Volume 100: The review article
Medical and scientific literature is so extensive that no single clinician can keep up with anything more than a very small fraction of it. We often, therefore, rely on papers which are a synoptic, or a sort of summary, of the available literature. These papers are the review articles of journals and are designed to educate, or to answer carefully conceived important clinical questions. Their format can either be a review, meta-analysis, or scientific editorial. They should be based, like all other papers, on a clearly stated question, hypothesis, or objective and include a thorough critical assessment of the literature and be written in a logical manner. They act as a useful resource since they are a synthesis of information from all (should be, but often a selection) of the relevant literature on the particular subject—they save others from the time-consuming task of sifting through the published material. They inform researchers of where their research stands on a particular issue and may suggest new directions for further research.1 Here, I define what a review is—or at least should be, discuss the recent trends in review writing and suggest some changes to the reviews published in the British Journal of Anaesthesia.Reviews are popular with readers because they are a synthesis of the current literature in a usually easily digestible form. They are also popular with journal editors because they are often some of the most cited articles and consequently enhance the impact factor of the journal. This latter fact often accounts for the very high impact factor of some journals which publish exclusively reviews.
There are two broad types of review: the descriptive or ‘narrative’ review and the systematic review. Systematic reviews are a newer type of review and are structured and formulaic and are most often associated with the requirements of evidence-based medicine and best practice. They require a critical appraisal of all the available literature and come to evidence-based conclusions—often after meta-analysis and statistical testing—that is, they require a systematic assessment. The descriptive review can be much more opinion-based, although it should still have a clear question and conclusion. Both types of review are subject to bias depending on what papers are included and what are excluded. Some of this bias is as a result of publication bias where editors prefer to publish research with positive results. However, much bias can also lie in the hands of the author who can exclude papers either because of never finding them in the first place (faulty searching) or because they do not meet some specified (often arbitrary) criteria of ‘worth’. This of course is a particular problem with the more structured reviews since these will conclude with advice regarding best practice which should be based on all the available evidence.
The first issues of the British Journal of Anaesthesia did not contain reviews as we know them today. They did have a section called ‘Abstracts of current anaesthetic literature’ which must have served a similar function. The very first of these was largely a discussion of the relative merits of ether, chloroform, and nitrous oxide.2 It contained a case report of hiccough which had persisted for 7 days after a chloroform anaesthetic; the then usual treatment was to no avail. Ether, given hypodermically, was then tried as it had been used previously with success in whooping cough. Ten cubic centimetre (as it was then—millilitre now) was given with good effect after a delay of 30 min, but the hiccough returned during the night. A further 6 ccm was administered and recovery was then complete and permanent. One of the articles from Le Progrès Médicale (May 12, 1923) bemoaned the Parisian surgeons of the time for being slow in adopting nitrous oxide which the author considered safer and better for the patient than either chloroform or ether. However, the same author deprecated the American and British custom of adding these latter agents to nitrous oxide as it introduced an unnecessary element of danger. Also in this section was the description of a novel technique of inducing anaesthesia (a somnambulic state) in dogs by raising the cerebro-spinal pressure—reported in the Japan Medical World of February 15, 1923.
Most of these early articles were narrative style reviews and even the original research papers were presented in a narrative rather than the now more usual ‘scientific’ style.
Using PubMed as search engine, I found that the first article published in the British Journal of Anaesthesia listed by them as a review appeared in the September 1963 issue—a post-graduate issue. This was published to coincide with the 21st anniversary of the introduction of arrow poison into anaesthesia. The first article classified as a ‘true’ review to be identified as such is consequently ‘The anatomy and physiology of the neuromuscular junction’ by Roberts.3
Using the same search engine, there are 856 articles in the British Journal of Anaesthesia identified as a review (accessed November 28, 2007)—there are actually 857 listed but I found a duplicate citation.
The National Library of Medicine has a standard definition of a review and of course this may be different to others' definitions.4
- Reviews are:
- analyses of the literature, primarily that of the past 10 yr;
- all articles from review publications (e.g. Annual Review series).
- analyses of the literature, primarily that of the past 10 yr;
- Reviews are not:
- collections of cases;
- articles primarily on the history of a subject;
- bibliographies.
- collections of cases;
The British Journal of Anaesthesia website hosted by Oxford University Press (http://bja.oxfordjournals.org/) lists the 50 most cited Journal publications. From this list, it would seem that the most cited review in the Journal is that of Mallet and Cox,5 which is a narrative review of thrombelastography published in 1992. This was a timely review of an educational nature and detailed not only the way the thrombelastograph worked along with the various traces that are obtained from it, but also suggested some procedures in which it may have a use. The apparatus has since gained a more widespread use and is now joined by other devices which operate in a somewhat different manner but provide the same information. The authors of the review conclude that there is still a gap in our ability to monitor coagulation adequately in the operating theatre and although there are many sophisticated haematological tests now available to assess haemostatic failure, these are still not generally available on an immediate, ‘at the bed-side’ basis. I would suggest that thrombelastography is the only feasible way of assessing activated fibrinolysis available to the anaesthetist. It is also a most useful device for assessing the effect of haemostatic drugs in vitro before administration to the patient. This review would thus fulfil the role of a good educational and narrative review and has certainly been well cited by others.
The first appearance of a systematic review in the British Journal of Anaesthesia is more difficult to find but the first one with ‘systematic review’ in the title was as recent as 2000—‘A systematic review of the efficacy of ginger in nausea and vomiting'.6 Six randomized controlled trials for or against the efficacy of ginger for nausea and vomiting were reviewed having fulfilled the pre-set inclusion criteria. Three on postoperative nausea and vomiting were identified; two suggesting that ginger was superior to placebo and as equally effective as metoclopramide. The pooled absolute risk reduction for the incidence of postoperative nausea, however, indicated a non-significant difference between the ginger and the placebo groups for 1 g ginger taken before operation [absolute risk reduction 0.052 (95% confidence interval –0.082 to 0.186)]. The authors suggested the need for larger studies.
