Why Some Medical Papers Move Us

I have loved reading medical papers for as long as I have practised medicine. Having worked mainly in obstetrics and gynaecology for nearly half a century, I have read them continuously, often amid heavy clinical workloads. Beyond the need to acquire knowledge, reading papers has long been a personal refuge. Even on the busiest days, it offered quiet satisfaction. Some papers moved me deeply and remain vivid in my memory. Until recently, however, I had not reflected on why certain papers affected me so strongly. The recent emergence of artificial intelligence (AI) in medical writing prompted me to reconsider this question—not in terms of efficiency or accuracy, but in terms of why we read and write medical papers at all. Looking back, the papers that moved me most fall broadly into three types. First are papers that overwhelm us with data and effort. Large randomised controlled trials that provide decisive knowledge belong to this category. The well-known trial by Kramer and colleagues [1], involving more than 17,000 participants, demonstrated a positive effect of breastfeeding on children's cognitive development and effectively settled a long-standing debate. Such papers command respect through their scale and clarity. Second are papers that move us through conceptual originality. Jauniaux and colleagues [2] proposed that placenta percreta—an extreme form of placental invasion—may represent a disease of the uterus rather than of the trophoblast. This was not merely a technical refinement but a genuine shift in perspective, akin to a small Copernican revolution within the field. Third are papers that move us because they are exceptionally well written and carry a distinctly human tone. These are often thought-expressing pieces—Perspectives, Opinions, or Letters—in which the author's voice is clearly present. They do not simply inform; they resonate. I find a parallel in novels. Type-1 papers resemble The Second World War by Winston Churchill: monumental in content and invaluable as a historical record. Type-2 papers remind me of Frankenstein by Mary Shelley, with its distinctive structure, built on interwoven letters and nested narratives, and its conceptual originality. Type-3 papers bring to mind The Old Man and the Sea by Ernest Hemingway, where each carefully chosen detail quietly and deeply moves the reader. What these papers offer is not only knowledge, but experience. We may forget the finer details of their content, yet the memory of having been immersed in their richness lingers. Such encounters encourage continued reading and, at times, a desire to conduct research or to write. In this sense, medical papers are not merely vehicles of information; they are part of what sustains medicine as a human endeavour. How, then, might AI influence this relationship between medical papers and our motivation to read and write them? Much of the current discussion focuses on visible advantages and risks—speed, accuracy, or error. Less attention has been paid to how AI may affect human satisfaction and values, which are less visible but no less important. The impact of AI is likely to differ across these three types of papers. For data-driven trials, AI may help present complex findings more clearly and efficiently, without diminishing their value. For concept-driven papers, the question is more open: originality largely arises from human experience and intuition, though AI may eventually contribute in unexpected ways. For thought-expressing pieces, however, the issue is more delicate. The tone, taste, and presence of such writing should remain genuinely human. Even when prose is polished, many readers can sense whether a piece carries a human voice [3]—a judgement that lies beyond scientific reasoning. One day, AI may be able to generate writing that closely resembles the style of great authors. If such writing becomes commonplace, we may need to ask what, exactly, we value when we read. Are medical papers merely objective records, or do they also possess something closer to an art? Medical papers ultimately exist for patients, and accuracy must always remain paramount. This is not a binary choice, nor should one element be placed categorically above another. It is a question of balance. At the very least, we should continue to respect the idea that medical papers—particularly thought-expressing ones—carry an artistic aspect [4], and that genuinely human writing deserves preservation. The author received no specific funding for this work. The author has nothing to report. The author declares no conflicts of interest. ChatGPT was used solely for language editing; all ideas, interpretations, and final wording are my own.