Specific tools of decision analysis, a set of mathematical rules for simplifying complex decisions, were applied to evaluate the impact of secondary leukaemia on the selection of initial treatment in Hodgkin's disease (HD). For this purpose, a combined 'expected utility' considering survival, relapse free survival, and secondary leukaemia was determined for different treatment strategies. Our analysis revealed that considerations of secondary leukaemia for initial therapy should include the a priori estimation of all possible events which may occur after initial treatment, e.g. the probabilities of recurrence and success of salvage therapy. In early and intermediate stage HD, for example, the minimal risk of leukaemia after successful radiotherapy (RT) must be weighed against the increased risk after treatment failure and subsequent salvage therapy. Thus, the difference of expected risk of leukaemia between RT and combined modality treatment (CMT) is within 4% for HD, stage II B and near to 0% in stage III A. In advanced stage HD, the addition of RT to chemotherapy has no adverse effect on the expected utility of initial treatment. These conclusions are only marginally affected by reported differences in rates of recurrence, salvage success, and secondary leukaemia. Subjective quality of life considerations, such as the latency period between treatment and leukaemia and patients' attitudes towards the occurrence of leukaemia, did not significantly affect expected utilities. In summary, our results strongly suggest that presently there is no sound basis for reducing the intensity of initial treatment in HD to avoid secondary leukaemia.