The Relative Citation Ratio in NIH iCite

How iCite works out an RCR

NIH’s Office of Portfolio Analysis described the method in PLOS Biology in 2016. It finds each paper’s field from its co-citation network: “the other papers that appear alongside it in reference lists”.

1. 1

   Count the paper’s citations per year since it was published.

2. 2

   Find its field from the papers cited alongside it, and that field’s citation rate.

3. 3

   Work out the expected rate: the citations per year that NIH-funded papers with the same field citation rate, published in the same year, receive.

4. 4

   Divide the paper’s rate by the expected rate. The result is its RCR.

The 2016 paper used R01-funded papers as its benchmark. iCite’s current RCR “uses all NIH-funded publications as a benchmark”. Because the field and the year are built in, iCite marks the RCR as field-independent and time-independent, and the h-index as neither.

What an RCR of 1.0 means

In iCite’s words, “Any paper with RCR 1.0 has an RCR higher than 50% of NIH-funded papers”. iCite also holds many papers that were not NIH-funded. Across all of them the median RCR is 0.37, so 1.0 sits between their 70th and 80th percentiles.

iCite’s NIH percentile ranks one paper’s RCR among NIH-funded papers: “95% indicates that this paper’s RCR is higher than 95% of all NIH funded publications”. So an RCR above 3.81 puts a paper in the top 10% of NIH-funded papers.

One paper or a set of papers

For a set of papers, iCite shows the mean, median and weighted RCR. The weighted RCR is the sum of the papers’ RCRs, and the mean divides it by the number of papers. The median is the middle paper’s RCR, so one very highly cited paper does not move it.

iCite’s example: 100 papers, one with an RCR of 101 and 99 with an RCR of 1, give a weighted RCR of 200, a mean of 2 and a median of 1. A set with more influence than average has a weighted RCR higher than its number of papers, and a set with less has a lower one.

The 2016 paper names two limits. Its method “was designed to assess neither the productivity of individual researchers nor the quality of their science”. And no citation metric, RCR included, “should be taken as a substitute for the actual reading of a paper in determining its quality”.

When a paper has no RCR yet

A new paper needs time to gather citations. iCite gives no RCR to papers from the current or last NIH fiscal year unless they have 5 or more citations, and it marks RCRs for papers from two years ago “provisional”. Its year moves with the NIH fiscal year every October. In iCite’s example for October 2026, papers from 2026 get a provisional RCR once they have 5 or more citations, and all papers from 2025 get a provisional one.

Clinical citations and APT

iCite’s Translation module counts citations from clinical articles, “generally defined as Clinical Trials or Clinical Guidelines” and found by PubMed’s publication types. These citations take time: “bench-to-bedside translation takes several years,” so recent papers have fewer of them.

Approximate Potential to Translate (APT) “uses the scientific community’s reaction to an article to estimate the likelihood that the knowledge from that paper will be used and cited in later clinical articles”. It takes one of five values: 0.05, 0.25, 0.50, 0.75 or 0.95. On average, 75% of papers with an APT of 0.75 are cited by clinical articles.

Look up your papers in iCite

iCite finds papers by their PubMed ID (PMID), so it covers papers listed in PubMed.

1. 1

   Open iCite at icite.od.nih.gov.

2. 2

   On the New Analysis screen, paste your PubMed IDs, separated by spaces, commas or semicolons. A spreadsheet column pastes as it is. You can also search PubMed or upload a file of PubMed IDs.

3. 3

   Click Process.

4. 4

   Read the Influence tab for each paper’s RCR, NIH percentile and citations per year, and the Translation tab for APT and clinical citations. Export downloads the table as a spreadsheet.

iCite takes up to 50,000 PubMed IDs at a time and updates its data every month.

How ScienceDex shows iCite’s figures

Impact reads the PubMed IDs in your CV and gets iCite’s figures for those papers in one click. ScienceDex sends iCite your papers’ PubMed IDs and nothing else. The figures join your CV as one change, which Undo takes back.

Its groups include these, each figure with a sentence on what it means.

• Compared with the field: papers with an RCR of 1.0 or more, the median RCR, and papers in the top 10% by RCR, at an NIH percentile of 90 or more.
• Clinical citations: papers cited by at least one clinical trial or guideline, and papers with no such citation yet whose APT is 0.75 or more.
• Citations: the citation count, the h-index and both per career year.

Each paper’s line gives its citations, its RCR with its NIH percentile, its clinical citations and your author role. A source line names iCite and the days of its data, and the Impact report prints the same figures for a promotion packet. Papers in your CV without a PubMed ID get no figures from iCite.

Faculty privacy sets out who can open your CV.