“…We thank Blumberg et al for their letter on our recent paper on ABO nonidentical platelet transfusions and mortality. 1 Regarding their questions about the use of hospital-days and Cox regression, we sought to investigate the role of multiple transfusions given to an individual patient by focusing on a time-varying measure of exposure to identical only, major, minor, or bidirectional mismatched platelets as opposed to a "switch on" with the first nonidentical exposure that does not take into account the type and number of transfused products. As illustrated in Figure 2 in our paper, person-time was allocated in a time-dependent manner upon receipt of a subsequent platelet transfusion, with each person-time interval categorized in an escalating classification from identical only to major or minor to bidirectional.…”