought, but while you are taking these drugs. It just isn’t clear which specific drug causes which kind of side effect.

Drilling deeper into the data, at King’s College London and his colleagues analysed results from 151 randomised-controlled trials and 17 US Food and Drug Administration reports. These covered 30 antidepressants whose effectiveness and side effects were compared against placebos over around an eight-week course of treatment for various psychiatric conditions.

“The beauty of these studies is that they were all placebo-controlled and randomised, so we can be confident that the differences that we are showing are due to the drug,” says Pillinger.

The researchers found that some of the drugs were linked with weight loss and others with weight gain. Maprotiline, for example, was linked to an average weight gain of about 1.8 kilograms, whereas agomelatine was linked to losing about 2.4 kg.

The study also showed that the drug fluvoxamine was linked to heart rate slowing by 8 beats per minute (bpm), on average, but nortriptyline increased it by 13 bpm. The latter also cut systolic blood pressure – your blood pressure when your heart is contracting – by between 3 and 7 millimetres of mercury (mmHg), while doxepin pushed it up by almost 5 mmHG, which could have knock-on effects on overall health. “Each 1 mmHg increase in your blood pressure, if you have high blood pressure, increases your risk of a stroke by 1 per cent,” says team member , also at King’s College London.

Free newsletter

Sign up to Health Check

Sign up to newsletter

Other drugs, such as paroxetine, duloxetine, desvenlafaxine and venlafaxine, were associated with increases in cholesterol. Whether these side effects worsen, improve or plateau beyond eight weeks is unclear.

Despite observational studies raising concerns that antidepressants affect levels of electrolytes, such as sodium, in blood, which could impact heart rhythms, “our study showed that actually there weren’t any problems at all with sodium”, says Pillinger.

Of the last year, about 85 per cent were either . Four of these were found to have a relatively good profile for most of the examined factors, but amitriptyline was seen to increase weight, heart rate and blood pressure, while venlafaxine raised heart rate, blood pressure and cholesterol.

“This is very informative for both physicians and patients,” says at epidemiology consulting group Epilytics in Vancouver, Canada. “This data can be put into an app, so, for example, if somebody has high cholesterol or cardiovascular problems, the app can tell the physician which one to avoid. This will give physicians more bandwidth to choose the antidepressant that best fits a patient.”

Why the causes of poor mental health may share a common root

The neat picture we once had about the causes of mental illness has turned out to be wrong, but we are building an understanding of a new single underlying factor

Pillinger says his team is developing a digital tool with a drop-down menu of an antidepressant’s common side effects, so medics can gauge which to avoid, or even which to choose, for example if weight loss would be beneficial. “It’s all about finding the right molecule for the right person and engaging in personalised prescribing,” says Pillinger.

But at Stanford University in California says the results may be biased if the team only included studies that showed side effects. Others may have not found side effects, which they didn’t report, meaning this data wasn’t included in the meta-analysis.

He also points out that people generally take antidepressants for longer than 8 weeks; sometimes for years. “Very different, long-term studies would be needed to show the presence and relevance of such outcome effects, if any, in the long-term,” says Ioannidis.

In response, Pillinger says there is “always the risk of bias” but “we can be relatively confident that we have gathered most of the data that is out there”.

Chronic inflammation messes with your mind. Here’s how to calm it

From depression to dementia, we are now realising the profound impacts of long-term inflammation on the brain. A better understanding of the underlying mechanisms is unlocking new treatments to protect our cognitive function and mental health

He accepts an analysis of longer-term studies is needed, which he says his team is working on, and points out that about 75 per cent of the participants in the existing trials were white. But he thinks the patterns the team spotted will show up in longer-term studies and will probably apply to other groups.

Previous studies suggest that some antidepressants, such as venlafaxine, , he says. Weight gain that occurs quickly can also last, he says. “Once you put the weight on, it’s hard to get rid of it.”

People on antidepressants who are worried about side effects shouldn’t stop their treatment, though, says Howes, because these medications can be very effective. “This is nuanced; it’s an individual decision, reflecting an individual’s risk profile and the benefits. If they are concerned, this is something that we would recommend that they talk to their doctor about.”