Potential of wearable cameras for dietary assessment in rural and urban mothers in Ghana

Objectives To compare the accuracy of Automatic Ingestion Monitor v2 (AIM-2; a micro-camera worn on eyeglasses), eButton (a micro-camera pinned to shirt or worn with lanyard around neck), and image-assisted 24-h dietary recall (IA24-HR) in measuring dietary intake from same meals, against researcher-conducted weighed food record (rWFR) among rural and urban Ghanaian mothers. Background Dietary assessment in low- and middle-income countries presents greater challenges than in high-income countries. Wearable cameras have the potential to provide a more objective approach to dietary assessment. Methods This was a cross-sectional study. Purposive sampling was used to recruit 60 mothers, but the analytical sample included 47 mothers (22 rural, 25 urban) with complete data. Images from AIM-2 and eButton, and data from IA24-HR and rWFR were used to measure dietary intake on 2 weekdays and 1 weekend day. Nutritional analysis was conducted using software programmed with food and nutrient composition databases. Intake estimates for food consumed weight (g), energy (kcal), carbohydrate (g), protein (g), and fat (g) were calculated for meal-specific comparisons. Agreement was assessed by calculating generalized concordance correlation coefficients (CCC) and performing Bland–Altman analysis. Intake estimates were compared across methods by conducting repeated measures analysis of variance. Results Mean (SD) ages of rural and urban mothers were 35.0 (10.9) and 44.7 (7.4) years, respectively (P < .001). Agreement varied between method pairs. Among rural and urban mothers, intake estimates from AIM-2, eButton, and IA24-HR were consistently and significantly lower than those from rWFR. However, intake estimates from AIM-2 and eButton were mostly closer to those of rWFR than IA24-HR. Conclusions Overall, agreements differed between method pairs. Compared to rWFR, intake estimates from AIM-2 and eButton were more accurate than those from IA24-HR in rural and urban mothers. Among the three test methods, AIM-2 performed best, providing the most accurate dietary intake estimates.