Introduction
What is an “academic” surgeon? Classically this means a surgeon affiliated with a university, who has an academic title, and who’s time is split between clinical activity, research, and education. A non-academic surgeon used to define one without an academic appointment, in part-time or full-time private practice, who devotes no time to teaching undergraduates or research. In my registrar days I was aware of a divide between these two groups of surgeons expressed as “town vs gown”. I always thought that this distinction was rather artificial, as all surgeons should be “academic”, to a greater or lesser extent. I suppose it depends on priorities.
After all, all surgeons must teach, if not to undergraduates, then at least to surgical residents and other surgeons. All surgeons have unique ideas and methods that deserve to be discussed, criticized, and evaluated. And all surgeons must have a sense of skepticism at what they are being taught. They wonder “why do I do it like this?”. This drives them to the literature to find out what others have written and to formulate their own approach based on what they were taught and what they have learned. So all surgeons are academic; some are “true academics” and work at a university department, which others are “clinical academics” who do not. If all surgeons are academic, it follows that they need a basic academic skillset. That is what this section is about.
Evidence Based Medicine
Evidence-based medicine has been the “buzzword” of medical practice for many years. Many of the things we do follow protocols established by various committees who research the literature, assess the quality of the relevant studies, and incorporate the findings into a set of guidelines, recommendations, or parameters. Thus, are patterns of clinical care established. More recently “care pathways” have been produced that proscribe how a condition is to be treated. The advantage of such a strategy is that there is uniformity of care between providers that elevates quality by ensuring that all patients will undergo appropriate workup before and after surgery, and that the indications for various combinations of treatment modalities are standardized. However, there is not always good evidence from high quality studies available for a certain clinical condition, and studies that are performed on large numbers of patients may have minimal relevance to an individual case, with the unique complexities that exist only in that patient. Therefore, evidence-based medicine is to be accepted as a worthy set of parameters to guide clinical care, but not as an infallible set of rules. Good clinicians recognize limitations due to lack of evidence and can individualize care strategies when this is indicated.
Building your Resume
A resume or curriculum vitae is a summary of your professional life and reveals many things about you. It lists your academic progress from as early as high school, with awards and accomplishments. It lists your career path including choice of university and the grades achieved there. It shows your extracurricular activities and interests and allows potential employers or selection committees to make a comparative assessment of your strengths and weaknesses. In our current environment, a career in medicine or related fields can be highly competitive. Competition for residencies, fellowships, grants, and staff positions is often intense and the curriculum vitae (CV) is an important aspect of that competition. A strong CV is a definite advantage, although strength in a CV can be defined differently.
An academic CV is characterized by involvement in academics: research projects, presentations at scientific meetings, scholarly articles in medical journal or books. Opportunities come in medical school and continue during residency. To an educated eye, quality is better than quantity. First authorships are better than the more anonymous and less significant “one name in the middle of many”. The ultimate evidence of scholarship is to see a project carried to completion and then showing up as one or more abstracts and one or more articles. This shows perseverance and hard work. Papers featuring the results of projects are not usually the first articles to be published in a career. It takes time to generate data. Reviews of important topics and even educational case reports are easier to write and can kick start a bibliography.
A socially admirable CV is more evident in extracurricular activities than the bibliography. Volunteering for a variety of community services is perfectly consistent with excellence of character and shows promise of the compassion needed to be an excellent physician. Involvement with sports and various past times is also noteworthy in as much as it is evidence of potential leadership qualities and a broad range of interests. Finally, other subjects already studied indicate the possible direction of the career. A minor in education can be expanded into an interest in surgical education. Exposure to environmental science may indicate a budding epidemiologist.
An ideal CV is laid out to be easy to read, and to clearly show the academic path of the owner. As time goes by and awards, projects and papers accumulate, a table of contents may be needed. An ideal CV is updated frequently and kept neat and tidy.
Doing Research
Research means finding out…finding out why diseases happen, how to prevent them, how to diagnose and treat them, and how to choose the best treatment out of a range of options. This is clinical research, mainly patient based and concerned with answering questions that are directly relevant to patient care. Not everyone is a researcher, and a surgical scientist is a hard row to hoe. However, what separates us from technicians is the thought that we put in to our work.
Why Research?
The obvious answer to this rhetorical question is simple: to find out the answer to a question. That is a very narrow view of the reasons research is important. A broader consideration of why we do research includes training in the rigor of scientific thought, instruction in the methodology of scientific effort, and learning how to be academically critical of research performed by others. These considerations are almost independent of the research performed. For example, in 1975 I spent a year investigating the connections of the nucleus reticularis tegmenti pontis in a rat brain. The specific knowledge I gained about the neuroanatomy of the rat brain was of zero benefit to me in my colorectal career. However, acquiring the qualities and abilities of a researcher has stood me in good stead for 35 years.
