A study by the Thames Cancer Registry highlights the confusion among doctors over the best way of treating breast cancer.

Researchers studied the hospital records of 334 women living in the four Thames regional health authorities in south east England who were diagnosed with invasive breast cancer early in 1990.

They concluded that there was a "lack of consensus" among clinicians over the best way to manage breast cancer and call for more patients to be included in clinical trials so that the success rates of the various approaches can be compared.

They found that doctors routinely ignore the treatment recommendations set out in guidelines published by the King's fund consensus conference in 1986. (These guidelines were produced after earlier research found wide variations in treatment between London hospitals.)

The results show that liver and bone scans and skeletal x-rays were often carried out.

"This is surprising, since the guidelines advised against the routine use of such investigations," say the researchers.

On the other hand, although guidelines strongly recommended axillary sampling (minor exploratory lymph gland surgery in the armpits) to diagnose how far advanced the cancer was, this was used in less than half of cases.

The researchers found that many hospitals failed to keep detailed case notes; three quarters of them had no information on the stage of the disease in individual cases.

In the quarter that did have at least some information about the stage of the illness, many contained no laboratory evidence to support the doctor's diagnosis.