The art and craft of scientific writing and critical appraisal of a research paper

Writing is an art and like any art form, it needs perseverance, dedication and practice. However, to write a good quality paper, the habit of reading scientific articles and analyzing them is very important. With the advent of internet and online publishing, we have access to colossal research articles on myriads of subjects making extraordinary conclusions. Evidence based practice requires us to rely on literature for our clinical practice, and we have abundant publications on all aspects claiming to justify all sides of the argument. In this context it becomes more important for all in clinical practice to be able to dissect an article and analyze it in details.


Introduction
With the advent of internet, the scientific world is flooded with publications.It has been almost 350 years since scientists have been publishing in journals. 1It is estimated that the number of journal articles published since that time is about 50 million. 2PubMed, an index of biomedical abstracts published by the National Center for Biotechnology Information has a collection of 19 million citations while PubMed Central, a full text archive of journal holds 1.7 million articles. 3Publication on PubMed amounts to one page per minute added on to the database on average and that excludes the large volumes of articles published but not added on the indexing service. 3ith such a large amount of publication, it becomes very difficult to identify good research articles.It requires the readers to be able to perform critical appraisal of the literature instead of taking them at their face value.Critical JSAN 2017; 4 (1)

Journal of Society of Anesthesiologists of Nepal
appraisal is regarded as a systemic process used to identify the strengths and weaknesses of a research article in order to assess the usefulness and validity of research findings. 4t has been defined as the " The application of rules of evidence to a study to assess the validity of the data, completeness of reporting, methods and procedures, conclusions, compliance with ethical standards, etc.The rules of evidence vary with circumstances." 5

Analysis of the Journal
It starts with the analysis of the journal where the article is published.There are various types of journal metrics such as speed, reach and impact factor.Speed implies review speed and online publication time.Unnecessary delay in publication might diminish the significance of the research.The term 'reach' corresponds to the geographic location and accessibility of the corresponding authors and the journal.
Impact factor of a journal is one of the widely used measures of assessing its quality.Conceived by Eugene Garfield in 1970's, Journal's Impact factor is from Journal Citation Report (JCR), a product of Thomson ISI (Institute for Scientific Information). 6JCR provides quantitative tools for evaluating journals.It is a measure of the frequency with which the "average article" in a journal has been cited in a given period of time.The impact factor for a journal is calculated based on a three-year period, and can be considered to be the average number of times published papers are cited up to two years after publication.(Figure 1) Impact factor helps to clarify the significance of absolute citation frequencies and eliminates bias associated with larger and older journals and those published frequently.None the less, the larger the number of previously published articles, more citations it will receive.The pattern of citation distribution has been found to be skewed that an analysis in 1992, showed only 15% of papers in a journal accounted for half the total citations implying that majority of the journals papers had less than average citation.Journal citation counts in JCR do not distinguish between letters, reviews, or original research.So, if a journal publishes a large number of letters, there will usually be a temporary increase in references to those letters.Self citation has been observed in around 13% of the citation a journal receives and to reduce its influence, a 'Revised Impact Factor' has been devised.However this might lead to biasness in article selected for publication in Journals.Hence, though impact factor is the only established measure of a journal's eminence, an article has to be assessed on individual merit.Other several impact metrics in use are SNIP, IPP, SJR and Eigenfactor.(Table 1)

Predatory journals
With the advent of online publishing, research and publication has been made accessible to all.However, it has also made easy for unscrupulous activities to flourish by establishing fake journals asking for hefty sums to publish just anything.These days email inbox is also filled with request for submission from a multitude of journals and for a price, publishing at the earliest.
It was in 2010 when librarian and researcher Jeffrey Beall from the University of Colorado Denver coined the term 'predatory publishing' and later came up with a list and criteria for evaluating publishers.

Journal of Society of Anesthesiologists of Nepal
submitting a fake and flawed paper which was accepted by majority of the journals (60%). 10y journal that frequently sends email requests for submission should raise suspicions especially if it charges a fee for publication.Identifying predatory journals involves scrutinizing the website for details such as invalid links, inappropriate address, the identity of the editors and editorial board which cannot be verified etc.Most fake publishers have a homepage that does not provide access to previous issues of the journal.Authors should practice caution while submitting articles to publishers and also while citing articles from these journals.

