Towards a doctoral thesis through published works.

Doctoral theses submitted in medical schools under a system dependent on publications (Sweden) and one which was not (UK) were compared. A subset consisting of UK theses containing papers (about 1/3 of all UK theses) was used. The publication-based theses gave candidates a significantly higher (P < 0.03) profile in terms of key authorship positions. Nevertheless, in 66% of the UK theses with papers the candidate was either the first or sole author. Swedish and UK theses with papers were of equal quality when assessed by the number of papers in journals: a) ranked in the top 100 (14% vs 10%) or 200 (26% vs 32%); or b) used more than once and either ranked in the top 1000 (median 224 vs 218) or in the top two thirds by subject section (98 vs 100%). UK theses benefitted from the greater impact of journals emanating from the UK compared to continental Europe (P < 0.001). An estimated 13% of UK PhD theses overall included three or more papers per thesis despite no requirement of publication. A publication-based doctorate should be introduced on trial in parallel with the existing systems to ensure efficiency and international comparability.