Wednesday 17 October 2012


Switch off the junk science, not the TV

The recent claim that too much television is bad for children is just another policy agenda dressed up as science.

At a recent Birmingham Salon event, ‘Trust Me: I’m a Scientist’, two University of Birmingham academics - Stuart Derbyshire, reader in psychology, and Joe McCleery, lecturer in developmental neuroscience - made some important points about the use and abuse of science to promote particular policy initiatives. Scientists today are under immense pressure to make discoveries ‘relevant’, and there are ample incentives – not least in terms of career progression - to ‘big up’ results.

I had spent the day of the Salon reading an Archives of Disease in Childhood journal article by Aric Sigman that was widely reported as showing that ‘doctors should curb amount of time children spend watching television to prevent long-term harm’. (See, for example, the Guardian and the Daily Mail.) According to the journal’s press release, ‘urgent action’ is needed to curb children’s daily screen time. We need to delay ‘the age at which [children] start the world’s favourite pastime’ in order to ‘stave off the risk of serious health and developmental problems’.

This journal article is a perfect example of scientists, researchers and clinicians making claims that go way beyond what the data actually shows. As usual, these claims are followed up with proposed policy initiatives designed to interfere in the minutiae of how we live our lives.

We are hectored by the government about the daily minimum amount of minutes we should exercise, maximum units of alcohol we should drink and minimum portions of fruit and veg we should consume. Now ‘in the face of mounting evidence, doctors’ leaders and government should take a stand and set clear guidelines on an activity that has so far eluded the scrutiny that other health issues attract’, we are told by Sigman.

The drive towards more and more ‘evidence-based’ meddling in how we live our lives is not only a UK phenomenon. As Sigman points out, the US Department of Health and Human Services now cites reducing [screen time] as one of its key ‘health improvement priorities’ in achieving its ‘national 10-year health promotion and disease prevention objective’. The Canadian Paediatric Society advises that: ‘No Child should be allowed to have a television, computer or video game equipment in his or her bedroom.’

So is there clear-cut evidence of deleterious effects on our health and well-being of screen-based technologies? And if governments were to succeed in reducing screen time, would we see an improvement in the health of the nation? I found no convincing evidence for either of these claims in Sigman’s paper.

When I remind undergraduate students that ‘correlation does not equal causation’, I usually get the don’t-you-think-we-know-that look. Yet, Sigman claims in his review article that there are proven ‘effects’ of television viewing on everything from ‘obesity’, ‘type 2 diabetes’, ‘cardiovascular disease’, ‘coronary heart disease’, ‘all-cause mortality’, ‘psychological difficulties’, ‘negative psycho-social well-being’, ‘attention problems’, ‘victimisation in middle childhood’, ‘peer-rejection experiences’, ‘lower empathy’, ‘sleep disorders’ and more – all on the basis of correlational data. That is, the studies reviewed showed associations not causal relationships.

It seems more likely that the causation is the other way around: that people whose health is failing are likely to watch more television, rather than television causing deleterious health effects. Equally, could it not be the case that those who suffer from sleep disorders or are rejected by their peers are likely to watch more television, rather than television viewing leading to peer rejection or poor sleep? Or maybe the association found is due to other, unknown factors? The truth is: we don’t know.

Take the example of attention difficulties. Sigman writes: ‘A longitudinal study of 2,623 children reported that children who watched television at ages one year and three years had a significantly increased risk of developing attentional problems by the time they were seven years old.’

This American Academy of Pediatrics (AAP) study, appearing in the journal Pediatrics, is widely interpreted as showing that, for every hour of television watched at one and three years of age, the children had almost a 10 per cent higher chance of developing attention-deficit hyperactivity disorder (ADHD) by the age of seven. But the study did not show that too much television at an early age increases the chance of developing ADHD. For a start, the children weren’t formally diagnosed with ADHD. Instead, their parents were simply asked whether their child had difficulty concentrating, was easily confused, was impulsive, had trouble with obsessions or was restless. And, most important, just because the children who were judged by their parents and teachers as having problems concentrating also watched more television, it doesn’t mean that television viewing was the reason they had problems paying attention. It could instead be that they already had ADHD, or just had difficulty concentrating, and got easily distracted by other activities, and therefore ended up watching more television.

