以24小時飲食回憶驗證懷孕婦女飲食頻率問卷之效度
No Thumbnail Available
Date
2015
Authors
Journal Title
Journal ISSN
Volume Title
Publisher
Abstract
本研究於台灣北中南三區特定醫院婦產科門診收案，採用樣本為245人。於懷孕期間以面談、電訪和郵寄等方式收集問卷，使用孕期所收集之5份24小時飲食回憶(24-hour dietary recalls, 24HDR)作為基準，檢測常用食譜為選項的餐次飲食頻率問卷(meal-based food frequency questionnaire, FFQ)在孕期婦女營養攝取之效度(validity)，並探討此FFQ在國內孕期婦女營養評估之可行性。師大食品營養素資料庫第二版(National Taiwan Normal University Food and Nutrition Management System 2，NUFOOD.2)進行運算，再用SPSS 19.0進行分析。
結果顯示，除了油鹽糖醬類無法比較外，兩方法的主要熱量食物分類一致，依序為：全榖根莖類、豆魚肉蛋類、湯飲奶類、水果類、蔬菜類、零食點心類。FFQ相對效度檢測，熱量及營養素除了維生素A和維生素B12外，FFQ攝取量皆高於24HDR。皮爾森相關係數從膽固醇0.13到鈣0.22，營養密度從飽和脂肪酸0.13到膳食纖維0.23，24HDR熱量校正FFQ後從飽和脂肪酸0.13到蛋白質0.73，大多呈顯著正相關。合併為3個餐次後，24HDR和FFQ的餐次熱量分布為:早餐23%和20%，午餐38%和40%，晚餐39%和40%；醣類分布最為一致，早餐21%、午餐40%、晚餐39%。餐次分布皮爾森相關係數從晚餐熱量0.13到早餐脂肪0.26，餐次營養素密度從午餐脂肪和晚餐熱量0.13到早餐脂肪0.31，餐次熱量調整從晚餐熱量0.13到午餐熱量0.42，其中24HDR熱量校正營養素為良好調整方法。Bland-Altman一致性分析熱量和營養素之散布點多落在可接受範圍內(limits of agreement, LOA= 95% confidence interval)，FFQ在較高及較低攝取量與24HDR差異較大，而營養素攝取百分比則較為固定。
經24HDR熱量校正FFQ調整後兩方法相比，熱量與營養素的皮爾森積差相關71%達顯著正相關，平均相關係數為0.27；淨相關74%達顯著正相關，平均相關係數為0.28；斯皮爾曼等級相關68%達顯著正相關，平均相關係數為0.30。餐次分布的皮爾森積差相關67%達顯著正相關，平均相關係數為0.20；淨相關67%達顯著正相關，平均相關係數為0.20；斯皮爾曼等級相關67%達顯著正相關，平均相關係數為0.21。FFQ雖然會高估大多營養素的攝取，但經24HDR熱量校正後之相對效度為佳，營養素所占熱量百分比亦和24HDR相似，故此FFQ能做為評估台灣孕期婦女營養攝取之有效工具。
關鍵詞：
飲食頻率問卷(food frequency questionnaire)、24小時飲食回憶(24-hour dietary recalls)、效度(validity)、懷孕婦女(pregnant woman)、營養攝取(nutritional intake)、餐次(meal times)
This study recruited 245 pregnant women from the obstetrics and gynecology clinic at hospitals in northern, middle and southern Taiwan and used the face-to-face interviews, telephone interviews and mail questionnaires to collect dietary data during pregnancy. We used the average of 5 24-hours dietary recalls (24HDR) as standard to assess validity of meal-based food frequency questionnaire (FFQ), and investigated the feasibility of the FFQ for assessing nutrition status during pregnancy in Taiwan. The database system entitled “National Taiwan Normal University Food and Nutrition Management System 2 (NUFOOD.2)” was used for data computing, and the SPSS 19.0 for data analysis. The results showed that the main source of foods groups classification in energy are the similar for the two methods, as following food groups: whole grains and roots, soybean-fish-meat-egg, soup-drinks-milk, fruits, vegetables, snacks and desserts except for oil, salt, sugar and sauce groups which are not comparable. Relative validity analysis of FFQ showed that most energy and nutrients intakes from FFQ are higher than 24HDR except for vitamin A and vitamin B12. Pearson correlation coefficients were from 0.13 for calcium to 0.22 for cholesterol, nutrient density from 0.13 for saturated fatty acids to 0.23 for dietary fiber, 24HDR energy adjusted (calibrated FFQ) from 0.13 for saturated fatty acids to 0.73 for protein, and most nutrients were significantly positive correlated. Combining into three meal times, energy distributions of 3 meals of 24HDR and FFQ are as follows: 23% and 20% for breakfast, 38% and 40% for lunch, 39% and 40% for dinner, respectively. The distributions of carbohydrate intakes were most consistent than other nutrients by 3 meal times, both methods were 21% for breakfast, 40% for lunch, and 39% for dinner. Pearson correlation coefficients of meal distributions were from 0.13 for energy of dinner to 0.26 for fat of breakfast. The correlation coefficients for nutrient densities were 0.13 for fat of lunch fat and energy of dinner to 0.31 for fat of breakfast. The correlation confficients for 24HDR energy adjusted (calibrated FFQ) were from 0.13 for energy of dinner to 0.42 for energy of lunch. The 24HDR energy adjusted (calibrated FFQ) suggested to be a better energy adjusted method. By using the Bland-Altman method to analyze the consistency of dietary intakes, most scattered points were within the acceptable range (limits of agreement, LOA = 95% confidence interval). Differences between FFQ and 24HDR were larger in the higher and lower dietary intake samples, but the percentages of energy for the related nutrients intakes were relatively fixed. Compared to the calibrated FFQ with 24HDR, the results showed that the 71% significantly positive correlations with the average Pearson correlation coefficient of 0.27. The results from the Spearman rank correlations showed 68% significant positive correlations with the average correlation coefficient of 0.30. In addition, the results showed the 74% significantly positive correlations by the partial correlation analyses with the average correlation coefficient of 0.28. For meals distributions, the results showed that 67% significantly positive correlation with the average Pearson correlation coefficient of 0.20. The results showed 67% significantly positive correlations by partial correlations with the average correlation coefficient of 0.20. The results showed 67% significant positive correlations by Spearman rank correlations with the average correlation coefficient of 0.21. Although FFQ tends to overestimate most nutrient intakes, the energy adjusted by 24HDR (calibrated FFQ) showed good relative validity, and the percentages of energy from protein, fat and carbohydrate were similar to the 24HDR. Therefore FFQ is an effective tool for assessing dietary intakes for pregnant women in Taiwan. Key words: food frequency questionnaire, 24-hour dietary recalls, validity, pregnant woman, nutritional intake, meal times