The Place of Research in a Medical Career
Involvement in some sort of academic exercise is almost inevitable for medical professionals. In medical school it is an extension of the labs; in postgraduate studies it is the main object of the degree; in residency it is often a requirement of the training program, and in clinical practice it is part of academic medicine. There are medical practitioners who just like to see patients, who are content with the science available in textbooks and journals and who have no urge to contribute to the world’s body of knowledge. They derive their professional satisfaction solely from patient care. There is nothing wrong with that and such practitioners will not be reading this section of the book. The fact that that you are reading this section is evidence that you are interested in academic pursuits, and you are the person for whom this section was written.
There has been an increasing emphasis on the importance of research, especially for medical students, residents, and junior faculty. This is most critical in an academic environment, where it is a legitimately expected part of an academic appointment. “Academic surgeons” usually have dedicated research time incorporated into their week (the temptation to have clinical care gobble up this time needs to be resisted) and must produce the goods. Most new academic surgeons are given some time and some support to establish themselves and their group. The other parts of the trifecta that constitutes an academic career are excellence in teaching and excellence in clinical medicine. Academic clinicians have a difficult job, but a rewarding one.
Surgeons in “non-academic” jobs can add some scientific thought to their patient care and to their residents and fellows. They tend to be amateur researchers but still can produce valuable work. My career exemplifies that of an amateur researcher. All that you need is enough of a sense of wonder and inquiry to generate the drive to answer questions.
Conceiving a Study
A study is an investigation into an unknown phenomenon and begins with an original thought. Coming up with a truly interesting and novel question is the key step to a good study; it is also the most difficult step, as most of the easy good ideas have already been taken. Producing a good idea in this age requires an extensive knowledge of what is known about a topic, because only then do we appreciate what is unknown. It demands an appreciation of current literature because that is where the knowledge rests. Attendance at a scientific meeting is also a very productive way of developing new research ideas. Listening to a series of projects being reported inevitably summarizes the literature for you and identifies studies that have either been badly done and need repeating, or studies that open new fields of enquiry to be explored. A good an idea for a study must ask a legitimate, novel and important question. Projects that repeat past work or extend it by emphasizing a series of slightly different aspects of the same topic (salami slicing) are less worthwhile.
Thoughtful physicians are always getting ideas; it’s part of the practice of medicine. There is so much that we don’t know and don’t understand and when we are confronted with a situation about which we have no clue, we are stimulated to investigate. Hopefully you will come up with a good idea. Because it’s your idea you are invested in it, jealous of it, and inclined to work hard to finish it. Someone else’s idea tends to be less compelling so if you have the interest and latitude either come up with your own idea or contribute to someone else’s. Research leaders who supervise multiple students, residents or fellows have a difficult task in supplying valid and novel ideas to everyone.
The Hypothesis
Turn your idea into a hypothesis…a question that sets the stage for a study that will attempt to prove the hypothesis or answer the question. A hypothesis focuses the study and provides direction, but it also states your “best guess” as a fact. For example you may develop a thought that says Australians are smarter than New Zealanders. The question of your study would be “are Australians smarter than new Zealanders?” Your hypothesis would be “Australians are smarter than New Zealanders”. A null hypothesis would be “there is not difference between Australians and New Zealanders in intelligence”.
The Protocol
Now that there is an idea it can be refined. This means sharing the thought with others…supervisor, partner, colleague. If it is your idea, you take the lead. Learning to delegate is an important aspect of collaborative research. The next step is a thorough literature review to make sure that the embryonic project is original, important, and feasible. This is the beginning of the study protocol.
Once the idea is validated a plan to perform the study is formed. The exact nature of the study depends on the question generated by the idea. There are several ways of conducting research. The study may be basic science, performed in a laboratory, with the help of contributors who have expertise in this setting. It may be a clinical study. Clinical studies can be prospective or retrospective, or a retrospective analysis of prospectively collected data. They may be descriptive or comparative. Comparisons can be made between matched controls, historical controls, case controls, or prospectively randomized groups. Designing studies demands expertise and here is where statisticians are helpful.
Studies involving patients usually require sample size calculations, so that the number of patients is appropriate to the size of the expected difference between groups, conferring enough statistical power for the study to achieve its aims. The most glamorous studies are prospective, randomized controlled studies that can either be blinded or open. These are favored because they are the most powerful way of determining a real effect of a treatment on a patient population. They take a lot of careful planning and time, and are not particularly suitable for a research fellow with a year at his or her disposal.
The best way for a clinician to learn about study design is to become involved in a randomized prospective study that is often multi institutional. Attending meetings where study design is discussed can be quite a revelation.
