Characterizing the Timing of Eating Among Urban and Rural Ghanaian Households Using a Wearable Camera, the Automatic Ingestion Monitor v2 (AIM-2)

Objectives: Eating timing studies have disproportionately been conducted in high income countries. Also, these studies have used participant self-report, which is subject to reporting bias. We aimed to characterize eating and energy intake timing in urban and rural households in Ghana more objectively using the AIM-2, a wearable camera attached to eyeglasses that automatically captures images of consumed food and beverages. Methods: A mother, father, adolescent and/or child ≤5 y from each household wore the AIM-2 during waking hours for 3 days. AIM-2 images were reviewed for start and end times of eating episodes, identification of foods and beverages consumed, and portion sizes. Nutritional analysis was conducted using the West African Food Database, the Food and Nutrient Database for Dietary Studies, and others. Eating episodes were designated meals or snacks based on time of day, context, and items consumed. Calculations of daily eating window (start time of last eating occasion minus start time of first eating occasion); eating episode duration (end time minus the start time of each eating episode); and eating frequency (total number of eating occasions) were performed. Results: AIM-2 data were available for 30 fathers (43.8±10.7 y), 30 mothers (43.4± 9.5 y), 18 adolescents (14.8±2.8 y), and 10 children (3.5±0.7 y) in urban households, and 30 fathers (37.3±12.4 y), 30 mothers (36.8±11.1 y), 7 adolescents (14.4±2.9 y), and 7 children (2.1±1.1 y) in rural households. Snacking was uncommon, with only 33% of household members snacking on one or more of the 3 days. Further, snack energy intake was 105±26 kcal/d higher in urban vs rural households. Meals were 452±21 kcal/d higher in energy, consumed more frequently (0.9±0.6 per d), and were 3.8±0.8 mins longer than snacks in both locations for all household members. Children had longer eating windows (1.4±0.47 h) than fathers and higher eating frequencies (0.73±0.17 per d) than other household members regardless of location. Eating frequency (2.4-2.7 per d) and window (6.5-9.5 h/d), and % who snacked (33%) was lower than reported for Blacks in the US (5 per d, 12 h/d and 90%, respectively). Conclusions: Eating timing in Ghanaians differs by location and household member and from Blacks in the US.