Portal de Periódicos da CAPES

Top 10 Tips for Undertaking Synthesis Research

2017; Wiley; Volume: 40; Issue: 3 Linguagem: Inglês

10.1002/nur.21790

1098-240X

Kathleen A. Knafl, Robin Whittemore,

Meta-analysis and systematic reviews

Knowledge in the form of best research evidence is a key component of evidence-based practice (Sackett, Rosenberg, Gray, Haynes, & Richardson, 1996). However, the voluminous, often unwieldy, body of research in many areas of nursing practice makes it difficult to draw conclusions about what is known. Similar studies may yield conflicting results, and researchers often study similar topics using different theoretical orientations and methods. With over 26 million publications listed in the PubMed database, millions with links to full-text content, and over 13,000 journals in the Web of Science database (including over 100 nursing journals), the volume of information, even in a focused area of inquiry, can be overwhelming. Over 30 years ago, in his bestselling book Megatrends, Naisbett (1982) famously stated, "We are drowning in information, but starved for knowledge" (p.17). His observation captures what many of us experience when faced with trying to make sense of a particular body of literature. The dramatic growth of research makes the need for synthesis more pressing than ever. For both clinicians committed to evidence-based practice and researchers wanting to build on and extend the current body of knowledge, synthesizing research provides a powerful tool for determining what we know. Yet, synthesizing the research can be a challenging undertaking. Recognizing the importance of synthesis, researchers have directed increased attention to developing and refining synthesis methods. As a result, multiple options are available for anyone proposing to undertake a synthesis project. Kearney (2016) addressed the importance of novice review authors being exposed to literature reviews that "serve as examples to support well-substantiated, nuanced claims about the state of the science" (p. 6), and in Table 1 we provide citations for strong examples of different types of reviews. Despite the many excellent resources available to support synthesis research, undertaking a research synthesis can be daunting. The increasing sophistication of methods and guidelines has contributed to the scope and quality of the work being done but at the same time poses new challenges. While there are numerous synthesis methods, the steps of the process are similar. In this guest editorial, we draw on our experiences conducting synthesis research and differentiating the various synthesis approaches to offer our top 10 tips for undertaking a synthesis study. In Table 2 we provide further elaboration on the tips by identifying some key questions that need to be addressed during the review process. Conducting top-quality research synthesis is clearly important for evidence-based practice and to guide future research. The journal Systematic Reviews is focused entirely on the publication of syntheses and the development of synthesis methods. All journals, regardless of their focus, benefit from publishing review articles, because they often are cited more frequently than articles reporting primary research. Review articles also can increase journal downloads and the impact factor of a journal. Authors benefit from writing review articles because published reviews establish the author's expertise in the field and can further interest in the importance of the author's field of inquiry. Vicki Conn, Editor of the Western Journal of Nursing Research, has written several editorials to encourage nurses in all stages of their careers to publish review articles (Conn, 2014a,2014b,2014c). Synthesis research is time-consuming and complex. There are numerous design decisions to be made, from finalizing the question to developing the research synthesis protocol. There are also challenges to implementing the research synthesis protocol. Therefore, having a qualified team who have diverse areas of expertise will greatly improve both the review process and the outcome. Likely members of a research synthesis team include a content expert, a methods expert with experience applying the design and analysis methods to be used, a librarian skilled in data-base searching, and a data manager. It takes time to finalize the research synthesis question. Often a preliminary scoping review is conducted to better understand the research that has been completed on the topic of interest and to determine the scope and magnitude of the body of research (Arksey & O'Malley, 2005; Levac, Colquhoun, & O'Brien, 2010). How much research has been published? Are there areas of research with conflicting results? Have research reviews been published on the topic? How long ago were the research reviews published? During the phase of refining the synthesis question, it is important to assess the potential for publication and the feasibility of the review, and it is equally important to be willing to redirect based on the existing research that has been published on the topic. For example, your original plan to complete a meta-analysis may need to be modified if a scoping study reveals that most of the research in the field is qualitative. In her guest editorial "Top 10 Tips for Successful Grantsmanship," Grey (2000) advised readers to "Dream big, start small," reminding us "Rome wasn't built in a day, and nursing science will not be built on one project" (p. 91). We think this is excellent advice for anyone doing synthesis research. It most likely will be impossible to synthesize an entire area of