Reviews, Systematic Reviews, Overviews: “What’s It All About, Cochrane?”

Abstract

In the fall of 2008, the Lienhard School of Nursing began a doctor of nursing practice (DNP) program. In the spring 2009 semester, I taught a course with two dear and experienced colleagues on epidemiology through an evidence-based practice lens. As such, the major course assignment was to conduct a systematic (integrative) review on a topic related to the course content. One of the intellectual arguments that ensued among the faculty was whether to include previous systematic reviews related to the topic that the students reviewed in their table of evidence (TOE), a standard part of every integrative review. There were proponents for each point of view. On the one hand, a previous systematic review conducted on the focused clinical question of the students' reviews was an important piece of high-level evidence. On the other hand, one could argue that a new systematic review should include only primary studies. Before discussing the data available on the issue, I should perhaps inform readers of how the DNP faculty at Pace define systematic reviews. There are two major types of systematic reviews of quantitative research. One, an integrative review, is in narrative format, and the other, a meta-analysis, is a statistical summary of evidence on a particular topic. Different from what we have previously known as a narrative literature review, systematic reviews, whether narrative or statistical, have the following characteristics in common: * Detail their search and retrieval methods to find evidence on a clinical question. * Use predetermined criteria to evaluate the level of evidence and quality of each study. * Present a synthesis of findings. * Address implications for practice and recommendations for further research (Levin, 2009, p. 39). Now that we are all on the same page (or at least this page), let's get back to the argument. I have read reviews that include other systematic reviews in their TOE (Oliver, Martin, & McMurdo, 2004) as well as those that include only primary (Nielsen et al., 2005). Taking a scholarly approach, I decided to see if there were any published standards on this question. First, I went to the textbooks and articles that we used in our course. One of those texts, by Melnyk and Fineout-Overholt (2005), contains the following relevant statement: Data are extracted from individual studies (p. 116). A second text (Haynes, Sackett, Guyatt, & Tugwell, 2006) details methods for conducting a systematic review. Haynes and colleagues tell us that the first line of attack in doing a systematic review is to search for and retrieve other systematic reviews on the topic of interest. They also discuss, however, how to extract data from the individual in a review. They do not mention whether to include the previous systematic reviews in the TOE even though they may be discussed in the narrative. Additionally, a relatively recent article by Tarraco (2005), coming from a human resource perspective, seems to indicate that the author of a systematic review would review primary and then integrate findings to generate new ideas and directions for the field (p. 364). Tarraco, however, does not specifically answer the question posed in this column, as he does not discuss how to develop a TOE. Now, I was getting a little frustrated for lack of an answer to the question of whether to include previous systematic reviews in a TOE of new systematic review. …