DEVELOPMENT OF PREDICTIVE RISK SCORE FOR QUALITY ASSESSMENT OF IMPORTED VEGETABLES AND RELATED PRODUCTS BY IMPORT AND EXPORT INSPECTION DIVISION

Abstract

Imported food surveillance plays an important role in consumer protection to reduce the potential risks from substandard products and their impact on consumer health. Developing predictive risk score can serve as an effective screening tool for unsafe products. This cross-sectional analytical study aimed to develop predictive risk score, using data from 2,468 imported vegetables and related products sampled under the Food and Drug Administration’s sampling plan during fiscal years 2019–2021. Data were recorded and categorized by relevant variables. Chi-square tests were used to identify factors significantly associated with the quality analysis results, followed by logistic regression analysis. Beta coefficients were used to assign weighted scores and develop predictive risk score for predicting the quality results of products. The results showed that factors significantly affecting the product quality (p < 0.05) include the sampling quarter, group of sampling port, food type, and the importer’s history of non-compliance. The total risk score was 7 points: 1 point for the quarter, and 2 points each for the port, food type, and importer’s non-compliance history. The Youden’s index showed the highest cutoff point at 3.5. The developed predictive risk score was moderately effective, with sensitivity of 66.67%, specificity of 82.21%, accuracy of 80.79%, positive predictive value of 27.32%, and negative predictive value of 96.09%. This predictive risk score can be used as a guideline for prioritizing food sampling that may fail quality standards. However, enhancing the data collection system is recommended to improve the predictive performance and accuracy of the predictive risk score.