Background: Public interest in herbal medicines has generated an increasing number of trials evaluating their efficacy. Trials with poor methodologic quality have exaggerated estimates of treatment effect, and incomplete reporting of trials causes difficulties in assessing trial methodologic quality. The objective of this project was to examine the quality of reporting of randomized controlled intervention trials of herbal medicine.
Methods: MEDLINE (1966 to September 2003) was searched for randomized controlled trials of 10 herbal medicines. Two individuals (J. G. and J. D.) independently assessed trials using the Consolidated Standard of Reporting Trials checklist. Disagreements were resolved by consensus. The mean number of checklist items reported across all and for individual herbal medicines was calculated. The influence of decade of publication and species of herbal medicine tested was explored using an analysis of variance.
Results: A total of 206 randomized controlled trials of herbal medicine were included. Interrater reliability on reporting quality assessment was high. A total of 45% of items were reported across all trials. The quality of reporting improved across decades from the 1970s to the 2000s. Individual herbal species differed in the total number of items reported, with echinacea, ginkgo, St. John's wort, and kava trials reporting the most items.
Conclusions: Important methodologic components of randomized controlled trials of herbal medicines are incompletely reported including allocation concealment, method used to generate the allocation sequence, and whether an intention-to-treat analysis was used. Also, key information unique to these trials may be missing, such as percentage of active constituents and type or form of the herbal medicine preparation. We suggest trialists consult a recent extension of the Consolidated Standard of Reporting Trials statement specific to herbal medicine trials when designing and reporting randomized controlled intervention trials of herbal medicines.