Health care literature, though better organized than other knowledge areas, is still cumbersome enough to make effective searching difficult and large retrievals commonplace. Subject searching (for drugs and diseases) is useful for searching an entire database, but a different approach is needed to limit retrieval to articles that health professionals should use for patient care.
Many different groups, such as basic scientists, clinical researchers, health practitioners, educators, and administrators, produce publications. Their written work ranges from idea papers, which include editorials and case reports, to wet laboratory and animal experimental studies, to small early studies with humans.
Another important category is the reports in the applied clinical literature about the large, carefully controlled trials done in hospitals and institutions with human participants. These trials are the studies that health care professionals should rely on to keep up-to-date and solve patient care problems. These trials have actual evidence that the changes in health care under investigation are proven advances. The applied clinical trials answer the following questions:
1. What caused the disease or disorder? (etiology)
2. How can I decide what the disease or disorder is? (diagnosis)
3. What happens to people with the disease over time? (prognosis/natural history)
4. How can I treat (or prevent) the disease/disorder?(therapy)
Only a fraction of the millions of citations in MEDLINE, CINAHL, EMBASE, and other related databases actually are of reports of applied clinical trials that address the four questions above. Most citations are reports of preparatory work completed before the applied clinical trials can be done. Applied clinical studies are unique in the manner in which they are produced (their study methods).
Authors refer, in their titles and abstracts, to the methods used in the research, and database indexers index these methodologies as keywords describing the articles. Searchers can use the methods textwords and subject headings to only retrieve papers that have sound research methods and can, therefore, be used directly in patient care. We will describe the study methodologies used in each category, list appropriate index and textwords (Table 1), and give search examples of how to use methodology terms for search filtering. (Table 1 omitted)
Reports of applied clinical research have these two common features:
1. They are comparative in nature. For example, a therapy trial compares patients taking one drug with patients taking another drug or placebo (aspirin versus heparin for stroke prevention). A diagnosis study compares two diagnostic tests (thallium scanning versus cardiac enzyme measurements for heart damage after myocardial infarction). Similarly an etiology study compares people exposed to some agent thought to cause a disease, with people not exposed to the agent (antacids with and without aluminum salts causing Alzheimer's disease). For prognosis studies, patients with a disease are compared to people without the disease.
2. Careful planning and execution go into these clinical trials. Protocol development includes very specific details describing all aspects of the study. Both concepts can be searched in MEDLINE and CINAHL.
As an example, CINAHL has 228 articles on hand washing from 1986 to 1992. When the first two CINAHL terms listed in Table 1 (comparative studies and clinical trials) are added to the search strategy, i.e., handwashing and (comparative studies or clinical trials), the retrieval is reduced to eleven studies that evaluate hand washing procedures. The eleven papers can be from any of the four purpose categories of therapy, diagnosis, etiology, or prognosis.
Therapy, prevention and control (can we reduce side-effects from ulcer medications?), and quality of care (can telephone …