Translation, Treatises, and Tweets
Address correspondence to David S. Pisetsky, MD, PhD, Durham VA Medical Center, Box 151G, 508 Fulton Street, Durham, NC 27705. E-mail: email@example.com.
As the pursuit of truth, science is a quest of risk and adventure especially as the subject of inquiry is life itself, the most complicated of all things. This pursuit is one of mankind's greatest achievements, demanding a hearty soul, a drive for creativity, and lots of moxie. The search for scientific truth is hard. Experiments fail. Surprises abound. Exhilaration alternates with frustration and dead ends seem more common than the open highway. The rewards are high, however, and science is the foundation of the modern world.
Given the excitement inherent in science, a record of its performance should be the stuff of great writing, producing rousing tales of high drama and roller coaster emotion. With “Eureka” the exultant cry of discovery, the story of getting there should be as stirring and suspenseful as a fine novel, with the scientist, like Ahab, shipping off in search of the Great White Whale. The story of flops—clinical trials missing their endpoints, P values maliciously ascending above 0.05, knockout strains pooping out—should also inspire poignant tales resembling that of the gifted athlete whose dreams of stardom are dashed by a blown-out knee or a face shattered by a screaming line drive.
If only that were the case and science writing was as rich and vibrant as the pursuit of science itself. Instead, the opposite seems to be true. Science writing, whether of the basic, translational, or clinical variety, unfortunately rarely stirs the soul as literature. Despite its functionality, this writing is notable for uniformity, even dreariness in style, rivaling that of a legal brief in its inherent pleasure as art. Personality and humor are lacking and the passive voice prevails, hiding the “I” or “we” who actually did the experiments.
The presentation of science in journals seems at variance with actual research. Few discoveries stem from the testing of hypotheses or rational searches. More often, scientists blunder into their discoveries by that venerable circumstance called serendipity. To cite just one example, the discovery of the toll receptors emerged from the cages of fruit flies not the culture of macrophages, the logical place to look for a receptor for lipopolysaccharide . After the fact, scientists create hypotheses to order (and often, reorder) their data, making their acquisition more “scientific” than it really was.
The Technology of Communication
In the conduct of biomedical research, communication is essential, with modern technology allowing it to soar. Indeed, the information content available for easy access is exploding because of the digital revolution and the extraordinary power of today's computers. A library can fit on a memory stick and War and Peace can show up as an e-mail attachment. This power, when applied to scientific publication, is shaking up the world of journals, changing the ground rules by allowing an unprecedented increase in the potential size of any paper (whether text or appendix). Propelling the move to online publication, this trend is pushing print versions precariously to the brink.
The implications of this revolution are enormous. Old print technology imposed sharp limits on the size of journals (and therefore individual papers) because of the high cost of printing. To have affordable journals with an adequate number of quality papers, editors had to enforce limits on the size of each in terms of word count as well as figure size and number. At present, such limits are fading since electronic publishing has few inherent constraints. The sky is the limit with respect to the amount of data and verbiage that can fill each paper, with appendices containing extraordinary amounts of data, often dwarfing those actually presented in the journal.
An opposite trend is always at play. Modern electronic media encourages communication packaged in small bites for instant delivery on a BlackBerry or iPhone. People crave communication via social media, thriving on their immediacy. In an era of texts, tweets, and twitters, a full-length paper is a massive tome poorly suited for the easily distracted minds of busy people or the miniscule screen of a hand-held device. Since these trends are only going to intensify, I think that it is time to consider the state of science writing and explore changes to make it more modern and stimulating. In this article, I will focus on language as well as the content of the introduction and discussion section. The challenge of reporting the voluminous data that current science generates is a topic for another day.
Language of Translation
The language of science is not set in stone but rather reflects its time and place in history. As such, a key force shaping science writing today is a growing focus on translation; in this context, translational research is directed to the development or improvement of new diagnostics, prognostics, and treatments. Certainly, the National Institutes of Health is moving in this direction, shifting from more basic to more applied approaches. This trend is having a strong impact on the conduct of science in academic institutions, which previously emphasized basic discovery, with the anticipation, even confidence, that someone else (i.e., industry) would pick up and carry the ball into the clinic.
