In order to use the EQ-5D questionnaire with bolt-on dimensions in economic evaluation studies, new value sets are needed. In this study, we explored the feasibility of a new approach called scaling factor model, which estimates bolt-on value sets using estimated EQ-5D dimensional weights. We designed a two-arm study, inviting university students to value health states with and without bolt-on items using the composite time trade-off (cTTO) method. We selected 25 health states from an orthogonal array and added 5 mildest EQ-5D states in the design. In arm 1, EQ-5D without self-care and standard EQ-5D states were valued, and in arm 2, standard EQ-5D states and EQ-5D with vision were valued. By arm, we compared the mean observed values of health states with and without bolt-on item. Next, by arm, we estimated value sets for the EQ-5D with bolt-on item states using both standard model and scaling factor model. Model performances were compared in terms of prediction accuracy and correlation with likelihood-based mean values. Adding a fifth-level bolt-on to EQ-5D resulted in statistically lower values. This effect is consistent across two arms and bolt-on items. The scaling factor models outperformed the standard models in all statistics. The scaling factor model offers a methodologically viable and low-cost option for producing value sets for EQ-5D, supplemented with bolt-on items. Future studies should further test this method using other bolt-on items and relevant study populations.