The target students were from the cohorts who had entered the course in 2000-2004 inclusive and had demonstrated unsatisfactory progress at any stage or who had left prematurely. Several methods were used to identify these students, none of whom had been included in our previous study [1
> The Minutes of the Academic Progress Committee
> Inspection of Part I examination marks (average over Years 1 and 2) for students scoring below 50%, and of Year 5 examination marks for students failing their Finals at first attempt.
> Lists compiled by the Clinical Sub-Deans of students who might require support or additional monitoring at the Foundation Deanery during their first post-graduate year
> Archived files of students who had failed to graduate
The course progress file of every student identified was then searched manually and all potentially relevant data extracted into a customised database. The data included:
> Available pre-admission factors. These were age at course entry, sex, fee status, declared disability, academic qualifications (number of A levels at Grade A, B or other), and type of offer made after interview (unconditional, conditional, reserve, or late).
> Key markers of course progress or difficulty in each year. These included exams or modules failed, extenuating circumstances brought forward, problems with health, disruptive personal/family/social difficulties, attendance at the Academic Progress Committee (APC), and interviews with senior pastoral care staff or Clinical Sub-Deans.
These categories of information were recorded as yes/no options with supplementary text boxes. We did not routinely include actual examination marks but recorded the number of exams failed in Years 1 and 2. Because some students had frequent low marks without actual fails, we did note the mark for Part I, which is a weighted average of all Year 1 and Year 2 summative assessments.
Additional File 2
provides a general description of the course.
> Summary options at the end of each year/course phase up to the end of Year 4 (normal progress, progress with resits, repeat year required, student left the course voluntarily, or course terminated)
> Final summary outcome for the final year (graduated BMBS, graduated after repeat exams, graduates after repeat final year, or failed to graduate)
> Additional fields for any repeated years
> The occurrence of any behaviour or attitudes that might constitute a 'concern' as currently defined by the Faculty, eg inappropriate behaviour to others, lateness or non-attendance at teaching, lack of commitment (see Additional File 1
). In view of the findings of Wright & Tanner (2002) and Papadakis et al (2005) we included failure to complete mandatory Hepatitis B vaccinations by the end of Year 2, without undue reminders, as a discrete category of unprofessional behaviour.
> Overall outcome summary. This was a semi-objective classification made after data extraction, and contrasts with our previous paper which used an all-inclusive definition of struggling [[1
]. We categorised students as:
○ 'struggler' with multiple problems throughout the course
○ 'preclinical' - problems largely confined to the early years
○ 'clinical' - problems largely confined to the later years
○ 'health-related' - problems largely related to ill health
○ 'borderline performance' - weak student, generally low marks throughout
○ 'no substantial problems'. Some students who were identified, for example, via APC attendance, had actually suffered only a minor or one-off drop in performance, and were subsequently eliminated from the database
○ left the course voluntarily
○ course terminated
Further variables were generated as required during the analysis, to create 'flags', and these are described in the Results section.
After checking and cleaning of the Access database, the file was transferred into SPSS v17 for descriptive analysis of categorical and yes/no fields. Text fields were handled by drawing up Reports in MS Access which were then scanned visually for indicators of difficulty (eg academic failures, disruptive physical or mental health problems), to generate semi-qualitative data.
In a secondary analysis, the data for students who failed to complete the course were examined, to determine why they left the course or failed to graduate, and this will be reported separately.