VOL: 98, ISSUE: 02, PAGE NO: 36
Louise Forward, MA, BA, researcher at the Health and Community Care Research Unit, University of Liverpool
Lindsay Hobby, BSocSc, RN, RM, researcher at the Health and Community Care Research Unit, University of Liverpool
This article offers practical information for nurses who are thinking of conducting a systematic review for the first time. It is based on our own experiences and details the compromises we had to make to produce a piece of work within the constraints of limited resources. It highlights factors that should be considered before deciding to conduct a systematic review, the problems we experienced and practical tips to simplify the process.
What is a systematic review?
Systematic reviews are not simply literature reviews. A systematic review enables researchers to make an objective assessment of primary research evidence and, in particular, of the outcomes of specific interventions or treatments. They are the result of a rigorous process (Box 1) and are considered the gold standard of literature-based research. According to strict scientific design, laid out in the NHS Centre for Reviews and Dissemination’s guidelines (The University of York and NHS Centre for Reviews and Dissemination, 2001), the available evidence is located, evaluated and synthesised. The aim is to achieve a comprehensive, unbiased and reliable summary of the evidence.
The principles of a systematic review can be used as a good practice framework to review the relevant literature for any piece of research, but a volume of literature reviewed in a systematic manner cannot be considered a systematic review. The key differences lie in the comprehensive nature and quality of the assessment: the searches are wide and the evidence is actively assessed. A checklist (Box 2) can be used to determine the quality of the literature.
Literature can be sorted into low and high quality, depending on the formality of the review. Less formal reviews do not attempt to be comprehensive or to sift out low-quality evidence. However, in robust reviews literature is often discarded on the basis of predetermined formats for inclusion and exclusion.
Examples of inclusion and exclusion criteria that we used to conduct a systematic review of the factors affecting the uptake of health and social care services by minority ethnic groups are detailed in Box 3.
Enough time must be allocated, whether the review is the primary focus or forms part of a larger project. A minimum of six months is recommended - any less might compromise the quality of the product.
The time available for the review will depend on resources and the volume of literature found. This can be assessed simply and quickly by conducting a preliminary search of two or three databases. The number of documents identified will give a rough idea of the volume of potentially relevant literature. Bear in mind that the NHS Centre for Reviews and Dissemination recommends that two researchers work on a systematic review to reduce any literature bias and increase the validity of the process.
Speaking to experts working in the particular field of study is also recommended and may identify unpublished studies. However, time limitations may mean this is difficult.
Although a systematic review may be the best research method to use, the available resources, including time, may be insufficient. If this is the case, establish what is acceptable to the funders, what is achievable by the researchers and the potential impact on the findings. An appropriate compromise may be to review a smaller amount of literature using the principles of systematic review.
Limiting the volume of literature can be achieved by:
- Reducing the time frames of searches to, for example, the past five years;
- Restricting the number of databases searched;
- Tightening the focus of the study by using only specific research methods, in-depth interviews or randomised controlled trials, or limiting group studies by age or sex;
- Restricting the search to specific journals (this may distort the conclusions as studies reporting positive effects as opposed to little or no effect are more likely to be published in leading journals);
- ‘Grey’ literature (unpublished) can be excluded but this may also result in an overoptimistic view;
- Limit searches to UK-only or regional areas.
It is important to assess the impact of the above restrictions on the review. A negative impact can be avoided by exploring the subject area before taking any decisions.
- Grey literature is particularly difficult to identify and retrieve. Some databases, such as the National Research Register, list unpublished work. However, these rely on researchers registering their work and may not be comprehensive. Once identified, unpublished work can be difficult to obtain;
- University library facilities are often busy during term-time, limiting access to databases and staff. During university holidays libraries are much quieter, but staff may be scarce and system upgrades often mean that databases are not in use;
- Interlibrary loans are invaluable for retrieving literature that is not stored locally, but these can take weeks or months;
- Tracking down missing journals is frustrating and time-consuming. Libraries may request that forms be completed for each article required before staff will start a search, causing further delay.
We recommend the following tips to help you manage the data, complete the review within your deadline and stay within your budget:
- Record references on index cards, making them portable and simplifying the process of retrieving literature;
- Order interlibrary loans as soon as possible;
- Go through the bibliographies of articles and reports to identify further references;
- Use the internet to find journals, saving time and money;
- Be aware that different papers published in different journals may report the same piece of research;
- Help and support is available from the NHS Centre for Reviews and Dissemination, including developing search strategies and conducting searches (for a small fee);
- Cochrane review groups have specific topic areas. See www.cochrane.org/cochrane/crgs.htm;
- Information on systematic review methodology is available on the Cochrane CD-ROM and various websites, including:
- For training courses see The Oxford Systematic Review Training Programme at www.ox.ac.uk
Carrying out a rigorous review is far from easy and, like any piece of research, a systematic review can be done badly. Despite its limitations a systematic review, or a review of the literature in a systematic manner, can provide a valuable insight into a subject area. It also gives the researcher useful skills, notably the organised identification and retrieval of literature and the ability to critically appraise research and assess its quality. Finally, bear in mind that systematic reviews are not an easy option and require time, patience and concentration.
Dickson, R. (1999)Systematic reviews. In: Hamer, S., Collinson, G. (eds) Achieving Evidence-based Practice: A Handbook For Practitioners. London: Ballière Tindall.
Dingwall, R. et al (1998)Catching goldfish: quality in qualitative research. Journal of Health Service Research Policy; 3: 3, 167-172.
Goodman, C. (1994)Literature Searching and Evidence Interpretation For Assessing Health Care Practices. Stockholm: Swedish Council on Technology Assessment in Health Care.
Mulrow, C.D. (1994)Rationale for systematic reviews. British Medical Journal; 309: 6954, 597-599.
Popay, J. et al (1998)Rationale and standards for the systematic review of qualitative literature in health services research. Qualitative Health Research; 8: 3, 341-351.