A Decision Tree Model: To Treat or Not to Treat Dental Caries in Primary Teeth

Objectives: The decision to leave decay or restore primary teeth is debated in the literature. The purpose of this study was to develop a prediction model (decision tree) to determine whether to restore or leave caries in primary teeth given the probabilities associated with future events.
Methods: An extensive review of the literature was conducted to obtain probabilities of outcomes associated with dental disease among children. A decision tree model was created using TreeAgePro© 2015 software to determine the optimal treatment for deciduous teeth with dental caries; “to restore” or “not to restore.” The following variables and outcome probabilities from the literature were used in the decision tree: leave tooth unrestored with infection (0.35), leave tooth unrestored with no infection (0.65), leave tooth unrestored with severe pain (0.50), leave tooth unrestored with moderate pain (0.17), tooth exfoliation with no pain (0.33), failed restoration (0.73), restore tooth and infection occurs (0.06), restore tooth with severe pain (0.13), restore with moderate pain (0.05) and restore tooth with no pain (0.82). Unlike previous models, the probability of “failed” and “successful” restorations were added to the model. Terminal branches of the decision tree reflected financial costs for hospitalization to manage dental pain and/or infection ($459), simple restoration on a deciduous tooth ($55) and exfoliation of a deciduous tooth with no pain ($0).


Results: The model predicted restoring a tooth has a 27% retention rate and a 73% fail rate. Even with a failed restoration, the probability of infection was 6%; whereas, leaving a tooth unrestored resulted in a 35% chance of infection.
Conclusions: When considering probabilities of failed or successful restorations, this prediction model indicated restoration of primary teeth as the best strategy.