Objectives: The aim of this investigation was to explore the effectiveness of search strategies developed to identify trials of specific complementary therapies in a range of clinical conditions.
Design: All primary studies included in a series of systematic reviews were identified. An analysis of the original source of the study and search term(s) by means of which the study had originally been retrieved was carried out. Each study was then searched for in each of 6 databases (AMED, Cochrane CENTRAL, MEDLINE/PubMed, EMBASE, CINAHL, PsycINFO). The proportion of studies located on each database was assessed and the indexing terms identified for each therapy were compared against the original search strategies.
Results: A total of 127 primary studies were identified from 35 systematic reviews. The number of studies on each therapy varied, but Cochrane CENTRAL listed the highest proportion for all therapies. No database listed all studies, and at least one unique study was listed on all databases except MEDLINE, whereas several studies were not found on any of the databases. Index terms were effective in locating studies on acupuncture, individual herbs, hypnosis, massage, and yoga. For the remaining therapies, use of text word search terms was important and particularly so for homeopathy, meditation, and reflexology. Variation in terminology for most of the therapies was encountered.
Conclusions: The small numbers of studies preclude firm recommendations, but several potential challenges in searching for complementary and alternative medicine (CAM) trials are highlighted. The findings suggest that a range of different sources is required for identifying relevant studies, particularly for certain therapies. The development of an optimum generic search strategy for each therapy is hampered by the variation in indexing of CAM studies. Possible optimum strategies are presented as a basis for discussion, and further testing of the effectiveness of these strategies is now a priority.