A Two‐Term MEDLINE Search Strategy for Identifying Randomized Trials in Obstetrics and Gynecology

Objective To develop and test a simple MEDLINE search strategy for identification of randomized controlled trials (RCTs) in obstetrics and gynecology. Methods To develop our search strategy, we asked clinicians in our department to indicate, from a list of search terms, the terms they would use to identify RCTs in MEDLINE. The two most common terms, controlled-clinical-trial (publication type) and randomized-controlled-trial (publication type), were combined with the link word, OR, and then used to identify RCTs in four obstetrics and gynecology journals for the years 1975, 1980, 1985, 1990, and 1995. Concurrently, a handsearch of these same journals and years was performed to identify RCTs. The sensitivity and precision of MEDLINE and handsearch were calculated using the total number of RCTs identified by both methods as a reference standard. Sensitivity is the RCTs identified by search strategy as a percentage of all RCTs identified by reference standard. Precision is the RCTs identified by a search strategy as a percentage of all articles identified by it. Results The overall sensitivity of our MEDLINE search strategy was 72.5%, and the precision was 83.4%. Over 2 decades, sensitivity of our MEDLINE search increased from 0% to 94.9% (P < .001), while its precision dropped from 100% to 75.5% (P = .003). For 1990 and 1995 combined, sensitivity and precision of our MEDLINE search strategy were 90.3% and 79.6%, respectively. Overall sensitivity for handsearch was 96.5%; its precision was 5.0%. Over 2 decades, the sensitivity of handsearch dropped insignificantly from 100% to 92.3% (P = .05), while the precision increased from 2.6% to 6.3% (P < .001). Conclusion Our simple MEDLINE search strategy has a high sensitivity and precision, especially in more recent years. Obstetricians and gynecologists may use it to search quickly for RCTs to guide patient care.

[1]  Jacob Cohen,et al.  Weighted kappa: Nominal scale agreement provision for scaled disagreement or partial credit. , 1968 .

[2]  C. Silagy.  Developing a register of randomised controlled trials in primary care. , 1993, BMJ.

[3]  R. Smith,et al.  Where is the wisdom...? , 1991, BMJ.

[4]  Elliot R. Siegel,et al.  Bibliographic-retrieval systems , 1990 .

[5]  R. Haynes,et al.  Searching MEDLINE for randomized clinical trials involving care of the newborn. , 1989, Pediatrics.

[6]  T. Poynard,et al.  The retrieval of randomized clinical trials in liver disease from the medical literature. A comparison of MEDLARS and manual methods. , 1985, Controlled clinical trials.

[7]  N. Wilczynski,et al.  How to search for and find evidence about therapy , 1996, Evidence Based Medicine.

[8]  D. Sackett.  Evidence-Based Medicine: How to Practice and Teach EBM , 2018 .

[9]  I. Olkin,et al.  Improving the quality of reporting of randomized controlled trials. The CONSORT statement. , 1996, JAMA.

[10]  S. McDonald,et al.  Identifying reports of controlled trials in the BMJ and the Lancet , 1996, BMJ.

[11]  S. Walter,et al.  Empirical evidence of bias in infertility research: overestimation of treatment effect in crossover trials using pregnancy as the outcome measure. , 1996, Fertility and sterility.

[12]  C. Adams,et al.  An investigation of the adequacy of MEDLINE searches for randomized controlled trials (RCTs) of the effects of mental health care , 1994, Psychological Medicine.

[13]  R. Brian Haynes,et al.  Developing optimal search strategies for detecting clinically sound studies in MEDLINE. , 1994, Journal of the American Medical Informatics Association : JAMIA.