Information of adverse events associated with dietary supplements or health food, which is collected by manufacturers or healthcare facilities, is inconsistent. Therefore, a method to collect essential information from patients or consumers and evaluate the causal relationship of adverse events is necessary. We previously modified a dendritic algorithm for evaluating medication-related adverse events (Jones JK. Fam Community Health 1982; 5: 58-67) for use in dietary supplements or health food. In this study, we improved the dendritic algorithm, especially in the temporal relation between taking dietary supplements or health food and onset of adverse events, and compared to a different algorithm based on the scoring scale developed by Naranjo et al (Naranjo CA, et al. Clin Pharmacol Ther 1981; 30: 239-45). Using both algorithms, eight raters (pharmaceutical science students) assessed 200 cases of adverse events provided by the manufacturer's customer inquiry center. The κ coefficient of multi-rater reliability was 0.51 for the modified dendritic algorithm and 0.35 for the scoring scale. The time required to complete the evaluation tended to be shorter using the dendritic algorithm. In conclusion, the present results indicate that the improved dendritic algorithm may be reliable and suitable for universal usage. Pilot studies using the modified algorithm during history taking of consumers or patients to collect information on adverse events are needed to assess the utility of the algorithm in clinical practice.
(Jpn J Clin Pharmacol Ther 2013; 44(5): 405-410)