Article type
Year
Abstract
Background: The science of research synthesis aims to systematically gather, examine and evaluate primary study reports that answer a clearly formulated question. However not all systematic reviews (SRs) reach the goal to minimize bias and some are furnished of methodological flaws. The revised and validated AMSTAR (R-AMSTAR) instrument has been generated to quantify the SRs quality.
Objectives: To describe the applicability and drawbacks of R-AMSTAR instrument.
Methods: We did and overview on the effectiveness of acupuncture in obstetric-gynaecological disorders. A comprehensive literature search was performed in CENTRAL, CDSR, DARE, MEDLINE databases up to July 2010 and with no language restriction. Two reviewers independently performed the literature selection and appraised the retrieved SRs using the R-AMSTAR instrument. This instrument contains 11 questions (each with 3 or 5 items) rated from 1 to 4 (total score range from 11 to 44). Although there were not a previous consensus between reviewers, they were experienced in the critical appraisal of SRs. Each item and overall agreement was tested using the kappa coefficient. The intraclass correlation coefficient (ICC) was used to test the agreement of R-AMSTAR sum scores.
Results: We identified 728 references and included 8SRs. The median score obtained by R-AMSTAR tool was 32 (range 21-41). Kappa coefficient value of each item differed widely and overall agreement was low (K = 0.338). The agreement in the sum scores was ICC 0.506 (moderate agreement). A greater agreement existed in the questions about 'a priori design provided’, 'a duplicate study selection and data extraction’ and 'the list of studies provided’.
Conclusions: A rather low agreement was obtained with the use of the R-AMSTAR instrument. Some items are ambiguous and may not be a direct attribute of SR methodological quality.
Objectives: To describe the applicability and drawbacks of R-AMSTAR instrument.
Methods: We did and overview on the effectiveness of acupuncture in obstetric-gynaecological disorders. A comprehensive literature search was performed in CENTRAL, CDSR, DARE, MEDLINE databases up to July 2010 and with no language restriction. Two reviewers independently performed the literature selection and appraised the retrieved SRs using the R-AMSTAR instrument. This instrument contains 11 questions (each with 3 or 5 items) rated from 1 to 4 (total score range from 11 to 44). Although there were not a previous consensus between reviewers, they were experienced in the critical appraisal of SRs. Each item and overall agreement was tested using the kappa coefficient. The intraclass correlation coefficient (ICC) was used to test the agreement of R-AMSTAR sum scores.
Results: We identified 728 references and included 8SRs. The median score obtained by R-AMSTAR tool was 32 (range 21-41). Kappa coefficient value of each item differed widely and overall agreement was low (K = 0.338). The agreement in the sum scores was ICC 0.506 (moderate agreement). A greater agreement existed in the questions about 'a priori design provided’, 'a duplicate study selection and data extraction’ and 'the list of studies provided’.
Conclusions: A rather low agreement was obtained with the use of the R-AMSTAR instrument. Some items are ambiguous and may not be a direct attribute of SR methodological quality.