Preliminary Reading

Reading general references such as textbooks before searching for primary sources is a recommended first step. Obtaining this type of reference is relatively easy, and general knowledge on a topic may be summarized well in a book. Although reading a book on the topic may not contain the most current information, it does provide an overview of an area. This overview may include a description of the CAM approach, the assumptions or theoretical basis of the approach, and the recommended indications or contraindications. Preliminary reading also helps identify important terminology that will be used to guide a search. Excellent CAM-specific resources are available for therapists with a thumbnail of the most recent evidence,⁷ or a more in-depth description of selected CAMs.⁸

Searching Databases

Textbooks are useful as a first resource but often are outdated. Searching the literature for review articles and meta-analyses is a recommended second step to gain an overview of a topic and begin to survey the literature. A review article is a description of one or more studies written by someone other than the original author. A meta-analysis is a type of review article in which a statistical method of combining results of a series of independent, previously published studies carried out for the same general purposes is used to summarize the literature.⁹ Many choices of databases to search exist, and an individual’s starting point may depend on their ability to access different resources. One route is through paid subscription services, such as Ovid, which offer a constellation of databases to select from; the other is through public access to PubMed. Both search routes are described here and an order is recommended.

Regardless of the search route one chooses, three suggestions exist for searching databases. First, each database has its own configuration and rules for including citations, and it is worth taking the time to learn them to become efficient at searching. This is especially the case for the way terms are organized in the database. In MEDLINE, for example, using the medical subject headings (MeSH terms) more efficiently extracts the relevant references than use of other terms. Useful references provide an overview of different databases and search strategies.¹⁰ Second, key words are selected based on the preliminary reading and the client’s condition or goals. A good start is to use the CAM as the keyword and then refine the search by including keywords related to client scenario. The usefulness of the key words may vary by database. Third, the reader may benefit from a survey of the literature and categorizing studies as they relate to outcome or mechanism. Mechanism studies validate an intervention by demonstration or description of how an intervention may work. Outcome studies address the efficacy of an intervention by a description of the changes that occur. The distribution of mechanism and outcome studies varies with the state of the literature. A well-developed literature, such as that of acupuncture, has a large number of outcome and mechanism studies.

Evaluating the Literature

After looking at review articles and then performing a more targeted search by narrowing it with key words, the therapist must evaluate the selected literature. There are many approaches to evaluating the literature. Hierarchies of evidence have been proposed by several authors.^1,14 These allow the reader to determine the levels of control as well as the ability to generalize the findings.

The balance of outcome and mechanism studies is another method to evaluate the literature. In some literatures, many studies may investigate the outcome of an intervention, such as pain reduction, but no studies may investigate how or why (mechanism) an intervention may effect change.

Another tool to evaluate the literature is to review the research designs used in the studies. This allows the health care clinician to understand the development of a particular area of research. Early on, an area may contain descriptive articles, consisting primarily of case studies; more developed literature may have articles with experimental designs (such as randomized clinical trials), in which questions of an inferential nature are asked. The latter is a more rigorous literature from which generalizations can be made. For example, in the well-developed pharmacological literature—preclinical (mechanistic) studies may form the basis for controlled clinical efficacy and effectiveness studies. In contrast, the study of CAM botanicals may contain anecdotal descriptions of botanical use before well-characterized preclinical studies are designed. In some instances, efficacy studies may precede the preclinical trials.

Analyzing the results of a search that uses multiple databases provides the therapist with an overview of the literature. The results of this search highlight the differences in number and type of references for each database. Using several databases helps identify which disciplines are studying questions of interest. Repeating a search over time allows the therapist to observe how the literature evolves. Additional citations and new areas of study may exist.