Coauthor Country Affiliations in International Collaborative Research Funded by the US National Institutes of Health, 2009 to 2017

IMPORTANCE The US National Institutes of Health (NIH) is the largest funder of biomedical and behavioral research in the world. International collaborative research—a subset of NIH’s portfolio—is critical to furthering the agency’s health research mission. OBJECTIVE To quantify the extent of the NIH’s international collaborations and the relative importance of this research through the lens of publications. DESIGN, SETTING, AND PARTICIPANTS This cross-sectional study used bibliometric data from the Web of Science database to analyze trends in the growth of NIH-funded publications from January 1, 2009, to December 31, 2017, and examined their importance using national affiliations of all coauthors listed, h indices, and citation impact scores. All countries with coauthor affiliations in NIH-funded publications during the study period were included. Data were analyzed from October 22 through November 16, 2018. EXPOSURES Country affiliations of coauthors’ institutions in NIH-funded publications indexed in the Web of Science database from 2009 to 2017.


Introduction
The US National Institutes of Health (NIH) is the largest funder of biomedical research in the world. 1,2ring fiscal year 2019, the NIH used its $39 billion budget to support scientific research in pursuit of improving human health and basic science. 3The NIH funds research at non-US and domestic institutions; much non-US-affiliated research is conducted with funding awarded to US investigators who collaborate with non-US-affiliated scientists or trainees. 4The NIH recognizes that science knows no borders and values the contributions of international collaboration for quality biomedical research.6][7] Clinical trials for infectious diseases, such as human immunodeficiency virus, Ebola virus, dengue, and Zika virus, for instance, require scientists in the United States to work with investigators in affected regions.][10][11] Rapid advancements in communication technology and the trend toward team-based research have facilitated such international collaborations, allowing scientists to freely exchange data and ideas. 12,13 assess the importance of international research at the NIH through one lens of scientific productivity-peer-reviewed publications-we evaluated the extent of NIH-funded international collaborative research and characterized trends during the past 9 years.5][16][17] By comparing trends in the number of publications over time and a metric of scores, such as the h index, researchers can describe the relative importance of publications with coauthors affiliated with different countries.To answer larger questions about the importance of US and non-US coauthorship overall, the category-normalized citation impact (CNCI; a normalized metric that describes how frequently an article has been cited compared to other articles in its discipline) 18 adds another dimension to the analysis.
We have conducted a bibliometric study to assess the growth of NIH-funded publications and characterize the scope of international collaboration supported by the NIH during a 9-year period when the NIH budget increased from approximately $30 billion to $35 billion but experienced an inflation-adjusted decrease. 19We evaluated the CNCI and the number of publications overall as a function of coauthors' country affiliations to achieve these goals.Insights into the global reach of NIH funding is critical to understand the current research landscape and inform future investments.
Collaborating institutions, including those from academia, government, and the private sector, can use these findings to guide their own decision-making to build productive international research partnerships.

Methods
The Web of Science (WOS; Clarivate Analytics) database was queried for NIH-funded publications using a comprehensive Boolean search string of the full names and acronyms of all NIH institutes and centers in the "funding text" field.We limited our analyses to the 9-year period from January 1, 2009, to December 31, 2017, because WOS only began indexing funding data in 2008, and records for 2018 were incomplete at the time of this analysis. 20The WOS "analyze results" feature was used to sort records by countries or regions and download aggregate counts of NIH-funded publications by country affiliation of coauthors.8 The CNCI used baseline citation rates by WOS subject area centered around a score of 1.00, with higher scores representing greater importance.All statistics were descriptive and presented counts and proportions of publications in the sample without reference to population parameters.The Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) reporting guideline was used in the reporting of this cross-sectional study.This analysis of publicly available publication data did not constitute human subjects research and did not require approval by an institutional review board or informed consent.
Data were analyzed from October 22 through November 16, 2018.The citation report feature in WOS was used to calculate 1-year h indices for each country affiliated with NIH-funded publications in 2017.The h index is a metric that reflects the number of publications and citation counts in a single number. 22For example, an author's h index would be 20 if he or she had published at least 20 articles that had been cited at least 20 times each.3][24] Previous bibliometric studies have found 1-year h indices to be highly correlated with those from larger citation windows, making the statistic a valid tool for comparison. 25,26The data were limited to publications in 2017 to match the time frame for annual publication volume.
Countries were ranked by number of publications and h index in 2017.Finally, records for all NIH-funded publications from 2009 to 2017 were sorted by citation count.

