Skip to main content
The BMJ logoLink to The BMJ
. 2003 Aug 30;327(7413):509.

Growing algorithms from seed

Colin Douglas
PMCID: PMC188417

A couple of weeks ago, along with half the trade across the United Kingdom, I braced myself for our annual influx of brand new preregistration house officers. Last month's students became, however tentatively, young colleagues on the ward.

The feast of the innocents passed uneventfully, as it usually does. No serious cause for concern and, as always, a process of sussing out the new batch: who's good, who's worried, who might just turn out to be a worry in their own right.

Worry comes into it a lot, as anyone who looks back honestly at their own first weeks in paid medicine will admit. So much is new and uncertain, so little to be taken for granted. Anything from a drug dose to a consultant's mood swing can bring disaster, and each patient is an unknown, so we worry.

We don't go on like that. We learn the basics, and as the years go by we worry less and even think less. The simple truth—and very much not for circulation beyond the strict confines of this column—is that most consultants don't have to think very much at all.

Instead, and each within our own specialties, we run every new case through a subcortical algorithm, built up over years of experience in the diagnosis, management, and outcomes of the common, the common variations, and the rare. To run a personal, continuously developing but essentially pre-conscious programme of “fine-grained probabilistic clinical reasoning,” as a researcher once helpfully described it to me, is not lazy or careless, but rather efficient and generally reliable: the hallmark of the seasoned professional.

Of course we must always be open to the truly exceptional, the case that—less than once a year or more—should rouse us from thalamic medicine. Not good to miss the things sent to try us; and hence of course our disproportionate zeal for the gee-whiz grand rounds case: the ordinary looking chap who turns out to be harbouring a grade 5 fascinoma.

But such stuff is rare, and probably best kept from our youngest colleagues, whose task for the moment is to work and look and listen and read and learn: a lot about what is normal, a great deal about what is common, and how such things are recognised, managed, and turn out. It is an investment for the long term: growing, as it were, one's own algorithms from seed.


Articles from BMJ : British Medical Journal are provided here courtesy of BMJ Publishing Group

RESOURCES