Once the design of the study is set a detailed description of what will be done is written. This must be detailed and exact enough to allow someone who has not been involved in the planning to repeat the study. Inherent in this “Methods” section will be definitions of variables and criteria used in the enrolment of patients. Definitions should follow those used by other studies to allow comparison with the literature. Exclusion and inclusion criteria are set and a rationale is given for each. The dates of patient accrual are specified and a list of datapoints to be collected is presented. Minimum length of follow up may be important.
If the study relates to a certain procedure, describe the procedure, or make reference to an accurate description. List the end points that will be abstracted from the data, and rank them, with primary endpoints and secondary endpoints. This guards against post hoc analysis. Finish with a description of the statistical approach to the data; the descriptive and the comparative statistics that will be used, and the level of power that defines significance.
Budget and Funding
Once a protocol has been written describing the rationale for the study, the aims of the study, the hypothesis, and the study design it is time to consider money. The context of every study in terms of the funding available and the predicted cost of the study will determine whether it goes ahead. Aspects of budget include personnel, opportunity costs, patient related costs, logistical costs such as publication and travel, and specific costs such as drugs and equipment. Funding support may be from non-specific departmental funds, Chair funds, Grants, or other sources. Sometimes there is no cost to a study, if it is just database research.
IRB Approval
The protocol and the budget then go to the Institutional Review Board for review and potential approval. Here the ethics of the study are examined, the potential for good and harm, and the practicality of the study design. There may be questions of informed consent to consider, and the wording of the consent form is crucial. Most researchers review a series of mandatory educational presentations generated by their institution prior to performing any research. In more complex projects there are often suggested changes to protocol generated by the IRB. Correction of the protocol and re submission are required. With IRB approval in hand, the study can begin.
Basic Statistics
I am not a statistician. I have not even taken a course in statistics, but I have used statistics all my professional life and have had more experience than most clinicians in designing, performing, and analyzing all sorts of clinical research from single investigator to multi-institutional international studies. So, I am well placed to present the level of statistical that could be defined as “basic”. All surgeons are likely to use statistics, either in their own research, or in reviewing research of others. Knowing about statistics is necessary when reading a scientific report as it allows you to decide whether it is a good study or bad, whether the author’s conclusions are valid or not, and whether you might want to change your practice on the basis of the study. Others will be critiquing your research in the same light.
Statistics is a science devoted to the collection and analysis of numbers, in particular numbers that define an aspect of multiple participants. Statistics are used to determine whether there are any differences between groups of participants and to state the probability of differences being real. Most medical research is comparative, involving descriptive data and outcomes data derived from multiple groups of patients exposed to different therapies or having different underlying properties. Basic statistics can be discussed under two headings: descriptive statistics and tests of significance.
Descriptive Statistics are used to describe a population or a sample. A population is any large number of people or objects such as everyone in America with Crohn’s disease, or even everyone in your institution with Crohn’s disease. A sample is a representative group drawn from a population, such as all Crohn’s patients having an ileocolic resection at your institution. Summary numbers drawn from a population are parameters, such as the population mean (µ) and the population proportion (p). Summary numbers from a sample are statistics, such as the sample mean (x). Before deciding on which statistics should describe your data it is necessary to know what type of data you are collecting. There are four types:
Categorical data: the values or observations fit into a category such as gender: the categories are “biologically male” or “biologically female”. Categorical data are described as number, or percentage, or proportion. The significance of differences between groups of categorical data is tested by Fisher’s Exact test when any subgroup has a number (n) less than 10, or by Chi Square test for groups greater than 10.
Nominal Data: where numbers are just labels to describe a group, for example where male is noted as “1” and female as “2”. Nominal data are treated in the same way as categorical data and cannot be added or summed.
Ordinal Data: exists where values or observations can be put in order. An example is a linear analog scale to measure pain, with values ordered from 1 to 10. These data can be described as mean and standard deviation, or median and range.
Continuous Data: the values or observations can be anywhere on a continual numeric range. These data are described as mean and standard deviation, or median and range.
Analysis of continuous data begins with an assessment of data distribution; how often values occur in a sample. Many natural human properties such a height or intelligence follow a normal distribution, or “bell shaped curve”. This can be seen by plotting the value measured on the X axis and the number of people with each value on the Y axis. The curve has a symmetrical appearance with most people having middle values and the curve “tailing” off toward the lower and the high extremes of the range. This normal distribution allows data summary by mean and standard deviation. The significance of differences between two normally distributed groups can be tested by Student’s t test, and for multiple groups by Analysis of Variance (ANOVA). In some datasets the data are skewed so that the curve is no longer bell-shaped but favors one end or the other. Non-parametric data are described by median and interquartile range (IQR), and the significance of differences between two groups is tested using Mann Whitney U test or Wilcoxon Rank Sum test.
Tips for the use of descriptive data
1. Using appropriate precision
Describe your data using the same precision you used to collect it.
