Projecting Critical Care Beyond the ICU: An Analysis of Tele-ICU Support for Rapid Response Teams

Abstract

Background: The rapid response team (RRT) concept was developed to improve care for decompensating patients outside of the intensive care unit (ICU) setting. The tele-ICU service (eICU^®) at Health First Hospitals (Brevard County, FL) has provided tele–critical care support for patients outside the ICU using a mobile platform (the eMobile platform) since 2012. In this study we sought to evaluate the ability of eMobile to support care administered by RRTs. Materials and Methods: A retrospective review evaluating mobile cart activations for RRT calls was performed. Data on mobile cart deployments were recorded over a 33-month period from January 2012 through September 2014. Results: Five hundred eighty mobile cart activations for critical care support were initiated by RRTs, and 577 were completed (>99%). For recorded gender, 223 patients (47%) were male, and 248 (53%) patients were female. Mean recorded age was 70 ± 16 years (median, 72 years). The most common patient conditions were respiratory distress (n = 190, 33%), altered mental status (n = 137, 24%) and hypotension (n = 59, 10%). The most common interventions were medication orders (n = 231, 40%) and laboratory studies (n = 92, 29%). For 566 eMobile calls with documented dispositions, 189 patients (33%) were managed without ICU upgrade. No adverse patient outcomes were recorded involving eMobile. Compared with the RRT program in 2009, the last year before testing of eMobile began (2010–2011), addition of tele–critical care support for calendar years 2012 and 2013 increased projected cost avoidance from unnecessary ICU transfers by a mean of 66% above the 2009 baseline. For Fiscal Year 2014, a projected cost avoidance analysis for unnecessary ICU transfers including costs of information technology (IT) support demonstrated a return on investment up to $1.66 for every $1 invested in IT support. Conclusions: Mobile critical care coupled with RRT is clinically effective and can generate meaningful cost avoidance.

Background

In these early years of the 21st century, the birth of rapid response teams (RRTs) and the evolution of tele–critical care stand as two notable innovations in the management of critically ill patients. Failure to rescue hospitalized patients from decompensations in vital organ function is a significant cause of in-hospital mortality outside of the intensive care unit (ICU). The RRT concept was developed to improve medical response for this patient population.¹ RRTs have been demonstrated to lead to declines in unexpected mortality^2,3 without incurring additional costs.³

Critical care telemedicine plays a growing role in modern ICUs. Initially seen as a new paradigm to deliver critical care,⁴ it is becoming increasingly commonplace. Today, up to 11% of critically ill patients in the United States receive care from tele-ICU programs.⁵ As with RRT, a growing body of evidence also supports critical care telemedicine as a means of both improving care and decreasing ICU utilization.^4,6,7

Health First is a not-for-profit healthcare system in Brevard County, FL. In 2003, Health First implemented an RRT program for its three community hospitals and regional hospital. Teams consist of an ICU nurse and respiratory therapist with simultaneous notification of the attending physician or the on-site hospitalist.

In 2004, Health First established the eICU^® tele–critical care program for its ICUs. The program monitors all ICU beds in the four hospitals of the Health First system, logging over 55,000 patient interactions annually.

Beginning in 2010, plans for a pilot program were initiated for mobile critical care support, including limited support of the RRT program. Concurrently, the Trauma Center at Health First became an active participant in the Florida Department of Health's Florida Emergency Trauma Tele-Network (FETTN). Developed by the late Dr. Jeffrey Augenstein, the FETTN is directed from the University of Miami Miller School of Medicine. Health First began an affiliation with the FETTN in 2009 and subsequently received funding to establish a telepresence hub for Secure Portable Accessible Redundant Reliable Operational Woven NETwork (SPARROWnet™).⁸ The SPARROWnet initiative of FETTN provides a medium for education, casualty triage, and virtual incident command for Florida's system of trauma centers.

Combining the telepresence resources of trauma with the expertise of tele-ICU led to the creation of the eMobile program for mobile critical care support of RRTs during overnight hours, defined as 1700–0700 h the following morning. After testing in 2010 and 2011, the eMobile program was formally launched in 2012, with critical care support provided to RRTs through a network of mobile carts communicating via Wi-Fi over a Health Insurance Portability and Accountability Act (HIPAA)–compliant and secure telepresence system operated and maintained by Health First.

In this study we evaluated the ability of the eMobile tele-ICU program to effectively provide support for RRT patients. Both clinical and financial outcomes were evaluated in order to better assess the value of the eMobile program.

