LAUSR

In Response: We thank Drs Cai et al for their interest in our publication, “One-year home-time and mortality after thrombolysis compared to nontreated patients in a propensity-matched analysis.” We agree that the effects of confounding variables need to be accounted for in observational studies. We used propensity score methods to address confounding. The purpose of propensity score matching is to ensure that the groups being compared, that is patients treated with r-tPA (recombinant tissue-type plasminogen activator) and the matched patients who were not treated with r-tPA, have similar distributions of measured baseline covariates. Indeed, we showed that the distributions of the propensity scores are balanced between patients who received r-tPA and those who did not. Whether the matched nontreated patients are similar or different from the nonmatched nontreated patients is irrelevant. To make an analogy with randomized controlled trials, patients in the control arm likely have similar baseline characteristics compared to the patients in the intervention arm, but they may be very different from all patients with disease in the population. Drs Cai et al also suggest that home-time, admission to nursing homes, and all-cause mortality are insufficient to evaluate the benefits of r-tPA. Our intention was not to establish the effectiveness of r-tPA in acute ischemic stroke, as this treatment is the current standard of care. Rather, we aimed to quantify the longterm effects of r-tPA using clinically relevant and patient-centered outcomes. For instance, we chose the one-year home-time metric because returning home is likely more important to patients than radiographic outcomes. We hope that our findings provide helpful information on the long-term effects of r-tPA and will encourage the use of pragmatic patient-centered outcomes such as home-time in future trials and observational studies of thrombolysis.