We studied whether an individualized digital decision aid can improve decision-making quality for or against knee arthroplasty. An app-based decision aid (EKIT tool) was developed and studied in a stepped-wedge, cluster-randomized trial. Consecutive patients with knee osteoarthritis who were candidates for knee replacement were included in 10 centers in Germany. All subjects were asked via app on a tablet about their symptoms, prior treatments, and preferences and goals for treatment. For the subjects in the intervention group, the EKIT tool was used in the doctor-patient discussion to visualize the individual disease burden and degree of fulfillment of the indication criteria, and structured information on knee arthroplasty was provided. In the control group, the discussion was conducted without the EKIT tool in accordance with the local standard in each participating center. The primary endpoint was the quality of the patient's decision on the basis of the discussion of indications, as measured with the Hip and Knee Quality Decision Instrument (HK-DQI). (Registration number: ClinicalTrials.gov:NCT04837053). 1092 patients were included, and data from 1055 patients were analyzed (616 in the intervention group and 439 in the control group). Good decision quality, as rated by the HKDQI, was achieved by 86.0% of patients in the intervention group and 67.4% of patients in the control group (relative risk, 1.24; 95 % confidence interval, [1.15; 1.33]). A digital decision aid significantly improved the quality of decision-making for or against knee replacement surgery. The widespread use of this instrument may have an even larger effect, as this trial was conducted mainly in hospitals with high case numbers.