A comparison of MEDLINE, EMBASE AND CINAHL on searches for systematic reviews in respiratory health

Tags: Poster
Dennis JA, Bara AI, Milan S

Objective: We have previously compared the efficiency of Medline and Embase with respect to identification of asthma RCTs published in Thorax and the American Review of Respiratory Disease. This showed an overlap between the two databases of 89%, 7% being unique to Medline and 4% unique to Embase. Another important area to investigate is the overlap of respiratory RCTs between Medline, Embase and Cinahl for specific systematic reviews.

Methods: Eighteen review searches were chosen from the areas of asthma, COPD and bronchiectasis, in the Cochrane Airways Group Register from 1995 to 1997. This register contains downloaded respiratory disease records from Medline, Embase and Cinahl. Trials are identified using a search developed by the Airways Group. The table below shows the number and percentages of RCTs found across the three databases.

Results:

Asthma reviews

MEDLINE ONLY EMBASE ONLY CINAHL ONLY MED/EM OVERLAP MED/EM/CIN OVERLAP MED/CIN OVERLAP

84 199 4 79 3 1

22.7% 53.8% 1.1% 21.4% 0.8% 0.3%

Bronchiectasis reviews

MEDLINE ONLY EMBASE ONLY MED/EM OVERLAP

6 59 4

8.7% 85.5% 5.8%

COPD reviews

MEDLINE ONLY EMBASE ONLY CINAHL ONLY MED/EM OVERLAP MED/EM/CIN OVERLAP MED/CIN OVERLAP

32 95 2 23 3 2

20.4% 60.5% 1.3% 14.6% 1.9% 1.3%

Discussion: The overlap for RCTs between Medline and Embase for review searches over the last three years is much lower than the overlap previously reported for the two respiratory journals over the period 1980 to 1993. The overlap is less than 25% for all three respiratory areas. Embase contained the largest number of unique studies, especially for bronchiectasis. This shows again the importance of using both Medline and Embase in reviews of respiratory health care.