connected with intensivists at the University Hospitals of Cleveland (Grundy, et al. ). After this proof-of-concept report, however, ICU telemedicine gained little traction for nearly 20 years, until Johns Hopkins Hospital established a continuously monitored ICU telemedicine service in a nonintensivist staffed surgical ICU. Their pre/post analysis suggested a 64% decrease in severity-adjusted ICU mortality and greater than 30% decrease in ICU length of stay, ICU complications, and costs (Rosenfeld, et al. ).
Along with better and less costly telemedicine technology, rapid adoption of electronic medical records, and a nationwide intensivist shortage, this and other evidence for the service’s clinical and cost effectiveness has spurred explosive growth in ICU telemedicine in the succeeding 2 decades, with at least 18% of hospitals and 28% of ICU beds supported by ICU telemedicine by 2018 (Ofoma, et al.).
Importantly, what “ICU telemedicine” represents varies substantially across hospitals and even across ICUs within systems. Two-way audiovisual technology is the defining feature, and at a minimum, programs provide intensivists and/or nurses who respond to consultation requests. Commonly, telemedicine clinicians directly connect with patients; monitor labs, hemodynamics, and alarms; and proactively contact on-site clinicians with recommendations or place orders directly into the electronic health record depending on whether the clinician acts as the patients’ primary, co-managing, or consultant provider. A centralized hub and spoke model with telemedicine personnel located at a single, remote center is the most common and best studied ICU telemedicine design. Additional staffing may include respiratory therapists, pharmacists, and advanced practice clinicians in coverage models that range from 24/7 to nocturnal and can also differ in whether patients are monitored continuously or on an as needed basis, triggered by alarms or clinician/nursing concerns.
On-demand services may extend to support for teams responding to medical emergencies inside and sometimes outside the ICU. Another equally important role that ICU telemedicine can provide is helping ensure facilities adhere to ICU quality metrics, such as ventilator bundles, DVT prophylaxis, and daily SAT/SBT.
Unsurprisingly, integrating ICU telemedicine into an existing system is very costly and complex, requiring substantial and thoughtful process redesign to maximize fiscal and clinical return on investment. One vendor of proprietary telemedicine technology, Philips eICU, estimates an implementation cost of $50,000 to $100,000 per bed with annual overhead, software maintenance, and IT staffing of ~20% of implementation costs in addition to clinician staffing of $1-2 million per 100 beds. However, some (but not all) evidence suggests that ICU telemedicine programs pay for themselves over time. An influential report from Sentara Healthcare, an early adopter of ICU telemedicine, described equipment costs of more than $1 million for a total of 103 critical care beds but attributed savings of $460,000 per month to decreased length of stay (Coustasse, et al.).
Cost savings are great, of course, but ICU telemedicine’s potential to improve clinical outcomes is the real priority. While Sentara’s early report included a 27% decrease in ICU mortality after telemedicine adoption, a 2011 meta-analysis of 13 studies, including 35 ICUs and over 40,000 patients, suggested decreased ICU mortality and LOS with a statistically significant effect on overall hospital mortality and LOS (Young, et al.). This highlights the Achilles heel of ICU telemedicine evidence: the pretest/posttest studies that dominate this field and likely contribute substantially to the inconsistencies in the evidence base.
In the absence of risk adjustment and control groups, many studies observed postimplementation changes that may reflect trends in patient mix or the effects of unrelated practice changes rather than the causal influence of ICU telemedicine. In fact, in studies using more robust methods, ICU telemedicine’s effect size has been smaller or nonexistent. For example, in 2016, Kahn and colleagues used CMS data to evaluate 132 ICU telemedicine programs using 389 matched controlled hospitals. There was a slight reduction in 90-day mortality (OR=0.96, CI 0.94-0.98) with only 12% showing a statistically significant reduction in mortality. Interestingly, hospitals in urban areas demonstrated greater benefit than rural facilities (Kahn, et al.).
The heterogeneity of the studied programs (e.g., primary vs consultative role, on-demand vs proactive involvement) and recipient ICUs (e.g., rural vs tertiary care facility, presence of bedside intensivists) further hinders a clear answer to the key question: Would ICU telemedicine benefit my hospital? Fortunately, some recent, well-designed studies have attempted to understand which attributes of ICU telemedicine programs provide results and which ICUs will see the most benefit. In a cohort of 118,990 patients across 56 ICUs, four interventions were associated with lower mortality and reduced LOS: (1) evaluation of patients within 1 hour of ICU admission, (2) frequent leadership review of performance data, (3) ICU best practice compliance, and (4) prompt response to alerts (Lilly, et al.). Kahn and colleagues have also investigated this issue, conducting an in-depth ethnographic evaluation of 10 hospitals identified in their 2016 study to have positive, neutral, or negative outcomes after ICU telemedicine implementation (Kahn, et al. ). They found that successful programs:
(1) provided consistent services matched to recipient needs;
(2) provided services both proactively and reactively without being obtrusive;
(3) embedded routine engagements unobtrusively into usual routines;
(4) had engaged leadership who set clear expectations and mediated conflicts; and
(5) had bedside clinicians who valued and sought out telemedicine participation in care.
The authors concluded that, “the true value of ICU telemedicine lies not in whether the technology exists but in how it is applied.” However, another recent analysis also suggested that, rather than telemedicine or recipient ICU design, targeting underperforming recipient ICU performance may be the key determinant of whether ICU telemedicine implementation improves outcomes (Fusaro, et al.). While the finding may reflect regression to the mean, the idea that ICUs with above-expected mortality derive greater benefit from ICU telemedicine support than already well-performing ICUs is certainly logical.
As COVID-19 strained health care systems across the country, we and others found ways to use ICU telemedicine to preserve optimal care delivery for critically ill patients. Our program at Intermountain Healthcare – already supporting 17 ICUs within our 24-hospital health system, as well as 10 external ICUs with experienced critical care physicians, nurses, respiratory therapists, and pharmacists – took on increased responsibility for ICU load balancing and interhospital transfers.
Leveraging telemedicine services also helped community ICUs care for sicker, more complex patients than usual and aided nonintensivist physicians called upon to manage critically ill patients in ad hoc ICUs at referral hospitals. While the pandemic certainly stressed ICU staff, we suspect that telemedicine’s ability to balance caseloads and distribute clinical tasks helped mitigate these stresses. At age 40, ICU telemedicine is both mature and still growing, with continued expansion of bed coverage and the range of services available. Looking ahead, as we confront a national shortage of intensivists, ICU telemedicine likely represents a cost effective and efficient strategy to maintain critical care capacity with the potential to ensure low-cost, high-quality care for all, regardless of location.
Dr. Graham and Dr. Peltan are with the Division of Pulmonary & Critical Care Medicine, Department of Internal Medicine, University of Utah School of Medicine, Salt Lake City, Utah; and Dr. Peltan is also with the Division of Pulmonary & Critical Care Medicine, Department of Medicine, Intermountain Medical Center, Murray, Utah.