Despite increasing attention on addressing socioeconomic disparities in diet quality, longitudinal studies are scarce. Furthermore, the effects of the COVID-19 pandemic on diet-related outcomes are yet to be fully understood. We examined changes in diet quality by educational level among adults in Madrid, Spain. We used data from recruitment (in 2017) and from 2021. At baseline, our sample included 1358 adults aged 40-75 years who were free of cardiovascular disease and completed a validated diet quality screener. Of them, 931 answered the survey in the follow-up visit in 2021. We used participants' diet quality index scores (range: 18-54; higher scores indicate better diet quality) as the dependent variable. As our independent variable, we assessed participants' educational levels (low, medium, and high). We fitted a multinomial regression using the categories of educational level as the main predictor, adjusting for age, sex, country of origin, and household composition. During the study period, 78.0% of participants sustained their diet quality, 11.6% improved it, and 10.4% moved away from a healthier dietary pattern. In descriptive analyses, we observed an increase in diet quality among less-educated females. Unadjusted multinomial models showed that a lower educational level predicted both increases and decreases in diet quality over the period. Even though the median diet quality scores did not change significantly, we observed heterogeneous changes over the four years. Variability within diet, with some improving and some worsening, seems to have increased among participants with lower educational levels. Future studies should look at the determinants of change in these population subgroups.
Keywords: Spain; adults; cohort; diet; education; health inequalities; longitudinal study.