PDF(Pubmed)
Abstract：
Nutrition service needs are huge in China. Previous studies indicated that personalized nutrition (PN) interventions were effective. The aim of the present study is to identify the effectiveness and feasibility of a novel PN approach supported by artificial intelligence (AI).
This study is a two-arm parallel, randomized, controlled trial in real world scenario. The participants will be enrolled among who consume lunch at a staff canteen. In Phase I, a total of 170 eligible participants will be assigned to either intervention or control group on 1:1 ratio. The intervention group will be instructed to use the smartphone applet to record their lunches and reach the real-time AI-based information of dish nutrition evaluation and PN evaluation after meal consumption for 3 months. The control group will receive no nutrition information but be asked to record their lunches though the applet. Dietary pattern, body weight or blood pressure optimizing is expected after the intervention. In phase II, the applet will be free to all the diners (about 800) at the study canteen for another one year. Who use the applet at least 2 days per week will be regarded as the intervention group while the others will be the control group. Body metabolism normalization is expected after this period. Generalized linear mixed models will be used to identify the dietary, anthropometric and metabolic changes.
This novel approach will provide real-time AI-based dish nutrition evaluation and PN evaluation after meal consumption in order to assist users with nutrition information to make wise food choice. This study is designed under a real-life scenario which facilitates translating the trial intervention into real-world practice.
This trial has been registered with the Chinese Clinical Trial Registry (ChiCTR2100051771; date registered: 03/10/2021).
This study is a two-arm parallel, randomized, controlled trial in real world scenario. The participants will be enrolled among who consume lunch at a staff canteen. In Phase I, a total of 170 eligible participants will be assigned to either intervention or control group on 1:1 ratio. The intervention group will be instructed to use the smartphone applet to record their lunches and reach the real-time AI-based information of dish nutrition evaluation and PN evaluation after meal consumption for 3 months. The control group will receive no nutrition information but be asked to record their lunches though the applet. Dietary pattern, body weight or blood pressure optimizing is expected after the intervention. In phase II, the applet will be free to all the diners (about 800) at the study canteen for another one year. Who use the applet at least 2 days per week will be regarded as the intervention group while the others will be the control group. Body metabolism normalization is expected after this period. Generalized linear mixed models will be used to identify the dietary, anthropometric and metabolic changes.
This novel approach will provide real-time AI-based dish nutrition evaluation and PN evaluation after meal consumption in order to assist users with nutrition information to make wise food choice. This study is designed under a real-life scenario which facilitates translating the trial intervention into real-world practice.
This trial has been registered with the Chinese Clinical Trial Registry (ChiCTR2100051771; date registered: 03/10/2021).
摘要：
背景：中国的营养服务需求巨大。先前的研究表明，个性化营养（PN）干预是有效的。本研究的目的是确定人工智能（AI）支持的新型PN方法的有效性和可行性。
方法：这项研究是双臂平行的，随机化，真实世界场景中的对照试验。参与者将在员工食堂享用午餐。在第一阶段,总共170名符合条件的参与者将按1:1的比例被分配到干预组或对照组.干预组将被指示使用智能手机小程序记录他们的午餐，并在用餐3个月后达到基于实时AI的菜肴营养评估和PN评估信息。对照组将不会收到营养信息，但会被要求通过小程序记录他们的午餐。膳食模式，干预后，预计体重或血压会优化。在第二阶段,该小程序将免费提供给学习食堂的所有食客（约800人），为期一年。每周至少使用2天的小程序将被视为干预组，而其他人将被视为对照组。预计在此期间后身体代谢正常化。广义线性混合模型将用于识别饮食，人体测量和代谢变化。
结论：这种新颖的方法将提供基于AI的实时菜肴营养评估和餐后PN评估，以帮助具有营养信息的用户做出明智的食物选择。这项研究是在现实生活中设计的，有助于将试验干预转化为现实世界的实践。
背景：该试验已在中国临床试验注册中心注册（ChiCTR2100051771；注册日期：2021年3月10日）。
方法：这项研究是双臂平行的，随机化，真实世界场景中的对照试验。参与者将在员工食堂享用午餐。在第一阶段,总共170名符合条件的参与者将按1:1的比例被分配到干预组或对照组.干预组将被指示使用智能手机小程序记录他们的午餐，并在用餐3个月后达到基于实时AI的菜肴营养评估和PN评估信息。对照组将不会收到营养信息，但会被要求通过小程序记录他们的午餐。膳食模式，干预后，预计体重或血压会优化。在第二阶段,该小程序将免费提供给学习食堂的所有食客（约800人），为期一年。每周至少使用2天的小程序将被视为干预组，而其他人将被视为对照组。预计在此期间后身体代谢正常化。广义线性混合模型将用于识别饮食，人体测量和代谢变化。
结论：这种新颖的方法将提供基于AI的实时菜肴营养评估和餐后PN评估，以帮助具有营养信息的用户做出明智的食物选择。这项研究是在现实生活中设计的，有助于将试验干预转化为现实世界的实践。
背景：该试验已在中国临床试验注册中心注册（ChiCTR2100051771；注册日期：2021年3月10日）。
