台灣糧食不安全與個人相關因素探討

Personal determinants of food insecurity in Taiwan

目標：檢視數種糧食不安全經驗量表（Food Insecurity Experience Scale, FIES）的個人化糧食不安全（Food insecurity, FI）分類方法（依序：安全、輕度、中度、重度），選取與FAO建議（FAO-FI）的族群估算結果一致性最佳之切點，據以探討影響台灣FI盛行率之個人因素。方法：使用2019年的台灣國民營養健康調查數據，納入之參與者為2266名16歲及以上且有回答完整FIES（0-8分）的國民。從數個FI計算方式中，選取與FAO-FI結果一致性最高者，進行個人層級分析。數據以SUDDAN加權並校正抽樣的設計效應，以序位羅吉斯迴歸探討FI之危險因子。結果：台灣2019年FAO-FI的「中重度」與「重度」FI盛行率分別為2.29%及0.44%。Ganhao-Arranhado等人的FI分類結果與FAO-FI差異最小。16-40歲、初中、專科或大學肄業、自覺經濟狀況困難，以及有吃檳榔習慣者有顯著較高的FI風險；偶爾抽菸者的風險比無抽菸習慣者低。調整所在縣市後，租屋或其他住屋狀況是FI的危險因子。結論：2019年的台灣糧食安全狀況較2014年進步，國人的FI與眾多的社會人口學及行為變項相關。

Objectives: To assess available food insecurity (FI) classification (ranked as secure, mild, moderate and severe) approaches used to personalize the Food Insecurity Experience Scale (FIES) and select a method most consistent with the population Food and Agriculture Organization (FAO)-FI measure, so as to explore FI status in Taiwan for individuals. Methods: From the 2019 Nutrition and Health Survey in Taiwan (NAHSIT), 2266 participants≥16 years who had completed the FIES (ranges 0-8) were recruited. To choose the best FIES classification fit, FIs derived by various cuts were weighted and adjusted for design effect by SUDAAN and compared with the FAO-FI. The associations between the chosen FI and socio-demographic data were explored further by ordinal logistic regression. Results: In 2019, the“moderate and severe”and“severe”FI prevalences in Taiwan were 2.29% and 0.44%, respectively. The cut-off proposed by Ganhao-Arranhado was most consistent with FAO-FI. The risk for FI was high for 16–40-year-olds; junior high school educated; junior college or university students; where self-perceived economic status was difficult; and where there was any betel nut chewing. However, with occasional smoking there was a lower risk. After adjustment for locality, housing instability became a significant risk factor for FI. Conclusions: Food security status, assessed as FIES in Taiwan, improved from 2014 to 2019. FI was associated with a range of potentially modifiable risk factors.