Miguel, José M. Pereira, 1947-Rito, Ana Isabel GomesKadosh, Maria Ana Silva Carvalho Cohen, 1986-2017-07-272017-07-2720162016http://hdl.handle.net/10451/28474Tese de doutoramento, Doenças Metabólicas e Comportamento Alimentar (Ciências e Tecnologias da Saúde) (Nutrição), Universidade de Lisboa, Faculdade de Medicina, 2016Accurate dietary assessment is critical for monitoring the nutritional status of children, examining associations between diet and health, and identifying dietary intake patterns and eating behaviours. The study of diets of children poses methodological problems relating to the accuracy of assessment. The 24-hour Dietary Recall (24-h DR) is among the most accurate methods to estimate total energy and nutrient intakes in school-age children. Traditional 24-h DR is expensive and impractical for large-scale studies. The application of technology to automate the more accurate 24-h DR is being conducted in multiple countries across the world, making them less expensive and easier to use. The objective of this study is to develop, validate and test a new web-based recall for dietary assessment in Portuguese school-age children: the Portuguese self-administered computerised 24-h DR (PAC24). The PAC24 is a self-administered web-based 24-h DR based on multiple pass method directed to second-, third- or fourth-grade Portuguese children (7-10 years old). In PAC24, children are first questioned about food and drink consumption on the previous day. Food entry is done via free text search, supported by a spell check application. For the majority of foods, amount consumed is estimated by selecting the closest portion size, served and leftover, if any, among seven different digital images. Data about time, place, television watching and computer use are assigned to each eating occasion. The food composition code and weight (g or mL) of selected items are automatically allocated and stored. A database of 380 food items is available. Food, energy and nutrient information is linked to a database that contains essentially information about Portuguese food composition table. The development of PAC24 was based on literature review, 21 focus groups (FG) developed in seven primary schools of the seven main regions of Portugal and input from national and international researchers with experience in computer-dietary assessment among children. A prototype of PAC24 was tested and its content was validated through an expert meeting. Accuracy was determined by comparison of PAC24 with lunch observations (‘gold standard’) in two schools in Lisbon and Tagus Valley (LTV) region. A pilot study was conducted to check the feasibility of PAC24 with respect to procedures, methods and data processing in one primary school in LTV region on two non-consecutive days with 15 days apart. The food consumption, particularly total energy and nutrient (protein, total fat, carbohydrate, fibre, calcium, sodium, potassium) intakes were studied. Wilcoxon signed-rank tests were used to compare the food consumption between the first and second measurement occasion. A p-value <0.05 level was considered as statistically significant. Two hundred and four children participated in the FG, where 3959 food items were identified. The main ways in which children report their previous day’s food consumption were revealed: they generally reported foods chronologically organized by the three main meals; snacks and some foods like sweets were only reported after specific prompts; not all children were able to report foods successfully. Finally, different meanings and labelling of some specific food items were identified and taken into account in the development of PAC24. In the usability test (n=12), it was shown that PAC24 is engaging for school-age children and therefore captures and maintains their attention while completing the questionnaire. This study also identified some technical problems that were simplified and modified before the validation study. Comparison of PAC24 against observations (n=41 children) at the food level resulted in 67.0% for matches (foods/drinks reported in PAC24 and by observers), 11.5% for intrusions (foods/drinks reported in PAC24 but not by the observers) and 21.5% for omissions (foods/drinks reported by the observers but not in PAC24). Use of the PAC24 led to underestimates of the weight of food on average by 32% of the actual intake. Children took on average 27 minutes to complete PAC24 on pilot study (n=48 children). Mean energy intake per day was 1799 kcal. Mean protein, carbohydrate and total fat intakes (as a percentage of energy intake) were 17.2%, 47.8% and 29.0%, respectively. Significant differences were not observed on food consumption between the first and second measurement occasion (p>0.05). In summary, PAC24 is a cost-effective, intuitive and an engaging method for Portuguese school-age children. The PAC24 could be used to estimate dietary intake on a group level, accurately. This method will provide useful information for epidemiological studies on the links between diet and health and contribute to the improvement of public health policies at national level.engCriançasAvaliação do consumo alimentarQuestionário às últimas 24 horasTecnologiaValidadeTeses de doutoramento - 2016doctoral thesis101489323