Author instructions of haematology and oncology journals: A cross-sectional and longitudinal study




Poster session 1 Wednesday: Evidence production and synthesis


Wednesday 13 September 2017 - 12:30 to 14:00


All authors in correct order:

Toews I1, Binder N2, Wolff R3, von Elm E4, Meerpohl J1
1 Cochrane Germany, Germany
2 Institute for Medical Biometry and Statistics, Medical Center - University of Freiburg, Germany
3 Kleijnen Systematic Reviews Ltd, UK
4 Cochrane Switzerland, Switzerland
Presenting author and contact person

Presenting author:

Ingrid Toews

Contact person:

Abstract text
Background: The debate about the value of biomedical publications led to recommendations for improving reporting quality. It is unclear to what extent these recommendations have been endorsed by journals.

Objective: We analysed whether specific recommendations were included in author instructions; which journal characteristics were associated with their endorsement; how endorsement of the domains changed; and, whether endorsement was associated with change of impact factor between 2010 and 2015.

Methods: We considered two study samples consisting of 'haematology' and 'oncology' journals of the Journal Citation Report 2008 and 2014, respectively. We extracted information regarding endorsement of the recommendations of the International Committee of Medical Journal Editors, of reporting guidelines, requirement for trial registration and disclosure of conflicts of interest. Data extraction was done by reading the author instructions and conducting a text search with keywords. We calculated a global generalised linear mixed-effects model for endorsement of each of the 4 domains followed by separate multivariable logistic-regression models and a longitudinal analysis. We defined endorsement as the author instructions saying that they approve the use of the recommendations.

Results: In 2015, the ICMJE recommendations were mentioned in author instructions of 156 journals (67.5%). CONSORT was referred to by 77 journals (33.3%); MOOSE, PRISMA, STARD and STROBE were referred to by less than 15% of journals. There were 99 journals (42.9%) that recommended or required trial registration, 211 (91.3%) required authors to disclose conflicts of interest. Journal impact factor, journal start year and geographical region were positively associated with endorsement of any of the 4 domains. The endorsement of all domains increased between 2010 and 2015. The endorsement of any domain in 2010 seemed to be associated with an increased impact factor in 2014.

Conclusion: Haematology and oncology journals endorse major recommendations to various degrees. Endorsement is increasing slowly over time and might be positively associated with the journal's impact factor.