Current issues of ACP Journal Club are published in Annals of Internal Medicine


Editorial

Beyond ACP Journal Club: how to harness MEDLINE for therapy problems

ACP J Club. 1994 July-Aug;121:A10. doi:10.7326/ACPJC-1994-121-1-A10b

Related Content in the Archives
• Correction: Beyond ACP Journal Club: how to harness MEDLINE for therapy problems
• Editorial: Applying the findings of clinical trials to individual patient
• Editorial: Beyond ACP Journal Club: how to harness MEDLINE to solve clinical problems
• Editorial: Beyond ACP Journal Club: how to harness MEDLINE for diagnostic problems
• Editorial: Beyond ACP Journal Club: how to harness MEDLINE for etiology problems
• Editorial: Beyond ACP Journal Club: how to harness MEDLINE for prognosis problems
• Editorial: Beyond ACP Journal Club: how to harness MEDLINE for review articles



The first editorial in the ACP Journal Club how-to-search series was published in the March/April issue and described using MEDLINE to aid clinical problem solving. We discussed clinician information needs, what information tools are available and used, reasons for doing MEDLINE searches, alternative MEDLINE products, and the economics of CD-ROM compared with online access. This second editorial in the series focuses on harnessing MEDLINE to retrieve citations on treatment.

Starting with a clinical example, we describe how to incorporate the various issues in therapy into MEDLINE searches. The patient, a 45-year-old woman, complains of frequent migraine headaches. Rescue therapy, especially sumatriptan, has been effective for her, but she does not like needles or the high price. Prophylactic ergonovine and methysergide have been only partially successful and she is afraid of adverse effects. She is otherwise healthy except for a remote history of asthma. She read in a magazine about biofeedback and relaxation therapy for migraines and asks for your advice on whether these are worth trying.

You could start looking for information in ACP Journal Club, which includes a 6-month index in the May/June and November/December issues each year. ACP Journal Club has 6 articles (3 abstracted and 3 listed in the Other Articles Noted section) appraising sumatriptan and aspirin treatment and homeopathy. Your patient, however, is interested in various behavioral therapies (biofeedback and relaxation techniques) for her migraines, and there are no abstracts on these types of treatments in ACP Journal Club so we must extend our search.

Scientific American Medicine, one of the first “Information Age” textbooks because of its periodic updating and evidence-based chapters, also has nothing on the effectiveness of behavioral techniques for migraine therapy. We need to consult other sources.

Let's try MEDLINE next. MEDLINE searching starts with breaking down the question into concepts that are combined with the Boolean operators, AND, OR, and AND NOT. The patient's question has three concepts: the disease, migraine headaches; behavioral therapies, relaxation and biofeedback; and, using ACP Journal Club scanning criteria for valid evidence, clinical or randomized controlled trials. The search strategy, using the searching language of the U.S. National Library of Medicine (NLM), would look like this:

*migraine (mh)

AND

explode behavior therapy

AND

clinical trial (pt)

The “*” in front of migraine indicates that we want only articles with migraine as the major focus of the paper. To make sure that we do not miss any earlier citations on the topic, we ask for the search to include material from 1985 to the present. For the 1991 to May 1994 file, there are 894 articles in which migraine is the focus of the paper. The “explode” before the behavioral therapy prompts the computer to group all behavioral therapy options (aversive therapy, biofeedback, cognitive therapy, desensitization, implosive therapy, and relaxation techniques) together for searching. The search term “clinical trial (publication type or pt)” is one of a new set of indexing terms introduced in 1990 by NLM that indicates the type of research design used in the study. Examples of publication types that indicate study design are clinical trial (pt), randomized controlled trial (pt), and multicenter study (pt).

Combining these 3 concepts—disorder, treatment type, and study design—and limiting the search to the English language literature result in the retrieval of 6 articles (1-6) that meet all search criteria. All 6 articles are relevant, a rare occurrence in searching. The reports provide evidence that alternative treatments and prophylaxis using behavioral therapies for migraine are at least as effective as other therapies and should be seriously considered.