Indeed, the results of a study by Ignacio Acevedo-Polakovich and colleagues at the University of Kentucky were ‘consistent with arguments suggesting that children who have difficulty paying attention may favour television and other electronic media to a greater extent.’

The alleged links between screen time and body fatness are equally tenuous. A meta-analysis – that is, a statistical technique for reviewing a large number of existing research studies – published in the International Journal of Obesity did find a statistically significant relationship between television viewing and body fatness for children between three and 18 years of age, but one that was ‘too small to be of substantial clinical relevance’.

Sigman does not rely solely on correlational research. He does cite a randomised controlled clinical intervention (that is, a study where the participants are randomly allocated to an ‘intervention’ and a ‘control’ group) where one group of four- to seven-year-olds had TV viewing reduced by half and the other group did not. According to Sigman: ‘After three years, there had been a significant reduction in the BMI of those who had halved their screen viewing, and relatively little in those who had not.’

Well, no. The results actually showed a rather moderate reduction in BMI (-0.24) for the intervention group and for the control group (-0.12). As McCleery pointed out during the Birmingham Salon debate, one mistake often made by researchers is to conflate statistically significance (which shows there is a low probability that differences are due to chance) and clinical significance (or in other words, a difference that may be meaningful).

The participants in the randomised controlled clinical trial referred to by Sigman started with an average BMI of 19.3. Therefore, we can work out that the BMI for the intervention group fell by 1.25 per cent and for the control group by 0.75 per cent. Surely if a programme that succeeds in reducing TV viewing by half only leads to a fall in BMI of less than one per cent for the intervention group relative to the control group, it indicates that TV viewing has a minimal effect on BMI?

Sigman does draw attention to dramatic changes in children’s lives over recent decades: ‘In Britain today, children by the age of 10 years have regular access to an average of five different screens at home. In addition to the main family television, for example, many very young children have their own bedroom TV along with portable handheld computer-game consoles (eg, Nintendo, Playstation, Xbox), smartphone with games, internet and video, a family computer and a laptop and/or a tablet computer (eg, iPad).’

Certainly, times have changed. I grew up in Norway in the Sixties and Seventies when there was only one channel available for six hours a day. My siblings and I were in awe of British people, who had access to three channels that could be watched at any time of the day. Now, of course, most people have access to dozens if not hundreds of TV channels, plus all the other media Sigman lists.

The media clearly play a more central part in young people’s lives than ever before. But is this good or bad? Of course, it is entirely legitimate for scientists and clinicians to ask what these changes mean for children and young people, and whether there is evidence of electronic media adversely affecting health and well-being. But a serious investigation would also look into the possible positive effects. We should also be asking what benefits children may get from engaging with new technology, and try to understand why these media are so enticing.

Sigman concedes: ‘The study of [screen time] as a public health subject is relatively new, and the associations between [screen time] and health risk cited in the observational studies do not prove direct causation.’ Yet he concludes that ‘a robust initiative to encourage a reduction in daily recreational [screen time] could lead to significant improvements in child health and development’.

This is another blatant example, in my view, of flawed ‘science’ being used to tell us what to do and how to bring up our children. Science cannot tell us how to live our lives. We make those decisions based on moral and political considerations. But science can inform those moral and political choices. That means scientists have a duty to analyse and report scientific data and information methodically and without bias – not manipulate statistics to promote a particular moral message.

As Derbyshire asked at the Salon event: ‘What happens to moral and political decision making and to science when scientists stray from providing information to telling us what to do?’ We need to consider what effect it will have on science – in the long term – if scientists, researchers and clinicians hide behind ‘science’ to continually hector us about how we live our lives.

First published by spiked