A Significant Decline in the American Domination of Research in Gastroenterology With Increasing Globalization From 1980 to 2005: An Analysis of American Authorship Among 8,251 Articles

Abstract

To determine whether the relative American contribution to gastroenterology research, as determined by authorship in journals, decreased from 1980 to 2005. Country of residence for authors was manually determined for 8,251 articles, encompassing every gastroenterologic article published in 1980 and 2005 in nine gastroenterology and four leading medical journals (8 American and 5 European journals). Country of residence was also determined for all editorial board members for the same years in the same journals. In all eight analyzed American journals, the relative frequency of non-American authors significantly and sharply increased from 1980 to 2005, and of American authors correspondingly decreased. The effect was invariant: (a) regardless of journal field: general medicine (e.g., N Engl J Med, from 10.1% in 1980 to 27.2% in 2005, odds ratio [OR] 3.34, OR confidence interval [CI] 2.40-4.62, P < 0.0001) versus general gastroenterology (e.g., Am J Gastroenterol, from 20.8% in 1980 to 51.9% in 2005, OR 4.29, OR CI 2.81-6.55, P < 0.0001); (b) regardless of journal gastroenterology subspecialty (e.g., Hepatology vs Endoscopy); (c) regardless of journal circulation: high circulation (e.g., Ann Intern Med, OR 3.50, OR CI 2.05-5.96, P < 0.0001) versus lower circulation (e.g., J Clin Gastroenterol, OR 6.88, OR CI 2.86-16.5, P < 0.0001); (d) regardless of journal impact factor: high impact (e.g., N Engl J Med, see above) versus lower impact (e.g., Dig Dis Sci, OR 3.06, OR CI 2.08-4.51, P < 0.0001); (e) regardless of article topic within a journal (e.g., colon vs upper gastrointestinal tract); (f) regardless of article length within a journal; (g) regardless of number of authors per article or order of authorship; and (h) regardless of whether international representation on the journal's editorial board increased (e.g., Am J Gastroenterol) or decreased (J Clin Gastroenterol) from 1980 to 2005. Contrariwise, in three of five analyzed European journals, the opposite effect occurred: the relative frequency of non-American authors statistically significantly decreased from 1980 to 2005 (e.g., Aliment Pharmacol Ther, from 87.9% in 1980 to 75.9% in 2005, OR 0.43, OR CI 0.20-0.94, P < 0.033), and of Americans authors correspondingly increased. The same trend generally held for the editorial boards of analyzed journals: the relative frequency of non-American board members significantly increased in five American journals (e.g., Am J Gastroenterol, from 4.2% in 1980 to 31.3% in 2005, OR 10.5, OR CI 1.68-63.5, P= 0.006), and decreased in all five analyzed European journals. Interstate and transnational collaborations (coauthorship) increased in publications from 1980 to 2005 (e.g., transnational, from 2.8% in 1980 to 14.1% in 2005 in N Engl J Med, OR 5.62, OR CI 3.25-9.73, P < 0.0001). From 1980 to 2005, American representation in American gastroenterologic journals significantly declined, whereas American representation in European gastroenterologic journals moderately increased. The latter finding, together with increased transnational collaboration and increased international representation in editorial boards, suggests that increasing globalization of information, partly due to computers and the Internet, may be an important etiologic factor. This may result in an improving quality of research performed outside America. Other potential etiologic factors include improving standard of living outside America and increasing clinical volume of academic gastroenterologists in America.