inquiry in one study. While it is important to dream big and have overarching ideas about the potential implications of a research synthesis, all research synthesis projects need to be feasible. We often get questions about the sample size for research synthesis projects. "How many articles are enough? How many are too many?" We always answer, "It depends." We have conducted reviews that were published with as few as seven articles and as many as 100 articles. The sample size depends on the research synthesis question, the type of data extracted, and how information-dense the articles included in the review are. For beginners, we recommend refining your research question to limit the scope of the project and thus the number of articles to synthesize. All of the approaches included in Table 1 can be used to conduct either large- or small-scale reviews. A modest project gives a novice an opportunity to master a synthesis approach without being overwhelmed by the data. Once you have finalized the research synthesis question, the method is determined based on the overarching synthesis goal and the existing evidence. Do you want to describe results, integrate the results into a thematic description or conceptual model, identify areas for future research, or assess the weight of evidence that certain types of interventions are efficacious or certain relationships are significant? Do you need to have a sample of articles with very homogenous methods, or with a variety of methodologies? In their article on methods for knowledge synthesis, Whittemore, Chao, Jang, Minges, and Park (2014) included a list of resources for conducting various types of reviews and emphasized the importance of aligning the research question, the available body of literature, and the proposed synthesis approach. There will be many challenging decisions in the development and implementation of the synthesis protocol. For example, deciding what quality appraisal tool(s) to use can be especially difficult. There is no gold standard for many synthesis designs, and there are numerous options. The National Heart, Lung, and Blood Institute (2014) have developed quality appraisal tools for quantitative designs, and a widely used quality appraisal tool for qualitative designs is the Critical Appraisal Skills Programme (CASP) checklist (CASP UK, 2013). Deciding what to do with quality scores is another important decision. Do you include or exclude reports from your sample based on quality? Do you consider quality when interpreting the results, by closely examining the quality scores of outliers? In screening articles to include in the review, categories of "yes," "no," and "maybe" will develop. How will you determine which of the "maybe" articles are included or excluded? Having a plan and a strong team is essential for making thoughtful, defensible decisions. There are many details and decisions in conducting a research synthesis. It is important to document these and maintain an audit trail of all decisions and challenges to the implementation of the protocol. Attention to detail is essential in documenting your search strategies and yields (terms, databases, dates, etc.), screening procedures, data needed to comply with PRISMA guidelines (Moher et al., 2009), data extraction procedures, quality assessment, and the process for verifying analyses. It also is essential to maintain a well-organized database with the results of all of the data extraction, analysis, and synthesis efforts (Havill et al., 2014). New software programs (e.g., EndNote, Covidence) continue to be developed to facilitate the management of sources and data in conducting systematic reviews. Conducting a quality research synthesis takes time, patience, and perseverance. It is important to carefully consider all of the time-consuming aspects of a research synthesis and to have a sufficient number of qualified team members, including research assistants, to assure that you are able to meet project deadlines. You cannot conduct a quality research synthesis in 2 or 3 months. Each step of the process is time-consuming, and you should never sacrifice quality for the sake of speed. Because synthesis research is time-consuming, it often is desirable to update your database search as you near completion of the project to enhance the currency of the review. Analysis is examining each study with the goal to understand specific results that address the research question, as well as factors within each study that may have contributed to the results. Further analysis is then conducted to compare results across studies. In comparing results across studies, the team should consider: Are the results of the studies fairly consistent? If not, can studies be grouped together based on their differences? Do factors such as setting, sample, measurement, or study quality help explain inconsistencies? Although you need to examine and understand the results of individual studies, avoid simply describing each study (e.g., "Study 1 reported this … Study 2 reported this"). After analysis, synthesis is needed. Combine the results of the multiple studies into a coherent understanding of the topic that answers the research question. Tables and figures often are used to highlight the results of the data synthesis. The process of conducting a research synthesis can be tedious and time-consuming. Thus, we encourage our teams (and ourselves) to keep calm, carry on, and celebrate milestones (big and small) along the way. This keeps us engaged and committed to the project over time. Plus, it adds some fun to the process. The authors declare no conflict of interest.