This study recruited 245 pregnant women from the obstetrics and gynecology clinic at hospitals in northern, middle and southern Taiwan and used the face-to-face interviews, telephone interviews and mail questionnaires to collect dietary data during pregnancy. We used the average of 5 24-hours dietary recalls (24HDR) as standard to assess validity of meal-based food frequency questionnaire (FFQ), and investigated the feasibility of the FFQ for assessing nutrition status during pregnancy in Taiwan. The database system entitled “National Taiwan Normal University Food and Nutrition Management System 2 (NUFOOD.2)” was used for data computing, and the SPSS 19.0 for data analysis. The results showed that the main source of foods groups classification in energy are the similar for the two methods, as following food groups: whole grains and roots, soybean-fish-meat-egg, soup-drinks-milk, fruits, vegetables, snacks and desserts except for oil, salt, sugar and sauce groups which are not comparable. Relative validity analysis of FFQ showed that most energy and nutrients intakes from FFQ are higher than 24HDR except for vitamin A and vitamin B12. Pearson correlation coefficients were from 0.13 for calcium to 0.22 for cholesterol, nutrient density from 0.13 for saturated fatty acids to 0.23 for dietary fiber, 24HDR energy adjusted (calibrated FFQ) from 0.13 for saturated fatty acids to 0.73 for protein, and most nutrients were significantly positive correlated. Combining into three meal times, energy distributions of 3 meals of 24HDR and FFQ are as follows: 23% and 20% for breakfast, 38% and 40% for lunch, 39% and 40% for dinner, respectively. The distributions of carbohydrate intakes were most consistent than other nutrients by 3 meal times, both methods were 21% for breakfast, 40% for lunch, and 39% for dinner. Pearson correlation coefficients of meal distributions were from 0.13 for energy of dinner to 0.26 for fat of breakfast. The correlation coefficients for nutrient densities were 0.13 for fat of lunch fat and energy of dinner to 0.31 for fat of breakfast. The correlation confficients for 24HDR energy adjusted (calibrated FFQ) were from 0.13 for energy of dinner to 0.42 for energy of lunch. The 24HDR energy adjusted (calibrated FFQ) suggested to be a better energy adjusted method. By using the Bland-Altman method to analyze the consistency of dietary intakes, most scattered points were within the acceptable range (limits of agreement, LOA = 95% confidence interval). Differences between FFQ and 24HDR were larger in the higher and lower dietary intake samples, but the percentages of energy for the related nutrients intakes were relatively fixed. Compared to the calibrated FFQ with 24HDR, the results showed that the 71% significantly positive correlations with the average Pearson correlation coefficient of 0.27. The results from the Spearman rank correlations showed 68% significant positive correlations with the average correlation coefficient of 0.30. In addition, the results showed the 74% significantly positive correlations by the partial correlation analyses with the average correlation coefficient of 0.28. For meals distributions, the results showed that 67% significantly positive correlation with the average Pearson correlation coefficient of 0.20. The results showed 67% significantly positive correlations by partial correlations with the average correlation coefficient of 0.20. The results showed 67% significant positive correlations by Spearman rank correlations with the average correlation coefficient of 0.21. Although FFQ tends to overestimate most nutrient intakes, the energy adjusted by 24HDR (calibrated FFQ) showed good relative validity, and the percentages of energy from protein, fat and carbohydrate were similar to the 24HDR. Therefore FFQ is an effective tool for assessing dietary intakes for pregnant women in Taiwan. Key words: food frequency questionnaire, 24-hour dietary recalls, validity, pregnant woman, nutritional intake, meal times
Description
Keywords
飲食頻率問卷, 24小時飲食回憶, 效度, 懷孕婦女, 營養攝取, 餐次, food frequency questionnaire, 24-hour dietary recalls, validity, pregnant woman, nutritional intake, meal times