Investigating dietary quality among individuals aged 15 years and over by diabetes status in South Africa

Analysis of the 2016 South African Demographic and Health Survey (SADHS) estimated that 11.7% of individuals aged 15+ years had poor glycaemic control, despite only 4.7% reporting a previous diabetes diagnosis, and a further 64.5% had prediabetes. Diet-related lifestyle change has an important role in diabetes prevention and management, however entrenched racial and socioeconomic inequalities and increasing urbanisation may present barriers to a healthy diet. Using data from the 2016 SADHS we investigated whether dietary choices differ by diabetes status defined by previous diagnosis and survey HbA1c, and whether the diet of people living with diabetes (PLWD) differs by age, gender, ethnicity and wealth quintile. Reporting of fruit, vegetable, sugar-sweetened beverage, fruit juice and fast-food consumption was used to create an index of healthy diet. Ordered logistic regression modelling considering the proportional odds assumption was used to investigate the effect of diabetes status and sociodemographic status on healthy diet among the general population and PLWD. Concurrent low consumption of fruit, vegetables and fruit juice was the most common dietary pattern among both the general population and people living with diabetes, with high consumption of fast-food and sugar-sweetened beverage less common. Among the general population, previous diabetes diagnosis, age [≥]55 years, non-black African ethnicity and being in the wealthiest quintile were significantly associated with increased odds of a healthier diet. Among PLWD, there was no association between previous diabetes diagnosis and healthy diet, high wealth remained significantly associated with a healthier diet, whilst female gender and having health insurance also became significantly associated with a healthier diet. Future public health interventions should focus on making fruit and vegetables more accessible to younger, black and socioeconomically poor populations, irrespective of diabetes status.