If you did not collect patient age to the week, don’t report age to the second decimal place, (e.g. the mean age was 56.78 years).
2. Reporting Percentages when the number of patients is <100
Percent means “per one hundred.”
Don’t use percentages at all for numbers below 20: use fractions (e.g. 1 out of 3 is not 33%, it is 1/3).
Report numbers greater than 100 as percentages.
Report numbers between 20 and 100 as both a fraction and a percent.
3. Summarizing Categorical Data
Specify denominators
Summarize in the text unless the no. of categories warrants bar graph
If continuous data are separated into ordinal categories, give cut points and their rationale
4. Reporting Ordinal Data as Continuous
Be careful of taking the mean of ordinal data when there are only a few categories. For a 10 point scale however, averages are OK
For small numbers of groups, report the number of responses for each category, or the category with the most responses (the modal score)
Summarizing Continuous Data
Measures of central tendency: mean, median, mode
Measures of dispersion: Use standard deviation, or interquartile range. Don’t use Standard Error of the Mean (SEM).
The normal distribution: If the SD is >0.5x the mean, the data are not normally distributed. If 2 x SD gives an impossible number (e.g a negative amount of years of age), the data are not normally distributed
Skewness: Degree of skewness - Mean minus the Median (for a normal this should be 0).
Coefficient of Variation A measure of dispersion. SD/Mean x 100.
Measures of dispersion and how to calculate them without a computer
1. Standard Deviation of the Mean (SD): the degree to which individuals within the sample differ from the sample mean. Use with the mean to describe normally distributed, continuous data.
How to find Standard Deviation
Calculate the Mean of your data set.
For each number: subtract the Mean and square the result.
Calculate the Mean of those squared differences. The result is called the Variance.
Calculate the square root of the Variance. The result is the Standard Deviation.
Important note- If the data set is only a sample of the whole population, change the formula to divide by one less when calculating the Variance. This is called Bessel's correction.
2. Standard Error of the Mean (SEM)
This is the standard deviation of the sample means and used as an estimate of the population mean; of how close to the population mean your sample mean is likely to be. It is always much smaller than the standard deviation of the mean of the individual which is the degree to which individuals within the sample differ from the sample mean. The formula for SEM is SEM = SD/√n. This statistic is often used inappropriately as a sample statistic.
3. 95% Confidence Interval
A confidence interval (CI) is a type of interval estimate of a population parameter and is used to indicate the reliability of an estimate. The 95% CI is the range of values where 95% of the population will fall. The formula is SEM x 1.96 (assumes normal distribution): SEM = SD/√n
If the 95% CI of two groups of data do not overlap there is a significant difference between them. For Odds ratio or Relative Risk, if 95% CI doesn’t overlap 1, the odds or risk is significant at 0.05% level.
Probability Testing
Hypothesis: H0 vs Halt
P = probability of getting a difference between groups as large or larger than that measured, by chance
Threshold of significance: ά = 0.05
Hypothesis Testing
State Hypothesis
Determine minimum difference that is clinically significant.
Set ά
Power calculation: Power = 1-ß (ß= probability of type II error: missing a real difference, 0.1 = 90% power; 0.2 = 80% power)
Writing
So your study is done. Now it is time to write it up, to report your results to the medical community and to accomplish the goals that you had when the study was conceived. In many cases the study is done with the aim of presentation at a meeting. This requires an abstract. An accepted abstract then is transformed into a presentation (poster or oral), and finally into a manuscript. Each of these steps requires mastery of a separate art form.
The abstract
Medical abstracts are the currency of medical research carried out by residents and fellows. Their purpose is to summarize a study on one page so that it can be evaluated, compared, and judged for presentation at a meeting. The selection of abstracts for presentation at medical meetings is inevitably competitive, and a polished abstract goes a long way to scoring well. Abstracts are usually of limited length; 250 to 300 words, or 3,000 characters. Because of these limitations, writing an abstract requires the most literary discipline of all medical writings. Sometimes hard decisions must be made about what to include, and there is absolutely no room for wasted words. The design of an abstract is related to the length of time for the presentation. In general, there will be time for one or perhaps two main ideas or points. Trying to cram more material will be counterproductive as a mass of data can obscure the main message of the study. Abstracts may do double duty as the abstract of a manuscript although the meeting submission comes first.
Before writing, check the requirements of the particular meeting or journal that is the abstract’s target. Make sure that the abstract structure, font, spacing, word or character count, and the use of tables, figures or illustrations is acceptable.
An abstract has a title, a list of authors, an introduction, a methods section, a display of the results of the study, and a conclusion. Of these the title is the most important.