By its nature, basic discovery science is marked by flashes and even skyrockets of insight and is well suited for cheering and hurrahs. The focus is the moment, the immediate exhilaration of finding something really new, something that no one else knows. Since such science does not care directly about relevance or consequence (do the travails of fruit flies fighting fungi really matter?), it lends itself to more lively and expressive thinking and writing. The “wow” factor can be high. After all, the toll receptors got their name when the gene was found and someone shouted “Das ist ja toll,” which means “That's great!”
Now, just about everyone is in the translation game and even traditional basic research is being reoriented and rebranded as translational. As a result, academicians play in the same arena as their colleagues in industry, with relevance to a disease setting or potential for practical application (no matter how far distant in the future) often a criterion for publication and, in turn, grant funding. In basic research, knowledge for knowledge's sake is often the goal as well as the outcome. In the realm of translation, however, the outcomes and measures of success are more restricted, often reduced into one of two categories: biomarker or therapeutic target. This restriction strongly determines the framing of a study and its conclusions. Describing every up-regulated protein or gene as a target or biomarker, however, is not that informative even if it is correct. Furthermore, with many papers pointing to a new treatment or a biomarker (to diagnose, stage, or subset patients), the pipeline is already overflowing with information, with few ideas on how to use it productively.
Structure of the Introduction
In fiction writing, the first sentence is often the most important, engaging the reader from the earliest moment possible with arresting words that grab the heart and mind. Every course in writing provides the classic examples like “Call me Ishmael” or “This is the saddest story I have ever heard.” These are great sentences because they come from craftsmen who understand the need to put a hook into the reader and pull him or her along into the story that follows.
Sadly, in science, the first sentence, and, for that matter, the first paragraph, do not function as hooks nor aspire to novelty or invention. Rather, the introductory material uses words and phrasing that are so standard that the danger of plagiarism (self or other) is constant. Like many investigators, I struggle to find different ways to say exactly the same thing, reconfiguring the words, adding an adjective here or there, fearful that some computer program will decide that the words I penned appeared somewhere else.
Reading like entries on Wikipedia, the current introductions often act as a space-filler, part of the formula to tell the reader what the paper will be about in a general way and demonstrate that the author has done his or her homework about the subject. Certainly, these introductions do not spark much interest or signal the arrival of a new talent. Wouldn't it be great to start a paper with a variant of Dickens (rather than me or another lupologist) and begin a paper with these words: “Whether I shall turn out to be the discoverer of something new about lupus, or whether that station will be held by anybody else, these pages must show.”
In addition to its uniformity, the current approach to composing an introduction is characterized by a viewpoint that reflects the focus on translation in contrast to basic discovery science. Thus, to heighten a paper's seeming importance, authors frequently resort to certain statements to position the new findings. At the top of the list is the proposition that prior understanding of a subject is limited, poor, or even woeful in such a way that progress is hampered. While an understanding of rheumatoid arthritis (RA) has a long way to go, as of the writing of this article, there are 114,504 PubMed citations on RA, suggesting that perhaps the understanding is better than “poor.” In fact, we have enormous knowledge.
Another way to position research, especially about a subject that could lead to a new target or treatment, concerns an assessment of current therapy. As such, authors have to accord current therapy the least credit possible, highlighting its inability to cure, its transient or uncertain benefits, and the high level of side effects. For many conditions, especially many cancers or neuropsychiatric disorders such as schizophrenia or autism, decent therapy is truly lacking and this construct is appropriate. The situation with rheumatology, however, appears different in a way that makes some of the language on current therapy disingenuous. The last two decades have witnessed remarkable progress especially in the treatment of inflammatory arthritis. Therapy is better now than ever before and remission in RA is a realistic goal. Nevertheless, reading introductions to many papers on arthritis, it is often unclear whether the year is 1953 or 2013 and whether the current therapy is gold or anti–tumor necrosis factor (anti-TNF). We are in a brave new world of therapy and should acknowledge that fact in framing studies.
While some authors will give current therapy a better grade, they nevertheless raise the issue of cost as a rationale for research, seeking an approach that will somehow be cheaper than current agents. Cost-saving is, of course, worthwhile but experience shows that the price of a medication is set by the marketplace not the cost of production, complexity of the synthesis, or the pathway inhibited. With current drug pricing, economics does not seem a compelling argument for translational studies, especially those involving very complicated interventions such as cell transfers of ex vivo expanded populations.
