OBJECTIVES: To quantify patient and physician preferences for therapeutic tradeoffs involving efficacy, side-effect risks, and evidence uncertainty in chronic hepatitis B (CHB) treatments.  METHODS: Physicians who treat CHB patients and adult patients with a self-reported physician diagnosis of CHB completed a web-enabled, discrete-choice experiment survey in Germany and Turkey.  Both patients and physicians answered 12 treatment-choice questions.  Each question required evaluating a pair of hypothetical CHB medication profiles defined by years the medicine has been studied, probability that patient’s viral load remains undetectable for 5 years with possible reversal of disease progression, 5-year treatment-related risks of a fracture and renal insufficiency, and monthly medication cost. Nested-logit and random-parameters logit models were used to estimate preference weights for all attribute levels and the mean relative importance of each attribute.  RESULTS: A total of 158 physicians and 118 patients completed the survey in Germany.  159 physicians and 117 patients completed the survey in Turkey. German patients ranked risk of renal insufficiency as most important while German physcians ranked efficacy as most important. Turkish physicians and patients disagreed on the relative importance of all treatment attributes. Turkish patients ranked years of evidence as the most important attribute, while Turkish physicians ranked risk of renal insufficiency as most important.  German physicians were willing to accept a 0.4% greater increase in fracture risk than patients in return for an additional year of evidence, while Turkish physicians were willing to accept a 3.2% smaller increase in fracture risk than patients for an additional year of evidence.  CONCLUSIONS: This is the first study to quantify patient and physician preferences for CHB treatment attributes and the first study to elicit physician and patient preferences for years of evidence. We observe different discrepancies between physician and patient preferences in Germany and Turkey.  Such discrepancies may interfere with optimal outcomes if not considered in patient-physician interactions.