Sir: I was delighted to read Sally Pidd's review regarding the College census and plans to establish an integrated database allowing production of statistics regarding career pathways of psychiatric trainees (Psychiatric Bulletin, October 1999, 23, 630-633).

I am interested in the retention of junior psychiatry trainees in the speciality and anecdotally, have been aware throughout my training of the ease with which many good trainees have left without the opportunity to tell their stories. I recently attempted to set up an audit project to identify how many basic level trainees in one teaching hospital scheme went on to pursue psychiatric careers, what became of those who did not and importantly identify the reasons given for leaving using a design method similar to that described by Harvey et al (Reference Harvey, Davison and Winsland1998).

Using medical staffing lists I was able to identify that 59% (27/45) of new recruits to the scheme over the years 1988-1990 were members of the Royal College of Psychiatrists 10 years later. Medical staffing lists are destroyed after 10 years and I was, therefore, unable to trace back any further than 1988. Unfortunately, no information was kept on individual doctors other than an initial and surname. It was evident that these handwritten lists were incomplete and inaccurate and made identifying and therefore tracing doctors who had left impossible.

I applaud the College's efforts to collate this type of information and would hope that individual schemes could be involved in auditing their retention of trainees. I suspect that the drop-out rates decrease exponentially throughout training and would welcome efforts to monitor why trainees leave before attempting MRCPsych Part I.

Unless we answer these questions quickly we all continue to face the prospect of working in understaffed departments in the future.