April 1, 2015

expert reaction to study investigating sedentary time and risk of developing diabetes

A paper published in the journal Diabetologia has examined the effect of sedentary time on diabetes risk. The authors, from the University of Pittsburgh, used observational data from over 3,000 participants and report a link between increased television watching and an increased risk of developing diabetes.

Prof. Kevin McConway, Professor of Applied Statistics, The Open University, said:

“As the press release highlights, the most newsworthy aspect of this research is the reported finding on ‘TV watching time’, but it is very important to understand the limitations of the study in relation to this conclusion.

“Superficially, the conclusion may appear firm because it is based on data from a randomised controlled trial. But the trial was not designed to investigate whether TV watching time had an effect on diabetes. As far as that possible relationship is concerned, the data effectively come from an observational study rather than a randomised trial, and it is always difficult in observational studies to sort out what causes what.

“Here, the researchers certainly observed that people who said they spent more time watching TV were more likely to develop diabetes, though in fact the increase in risk was rather small. Type 2 diabetes – the kind that generally develops in adulthood – is a relatively common (and increasingly common) disease in the UK. The number of new cases annually, per 100,000 people, is about 500. If the results of this study are taken at face value, and everyone in the UK watched an extra hour of TV daily, then the number of new cases per year would go up by about 17 per 100,000.

“But the results shouldn’t be taken at face value, not without rather a lot of questioning. Because it’s observational, the study doesn’t tell us directly whether extra TV watching actually causes the extra cases of diabetes. There might be some confounding factor that causes people to watch more TV and, independently, causes them to be more likely to have diabetes. The researchers allowed for the most obvious possible confounding factors – age, sex and a measure of general physical activity level – in their calculations, but that’s not everything.

“Importantly, when they also allowed for the participants’ weight, the extra risk of diabetes was considerably smaller, and was not statistically significant (so that it might well just be due to chance). The study report argues that this does not necessarily negate their main finding on TV watching. This is basically on the grounds that, perhaps, watching more TV might make people put on weight, and it might be the extra weight that increases the diabetes risk, so that the risk increase is still an indirect effect of sitting in front of the TV. But the adjustments that the researchers made for weight did not only look at changes in weight over the study, but also weight at the beginning. Perhaps people who weighed more throughout the study were less active and watched more TV right from the start, because of their weight, and independently were also more likely to develop diabetes because of their weight, and maybe that is why the adjustment for weight reduces the apparent effect of TV watching on diabetes. We just can’t tell from this study exactly what’s going on here.

“All this really illustrates is how difficult it is, from this sort of study, to make any sense of what causes what. It’s certainly plausible that spending longer in front of the TV does increase your risk of diabetes, but this study certainly hasn’t established that that is true.”

‘The impact of lifestyle intervention on sedentary time in individuals at high risk of diabetes’ by Bonny Rockette-Wagner et al. was published in Diabetologia at 23:01 UK time on Wednesday 1 April 2015

Declared interests

None declared