The management of the patellar articular surface at the time of primary total knee arthroplasty (TKA) is controversial. We used expected-value decision analysis to determine whether the patella should be resurfaced in TKA, and also whether secondary resurfacing on an unresurfaced patella is worthwhile. Outcome probabilities and utility values were derived from randomized controlled trials only. A decision tree was constructed and fold-back analysis was performed to ascertain the best treatment path. Sensitivity analyses were performed to determine the effect on decision-making of varying outcome probabilities and utilities. Our model showed patellar resurfacing is the best management strategy for the patella at the time of primary TKA. This decision is robust to changes in the specific data: the best path would remain the same as long as the incidence of persistent anterior knee pain (AKP) with resurfacing remains less than 29% (current mean, 12%) or the incidence of AKP after nonresurfacing falls below 12% (current mean, 26%). Delayed (ie, secondary) patellar resurfacing for ongoing patellar pain provides inferior results for the majority of patients.