A controlled analysis of authorship trends over two decades

OBJECTIVE: The purpose was to assess authorship trends over time by taking account of publication type and controlling for number of investigating cen ters and funding status in a multivariable analysis. STUDY DESIGN: A database of 403 randomized studies and 193 controlled obser vational studies was assembled by means of a combination of electronic and hand search of 4 generic obstetrics and gynecology journals for the years 1 975, 1980, 1985, 1990, and 1995. A multivariable logistic regression model was built for evaluating the effect of time on authorship with multiauthore d articles as the binary outcome variable (articles were classified as eith er those with <6 authors or those with greater than or equal to 6). The ana lysis was performed separately for the 2 types of publications, and it was adjusted for the confounding effects of number of centers, funding status, and journal of publication. The beta coefficient (and its exponent) associa ted with the time term in the logistic model provided a measure of the tren d in publication of multiauthored articles. RESULTS: In randomized studies the odds of publishing a multiauthored artic le, given the number of centers, funding status, and journal, were increase d on average by 6% with every 5-year increment in time (odds ratio, 1.06; 9 5% confidence interval, 1.02-1.10; P = .007). Similarly, in controlled obse rvational studies, there was a 10% increase in the odds of publishing multi authored articles (odds ratio, 1.10; 95% confidence interval, 1.01-1.20; P = .03). CONCLUSION: There is an inflationary trend in authorship that is not explai ned solely by the increased collaboration between centers and funding for r esearch.