Leukemia after Treatment of Ovarian Cancer or Hodgkin's Disease

To the Editor: When the results of case–control studies differ dramatically from the results of large cohort studies, it is important to keep in mind the fundamental differences in methods that affect the reliability of their inferences. In the case–control method, the relative risk of disease resulting from exposure to a hypothetical cause must be estimated indirectly. Furthermore, because samples for comparison are selected retrospectively in case–control studies, there is less assurance than in cohort studies that case patients and controls do not differ except in their exposure to the hypothetical cause. Even though they are clinically relevant, the matching . . .