Materials and Methods

The eMobile program was available to support RRTs 365 days a year from 1700 to 0700 h the following morning. VitalWatch utilizes the eICU (Philips, Andover, MA) system for monitoring of ICU patients. The mobile telepresence carts (Cisco Ex90, San Jose, CA) are connected by secure Wi-Fi. The carts include a high-definition pan-tilt-zoom camera that can be controlled remotely and allows for live two-way audio and video interaction. Communication is enabled by commercial telepresence software (JABBER^®; Cisco, San Jose, CA) that meets U.S. standards for patient privacy (HIPAA). Telepresence calls are routed through a secure system maintained and operated by Health First. RRT patients were remotely assessed and managed by a board-certified intensivist with privileges for tele–critical care and a critical care–certified nurse. For eMobile calls by RTTs, data including patient complaint, interventions, and disposition were documented by the tele-ICU team.

This was a retrospective review of eMobile call documentation using descriptive statistics. The study received Institutional Review Board approval from the Western Institutional Review Board. Data on mobile cart deployments were recorded for a 33-month period. Patients were sorted by gender, age, hospital, and complaint at time of RRT activation. For outcomes, we looked at patient disposition, eMobile interventions, patient status following interventions, on-site feedback, and technical issues.

Two separate financial analyses were undertaken, both focusing on cost avoidance in reduction of ICU transfers. A cost avoidance analysis for treatment without status upgrade to the ICU was undertaken for before and after introduction of the mobile cart to the RRT program. Cost avoidances on ICU admission for calendar years 2012 and 2013 were compared with that of calendar year 2009, the last year before the initial pilot study. For the analysis, projected cost avoidance was calculated for an ICU length of stay (LOS) of 1, 2, and 3 days, respectively. Transfer costs included costs for nursing, hospital transport, environmental services, and centralized patient logistics. Costs for an ICU was defined as the average calculated cost for a patient care room at that level of care and those charges associated with use of resources, including nursing, supplies, and medications.

A separate cost analysis was performed for Fiscal Year (FY) 2014, evaluating cost avoidance for ICU admission and factoring in costs of information technology (IT) support of the eMobile system during FY 2014. Cost avoidance was based on the average ICU LOS at Health First hospitals for FY 2014.

The portion of the IT telepresence budget for eMobile included costs assigned to eMobile for hardware maintenance, software upgrades, licensing fees, and labor as opposed to those costs assigned to telepresence for trauma, stroke neurology, psychiatry, pediatrics, wound care, executive leadership, and continuing medical education.

The purpose of these analyses was to evaluate return on investment in reducing the need for ICU admissions specifically for the eMobile system. Costs for VitalWatch itself were not taken into consideration. All outside sources of funding, including grant funding for SPARROWnet, were excluded from the calculation for IT costs.

Results

There were 580 eMobile activations for critical care support during the period from January 1, 2012 to September 30, 2014. Mean calls per month were 18 (range, 8–26). The three community hospitals accounted for 315 (54%) and the regional hospital for 264 (45%) of calls. Gender was recorded for 471 calls (81%) and age for 435 calls (75%). Of these patients, 223 (47%) were male, and 248 (53%) were female. Mean recorded age was 70 ± 16 years (median, 72 years).

The most common presenting symptoms for mobile cart activations by RRTs were respiratory distress (n = 190, 33%), altered mental status (n = 137, 24%), and hypotension (n = 59, 10%). Tele-ICU team interventions were recorded for 577 calls. The most common interventions were medication orders (n = 231, 40%), laboratory studies (n = 214, 37%), diagnostic studies (n = 174, 30%), patient monitoring pending ICU transfer (n = 123, 21%), respiratory treatment orders (n = 121, 18%), and fluid administration (n = 84, 14%).

Patient disposition was documented for 566 encounters. Of these, 332 (59%) were upgraded to an ICU setting, and 41 (7%) were upgraded to a progressive care unit. An additional 189 patients (33%), consisting of 185 patients admitted to a medical/surgical wing and 4 patients admitted to a progressive care unit, were managed by eMobile at the bedside and did not require upgrade to ICU or progressive care unit. One additional patient was managed by eMobile at the bedside and transferred directly to the operating room for unplanned emergency surgery. The eMobile team responded to four Code Blue events for in-hospital cardiac arrest. Two patients died, whereas 2 responded to resuscitation and were transferred to an ICU.

