Compare Topics

Comparing topics helps researchers choose between proposal directions, identify which research areas are gaining or losing NIH attention, and understand how funding landscapes differ across closely related subfields. The side-by-side view is most useful when you are deciding between two viable framings of the same underlying question, not when you are trying to rank unrelated topics.

Look at the overlay charts to see whether one topic is growing faster than the other. Then check the institute and mechanism breakdowns to understand who funds each area and how they fund it. A topic dominated by R01s at NCI requires a different application strategy than one funded through U01s at NIGMS, even if the award counts look similar on the chart.

For the most meaningful comparisons, pick related but distinct terms (for example, "CAR-T therapy" vs "checkpoint immunotherapy") and keep the year range identical. Broad terms like "cancer" will return too many results to support strategic decision-making, and mismatched year ranges will create apparent differences that are actually just reporting-window artifacts.

Comparing two topics is valid when both terms describe work that is plausibly funded by the same institutes through similar mechanisms. Comparing "single-cell transcriptomics" and "spatial transcriptomics" is valid because both sit within the same genomic methods space and both attract similar reviewers. Comparing "Alzheimer disease" and "RNA splicing" is less useful because the populations of institutes, mechanisms, and reviewers hardly overlap.

A comparison also loses meaning when one term is much broader than the other. If you place "immunotherapy" against "anti-PD-1 immunotherapy," the broader term will dominate every chart because it subsumes the narrower one. When this happens, the right move is to pick two narrower terms that are genuine peers, not a parent and a child term.

Finally, take caution when the two topics differ meaningfully in their typical lag profile. A fast-moving method topic may show publications and new awards within a year of a breakthrough, while a slower-moving clinical topic may take three to five years to show a comparable response. A side-by-side chart flattens that difference and can make a slow-moving field look inactive when it is simply on a different cycle.

The most common misread is treating a one-cycle difference as a real trend. NIH funding has enough variance that one topic can look like it jumped or dropped in a single year purely due to sample noise. Use at least three years of data before calling a difference real, and compare against the broader funding base for the institute rather than against zero.

A second misread is attributing all of a topic's growth to scientific momentum when some of it reflects terminology drift. A topic that shifts its preferred keyword over time can look like it is shrinking in searches that use the old wording. When you see a topic declining sharply, verify with a related-term search before drawing conclusions.

A third misread is comparing totals without normalizing for the overall NIH budget trajectory. If both topics are growing at 5% per year and the overall NIH budget grew by 5% in the same window, neither topic is actually gaining share. Always keep institutional context in mind when drawing strategy conclusions from a chart.