As often practiced, this mode of writing produces an introduction of the following kind, which I created for this purpose although I could have easily used it for a paper I am writing. “RA is a chronic inflammatory disease that results from a combination of genetic and environmental factors. This disease causes severe disability and leads to early mortality. Despite extensive investigation, the pathogenesis of this disease is poorly understood, limiting the development of new therapies. Although recent advances in treatment with aggressive disease-modifying antirheumatic drug approaches have improved outcomes, nevertheless, many patients do not have adequate responses, remission is rare, and serious side effects such as infection limit treatment. Furthermore, biologic agents are expensive, preventing their widespread utilization. Therefore, there is an urgent need for new pharmacologic agents targeting novel pathways.”
Please note the use of the word “urgent,” which I see commonly in papers as a further sign of the serious state of affairs. The road to translation is long and arduous and, since the development of new drugs is prone to failure, the need will likely remain urgent far into the future. Furthermore, unmet needs in medicine will always be there because of side effects or inadequate responses even for an otherwise very effective agent. Better therapy is therefore always needed, a point that does not require constant restatement.
Following such an introduction, many current papers go on for 6, 8, or even 10 paragraphs (often enough to fill 2 journal pages) explaining the basis for the work at hand. By the time the reader gets to the Materials and Methods section, fatigue, not to say a sense of déjà vu, has likely set in. As someone who has written ponderous introductions for decades, I would like to propose a more parsimonious writing style that is short and sweet, getting to the bottom line quickly, albeit not as quickly as a tweet. I would cite an analogy to the elevator pitch that students at business schools practice: 1 or 2 minutes of rapid-fire delivery to jolt a prospective client to make an investment or purchase by the time the elevator stops and the doors open.
For my elevator pitch introduction, I suggest a maximum of 2 or 3 paragraphs. I would also suggest reducing references to the shortcomings and cost of current therapy since everyone making the pitch will say exactly the same thing. As in a political campaign, negative ads can get a candidate only so far and, like voters, readers want to hear the positive. Personally, I am more likely to read on when the science seems novel and is presented with a little zip. At this point, we all get the point about translational research. It does not need endless repetition.
Structure of the Discussion
Like introductions, discussions are getting longer and longer especially as the current technologies, especially the “omics” kind, produce staggering amounts of data, the veritable “billions and billions” of bits that make papers like dark and dusty archives more than enjoyable expeditions. While the discussion of such huge amounts of data is itself a big problem, I will consider something else that results from pressures on scientists in a competitive environment. Thus, authors have to engage in a kind of self-promotion to declare that their work is significant beyond the usual assertion that the studies show something for the first time. For the usual scientist who is modest and tentative by nature and fears overstatement, this type of self-promotion may be unsettling. Nevertheless, to catch the attention of the reviewer, editor, and reader and eventually a granting agency, the language can get fevered and inflated.
In the discussion section, many authors start off with a sentence of the following kind: “The findings presented herein provide important new insights into the pathogenesis of RA.” The sentence is intriguing in the implication that the current paper accomplished something that the previous 114,504 citations failed to do: lift RA from out of the mire of poorly understood diseases. (By way of disclosure, I have contributed more than my share of such sentences and I have sometimes suggested that my new insights are not just “important” but are actually “striking.”)
A related issue concerns the authors' effort to extend the relevance of their findings beyond one experimental system or disease focus. Consider the language I encountered in an outstanding paper on the role of an ion channel in neutrophil chemotaxis . While this paper involved an elegant analysis of signaling, the authors nevertheless noted that the mechanism described could affect “the development of restenosis, diabetic microvascular disease, chronic allograft rejection, pulmonary hypertension, as well as neurological disorders.” In many respects, this assessment is correct since immune-mediated diseases involve pathways that operate generally, with neutrophils likely involved in all inflammatory diseases. Nevertheless, if juxtaposed with an assertion about the “poor understanding” of pathogenesis in the introduction, positing such a broad applicability may be excessive, however tempting. (I should note that the authors of the paper on chemotaxis framed their paper in an admirable way, emphasizing the basic biology.)
Usually, in science, the significance of a finding is learned over time, with extensive confirmatory work needed to determine whether the findings reported indicated that the paradigm has shifted or merely been budged or tweaked. I am therefore wary of any author's claim about great significance from findings in a single paper, especially one of the first on a subject. From my perspective, I like understatement of which Watson and Crick's is one of the most famous: “It has not escaped our notice that the specific pairing we have postulated immediately suggests a possible copying mechanism for genetic material” . Imagine how that sentence would be elaborated today.
