A roadmap for the development and validation of a quality scoring method for herbal tea – honeybush tea as case study

· Honeybush industry currently lacks a generic sensory quality grading system. · Beneficial for Protected Designation of Origin (PDO) certification. · Comprehensive baseline data was used to identify key quality parameters. · Industry input essential for the development of a valid quality grading system. · Statistical validation is essential when developing a quality scoring method. A scientifically founded quality scoring method was developed to evaluate honeybush infusions based on their sensory quality. A systematic approach was followed: a) Establishment of a comprehensive sensory and physicochemical dataset; b) Determination of the pre-requisites of a quality grading system for honeybush tea through industry consultation, followed by scorecard development by a group of experts and c) Scorecard validation by an industry panel. Appearance, aroma and palate attributes were identified as key parameters. Four-point ordinal scales were used to facilitate scoring of these parameters. Score values (0 to 3) were allocated to the ordinal scales and weights (%) were assigned to the respective parameter categories to obtain a total score (%). The scorecard also included checkboxes for citation of ‘dry leaf appearance’, ‘infusion haze’ and the presence of specific positive and negative attributes. Based on the total score (%), ‘high’, ‘moderate’, ‘low’ and ‘poor’ sensory quality classes were defined. The validity of the scorecard was tested to determine whether an industry panel could accurately categorise batches of commercially processed honeybush tea ( n = 18) as ‘poor,’ ‘low,’ ‘moderate,’ or ‘high’ quality. Prior to validation, each of the 18 validation samples was pre-assigned one of the quality classes based on the findings of descriptive sensory analysis (DSA). The discrimination capacity of the quality scoring method was evaluated by comparing the DSA results with the scorecard results. The findings showed that while the industry panel was able to distinguish between ‘moderate-high’ and ‘poor-low’ quality groupings, they had less success distinguishing between ‘low’ and ‘poor’ quality samples, especially when it came to characterising the ‘low’ and ‘poor’ quality samples in terms of sensory attributes. The need for industry assessor training in parameter and scale recognition and calibration to facilitate quality classification using the scorecard, was identified.