Friday 23 June 2017

Automatic academic: working myself out of a job

Guest post by Emma Foster, Lecturer in Public Health Nutrition, Human Nutrition Research Centre, Newcastle University

Since I started working in dietary research I’ve been fascinated by how and why people misreport their dietary intake. Lots of excellent research (by others) has gone into understanding how the hassle of recording food intake, problems with memory and attention (if you are busy doing something else at the same time you may not be paying attention to what you are eating) along with social-desirability bias (am I really going to admit to the nutritionist interviewing me how many doughnuts I ate yesterday!) together tend to result in an under-estimate of energy intake and an over-estimate of those foods seen to be “healthy”.

Much of my research has focused on how we can make it less of a burden and perhaps even an enjoyable experience for volunteers taking part in nutrition research studies. I developed food photographs for portion size estimation with children, so participants don’t need to weigh everything their child eats….and more importantly doesn’t eat but ends up wearing!

Food photographs estimate portion size with children, so participants don’t need to weigh everything their child eats (or ends up wearing!)


More recently I’ve been developing an online 24-hr recall system, which sometimes feels like I’m making myself and other nutrition researchers surplus to requirements! In the “olden days”, when I first joined the Human Nutrition Research Centre at Newcastle University, all dietary data was collected by a researcher who went out to people’s homes to interview them about their dietary intake (something I really quite miss). This was followed by day after day sitting at a computer linking the foods and drinks reported to food composition data and weights (which I don’t miss quite as much!). Now with the online recall we are able to collect the data remotely. We send people a URL and login details and the computer system does the rest. It takes them through the previous day, asking for details on foods consumed, getting people to estimate portion size using photographs and checking for forgotten items like butter on toast or sugar in tea. The system automatically does the linking to the food composition data and the weights consumed and the researcher can download the data as soon as the volunteer has submitted their recall.

More beans please. A screenshot from INTAKE24


But surely it doesn’t do as good a job as a highly skilled nutrition researcher such as myself….right? Well it’s not actually that far off! When compared with a traditional face-to-face interview with 180 people the system was found to underestimate energy intake by just 1% on average and average intakes of protein, carbohydrate, fat, vitamins and minerals were all within 4% of the interviewer-led recall. Perhaps most amazingly people reported enjoying completing the system!

So if you would like to measure food intake as part of your research but can’t afford to employ a nutritionist/dietitian as part of your research team (we’re not cheap) then have a look at our demo on https://intake24.co.uk/demo and drop us an email at support@intake24.co.uk and we can set you up a survey straight away – and it really is free.

No comments:

Post a Comment