MORE THAN 25 YEARS HAVE PASSED SINCE THE original description of intensive care unit (ICU) telemedicine, a technological strategy to improve critical care outcomes by expanding the reach and availability of intensivist clinicians. Since then, the understanding of evidence-based practice and the role of information technology in the ICU have substantially increased. Multiple commercial applications of ICU telemedicine now exist, and telemedicine is widely touted as an all-encompassing strategy to improve ICU outcomes. Yet even after 25 years, the optimal role of telemedicine in the ICU remains uncertain. A large, multicenter study published recently showed no demonstrable clinical benefit, and a recent meta-analysis found no beneficial association between ICU telemedicine and in-hospital mortality. These results have left clinicians, hospital administrators, and policy makers wondering how to best use this technology, if at all. In this issue of JAMA, Lilly and colleagues report a single academic center investigation that is of similar size and scope as the previous multicenter study, but has dramatically different results. Lilly et al examined the outcomes associated with implementation of ICU telemedicine among 6290 adult patients admitted to 7 ICUs using a stepped-wedge time series design. Tele-intensivists acted in complement with local clinicians to enforce daily goals, review adherence to evidence-based practices, and respond to bedside alarms. Admission to an ICU operating under the telemedicine model was associated with increased receipt of evidence-based preventive strategies and lower rates of ICU-acquired complications. Hospital mortality, ICU length of stay, and duration of mechanical ventilation also decreased. Importantly, telemedicine was associated with lower mortality both within ICUs over time and across ICUs during the same periods, strengthening the inference that the results are not due purely to time trends. The current study by Lilly et al and the previous study by Thomas et al included roughly the same number of ICUs and hospitals, used the same proprietary telemedicine system, and were both directed by academic leaders in the field. So how can these discordant results be reconciled? In the study by Thomas et al, there was limited physician buy-in and less than one-third of the patients received the full discretion of the telemedicine team for clinical care. In contrast, the telemedicine team in the study by Lilly and colleagues was allowed full discretion for all patients. Increased discretion may have led to more interventions directed at improving outcomes. Perhaps more importantly, the telemedicine program in the study by Thomas et al was not linked to any specific quality improvement programs. Instead, the teleintensivists provided mostly remote monitoring, intervening only when necessary—such that care via telemedicine was reactive rather than proactive. In contrast, the telemedicine program in the study by Lilly et al was tightly linked to specific quality improvement activities. The teleintensivists conducted real-time best-practice audits for selected evidence-based practices, reviewed care plans and daily goal sheets, and ensured adherence with the local intensivists’ plans. These interventions worked, as evidenced by the patients in the telemedicine program being more likely to receive evidence-based practices for prevention of ICU complications, demonstrating a plausible mechanism for the decrease in mortality. For this reason, the study by Lilly et al provides the first convincing evidence that ICU telemedicine can be an effective complement to bedside care in some settings. However, this study is not without limitations. The size of the adjusted mortality reduction is implausibly large, leaving open the possibility that some of the apparent benefit is due to differences in how severity of illness was measured between the 2 study groups. Moreover, the 7 ICUs in this study are part of one relatively well-resourced academic medical center that has a strong culture of quality improvement. It is unclear if these results could be replicated in hospitals with fewer resources to devote toward ICU quality. Additionally, all of the telemedicine physicians also worked in the target ICUs, which may have served to increase buy-in among local practitioners. These