RESUMEN
OBJECTIVE: Our aim was to analyse the association of change patterns on TV-viewing and computer/tablet use and incidence of elevated consumption of ultra-processed food consumption and lower consumption of fruits and vegetables during the COVID-19 pandemic. DESIGN: Data of 39 208 Brazilian adults from a Behaviour Web Survey were used. Unhealthy nutrition habits were eating fruits or vegetables for <5 d/week and ultra-processed food (sugary foods, snacks, ready-to-eat frozen foods and embedded foods) for ≥5 d/week. For incidence indicators, we only considered participants without unhealthy behaviour before the quarantine. We created four categories of change in TV-viewing and computer/tablet use, considering a cut-off point of 4 h/d for each behaviour (1 - consistently low, 2 - become low during the quarantine, 3 - become high during the quarantine or 4 - consistently high). Analyses were adjusted for sex, age group, highest academic achievement, per capita income, working status during the quarantine, skin colour and adherence to the quarantine. SETTING: Brazil. PARTICIPANTS: Brazilian adults (nationally representative). RESULTS: Logistic regression models revealed that high TV-viewing and computer/tablet use incidence were associated with higher odds for elevated frequency of ultra-processed food consumption (TV-viewing: OR 1·70; 95 % CI 1·37, 2·12; computer/tablet: OR 1·73; 95 % CI 1·31, 2·27) and low consumption of fruit and vegetables (TV-viewing: OR 1·70; 95 % CI 1·29, 2·23; computer/tablet: OR 1·53; 95 % CI 1·08, 2·17) incidence. Consistent high computer/tablet use also presented higher odds for incidence of elevated frequency of ultra-processed food consumption. CONCLUSIONS: Participants with incidence of sedentary behaviours were also more likely to present incidence of unhealthy diet during the COVID-19 pandemic quarantine.