The continuing growth in usage and popularity of complementary and alternative medicine (CAM) [1
] has resulted in an increase in scientific research, including systematic reviews, in this field [3
]. Some hierarchies of evidence suggest that high-quality systematic reviews of randomized controlled trials (RCTs) are the gold standard of evidence-based medicine since they attempt to synthesize research using a rigorous methodology to limit bias [5
]. Part of this methodology includes identifying a comprehensive evidence base of controlled clinical trials (CCTs—including randomized controlled trials) through a highly sensitive search of the literature [7
]. Such broad searches have resource implications for the conduct of systematic reviews, so knowledge of various bibliographic databases' (db) productivity, overlap and features is important. In addition, practitioners, researchers and librarians searching for high-quality evidence of CAM may not know which or how many databases to search in order to obtain a totality of evidence.
The objective of this project is to compare 15 bibliographic databases relevant to CAM in order to ascertain the unique contributions of each; as well as a study of other features of these databases, such as searchability and cost. Such a broad comparison of CAM databases has not been undertaken to date. A literature review of similar research provided the findings summarized below.
A recent study by Pilkington (2007) analysed original sources of controlled trials included in 35 CAM systematic reviews and then determined whether the 127 trials identified were indexed in MEDLINE, EMBASE, Cochrane CENTRAL (Cochrane Controlled Trials Register or CCTR) and three other databases (AMED, CINAHL and PsycINFO) [8
]. The study found that 27 (out of the 127) trials were uniquely found in one of the databases and AMED produced the highest number of unique citations (n
= 5) [8
]. Sampson et al. [3
] analysed 13 databases for pediatric CAM RCTs, identifying over 900 relevant records from over 300 journals. The study concluded that for an exhaustive search for pediatric CAM RCTs, CAB Health, CINAHL and AMED should also be searched in addition to MEDLINE, EMBASE and CENTRAL (CCTR) [3
]. Murphy et al. [9
] analysed 13 databases for trials to be included in a systematic review on the reliability of spinal palpation. The study found that of the 49 trials (retrieved from the databases) that were determined to be relevant for the systematic review, MANTIS was the most productive database (n
= 35), followed by PubMed, CINAHL and MD Consult (n
= 19 each) [9
Sood et al. [10
] analysed databases used in 10 acupuncture systematic reviews. The study found that PubMed indexed 69% of the total 108 included acupuncture trials and the number of databases searched varied considerably from 3 to 12 (median 5) [10
]. Vickers [11
] analysed The Cochrane Collaboration Complementary Medicine Field's registry of 3774 complementary medicine RCTs from 965 journals. The study found that 81% of the RCTs were indexed in MEDLINE and only about one-third of these could easily be found by a MEDLINE search [11
]. Aker et al. [12
] studied three databases for chiropractic literature, identifying 385 included citations. The study found that MEDLINE retrieved 68% of the citations, CHIROLARS retrieved 23% and Index to Chiropractic Literature retrieved 10% [12
]. McPartland and Pruitt [13
] conducted an extensive literature search on a herbal medicine topic (saw palmetto) that included searching MEDLINE and four additional databases (EMBASE, Cochrane, AGRICOLA and IBIS) along with hand searching of non-indexed herbal journals. They found that MEDLINE yielded only 33% of the 58 relevant clinical trials identified [13
Moher et al. [14
] analysed the reporting in pediatric CAM systematic reviews, and found that they were ‘particularly weak in terms of the comprehensiveness in their search to identify primary studies', (page 9) with only 40% reporting a search that was reasonably comprehensive. Shekelle et al. [15
] analyzed the sources and methods used to identify CAM evidence in 21 Evidence-based Practice Center reports on CAM interventions. The study concluded that CAM systematic reviewers should include CAM specialized databases in their search methods [15
]. Additional prior research comparing databases and analyzing the bibliometrics of CAM literature has either studied a specific area of CAM or focused on a small number of databases [16
As can be seen from the literature review above, a broad comparison of CAM databases has not been undertaken to date. This project studies 15 databases beyond the three major databases (MEDLINE, EMBASE, and CENTRAL/CCTR) in an attempt to identify which additional databases might be useful sources of CAM controlled trials. The results of this project may benefit systematic reviewers, researchers, practitioners and librarians involved with CAM.
This project was part of a larger program to update The Cochrane Collaboration Complementary Medicine Field's trials registry, which is a database containing information on CAM controlled clinical trials. The definition of CAM used in this study was based on the one provided by the National Center for Complementary and Alternative Medicine (NCCAM), which defines CAM as health care practices and products that are “not presently considered to be part of conventional medicine” [27
]. Examples of the CAM therapies and modalities covered in this study include, but are not limited to
: acupuncture, chiropractic, hands-on healing, herbal medicine, homeopathy, hypnosis, magnetotherapy, massage therapy, mind-body techniques, naturopathic medicine, osteopathic manipulation, prayer, yoga, etc. Nutritional supplements and diet therapy were considered to be CAM if they are not used in conventional medicine; for example, high-dose B-vitamin therapy was considered to be CAM, whereas a (routine-dose) prenatal multivitamin was not considered to be CAM.