Skip to main content

Table 1 Variables included in the selection and performance datasets and the multiple imputation

From: Would changing the selection process for GP trainees stem the workforce crisis? A cohort study using multiple-imputation and simulation

Selection dataset: GMC number, application Round, selection scores and progression through each Stage of the selection process, applicant decisions (withdrawal from the selection process and whether a post offer was accepted or declined) and personal characteristics (including gender, ethnicity and country of primary medical qualification).
Performance dataset: GMC number, date training commenced, Annual Review of Competence Progression (ARCP) outcomes (as detailed in The Gold Guide [12]), indicators of less than full time (LTFT)/Out of Programme (OOP) status, MRCGP examination performance on the AKT and CSA including date taken and score achieved relative to the pass mark and date of GP Registration (if obtained).
Used in multiple imputation as part of the prediction algorithm (available for all applications): Round of application (those applying in Round 3 in 2014 were re-coded into Round 2; data for 2010 were not available), Stage 2 scores, gender, ethnicity (coded as white/black and ethnic minority), and country of primary medical qualification (coded as UK/non-UK).
Imputed for each application where actual data were missing: withdrawal from selection process, Stage 3 total score (the sum of the competency scores across the three scenarios and written assessment), offer accept/decline, LTFT, OOP, ARCP Outcome 4, and FTE-equivalent actual time to GP Registration if no ARCP Outcome 4.