Work Package 1 is led and managed by Mark Green at University of Liverpool with input from Bernard Corfe and Elizabeth Williams from University of Sheffield. Existing consumer behaviour data will be provided by a well known supermarket. 

The work package will analyse three distinct secondary data sources to profile population-level behaviours surrounding protein intake. It will identify patterns in behaviours by demographic characteristics to identify important differences within the population that will help to inform future interventions. The first data source will be the National Diet and Nutrition Survey (NDNS) (2011-2012) which is a nationally representative survey of dietary behaviours. The NDNS will allow us to explore patterns in overall protein levels, as well as where individuals acquire their main sources of protein from. These measures will be calculated for 5-year age groups and will be stratified by sex and socioeconomic status (SES). We will report summary statistics for these measures, and use regression models to analyse the associations of these factors to our measures of protein intake. We will then use Latent Class Analysis (LCA) to examine the existence of similar ‘types’ of food choices within the population. This will provide a typology of behaviours, which is useful for profiling similarities and differences among individuals. The LCA will be undertaken for individuals aged 40-54, 55-69 and 70+ years separately, and we will use regression analysis to explore how the typologies relate to overall protein intake.

Analyses of NDNS will be complemented by interrogating consumer data on purchasing behaviours and consumer behaviour an attitudes data from Bradgate Bakery. While these data are less representative of national populations, they provide novel information on objective food purchasing behaviours. Consumer data are rarely used throughout the literature due availability issues and therefore, represent a novel aspect of our study. The data will be used to validate the observations from the NDNS on how protein intake varies by age. We will use the same statistical methods when analysing both the Sainsbury’s consumer data and the NDNS data to ensure greater comparability of our results.

In a third and complementary approach we will use protein intake data from two trials with older adults centred in Sheffield (FIT and Nana). In total these yielded 300 four-day food diaries, which offer a more detailed and fine-grained snapshot of a regional population. The data will be analysed using similar approaches to examine the quantity, quality and timing of protein intake in the population not afforded by our other sources. Analyses will be stratified by gender, age and SES.

Outcome and Deliverable: By cross-referencing our complementary analyses we will establish a profile of protein intake that characterises the nature of intake (quality and course), the quantity of intake, the timing of intake (important in targeting novel products) and how these attributes alter with the ageing process, and differ according to gender and SES. Report will be written for lay and scientific publications to support the next project phases