"WILL IT mean an operation, doctor?" asks the nervous patient, expecting the surgeon to say yes only if surgery is absolutely essential. Yet in practice, a surgeon's decision to operate is almost always a matter of opinion; only rarely is there no alternative treatment (including doing nothing at all). Removal of a breast cancer, for example, may not be the only option for an elderly woman with a slow-growing tumour likely to respond well to hormonal treatment.

One factor being recognised more widely in the decision whether or not to operate is the effect of fashion. Surgeons tend to be optimistic enthusiasts (otherwise they would not have become surgeons), and when a new operation or procedure is introduced they hurry to learn the technique and try it out on their patients. One recent example of the rise and fall of a surgical fashion is the operation for glue ear in children, which became popular around 20 years ago but is now fading, as opinions have changed.

Glue ear is a complication of infection of the middle ear, the chamber behind the eardrum through which sounds are conveyed to the inner cochlea, where they are converted into nerve impulses. Almost all small children have one or more ear infections, and though most clear up with antibiotic treatment, in some cases the middle ear is left clogged with a sticky fluid, which makes the child partly deaf. Glue ear will clear up without treatment in most cases, but an operation to drain the fluid might be expected to speed up the recovery. In the 1970s and 1980s, parents became convinced children with untreated glue ears would fall behind at school.

This anxiety corresponded with a surge of enthusiasm among ear, nose and throat surgeons for operating on children with persistent deafness and ear discomfort. A small hole was made in the eardrum and a drainage tube or grommet inserted to keep the hole open, allowing fluid to pass out. The operation was often combined with removal of the adenoids at the back of the nose, which also helped the ear to drain through the Eustachian tube, its natural connection to the throat. Unfortun- ately, the grommets had a tendency to fall out and further operations would be needed to insert new ones.

Between 1975 and 1982 the number of glue ear operations on children doubled in many parts of England. But then research studies began showing that when children were allocated at random to surgical treatment or observation (waiting and seeing), there was no difference in the eventual outcome. The results of these studies percolated through to parents, general practitioners and surgeons, and the operation became far less frequent.

In an ideal world, surgeons would not embark on new treatments until clear evidence was available from controlled trials. In reality, though, such trials are so expensive and so time- consuming that they are unlikely to be organised until an operation has already become fairly popular.