The Centers for Disease Control and Prevention established the US National Tuberculosis Genotyping and Surveillance Network to study the utility of genotyping Mycobacterium tuberculosis isolates for prevention and control. From 1998 to 2000, four sites performed conventional contact investigations and epidemiologic investigations of cases with genotypically matched M. tuberculosis isolates, called cluster investigations. The authors compared cluster pairs (two cases with M. tuberculosis isolates having identical genotypes) whose epidemiologic linkages were discovered only during cluster investigation with those whose epidemiologic linkages were discovered during conventional contact investigation. Among the 2,141 reported culture-positive tuberculosis cases, 2,055 (96%) M. tuberculosis isolates were genotyped. By itself and at a minimum, cluster investigation added 43 (38%) of the 113 total epidemiologic linkages discovered. Of the epidemiologic linkages discovered during conventional contact investigation, 29% of tuberculosis case pairs were not supported by genotyping data. The linkages discovered only during cluster investigation were more likely discovered in nontraditional settings and relationships and among larger clusters (cluster size of >5: adjusted odds ratio = 57.6, 95% confidence interval: 31.8, 104.6). Information gained from genotyping M. tuberculosis isolates should initiate cluster investigations of tuberculosis cases not previously discovered as being epidemiologically linked during conventional contact investigation. Cluster investigations will play a crucial role in predicting recent tuberculosis transmission more accurately, as we move toward tuberculosis elimination in the United States.