Abstract
Diet monitoring has been linked to improved eating habits and positive health outcomes. However, traditional methods place a high burden on both participants and researchers through pen and paper recordings, and manual nutrient coding respectively. The digitalization of dietary monitoring has greatly reduced these barriers. This study aims to develop and test the usability of a diet application with novel food recognition system and assess diet quality among adults in a public university setting. The study was conducted in four phases. The first three phases focused on gathering information, planning and determining features, and development of MyDietCam. Phase 4a involved the participation of 31 individuals, who engaged in real-world food recognition testing, evaluating the AI’s accuracy. In Phase 4b, 65 participants from another local university utilised MyDietCam for three days to assess the usability of the app while their diet quality scores were calculated using the Malaysian Healthy Eating Index (M-HEI). The final iteration of MyDietCam incorporates three primary components: food logging, goal setting, and nutrient analyses. The AI revealed a 72% recognition accuracy in real world scenarios. Usability testing yielded a score of 5.15 out of a possible 7, signifying acceptable usability. However, the M-HEI scores indicated poor diet quality, with an average score of 42.54. The development of MyDietCam demonstrated its usability and potential for facilitating dietary monitoring. Future work should focus on addressing the
issues raised before testing the effectiveness of prolonged app usage on improving diet quality and other health outcomes.
Links
BibTeX (Download)
@mastersthesis{mmedsc-2023-kong,
title = {Development and Usability Testing of a Mobile Diet Application and Assessment of Diet Quality among Malaysian Adults.},
author = {Nadine Kong},
year = {2023},
date = {2023-12-20},
school = {Department of Social and Preventive Medicine, Faculty of Medicine, Universiti Malaya},
abstract = {Diet monitoring has been linked to improved eating habits and positive health outcomes. However, traditional methods place a high burden on both participants and researchers through pen and paper recordings, and manual nutrient coding respectively. The digitalization of dietary monitoring has greatly reduced these barriers. This study aims to develop and test the usability of a diet application with novel food recognition system and assess diet quality among adults in a public university setting. The study was conducted in four phases. The first three phases focused on gathering information, planning and determining features, and development of MyDietCam. Phase 4a involved the participation of 31 individuals, who engaged in real-world food recognition testing, evaluating the AI’s accuracy. In Phase 4b, 65 participants from another local university utilised MyDietCam for three days to assess the usability of the app while their diet quality scores were calculated using the Malaysian Healthy Eating Index (M-HEI). The final iteration of MyDietCam incorporates three primary components: food logging, goal setting, and nutrient analyses. The AI revealed a 72% recognition accuracy in real world scenarios. Usability testing yielded a score of 5.15 out of a possible 7, signifying acceptable usability. However, the M-HEI scores indicated poor diet quality, with an average score of 42.54. The development of MyDietCam demonstrated its usability and potential for facilitating dietary monitoring. Future work should focus on addressing the
issues raised before testing the effectiveness of prolonged app usage on improving diet quality and other health outcomes.},
type = {masterthesis},
note = {Master of Medical Science (Research) Dissertation},
keywords = {},
pubstate = {published},
tppubtype = {mastersthesis}
}
