For practical decisions on common recurring maintenance actions, the information from routine inspections form a decision basis for the bridge manager. It is often difficult to assess whether this information is sufficient for deciding on a repair action, or if more information is needed. For many bridges the information for supporting decisions may be limited, although those bridges cause large yearly maintenance costs for the society. The purpose of the study presented in this paper is to show how two different models for decision making based on Bayesian decision theory, a point-in-time decision model and a sequential updating decision model, can be used to improve the decision-making process for common maintenance decisions. The models use information from routine inspections and incorporates time dependent aspects such as material degradation and time value of money to improve the decision-making process. The focus is on presenting the methodology with a case study of a concrete bridge in Sweden where the edge beams may have to be replaced. Three assessment approaches are considered: (i) no assessment, (ii) desktop evaluation and (iii) measurements. The main finding is that sequential updating decision making will provide a higher benefit than a point-in-time decision, and thus give higher Value of Information. This value becomes even higher when the measurements are selected for the assessment. The results also show that the edge beams should be replaced. The general approach presented can be applicable to many decision scenarios related to maintenance of deteriorating structures.