The Title
A catchy title can attract the attention of a tired or bored reviewer or judge and invite a more thorough reading than might otherwise be the case. Titles can summarize the question or hypothesis of the study (Is it true? Are Australians smarter than Kiwis??) or the result of the study (Kiwis are smarter than their Australian Cousins). Other titles can be more frivolous but still catchy (Polishing the crystal ball: a study of desmoid risk factors in patients with FAP), or (Morituri te salutant: the outcomes of emergency colorectal surgery in patients who are ASA V).
The background
The background sets the scene for the study, establishing why it is being done. In 2 or 3 sentences state the problem or question that the study addresses, the reason it has not yet been adequately addressed, and the hypothesis that you have formulated.
(Desmoid disease is the second most common cause of death in patients with FAP, and 80% of desmoids occur after abdominal surgery. No study has shown whether ileoanal pouch surgery is followed more often by desmoids than ileorectal anastomosis. We hypothesized that desmoids would be more common after ileoanal pouch because of the increased tension on the small bowel mesentery).
Methods
This section describes how the study was performed. This must include the selection criteria for entry into the study, exclusion criteria, what was done to the patients, the outcomes that were measured (the primary and secondary end points), and the methods of analysis of the data.
(All patients undergoing ileoanal or ileorectal anastomosis for FAP at the Cleveland Clinic from 1995 to 2015 were included. Those with abdominal desmoids at the index surgery were excluded. Patients were followed up yearly for endoscopy of the pouch or rectum, and abdominal examination was routine. CT scans were done for symptoms. The primary endpoint was the incidence of symptomatic abdominal desmoid tumors. Secondary endpoint was incidental (asymptomatic) desmoids found on laparotomy or CT for unrelated reasons.
Results
This is the key section of the abstract, where the results of the study are presented. First, describe the study sample(s) or cohort(s). This means demographics (ages, genders, ethnicities where appropriate), details of patient selection for either treatment groups or disease groups or both, outcomes, and end points. In a short abstract, report only the key analyses of the primary end points. Subtle details of sub analyses can be reported in the manuscript. One simple table, and/or a graph or illustration, is usually allowed. Do not state results in text AND table. Data are described only once.
Conclusion
Please don’t summarize the study in the conclusion. The conclusion is a statement of the significance of the study for clinical practice or for further research.
Read your abstract. Read it again. Wait a while and read it again. Correct your grammar. Make sure you have followed the instructions published by the meeting organizers. Check the word count. Early versions of the abstract will usually be too wordy, so subsequent edits chip away at the verbiage like a sculptor chips a block of marble. Gradually, the finished work will emerge. Making changes to the abstract after it has been accepted should not be routine and should be done only with permission of the program committee.
Making a Poster
Posters are often regarded as a “second best” outcome of submission of a paper to a meeting, and yet presentation of a poster can be very rewarding. Posters allow viewers to read the work in their own time, and therefore can be more detailed than a podium presentation. Viewers can also have an individual discussion with the presenter and ask detailed questions on a one-to-one basis. While the entire audience may not see the work, those with a particular interest in your topic will come.
It is important to put the effort in to making an attractive, legible, interesting poster that is easy to read from a distance of 3 or 4 feet, highlights the work, will draw the attention of attendees, and reflect well on its authors and their department. The first step is to read the instructions to the authors and use the template that is often provided. Sometimes your institution will want a uniform template to be used so that all posters from one department or institution can easily be identified by their pattern of colors and design.
A simple way to start a poster is to reproduce the abstract on a poster-sized layout. The title and authors should be the same as the abstract, as this is how the work will appear in the book of the meeting. The abstract can occupy a corner of the poster, and be followed by an introduction, methods, results, and conclusion.
The usual way to read a poster is from top left to bottom right (when you are facing it) and this is the way that the text should flow. There are usually two or three vertical columns, and the text flows down them. There should be room to add tables, graphs, and illustrations. These make the poster more attractive (especially illustrations).
Text in a poster is much more readable when it is in bullet points. A solid block of text is quite off-putting to a reader and requires much more effort to read. It is very important not to overcrowd a poster as this is unattractive for a potential reader. Photographs or cartoons can be add humor. Making good posters is an art. And remember … only present the data once.
Writing a Manuscript
At least 50% of abstracts and presented papers are never published as manuscripts. This is such a shame as it represents wasted effort and the loss of potentially valuable knowledge to the scientific community. The extra work required to turn an abstract into a manuscript and for the manuscript to be published is daunting, and the writing process is not everyone’s cup of tea. However, it should be done. Here the steps to be taken to turn an abstract into a manuscript are described.
Step 1. Choose your Journal.
