You have convinced me! GP health including prior sickness absence, ethnicity and available pension also seem relevant. J At 11:53 04/02/2014, you wrote: >Given that it seems to be multifactorial we probably need a >QResignation or QVacancy formula for NHSE to assess local risk. > >I would suggest factors such as patients/GP, remuneration/patient, >age, sex, changes in income, population density > >Started writing this tongue in cheek, but I am gradually convincing myself!