Given the amount of data in current papers, discussions are now like treatises as authors try to explain the myriad of findings that emerge from a gene array or genetic study. In many instances, the authors can do little more than speculate, trying to establish some kind of link with a currently proposed mechanism for pathogenesis for one or a few of the hundreds or even thousands of up-regulated genes found in an array. I think that the value of such discussions is limited unless a finding provides a clear and compelling clue to a mechanism or indeed shakes and rattles a paradigm. The impact of a discussion can dwindle if the author tries to be too comprehensive and consider too many findings especially if they have not been confirmed.
One place where I think more rather than less discussion would be better concerns papers on clinical trials. In my experience, these discussions tend to be shorter than those for more basic or translational papers and much more circumspect in their conclusions. Since studies of trials may be associated with applications to the Food and Drug Administration or a comparable agency, sponsors prefer a more conservative approach. In this regard, I am particularly disappointed in the manner in which the issue of safety is approached. Who would have thought that people can get along without TNF-α, interleukin-6, or peripheral B cells and not succumb to infection more readily? I would like to read more about these interesting experiments of nature so to speak, since the lessons learned for basic biology are likely to be great.
To counteract any further growth of discussions, I would also recommend an elevator pitch rule for that section of the paper, although I would allow a few extra floors in comparison to the introduction. Given the duration of an elevator ride, time is of the essence and writers should learn to use their minutes wisely and deliver their messages in as incisive and trenchant a way possible. In this regard, given the state of discussions, it is not surprising that editorials are now so popular. The style of editorials is looser, the language is freer, and the title (which almost serves as a first sentence) shows creative spark. (I could have titled this paper “Trends in Language Utilization and Data Presentation in Scientific Publication: Implications for Translation Research,” but I doubt you would have looked at it). Furthermore, editorials summarize the key concepts of a paper concisely and highlight the shortcomings as well as the strengths in a way that a paper's authors would be reluctant to do. If an editorial can provide an enlightening perspective in about 1,000 words, so should a well-constructed and engaging discussion.
At present, science writing is challenged by the volume of data that must be presented and the pressure to place these data in a framework that emphasizes the dyadic outcomes of translation: biomarker or therapeutic target. In science, there are many other outcomes that are valuable: unexpected and unusual findings, a fresh way to think, and the sheer excitement to find about how a leukocyte navigates to the joint or a cell crumbles during apoptosis. New knowledge is special and needs attention and visibility whatever its consequences.
Since I have criticized a tendency to be negative, I would like therefore to be positive and make the following recommendations:
- Shorten all papers and place limits on the size and content of the introduction and discussion.
- Reduce the data content of papers, including the number of figures. A figure with 20 panels is not one figure. It is 20 figures. Limit appendices in size and scope or make the data freely available without a necessary link to a paper.
- Encourage the use of the first person in writing.
- Put more personality in the writing. Adjectives are good, although I am not talking about using “gnarly” to describe a complicated dose-response curve. Use the thesaurus function on your word processor to come up with something different.
- Make the discussion section more like an editorial and add some punch and controversy. Tell me the upside as well as the downside and why a target may disenchant and your biomarker fizzle.
- Emphasize fundamental discovery more and translation less.
- Describe abnormalities discovered as clues to pathogenesis. As anyone who likes crime fiction knows, clues are very important since they can reveal the perpetrator and weapon.
- Frame studies in terms of biomarkers only when there can be a genuine attempt to operationalize the finding into a useful test.
- Be cautious in postulating findings as the basis of therapy. Blue sky thinking is fine but clouds and rain showers are always in the forecast.
- Discourage hype and overreach. While inevitable in competitive times, they can detract from the spirit of science, which knows well the enormity of the task.
While these recommendations are modest, I think they represent a good first step and provide the basis for energizing science writing and making it better. And, to violate one of my own rules, I think that better writing is something that is urgently needed.
Note: As an author, I am guilty of everything that I criticize. As an editor, I have been a traditionalist leery of informal language or colloquialisms. I like the use of the first person, however, and am vehement on having two spaces after a period.
Dr. Pisetsky drafted the article, revised it critically for important intellectual content, and approved the final version to be published.
This article had its origin in a lecture entitled “Reflections on Medical Writing,” which I gave as the Wallace Epstein Visiting Professor at the University of California, San Francisco. I would like to thank Dr. David Wofsy for inviting me to deliver this lecture and Dr. Wallace Epstein for encouraging original thinking and for very illuminating and provocative discussions.