I sometimes find myself reflecting when my well-cited journal articles begin to show a decline, whether gradual or sudden. In academic medicine, citations have become a form of currency, used not only to trace ideas but also to measure worth. Yet what gets cited is not always what matters most. In biomedical research, citations support arguments and help measure impact. In principle, they show how knowledge develops over time. In practice, however, the way citations accumulate and decline for a given research article is far less straightforward. Over the years, certain patterns begin to emerge, with important implications for how work is evaluated in academic medicine, shaping promotion, funding, and professional recognition.
The hidden biases of search engines and accessibility
Researchers now depend heavily on search engines. What appears first is more likely to be cited. Newer papers, along with those that are already well cited, tend to be selected more often. Papers from widely recognized journals may also be more visible. Familiarity still plays a role, but visibility through search engines has become increasingly important.
While writing a paper, citations are often added where needed. One identifies where references are required and fills those places with appropriate sources. When several papers could serve the same purpose, the choice is influenced by how recent they are, their visibility, and, last but not least, their availability and convenience. With much of the literature now digitized and often placed behind paywalls, even work that was once freely available, the original source is not always selected. I often find it difficult to access my own work without paying to download it, somewhat ironic, but true. Experienced writers generally know what should be cited, yet in practice they often write first and insert references afterward.
How time and review articles obscure original research
Older papers are cited less frequently, even when they remain valid. There is a tendency to avoid older references, assuming they are outdated or that their content is already well known. Review articles influence how citations are made. They shape how a field is presented and often become the default reference in place of multiple original papers. Over time, the original work itself may become less visible, even when it remains foundational.
There are also cases where earlier work is repeated with some modifications and published again. When an idea begins to take hold, it may appear in different forms, sometimes with incremental additions. Such work can accumulate citations, some of them self-citations, while the original contribution is cited less. At the same time, repetition in research is not inherently problematic. Reproducibility and revisiting earlier work with new methods or perspectives are essential, helping to confirm, refine, or even challenge original results. The concern arises when the original work is sidelined and citations shift disproportionately toward repeated or derivative versions.
The transactional nature of modern citation metrics
Before citation indexing became widespread through the work of Eugene Garfield in the 1960s, there was less emphasis on counting citations, and scientific recognition depended more on contribution and reputation. Over time, citations became measurable and increasingly central to evaluation. The concept of “citation classics,” introduced by Eugene Garfield, contributed to the emergence of highly cited scientists as markers of recognition within the scientific community. This also introduced a more transactional dimension to the assessment of research. Citations are now widely used in evaluating work, and they inevitably influence behavior.
When a new journal is launched, citation metrics quickly become a major concern in establishing its standing and achieving recognition as a top-tier publication. Editors and their staff work to improve these metrics, not unlike the pursuit of television ratings. This dynamic is not widely appreciated outside the publishing community. Citation metrics are often promoted as indicators of influence, much like ratings for ongoing television programs or serials.
Re-evaluating the true value of foundational research
Journals also contribute to this process indirectly. When a manuscript reaches an editor or section editor, the selection of reviewers is often guided by systems that identify researchers working in the area, frequently based on citation patterns or the reference list of the manuscript. Reviewers, upon receiving a manuscript, may check whether their work has been cited. This may not deliberately influence their decision, but it can still affect how the manuscript is perceived. In my own experience, these patterns become increasingly visible over time. For many in academic medicine, they shape not only how research is disseminated but also how careers evolve.
A decline in citations does not necessarily indicate a loss of importance. It may reflect changes in how citations are selected and how the literature is accessed. It also reminds us that important contributions can continue to shape later work even when they are cited less often. Such shifts do not diminish the value of earlier work; rather, they change how it is seen and engaged with over time. Giving due recognition to original contributions, even when relying on review articles, can help promote more careful and balanced citation practices.
Rao M. Uppu is a professor of environmental toxicology and chemistry.
