Remote food photography for the real-time measurement of children's food intake

描述(申请人提供)：儿童和青少年中超重和肥胖的流行率正在上升。令人担忧的是，超重和肥胖儿童在成年前正经历与体重相关的健康问题，例如2型糖尿病。体重增加和肥胖是正能量平衡的结果，能量平衡是摄入的能量多于消耗的能量，而能量摄入是体重增加的重要预测因素。儿童和青少年吃的食物类型也会影响健康。儿童和青少年每天消耗超过500千卡的低营养、高能量食物，即使在能量过度消耗的情况下，也可能发生营养不足。例如，肥胖儿童缺铁的比率更高。需要准确的方法来衡量儿童和青少年的食物摄入量，以促进促进健康饮食和体重管理的研究和干预。现有的方法，如自我报告(例如，食物记录)，与重大参与者和较小儿童的照顾者负担有关。此外，这些方法对食物摄入量的估计不准确(低)，特别是对体重较高的人和青少年。在使用这些方法时，一个很大的误差来源是由于参与者对部分大小的估计不佳。因此，有必要开发和验证方法来估计儿童的食物摄入量，即不依赖参与者准确估计份量的能力的青少年的食物摄入量。我们小组开发了远程食物摄影方法(RFPM)来近乎实时地测量成年人的食物摄入量，有效性数据是肯定的。与黄金标准、双标签水相比，RFPM低估了自由生活条件下7天的能量摄入量不到7%。当使用RFPM时，参与者使用智能手机拍摄他们选择的食物和盘子浪费的照片。这些照片会立即通过无线网络发送给研究人员。参与者在智能手机上收到符合生态瞬时评估(EMA)方法的自动提示(电子邮件)，提醒他们为自己的食物拍照。开发了计算机应用程序来管理EMA方法，并使用计算机成像算法自动识别食物和估计食物摄入量。数据被快速高效地收集和分析，从而实现了近乎实时的食物摄入量测量。在拟议的研究中，RFPM将被修改为测量青少年的食物摄入量。然后将RFPM的有效性与纸笔食品记录的有效性进行比较。在对RFPM进行修改和对4名青少年进行初步测试后，RFPM和食物记录的信度和效度将在青少年(12-18岁)的样本中进行测试。在参与者之间的平衡顺序中，RFPM将在一周内使用，而食物记录将在另一周内使用。用这两种方法估计的食物摄入量将与三个黄金标准进行比较：1)在自由生活条件下通过双标签水(DLW)测量的能量摄入量，2)在学校食堂午餐时测量的食物摄入量(能量和营养摄入量)，以及3)在基于实验室的测试餐中测量的食物摄入量。还将比较不同方法(RFPM和食品记录)之间的用户满意度和负担，成本效益分析将确定一种方法是否以及在什么条件下比另一种方法更具成本效益。我们还将测试RFPM的有效性是否与能量摄入水平有关。

DESCRIPTION (provided by applicant): The prevalence of overweight and obesity among children and adolescents is increasing. Alarmingly, overweight and obese children are experiencing weight-related health problems before adulthood, such as type 2 diabetes mellitus. Weight gain and obesity are the consequence of a positive energy balance, which occurs when more energy is ingested than expended, and energy intake is an important predictor of weight gain. The types of foods that children and adolescents eat also affect health. Children and adolescents consume over 500 kcal/day of low-nutrient, energy dense foods, and nutritional deficiency can occur even in the presence of overconsumption of energy. For example, obese children have higher rates of iron deficiency. Accurate methods to measure the food intake of children and adolescents are needed to facilitate research and interventions to promote healthy eating and weight management. Existing methods, such as self- report (e.g., food records), are associated with significant participant and, with younger children, caregiver burden. Moreover, these methods produce inaccurate (low) estimates of food intake, particularly for people of higher body mass and adolescents. A large source of error when using these methods is due to participants' poor estimation of portion size. Hence, there is a need to develop and validate methods to estimate the food intake of children, namely adolescents that do not rely on the participant's ability to accurately estimate portion size. Our group developed the Remote Food Photography Method (RFPM) for measuring the food intake of adults in near real-time, and the validity data are positive. The RFPM underestimated energy intake by less than 7% over 7 days in free-living conditions when compared to the gold standard, doubly labeled water. When utilizing the RFPM, participants use a Smartphone to take pictures of their food selection and plate waste. These pictures are immediately sent to the researchers via a wireless network. Participants receive automated prompts (emails) on the Smartphone that are consistent with Ecological Momentary Assessment (EMA) methodology reminding them to take pictures of their foods. Computer applications were developed to manage EMA methods and to automatically identify foods and estimate food intake using computer imaging algorithms. Data are collected and analyzed quickly and efficiently, allowing near real-time food intake measurement. During the proposed research, the RFPM will be modified to measure the food intake of adolescents. The validity of the RFPM will then be compared to the validity of pen-and-paper food records. Following modification of the RFPM and pilot testing with 4 adolescents, the reliability and validity of the RFPM and food records will then be tested in a sample of 64 adolescents (age 12-18 years). In a balanced order across participants, the RFPM will be used during one week, and food records will be used during another week. Food intake estimated with both methods will be compared to three gold standards: 1) energy intake measured by doubly labeled water (DLW) in free-living conditions, 2) food intake (energy and nutrient intake) measured during lunch in a school cafeteria, and 3) food intake measured during a laboratory-based test meal. User satisfaction and burden also will be compared between methods (RFPM and food records), and a cost- effectiveness analysis will determine if and under what conditions one method is more cost-effective than the other. We will also test if the RFPM's validity is associated with levels of energy intake.