Aim: To identify the predictive factors for the publication of theses defended at the Faculty of Medicine of Sousse (FMSO) in Tunisia.
Methods: The development of the dissertation was studied following the interrogation of three databases "Medline", "Scopus" and "article @INIST" as well as "Scholar Google". The drafting quality of the thesis summary was assessed using a grid of 20 iso-weighted items. The predictors of publication were studied by binary logistic regression, with a significance level of 10%.
Results: Out of the 670 theses defended at the FMSO, the mention "proposal for the thesis prize" was awarded for 22% of the thesis students. These theses were of the "clinical" type in 68% of cases, 80% of which were in "case study" format. The writing quality was deemed satisfactory in 47% of the theses. The publication rate was only 10.3%. The univariate analysis found three factors determining the publication of theses: the field of non-clinical research: epidemiological or fundamental, good editorial quality and the mention "proposal for the thesis prize". After adjustment, the latter was the only independent predictor of the publication of the thesis (ORa=1.60[1.007-2.559]).
Conclusion: This low rate of publication of theses at the FMSO illustrates the difficulties of thesis students and their directors in research methodology and scientific writing. Theses accepted with distinction must be better accompanied to facilitate their publication.