Analysis of the research setting
When and where the study was done bears an immense significance.Research publications from an academic institute might be less biased than those from private institutions.Delay in the publication of a paper from the time the study was concluded might indicate the lack of importance given by the authors to the study.Time between acceptance and publication of articles depends on the journal.Analysis of 2700 papers published in 135 journals sampled from Scopus citation index found the delay to be 9 to 18 months and varied according to the subject matter with the shortest delay in science technology and medical field. 7

Relevance of the research question
The most important reason for any publication is its relevance to mankind and its contribution to the knowledge.A study of utmost methodical rigor has little significance if it has no bearing to its own field of work.Since relevance is a subjective opinion, the reader should understand the research question of an article.Landmark papers leading to paradigm shift in science are a rarity.Most papers tend to validate previous studies and make incremental advancement in research by extending research findings to new population or clinical context.
An ideal research question identifies three components: the group or population of patients, the studied parameter (e.g. a therapy or clinical intervention) and the outcomes of interest. 11In general, clinical research questions fall into two distinct categories: 1. Effectiveness of treatment: This relates to whether one treatment is better than another in terms of clinical effectiveness (benefit and harm) or costeffectiveness.

Frequency of events:
This refers to the incidence or prevalence of disease or other clinical phenomena, risk factors, diagnosis, prognosis or prediction of specific clinical outcomes and investigations on the quality of health care.

Structure of the research paper
Over the last century, in health sciences, research has evolved from descriptive to well structured form.
Articles were organized like a chapter in a book with headings associated with the subject matter until 1945.Since 1950's the IMRAD (Introduction, Methods, Results and Discussion) came into use and has been widely adopted after the guidelines were set by the International committee of Medical Journal Editors, formerly known as the Vancouver Group. 17The IMRAD format promotes uniformity and facilitates modular reading such that the reader can browse to the area of interest.It also facilitates the peer review process.Though critics of the IMRAD structure claim that it is too rigid and simplistic as it limits creativity and may not give realistic representation of the thought process of the scientist, however, at present it is the only universally accepted and adopted format in biomedical publication. 18tle of an article entices the readers.It should be catchy, reflect the content clearly and be specific.An ideal title identifies the article's main issue, begins with the subject matter, and is short but complete, accurate and unambiguous.The language should be simple with preferably short sentences and active writing when possible.Tense is important in scientific writing.All known facts and hypothesis should have present tense and past tense should be there for description of the results.Illustrations and figures should be contextual, legible and properly labeled.

Appropriateness of the study design
The main strength of any research is its study design.
With the journals flooded with hundreds of articles on the same subject, it is always difficult to identify the important ones.Careful scrutiny of the study design helps to isolate the researches whose conclusions can be relied upon.With the numerous types of researches published ranging from retrospective series to randomized controlled trials, there are wide variations in the format of the study design.To bring uniformity in these designs and to avoid biasness, guidelines have been proposed for these studies.These guidelines specify the minimum information that should be included in a research report to allow readers to judge the quality of the study.(Table 2) JSAN 2017; 4 (1)

Biasness in a research
In simple terms, it means that the results of a study have deviated from truth.It can be attributed to chance (e.g. a random error) or to the study methods (e.g.systemic bias).Random error affects the precision of the study but does not influence the results in any particular direction.
The systemic bias results during participants' selection, data collection, analysis of results and its interpretation.
Bias can emanate from the author's background, institutional affiliation, grant or funding etc. Publication bias leads to failure of publication of important research with negative results that might have contributed a lot for the science.

Statistical analyses performed
While critically analyzing, for non statisticians, it might be difficult to assess the appropriateness of the statistical tools used in the study.Factors such as appropriate sample size, data collection tools, data analysis methods chosen should be properly stated.The results should be stated clearly and fully supported by the analysis.
Missing data and the loss to follow up that can occur in a study should be properly mentioned since these might greatly influence the results and add bias to the study.In RCT, patients' data should primarily be analyzed on the basis of random allocation regardless of them receiving the treatment or not according to the principle of intention-totreat (ITT).Any protocol violation in the study should be properly stated.Data should be presented in such a way that a reader can verify the statistical accuracy if required.
In a research, statistical significance may not always translate to clinical significance and the reader has to be aware of the implications.There is always a tendency to extrapolate elaborate conclusions not backed up by

Figure 1 :
Figure 1: Calculation for Journal Impact Factor