Abstract
Access to unhealthy commodities is a key factor determining consumption, and therefore influences the prevalence of non-communicable diseases. Recently, there has been an increase in the availability of food 'on-demand' via meal delivery apps (MDAs). However, the public health and equity impacts of this shift are not yet well understood. This study focused on three MDAs in New Zealand and aimed to answer (1) what is the health profile of the foods being offered on-demand, (2) how many food outlets are available and does this differ by physical access or neighbourhood demographics and (3) does the health profile of foods offered differ by physical access or neighbourhood demographics? A dataset was created by sampling a set of street addresses across a range of demographic variables, and recording the menu items and number of available outlets offered to each address. Machine learning was utilised to evaluate the healthiness of menu items, and we examined if healthiness and the number of available outlets varied by neighbourhood demographics. Over 75% of menu items offered by all MDAs were unhealthy and approximately 30% of all menu items across the three MDAs scored at the lowest level of healthiness. Statistically significant differences by demographics were identified in one of the three MDAs in this study, which suggested that the proportion of unhealthy foods offered was highest in areas with the greatest socioeconomic deprivation and those with a higher proportion of Maori population. Policy and regulatory approaches need to adapt to this novel mode of access to unhealthy foods, to mitigate public health consequences and the effects on population groups already more vulnerable to non-communicable diseases.