To compare fitness for work assessment outcomes and written advice across current and trial medical statements. To examine the use of and suggestions to improve the trial medical statement.

Comparative study with a two-way mixed design using questionnaire-based vignettes presenting GPs with three hypothetical sick leave cases (back pain, depression, combined back pain and depression) and medical statements (current or trial). The questionnaire also gathered GP views of using the trial Med 3.

Nine primary care organisations (PCOs) in England, Scotland, and Wales.

Five hundred and eighty-three GPs employed in PCOs in summer 2008 were randomised to receive a current or trial Med 3 postal questionnaire. GPs assessed vignette patients' fitness for work using the questionnaire medical statements.

GPs using the trial Med 3 were less likely to advise refraining from work and more likely to provide written fitness for work advice compared to GPs using the current Med 3 form. Date sections of the trial Med 3 were used inconsistently, and a return to work date was unclear. GPs wanted further clarification of the implications of assessing a case as ‘fit for some work’ and its relationship to employers' willingness to follow GP advice about work.