The target of your submission is usually determined by the topic of your manuscript. Colorectal surgeons publish in a colorectal journal because the manuscript will be of zero interest to orthopedic surgeons or dermatologists. Sometimes journals are mandated by the process of meeting submission: presenting at American Society of Colorectal Surgeons means that the manuscript goes to Diseases of the Colon and Rectum; presenting at American Surgical means that the manuscript goes to Annals of Surgery. If there is no mandate, choose a journal who’s readership is most appropriate for the subject matter concerned, and if there are choices, usually the journal with the highest impact factor is chosen. (Impact factor = number of citations in a years divided by the total number of articles published in the two previous years). Some impact factors are: New England Journal of Medicine 26.015, The Lancet 14.607, Annals Surgery 13.787, Diseases of the Colon and Rectum 4.657, New Zealand Medical Journal 1.02.
Once the journal is chosen, read the instructions to authors very carefully. Complying with the journal requirements will facilitate a smooth submission process. Many journals have different rules for different types of contribution. For example, Diseases of the Colon and Rectum doesn’t accept case reports, and technical papers need a video. Original articles usually require a video abstract. Reviews are a popular submission because they are usually highly cited.
Step 2. Decide on the authorship.
There are rules that define requirements of authorship that are accepted by most journals. The purpose of these rules is to prevent fraudulent padding of a CV by having as an author a person who has done no work for the manuscript. Every author should be conversant with the manuscript and be able to sign off on what has been written. Having patients in the study is not a reason to be an author. Being the Chair of a department is not a reason to be an author. In 2008 the International Committee of Medical Journal Editors published a statement on authorship requirements that included the following:
substantial contributions to conception and design, acquisition of data, or analysis and interpretation of data
drafting the article or revising it critically for important intellectual content
final approval of the version to be published.
Authors should meet conditions 1, 2, and 3.
In general, the first author is the person responsible for the idea or the one who did most of the work on the study. The last author is the senior author. Decisions on the order of authors can be political, which is a shame.
Step 2. Writing.
A manuscript has a structure, usually consisting of
Title
Abstract
Introduction
Methods
Results
Discussion and Conclusions
Acknowledgements
References
Figures and Tables
Title
A title is the most important part of the manuscript as it either attracts a potential reader or causes the reader to pass by. Many people scan tables of contents in journals, looking for articles that are interesting, topical, and informative. They see the titles and make decisions to read or not to read based on what they see. In addition, titles and key words will determine if your article appears in a google search or a PubMed search. Titles also need to be catchy, without being forced or trite. You can name your article based on the question under study (Does radiation improve the outcomes after local excision for rectal cancer?), or on the results of the study (Radiation improves outcomes after local excision for rectal cancer), or on the procedure being studied (Adjuvant radiation for locally excised rectal cancer). Sometimes there is a title and a subtitle. (Rectal Cancer: results of local excision with adjuvant radiation).
Abstract
The purpose of the abstract is to summarize the study so that readers who scan the journal can determine if they want to read the whole article. It is a mini version of the manuscript itself, governed in form by the requirements of the journal. The introduction explains the reason for the study, summarizing what is known and what is unknown and pointing out why the proposed study is important. Methods describes how the study was done, the exclusions and inclusions, the study subjects, and the primary and secondary outcomes. The results section describes details the results, describing study subjects, interventions, and outcomes. There is usually no discussion in the abstract, where the results section is followed by a conclusion. It’s important to realize that conclusions should be just that…the implications of the study results for medical practice or for the status of knowledge in general. Conclusions should not just summarize the results. Abstracts generally cannot contain abbreviations or references, and unlike meeting abstracts, figures or tables are not usually allowed.
Common Mistakes:
Too much background or methods information
Figures or images
References to other literature, figures or images
Abbreviations or acronyms
Introduction
The introduction to a manuscript explains the deficiency in the current base of knowledge that the study has been planned to correct. It presents the results of the extensive literature review that identifies the lack of knowledge, the importance of the gap to patient care of understanding of disease, and the reasons why other studies have failed to adequately address the issue. The introduction finishes by stating the aim of the study and often by generating a null hypothesis. The introduction should be concise and clear, distilling the current literature to identify the knowledge gap clearly and with focus. If the study repeats or extends a previous study, explain why the extension is needed and how it will add to current knowledge.
Common mistakes:
Too much or not enough information
Unclear purpose
Lists
Confusing structure
First-Person anecdotes
Methods
The methods section describes how the study was performed in enough detail to allow a different set of investigators to reproduce the work. It begins by defining the setting of the work (your department) and the use of an ethical committee or institutional review board to approve the proposed work. This followed by a narration of the procedures to be performed or the way in which the data will be accessed and acquired. Aspects that need to be provided depend on the nature of the study but may include the following:
Subjects
Sample preparation techniques.
Sample origins.
Field site description
Data collection protocol
Data analysis techniques
Any computer programs used.
