Mismatch between overall and pairwise overlap analysis in a sample of overviews: A methodological review

Article type
Authors
Bracchiglione J1, Meza N2, Pérez-Carrasco I3, Vergara-Merino L2, Madrid E2, Urrútia G1, Bonfill X1
1IIB SANT PAU, CIBERESP, Barcelona
2Interdisciplinary Centre for Health Studies (CIESAL), Universidad de Valparaíso, Viña del Mar
3School of Medicine, Universidad de Valparaíso, Viña del Mar
Abstract
Background: Overlap of primary studies among systematic reviews (SRs) is a key methodological challenge for overviews. There are limited reports of the methods being used to deal with overlap by overview authors. The corrected covered area (CCA) is a widely accepted method to assess overlap, but no detailed CCA assessment of a representative sample of overviews has been made using an overall and a pairwise approach.

Objectives: To describe the approaches used to deal with the overlap within overviews, and to estimate and compare the assessment of overlap using an overall and a pairwise approach.

Methods: We searched PubMed for overviews published in 2018. Two independent authors conducted the screening process. We described the strategy used for assessing overlap and calculated overall and pairwise CCA for each overview using the GROOVE tool.

Results: We analysed a random sample of 30 out of 89 eligible articles (Figure 1). Eleven did not address the overlap. Of the remaining 19, the most common strategies used were visual representation (11/19) and discussion of overlap as a limitation (10/19). Median overall CCA among the included overviews was 6.7%. The pairwise analysis showed that 52.8% of SR pairs had slight overlap, whereas 28.3% had very high overlap. Our analysis showed that 8/30 overviews with overall slight or moderate overlap had at least 20% of pairwise nodes (pairs of SRs) with high overlap, and 2/30 overviews with high overlap had at least 20% of pairwise nodes with slight or moderate overlap (Figure 2). Figure 3 provides an example of an overview with moderate overall overlap and a high proportion of highly overlapped pairs of SRs.

Conclusions: Reported strategies for addressing overlap vary considerably among overview authors. The pairwise approach for assessing the CCA revealed highly overlapped pairs of SRs in overviews with overall slight overlap and vice versa. We encourage authors to complement the overall CCA assessment with a pairwise approach.

Patient, public and/or healthcare consumer involvement: No patients or consumers were involved in this research.