Article    Table of Contents                    Full Text of this article Free    Figures/Tables List  Services    Send comment/rapid response letter    Notify a friend about this article    Alert me when this article is cited    Add to Personal Archive    Download to Citation Manager    ACP Search                            Get Permissions  Google Scholar    Search for Related Content  PubMed Articles in PubMed by Author:    Mulrow, C.    Grimshaw, J.    Related Articles in PubMed    PubMed Citation    PubMed

ACADEMIA AND CLINIC

Integrating Heterogeneous Pieces of Evidence in Systematic Reviews

Cynthia Mulrow, MD, MSc; Peter Langhorne, PhD, MRCP; and Jeremy Grimshaw, MBChB, MRCGP

1 December 1997 | Volume 127 Issue 11 | Pages 989-995

Researchers preparing systematic reviews often encounter various types of evidence, which can generally be categorized as direct or indirect.The former directly relates an exposure, diagnostic strategy, or therapeutic intervention to the occurrence of a principal health outcome. Evidence is indirect if two or more bodies of evidence are required to relate the exposure, diagnostic strategy, or intervention to the principal health outcome.

Heterogeneity of data sources complicates integration of both direct and indirect evidence.Participants in different studies may have a wide spectrum of baseline risk and sociodemographic and cultural characteristics. A variety of formulations and intensities of exposures, diagnostic strategies, and interventions, as well as diversity in the selection and definition of control groups, may be encountered. Outcome measures may be different, and similar outcomes may be measured or reported differently. Heterogeneity of study designs and of methodologic features and quality within a given design may be found. The effective integration of direct and indirect evidence requires development of explicit models that serve as analytic frameworks for linking the important pieces of evidence. A model can be viewed as a series of subquestions, with each important subquestion warranting a systematic review. Several subjective and quantitative methods can then be used to integrate the evidence. Tabular displays of major findings and strength of evidence for each subquestion can help reviewers, patients, and providers to integrate the differing research findings and draw reasonable conclusions. Various quantitative techniques, such as decision analysis and the confidence profile method, are also available. No single integration approach is clearly superior, none obviates uncertainty, and all underscore the role of careful judgment in integrating evidence.

Author and Article Information

From the University of Texas Health Science Center at San Antonio and Audie L. Murphy Veterans Affairs Hospital, San Antonio, Texas; Royal Infirmary, Glasgow, Scotland; and the University of Aberdeen, Aberdeen, Scotland.
Acknowledgments: The authors thank Drs. Robert Fletcher and Brian Haynes for their critical reading of the manuscript. They also thank the clinical reviewer, Norman J. Wilder.
Requests for Reprints: Cynthia D. Mulrow, MD, MSc, Audie L. Murphy Memorial Veterans Hospital, 7400 Merton Minter Boulevard (11C6), San Antonio, TX 78284.
Current Author Addresses: Dr. Mulrow: Audie L. Murphy Memorial Veterans Hospital, 7400 Merton Minter Boulevard (11C6), San Antonio, TX 78284.

This article has been cited by other articles:

				J. van der Steen, R. Kruse, K. Szafara, D. Mehr, G van der Wal, M. Ribbe, and R. D'Agostino Benefits and pitfalls of pooling datasets from comparable observational studies: combining US and Dutch nursing home studies Palliative Medicine, September 1, 2008; 22(6): 750 - 759. [Abstract] [PDF]

				S. Shepperd Hospital at Home: The Evidence Is Not Compelling Ann Intern Med, December 6, 2005; 143(11): 840 - 841. [Full Text] [PDF]

				D. Reed, E. G. Price, D. M. Windish, S. M. Wright, A. Gozu, E. B. Hsu, M. C. Beach, D. Kern, and E. B. Bass Challenges in Systematic Reviews of Educational Intervention Studies Ann Intern Med, June 21, 2005; 142(12_Part_2): 1080 - 1089. [Abstract] [Full Text] [PDF]

				R. Kaushal, K. G. Shojania, and D. W. Bates Effects of Computerized Physician Order Entry and Clinical Decision Support Systems on Medication Safety: A Systematic Review Arch Intern Med, June 23, 2003; 163(12): 1409 - 1416. [Abstract] [Full Text] [PDF]

				J. E. Bekelman, C. P. Gross, and Y. Li Effect of Industry Sponsorship on the Results of Biomedical Research--Reply JAMA, May 21, 2003; 289(19): 2503 - 2503. [Full Text] [PDF]

				D A W M van der Windt and L M Bouter Physiotherapy or corticosteroid injection for shoulder pain? Ann Rheum Dis, May 1, 2003; 62(5): 385 - 387. [Full Text] [PDF]

				J. E. Bekelman, Y. Li, and C. P. Gross Scope and Impact of Financial Conflicts of Interest in Biomedical Research: A Systematic Review JAMA, January 22, 2003; 289(4): 454 - 465. [Abstract] [Full Text] [PDF]

				M. J. McQueen Overview of Evidence-based Medicine: Challenges for Evidence-based Laboratory Medicine Clin. Chem., August 1, 2001; 47(8): 1536 - 1546. [Abstract] [Full Text] [PDF]

				C. D. Mulrow and K. N. Lohr Proof and Policy from Medical Research Evidence Journal of Health Politics Policy and Law, April 1, 2001; 26(2): 249 - 266. [Abstract] [PDF]

				A. S. KLIGER and W. E. HALEY Clinical Practice Guidelines in End-Stage Renal Disease: A Strategyfor Implementation J. Am. Soc. Nephrol., April 1, 1999; 10(4): 872 - 877. [Abstract] [Full Text]