Which strategy is more promising to improve the outcome of primary conventional chemotherapy in advanced Hodgkin's disease (HD): a moderate dose escalation or treatment intensification by shortening cycles? The answer generally depends on two factors: the tumour growth velocity and the chemosensitivity of the tumour. A simple mathematical model of tumour growth and chemotherapy effects was developed to quantify this dependency. The model allows to estimate the distribution of latency times (i.e., the time a tumour requires to grow from one cell to clinical detection) and the distribution of chemosensitivity in a patient population on the base of clinical data on tumour control and treatment given. The model was fitted to the data of 705 stage IIIB/IV HD patients of the German Hodgkin's Lymphoma Study Group (GHSG). The model reveals considerable heterogeneity in chemosensitivity and a significantly positive slope of the dose-response relationship at the standard treatment dose level. The model can be used to simulate the effect of various treatment escalation and intensification strategies. On the basis of such simulations we predict only small benefits (about 3% in 5-year tumour control rates) with shortening cycle intervals from 4 to 3 weeks. In contrast, we predict that a moderate dose escalation by 30% of a standard chemotherapy will lead to a potential benefit in the order of 10% in tumour control at 5 years. The presently ongoing HD9 trial of the GHSG is designed to demonstrate this effect.