Conventional estimates of survival from cancer may be too pessimistic, say doctors in Germany.

The expected survival rate from a particular form of cancer is vital information for those who've recently been diagnosed. There are two main ways of estimating survival rates. The more conventional approach is called the cohort method, and it is based on survival experience of patients who may have been diagnosed many years previously.

The cohort method does not take into account recent improvements in detection and treatment. So it may give an unduly pessimistic view. Researchers at the German Centre for Research on Ageing, Heidelberg, Germany have shown this is the case. They've compared cohort analysis with another approach, called period analysis. In period analysis, one looks at survival data from a specific time period, like a particular year.

The researchers compared five, ten, 15 and 20 year survival rates for a range of cancers, obtained from cohort and period analysis. The cohort analysis involved patients diagnosed between 1978 and 1993, while the period analysis was for 1998.

In all cases, period analysis gave better survival rates. For instance, the 20 year survival rate - for all forms of cancer - was 11 per cent higher when using period analysis. Similar adjustments are found for breast, thyroid, testicular and many other forms of cancer. The study suggests that maybe newly diagnosed patients should be made aware of how survival data depends on the method of calculation - this could prevent them from being too pessimistic.

Source

The Lancet 12th October 2002