It’s pretty clear that diet has an impact on health. Exactly how much impact, and which bits of diet, is less clear. I analysed two studies this week for the excellent NHS Choices’ Behind the Headlines feature, which takes apart health studies in the news to see whether they stand up to the hype.
The first was one of those papers that gets the Daily Mail rubbing its hands with glee. ‘Obese people are in denial about the amount of sugar they eat,’ their headline ran. The study was intriguing. It questioned whether people who ate more sugar (as opposed to just more calories generally) were more likely to be overweight. The answer was yes, if you measured people’s sugar consumption by the sugar concentration in their urine. Or no, if you asked people how much sugar they ate, via a food diary.
We don’t actually know whether people were ‘in denial’. The researchers suggested the difference in outcomes was because overweight people under-reported how much sugar they ate. That’s one explanation; another might have been issues with the urine tests, or the food diary design. But it’s all too believable that people simply ‘forgot’ to mention that sneaky biscuit or chocolate bar when they reported their food intake to researchers. Here’s my analysis.
The second study was interesting – a rare randomised controlled trial of dietary guidelines. Most of our evidence about healthy diets comes from observational studies – basically, looking to see what healthy people eat, compared to less healthy people. That’s subject to all sorts of bias which can be avoided by an RCT.
The three-month study showed that people who followed healthy eating guidelines on salt, fat, vegetables, wholegrains and oily fish reduced their blood pressure by a significant amount, enough to cut their chances of having a heart attack or stroke. My only reservation with this study was that I couldn’t find the researchers’ calculations for the widely-reported headline figure of a one-third reduction in risk. I’m sure they were fine, but I like to see them. My analysis is here.