The inclusion of indirect medical costs is a topic of ongoing discussion in the literature. Actual practice seems to be to include only medical costs of so-called related diseases. However, this criterion is not unambiguous since health gains in life years gained may also depend on unrelated medical care. OBJECTIVE: To compare different ways to include both direct and indirect medical costs and health effects in cost effectiveness ratios. METHODS: Smoking cessation interventions were evaluated using a dynamic population model, the RIVM chronic disease model. This is a multistate transition model that links prevalence of risk factors to the incidence of 28 chronic diseases. Three different cost-effectiveness ratios were compared: 1) all health effects were ascribed to the smoking cessation interventions while only costs of smoking related diseases were taken into account; 2) only the minimum gain in QALY's and life years that can be attributed to the interventions were included while only costs of smoking related diseases were taken into account; and 3) all health effects and all health care costs in life years were included. RESULTS: Ratio 1) equals €2650 per QALY gained. Exclusion of health effects on competing diseases increases the ratio to €3600 (ratio 2). Finally, if all costs and effects are included, ratio 3) equals €8560 per QALY gained, which demonstrates that the cost-effectiveness ratio increases enormously when health care costs of competing diseases are taken into account. CONCLUSION: The large differences in outcomes urges one to think about the interpretation of cost-effectiveness ratio's and what ratio to use. We argue that for the evaluation of preventive interventions in a population model, the third ratio is the best, since it seems impossible to isolate the precise effects of an intervention.