An increase in the number of reported melioidosis cases was observed in the first 4 months of 2004. These cases were associated with a significant increase in case-fatality rate compared with the past 5 years. In order to exclude the possibility of a single source, including the possibility of intentional release of Burkholderia pseudomallei, we applied a multiplex PCR-based multilocus variable-number tandem repeat (VNTR) assay to determine the clonality of the clinical isolates. Our investigation indicated that a total of 30 different VNTR types could be distinguished in the 32 clinical isolates of B. pseudomallei obtained during this period, thus indicating that infection was unlikely to have occurred from a single source. Our experience underscores the usefulness of a rapid strain typing method in augmenting an epidemiological investigation into an infectious disease outbreak, particularly at a time where the intentional use of biological agents is a potential threat to public health.