Development and methodological user-validation in industry of a 3D-printed biosensing toolkit for tropane alkaloid detection.

Abstract

The number of biosensors for food contaminants are growing, yet few are tested in real-world settings. Here, we developed and validated a biosensor for detecting tropane alkaloids in buckwheat, emphasizing in-field accuracy and user-experience. Sixty participants from a major food producer tested the biosensor in a pre-structured manner. All quality control employees (n = 31) correctly identified blank or contaminated buckwheat. However, non-analytical-users (n = 29) showed lower accuracy when visually interpreting icLFIA-results (Kappa-score = 0.5/1), which improved substantially when using a digital reader (Kappa-score = 0.9/1). Post-test questionnaires yielded an excellent usability score (88/100 ± 8). A difficulty score, based on how many participants struggled and the corresponding perceived difficulty, allowed identification of challenging workflow steps. Notably, most challenging steps could be resolved by adjusting the instructions. This user-validation represents a pioneering effort in food safety analysis, especially point-of-need detection, to offer quantitative evidence for claims related to stakeholder usability and provides a framework for other innovative (bio)sensing approaches.