New OA paper (with @JeffBrunstrom @corferesearch @LizWilliamsRNut @Dr_A_Johnstone @EmmaJStevenson and Ant Watson) exploring protein consumption/purchasing in ageing adults using survey data and supermarket loyalty card data https://nutritionj.biomedcentral.com/articles/10.1186/s12937-020-00602-3 - a thread of what we found

We were interested in supporting healthy ageing. Nutrition is a key part of this, especially in offsetting the effects of sarcopenia (the natural decline of muscle mass and strength with age) which remains a key driver of old age frailty and mortality risk.

Protein consumption declines with older age groups in men, but not women. Older adults may benefit from higher protein intake to minimise the impact of sarcopenia and this may represent a key area to ensure healthy ageing.

Protein purchasing is lower compared to consumption - triangulating the reasons behind this is tricky due to the descriptive nature of our data.

People consume most of their protein in evening meals (and a lesser extent lunch). The low levels of protein consumption in breakfast, as well as snacking, may be important opportunities to intervene/design products to increase overall consumption.

Sources of protein consumption and purchasing were similar, but not consistent across age groups. Meat, fish and poultry, followed by diary and bakery (bread) goods were largest sources, but increasing importance of tinned, packaged and frozen goods with increasing age.

We also looked at the proportion of age groups meeting recommended daily protein levels - while the majority of adults meet the RNI 0.75g protein per kg body mass, few older adults meet higher recommendations (>1.0g/kg) which they might benefit from the most.

Our study has several novelties: it is the most detailed national study on protein-related behaviours, using rarely accessible supermarket loyalty card data to supplement traditional surveys. We need both sources to reveal the nuances of population patterns.

There are several limitations - it is descriptive and the lack of longitudinal linked data limits the quality of conclusions. As a first step to identify opportunities to inform intervention development, it represents a useful application of #BigData for studying nutrition

Thank you to @BBSRC for funding our work, as well as everyone who has supported or discussed our work @geodatascience @livunigeog @CDRC_UK