Another way of doing the same search is to include all headache types, including migraine, as follows:

explode *headache OR *migraine

AND

explode behavior therapy

AND NOT

explode child

limit to English language only

limit to only core clinical journals

Using the Boolean OR would pool the headache terms (general headache, cluster headache, or vascular headache) with migraine. Note that the indexers at NLM do not consider migraine a type of headache so we must put both on the same subject line linked with an OR. The OR can be used to combine terms that mean the same thing (e.g., AIDS or acquired immunodeficiency syndrome) or similar terms (e.g., amoxicillin or ampicillin or erythromycin). An example of the Boolean AND NOT is also included with this search. The first search on migraine alone retrieved articles that dealt with children and adolescents. To eliminate these articles from the retrieval, we apply the Boolean operator AND NOT. The “explode child” term gathers up the various pediatric age categories (infant, newborn; infant; child, preschool; child; and adolescent), and stipulating AND NOT exp child ensures that any articles with these terms present in the indexing will not be retrieved.

The search is also limited to 125 “core” clinical journals. MEDLINE includes over 4000 journals and a search that is limited to the core clinical journal subset, including JAMA, Annals of Internal Medicine, Lancet, BMJ, and The New England Journal of Medicine, and so on, not only produces retrievals that are smaller, but the articles are more likely to be in smaller hospital libraries or even on office shelves.

This search retrieves 1 citation from the current MEDLINE file and a total of 5 citations back to 1985. None of the citations are the ones retrieved in the first search. They are “on topic” from a content point of view but they do not “pass criteria” for inclusion in ACP Journal Club. Our earlier approach was therefore needed to retrieve the sound ready-for-clinical-application articles.

We recently evaluated various search terms denoting study method for their ability to retrieve articles about sound clinical studies (e.g., randomized controlled trials for therapy) (7), compared with a “gold standard” based on manual searching of 10 core internal medicine journals for 1986 and 1991. The most effective searching strategies are shown in the Table below.


Table. Most Effective Searching Strategies*

Time period for searching Best single-term search Term sensitivity Term specificity
1990 on clinical trial (pt) 93% 92%
1989 and before random: 82% 95%

*The “:” is a “wildcard” that indicates all possible endings of a search term.

New tailor-made look-up options for finding evidence-based information are emerging quickly and will make searching for information on the fly much easier for internists. For example, we expect that an electronic version of ACP Journal Club will be commercially available in early 1995. In the meantime, MEDLINE has started to index ACP Journal Club commentaries and editorials, beginning with the first issue in 1994, so you can use MEDLINE to hunt for the most important studies of the treatment, diagnosis, cause, course, economics, and quality of care of problems in internal medicine from now on. To designate only articles from ACP Journal Club, use the form ACP J Club (cm) in your search.

Subsequent editorials in the series will discuss strategies for searching for sound studies of diagnosis, prognosis, or etiology of disorders, as well as for systematic review articles and studies of quality improvement. Good searching!

K. Ann McKibbon, MLS
Cindy J. Walker-Dilks, MLS


References

1. Osterhaus SO, Passchier J, van der Helm-Hylkema H, et al. Effects of behavioral psychophysiological treatment on schoolchildren with migraine in a nonclinical setting: predictors and process variables. J Pediatr Psychol. 1993;18:697-715.

2. Kroner-Herwig B, Fritsche G, Brauer H. The physiological stress response and the role of cognitive coping in migraine patients and non-headache controls. J Psychosom Res. 1993;37:467-80.

3. McGrath PJ, Humphreys P, Keene D, et al. The efficacy and efficiency of a self-administered treatment for adolescent migraine. Pain. 1992;49:321-4.

4. Holroyd KA, Penzien DB. Pharmacological versus non-pharmacological prophylaxis of recurrent migraine headache: a meta-analytic review of clinical trials. Pain. 1990;42:1-13.

5. Blanchard EB, Appelbaum KA, Radnitz CL, et al. A controlled evaluation of thermal biofeedback and thermal biofeedback combined with cognitive therapy in the treatment of vascular headache. J Consult Clin Psychol. 1990;58:216-24.

6. Gauthier J, Lacroix R, Cote A, Doyon J, Drolet M. Biofeedback control of migraine headaches: a comparison of two approaches. Biofeedback Self Regul. 1985;10:139-59.

7. Wilczynski NL, Walker CJ, McKibbon KA, Haynes RB. Assessment of methodologic search filters in MEDLINE. Symp Comp Appl Med Care. 1993;17:810-6.