Description of equipment and its use
The methods section also includes definitions of terms, subjects and procedures such as:
Define terms (e.g. local recurrence, patient quality of life)
Define Primary and Secondary End Points (important to provide a valid analysis of outcome data)
Define Groups and Controls
Define Inclusions and Exclusions
Define Complications
Define length of follow up
Define Statistics
Sometimes methods can include flow charts that illustrate changing numbers of patients when there have been exclusions for a variety of reasons. Often descriptive tables are provided to characterize demographics of groups of patients.
Common Mistakes:
Too little information
Information repeated from the Introduction.
Verbosity
Results/ sources of error not reported.
Statistics
The methods section usually ends with a description of how the data are to be analyzed. This usually involves listing the descriptive statistics based on the type of data analyzed, and the ways in which the data will be compared. This can be a very complicated section and if so, is better written by a statistician. Simple studies can usually be done by you. Problems with amateur statisticians include:
1 in 10 studies did not explain analyzed variables
1 in 9 studies did not describe statistical analysis
1 in 2 did not report units of measurements
Results
The results section provides the “meat” of the paper. The results. It begins by a description of the study participants, the exclusions that were necessary, and the numbers of experiments or procedures that were done successfully over what period. Potential sources of variation are listed and accounted for by various techniques. It is important to state what results are most important and here the identification of the primary end points of the study is helpful. Clarity is essential in the results section and presenting the data as tables or graphs may help. Do not describe the results twice…once in text and once in a graph. Only one presentation of the data is needed.
Common mistakes:
Verbosity
Redundancy (presenting the same data multiple times)
Discussion and interpretation of data…(don’t do this in the results section!!!!)
No figures or tables
Methods/materials reported…(don’t do this in the results section!!)
Discussion
Here is where the results of the study are interpreted in the light of the limitations and deficiencies of the study itself, and the relevant literature. It begins with a broad re-statement of the findings of the study. Did the study confirm/deny the hypothesis? If not, did the results provide an alternative hypothesis? What interpretation can be made? Next the results are set within the available literature, comparing the relative validity of the current study with that of those already published. Does the data agree with the literature? Does it address something new and provide new information or does it just confirm what we expected. If the data are at odds with what has been reported, why?
The next step is to critique your own study. What are its limitations? What are the problems with the data? What are the strengths of the study? Bearing the limitations and strengths in mind, what do the results of your study mean for patients, for knowledge, and for the future of patient care?
Common Mistakes:
Discussion combined with re-presentation of results.
New results discussed.
Making statements that are too broad.
Incorrectly discussing inconclusive results
Ambiguous data sources
Missing information
Conclusion
This is the “take home message” from the paper. It is NOT a place to summarize data but rather a place to infer from the data what should be done to improve outcomes. For example, in a study where 10 people lay out in the sun in bathing suits for 60 minutes without sunscreen and all of them got burned the conclusion would not be “lying out in the sun for 60 minutes causes severe burns”…rather it would be “do not lie out in the sun for 60 minutes without sunscreen”.
Medical Writing
For many of us, writing is a chore, and good writing is out of reach. The triaging process of High School, University, and Medical School tends to favor those with a scientific bent, while English and History majors choose other professions. Describing surgical studies in a way that is understandable and helpful is a gift, and when a gifted individual is found within a surgical department they are prized. As a student who was gifted in English and History, I will use this chapter to pass on some of the techniques that can make your manuscripts shine.
Clarity: writing a scientific article for a medical audience requires clarity above all. There is no point in confusing the readership when your primary aim is to describe a new concept, a new technique, or a new discovery. The purpose of medical writing is to transfer knowledge but if that knowledge is clouded in confusing language, then you have not served your colleagues well. The features of clear writing are precision, brevity, a well-ordered sequence of ideas, objectivity, a detached writing style, and compelling prose that invites attention. The enemies of clarity in medical writing are verbosity, pomposity, cliches, and repetition.
Correct grammar: there is little more off-putting to a reader than poor English. Unfortunately, the ability to write good English, with acceptable punctuation, appropriate syntax and correct grammar seems to be a rarity among doctors who write. While the autocorrect that is automatic in current word processing programmes helps in most cases, it is not foolproof. Some academic departments employ a copy editor to overcome this issue.
Avoid the first person: Medical writing is much better in the third person..”60 patients were consented to enter the study” rather than the first person “I consented 60 patients to enter the study”.
Don’t boast: Authors love to tell their readers that “this is the first study to report on hammer toes in left-handed American immigrants to Huntly”. Occasionally it is relevant to note that no other paper has described a single port appendectomy but there are humble ways of saying this, and bragging about it is objectionable.