In 1997, there was an analysis of review articles published in four anaesthesia journals, including this Journal, to assess their validity.7 It had already been 10 yr since reviews in general medical journals had undergone a similar scrutiny.8 This latter article found that medical reviews at that time did not routinely use scientific methods to identify, assess, and synthesize information. It further suggested that had these methods been used it would have resulted in a more structured review of improved quality. Similarly, the reviews specifically in the anaesthetic journals during 1995 also found that most were not conducted to more rigorous standards. Evidence for this statement comes from the fact that most did not have an explicit purpose stated at the outset; did not specify the search strategy used; nor the inclusion and exclusion criteria used for the papers included; and did not have any quality assessment performed. Moreover, very few suggested any future research directions. Also, interestingly, only 10% in this Journal were considered to be systematic reviews.7
I will now state my personal opinion as to what should be the future of reviews in the British Journal of Anaesthesia. The narrative style review is fine but it is subject to inevitable bias and to my mind often falls into the category of a general educational type article—they often do not pose or answer a question and do not suggest further areas of research. These articles could perhaps be included in a section of the Journal called ‘Continuing medical education articles’ such as is seen in the journal Critical Care Medicine. The following references provide examples of recently published review articles in the British Journal of Anaesthesia that I would categorize as such.9–12 In the past year, a number of such reviews in the British Journal of Anaesthesia have been the subject of our on-line CME development.
This then leaves the systematic reviews and while appreciating the problems of introducing evidence-based medicine into actual practice, I wonder if this Journal should be more pro-active in promoting this type of review article. A section specifically called ‘Systematic reviews’ would go part of the way to achieve this and is a common feature of another critical care journal, Intensive Care Medicine.13–15 There are similar reviews published in the British Journal of Anaesthesia,16 17 but they are often not associated with clear-cut detailed recommendations along with suggestions for further work—as would be dictated by an evidence-based approach. We should seek to publish reviews that are perhaps more limited in scope, that pose specific and verifiable questions, have a structured and defined search and inclusion strategy, that assess the strengths and weakness of the evidence, that produce conclusions and pose further research questions. This approach would allow a more comprehensive review of the literature. It would also hopefully help to provide answers to the controversies of the day and thus be of benefit to our patients.
Department of Anaesthesia and Intensive Care
Institute of Medical Sciences
Foresterhill
Aberdeen AB25 2ZD
UK
E-mail: n.r.webster{at}abdn.ac.uk
References
1 Huth EJ. The review article and the metaanalysis. In: Writing and Publishing in Medicine (1999) 3rd Edn. Baltimore: Williams and Wilkins. 93–101.
2 Annonymous. Abstracts of current anaesthetic literature. Br J Anaesth (1923) 1:41–4.
3 Roberts DV. The anatomy and physiology of the neuromuscular junction. Br J Anaesth (1963) 35:510–20.
4 Charen T. MEDLARS indexing manual part II. (1983) Bethesda, MD: National Library of Medicine. section 12.1.
5 Mallet SV, Cox DJA. Thrombelastography. Br J Anaesth (1992) 69:307–13.
6 Ernst E, Pittler MH. Efficacy of ginger for nausea and vomiting: a systematic review of randomized clinical trials. Br J Anaesth (2000) 84:367–71.
7 Smith AF. An analysis of review articles published in four anaesthesia journals. Can J Anaesth (1997) 44:405–9.[Web of Science][Medline]
8 Mulrow CD. The medical review article: state of the science. Ann Intern Med (1987) 106:485–8.
9 Pirracchio R, Cholley B, De Hert S, Solal AC, Mebazaa A. Diastolic heart failure in anaesthesia and critical care. Br J Anaesth (2007) 98:707–21.
10 Wong GT, Irwin MG. Contrast-induced nephropathy. Br J Anaesth (2007) 99:474–83.
11 Mols G, Priebe HJ, Guttmann J. Alveolar recruitment in acute lung injury. Br J Anaesth (2006) 96:156–66.
12 Podgoreanu MV, Schwinn DA. Genomics and the circulation. Br J Anaesth (2004) 93:140–8.
13 Stevens RD, Dowdy DW, Michaels RK, Mendez-Tellez PA, Pronovost PJ, Needham DM. Neuromuscular dysfunction acquired in critical illness: a systematic review. Intensive Care Med (2007) 33:1876–91.[CrossRef][Web of Science][Medline]
14 Griffiths J, Fortune G, Barber V, Young JD. The prevalence of post traumatic stress disorder in survivors of ICU treatment: a systematic review. Intensive Care Med (2007) 33:1506–18.[CrossRef][Web of Science][Medline]
15 Gao H, McDonnell A, Harrison DA, et al. Systematic review and evaluation of physiological track and trigger warning systems for identifying at-risk patients on the ward. Intensive Care Med (2007) 33:667–79.[CrossRef][Web of Science][Medline]
16 Naguib M, Kopman AF, Ensor JE. Neuromuscular monitoring and postoperative residual curarization: a meta-analysis. Br J Anaesth (2007) 98:302–16.
17 Pandit JJ, Satya-Krishna R, Gration P. Superficial or deep cervical plexus block for carotid endarterectomy: a systematic review of complications. Br J Anaesth (2007) 99:159–69.
CiteULike 
Connotea 
Del.icio.us What's this?
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||