Results
From 2009 to 2017, the number of NIH-funded publications increased 46.2% from 67 041 to 98 002 (Figure 1).In 2017, the coauthors most represented in NIH-funded publications were from China, the United Kingdom, Canada, and Germany (Figure 2A).Although all countries experienced increases in their annual publication rates, the greatest increases were from publications with coauthors from China (from 1976 to 6982; 253.3%), the United Kingdom (from 2733 to 5582; 104.2%),Canada (from 2614 to 4260; 63.0%), and Germany (from 2059 to 3780; 83.6%).China was the most frequent country affiliated with publications that included US and non-US authors (n = 6982).With the exception of China, the numbers of authors from low-and middle-income countries (LMICs) all experienced sustained lower growth in the same time (Figure 2B).
To examine our most significant global partnerships in research, we identified the country affiliations most represented among NIH-funded publications in 2017.These were ranked by the total number of publications and their importance as measured by the h index (Figure 3).Of the top 20  The same analysis was repeated using only LMICs (Figure 3B).The h indices for China and Brazil were 50 and 39, respectively, ranking them as the top 2 LMICs by publication output and h index in 2017.Uganda and Tanzania, the only 2 low-income countries ranked among the top 20 LMICs for publication output, did not fall among the top 20 countries by h index.Similarly, Nigeria did not rank  Of the top 20 countries ranked by h index, 12 were European countries and 4 were high-income

JAMA Network Open | Global Health
East Asian and Pacific countries as classified by the World Bank. 21Internationally collaborative publications tend to be most productive and have greater h indices and CNCI among European partners and Canada as well as various high-income Asian countries and China.China stood out for its most rapid growth in the same period and as the only LMIC among the top 10 represented countries.
Marked increases have occurred in NIH-funded publications among LMICs, especially in Africa and Latin America, although these rates still lag behind rates of the HICs.

Limitations
This analysis has several potential shortcomings inherent to the methods used.Databases such as WOS do not index all publications and thus cannot provide a complete data set for bibliometric analysis. 14,54However, of the various databases available, WOS captures the most publications in biomedical research with full funding information (29%) and accordingly was used for this study. 54rthermore, publication data, including funding sources and coauthor country affiliations for each study, can be missing or inaccurate if entered into the database incorrectly. 14,55All authors in the database had country affiliations listed, but publications not captured by the database could nevertheless limit the internal and external validity of the study.The metrics themselves-h index in particular-must be interpreted in context and with the recognition that the country of publication can influence these statistics. 14,24The analysis also does not account for the balance or authorship order of US-affiliated and non-US-affiliated coauthors on a given report.Finally, coauthors' contributions do not factor into the analysis, eliminating any potential differences among the distribution of those who write the manuscripts, for instance.The distribution of coauthors' roles by country and country income merits further research.In addition, national affiliation does not necessarily imply different geographical locations for coauthors.International fellows working at NIH-funded facilities in the United States still qualify that publication as an international collaboration if those coauthors maintained their affiliations with non-US institutions.Future research on this topic should consider investigating the different roles and breakdown of authorship by country affiliation.
Previous studies have addressed some of these concerns but often on the institutional level rather than on a global scale.Developing a more robust profile of NIH-funded work at this level may help scientists and policy makers support important biomedical research.

Conclusions
International collaborative research offers investigators new opportunities to maximize importance and interact with non-US investigators by sharing opportunities from populations having unique features and genetics, exposures, or health care delivery and bringing new technologies to emerging or existing health problems.As biomedical research continues to grow, scientists will need to work together across borders to solve increasingly complex health issues.Improving human health across the globe depends on it.

Figure 2 .
Figure 2. Annual National Institutes of Health-Funded Publications for 2009 to 2017

Figure 3 .
Figure 3. National Institutes of Health-Funded Publications and h Indices for 2017

JAMA Network Open | Global Health
Additional search strings were used to query the WOS database for publications by institutes and centers and create reports on coauthors' national affiliations as

Downloaded From: https://jamanetwork.com/ on 09/17/2023 described
above, using 2017 as the sample year.Publications with funding from multiple institutes and centers were included in each relevant report.InCites (Clarivate Analytics), a WOS tool, was used to calculate the CNCI for all NIH-funded publications in 2017 grouped by coauthors' national affiliations (US coauthors only, non-US coauthors only, or US and non-US coauthors).
most represented country affiliations, 16 were classified by the World Bank as high-income countries (HICs), 3 as upper middle-income countries, 1 as a lower middle-income country, and none as low-income countries.Of the top 20 LMIC affiliations, 12 were upper middle-income countries, 6were lower middle-income countries, and 2 were low-income countries (Uganda [301 publications] and Tanzania [151 publications]).Using the h index as the primary metric for ranking the importance of a country's publications, however, elevated the United Kingdom (h index, 66) and Germany (h index, 59) to the top of the list (Figure3A).Canada (h index, 56) remained third from the top, and China (h index, 50) moved down to the sixth place by h index.India, South Africa, and Taiwan no longer appeared among the top 20 countries, replaced by Austria (h index, 34), Norway (h index, 31), and Singapore (h index, 31).In other words, the latter countries' publications had a greater importance by this metric despite totaling fewer publications overall.
among the top 20 countries in the same group.Russia (h index, 30), Colombia (h index, 22), Iran (h index, 21), and Malaysia (h index, 17) ranked substantially higher in the list by h index.

Table .
Top 10 Most Cited National Institutes of Health-Funded Publications With Non-US Coauthors, 2009-2017 38breviations: HIC, high-income-country; LMIC, lowand middle-income country.aAllpublicationsexcept for that of Uhlén et al38included at least 1 US-affiliated coauthor.Data in parentheses indicate percentage of authors from LMICs.