Verbosity: there seems to be a belief among many authors that the longer a manuscript is the better it must be. This belief is not held by the editors of medical journals, nor by those who review submissions. However, it is widespread and causes a temptation among inexperienced writers to write more rather than less. This leads to using three words when one will do, to repeating information with different words, and the use of cliches. Two of my pet cliche peeves of which surgical residents seem particularly fond are “surgical intervention” and “operative intervention”, when a simple “surgery” would do for both. Sometimes metaphors can also be cliches such as “polyp burden”, which is less annoyingly written as “number of polyps”.
Read and re-read: critically editing your own work is an essential part of the process of producing a worthwhile paper. The first pass is rarely faultless, and the manuscript needs to be read multiple times. It is usually good to “sleep on it” and read the work again the next day. With each reading changes are made until the number of changes diminishes and the writing assumes its final edition.
Peer Review
Peer review is at the heart of quality in medical literature, guarding against inadequate, fraudulent, and even dangerous publications. In fact, the main responsibility of the Editor and the Board of Reviewers is to sort the “wheat from the chaff”; to make sure that the articles that appear in their journal contain worthwhile, helpful information that derives from well designed and well performed studies. You may be asked to serve as a reviewer for a journal. This is a compliment and a responsibility. The editor is relying on you to take a careful look at the assigned manuscript, to comment on any deficiencies that you have found, to make a recommendation about whether the manuscript should be published, and, if not, whether changes may improve it enough to be acceptable. I have found the following approach helpful.
What is the Journal? Journals are rated according to Impact Factor, the number of citations of Journal articles in the previous 2 years divided by the number of citable articles published. Authors generally submit their work to journals with higher Impact Factors for the extra prestige, the more “academic” readership, and the increased credibility of their work. In general, the higher the Impact factor the more strict are the requirements for publication. If you are reviewing for a high impact journal this is a particular compliment. Editors recruit reviewers that know something about the topic of the manuscript, selecting 3 or 4 from their pool, and from a review of the literature to see who has been publishing in the field.
Who wrote the manuscript? Look who the authors are and the Institution from whence the manuscript came. Although a high level of notoriety doesn’t guarantee that the manuscript will be acceptable, well published authors and Institutions with a history of excellence tend to produce better work. Having said that, it is important to keep an open mind when reviewing the work. Let the objective quality speak rather than preconceived expectations based on the authors and their Institution
Try and do your review in the time allowed. Editors are conscious of the need to provide a decision on the suitability of the work to the authors without undue delay. They are waiting for you and if you don’t get your review back in time the editor may let you know that you are no longer needed. You may not be asked to review for that journal again.
Read the manuscript through carefully. Is this novel work, or it is repeating prior studies? If it is not novel, does it add anything to the current state of knowledge? If it is not novel and doesn’t add anything, then it probably should be rejected. The introduction should tell you the rationale for this study. Is this valid? Is the review of what is currently known adequate? Does the method section contain enough detail to allow you to follow the conduct of the study? Has patient selection been well described? What sort of study design is used? Is it descriptive or comparative, retrospective or prospective, matched or randomized? Has randomization been done properly? Do the authors define the variables adequately? Do they define primary and secondary endpoints? The results section should describe the sample adequately and the endpoints in some detail. If there is a matching or a control group are the matching criteria adequate? Do they make sense? Have the correct statistics been used? Have the data been presented correctly? Should there be more or less tables, graphs, illustrations? If there are complex statistics outside your ability to review recommend that the editor involve a statistician to check them. The discussion should reiterate the results of the study, place them in the context of the available literature, and describe their significance in the light of this review. Then the limitations and advantages of the study should be considered, before the results are summarized and a conclusion is reached. This this a novel finding? Is it a valid finding? What are its implications for care of patients or for our knowledge of the disease involved?
Based on this review, which may take multiple readings of the manuscript, you will start to formulate an impression of the work. This ranges from” Wow, this is great stuff. We must publish,” through “this is pretty good but they could have made it better by doing these things” to “ this is terrible. They used the wrong statistics and didn’t have enough patients”. Make a list of comments, try to be helpful.
Writing the report is important and not to be rushed. You will write comments to the authors and then to the editor. The comments to the authors should be polite, even if they are critical. Please don’t make insulting or derogatory comments. These authors are your colleagues and even if the manuscript is poor you must be polite. Many reviewers point out issues that they have found under Major and Minor headings. The major issues are important and may make the manuscript unacceptable. Minor issues, if corrected, will improve the manuscript quality, but they are not deal breakers against publication. Sometimes you will have questions for the authors and sometimes suggestions. If you like something about the manuscript you can tell the authors. Positive feedback is appreciated.
When you write to the editor your job is to say what you think of the work, and whether it should be published in the journal. Whether you recommend publication or rejection, tell the editor why. You can refer to your comments to the authors but you don’t need to repeat them. If you think that the manuscript is worthy of an editorial, suggest that. If you see problems with statistics, suggest such a review. Put yourself in the position of the editor…hoping for good reviews from his reviewers.