No adverse patient events were reported for care administered via the eMobile program. Patient status after intervention was documented for 152 patients. Of these, 127 (84%) were judged by bedside caregivers to be have an improved clinical status. Caregivers expressed a positive opinion of the mobile cart in 80 of 81 documented instances (99%).

There were no technical issues reported for 430 eMobile calls (74%). Three eMobile calls were aborted for issues related to connectivity. Audio and/or video issues affected 94 (16%). Wireless connectivity issues were noted for 39 calls (7%). Educational deficits in mobile cart use by RRTs were noted for 7 calls (2%). Low mobile cart battery power affected 5 calls (1%).

Two separate cost analyses were undertaken. The first compared cost avoidance from reduced ICU transfers for the calendar year 2009 with calendar years 2012 and 2013. Calendar year 2009 was the last year without tele-ICU support from an eMobile pilot program for RRTs. Calendar years 2012 and 2013 were the first 2 years of the eMobile program proper.

For the analysis, fixed and variable costs for an ICU bed compared above a medical/surgical ward bed were calculated for projected ICU LOS of 1, 2, and 3 days. Transfer costs to the ICU were also calculated and included. The results are presented in Table 1.

Table 1.

Cost Avoidance for Rapid Response Team

				ICU LOS
	EMOBILE (Y/N)	TOTAL NO UPGRADE	EMOBILE NO UPGRADE	1 DAY	2 DAY	3 DAY
2009	N	87	—	$82,700	$165,300	$250,000
2012	Y	145	70 (48%)	$131,500	$261,000	$392,000
2013	Y	166	63 (38%)	$143,600	$287,200	$430,800
Mean CA 2012 and 2013				$137,550	$274,100	$411,400
% mean CA 2012 and 2013 above 2009				+66%	+66%	+65%

CA, cost avoidance; ICU, intensive care unit; LOS, length of stay; N, no; Y, yes.

Compared with calendar year 2009 and coincident with the introduction of the eMobile program, there were large increases in the number of RRT patients managed without upgrade in 2012 (+67%) and 2013 (+90%). RRT patients managed by eMobile without ICU upgrade accounted for 48% of patients in 2012 and 38% of all RRT patients not upgraded in 2013. Taking transfer costs and total costs per day for an ICU bed into account, the mean cost avoidance on ICU transfers for 2012 and 2013 were 65–66% above the cost avoidance of 2009 for the projected ICU LOS.

A second analysis evaluated cost avoidance for FY 2014, defined as October 1, 2013–September 30, 2014. This calculation was based on transfer costs, ICU bed costs per day above a medical/surgical ward bed, and the IT cost for support for the eMobile program. Cost avoidance was calculated for the average ICU LOS for all Health First ICU patients in FY 2014. Results are given in Table 2.

Table 2.

Calculation of Cost Avoidance and Net eMobile Savings Fiscal Year 2014

	CALCULATION	VALUE
Patients managed without ICU upgrade		78
Savings on transfer	(78 × $500)	$39,000
CA on mean ICU stay	Patients × ICU cost ×2.6 days	$405,600
Gross savings on eMobile	Transfer + ICU stay	$444,600
IT eMobile budget	IT costs	$167,000
Net savings FY 2014	(Transfer + ICU) − IT	$277,600
Return on investment FY 2014	Net/IT cost	1.66/1

CA, cost avoidance; FY, Fiscal Year; ICU, intensive care unit; IT, information technology.

For FY 2014, 78 patients were managed by eMobile without upgrade to the ICU. Average cost for an ICU bed was $2,000/day. Average ICU LOS for Health First ICU patients was 2.6 days. Total IT cost for the Health First telepresence network in FY 2014 was $238,500. Based on network use and labor for equipment and software support, the cost of IT support for eMobile was held to be 70% of total budget, or $167,000. Subtracting IT cost from total cost avoidance on transfers and average ICU stay yielded a net savings of $277,600 for FY 2014. The return on investment in IT allocation demonstrated a net cost avoidance at a ratio of $1.66 saved for every $1 spent on IT support.

Using this methodology, a separate analysis was done to identify where eMobile would demonstrate a break-even point in return on investment of IT resources. Setting ICU LOS to 1 day instead of the mean of 2.6 days, gross cost avoidance was calculated to be $195,000. IT cost for eMobile support would then need to approach 82% for no net cost avoidance in FY 2014. Conversely, maintaining a projected ICU LOS of 2.6 days and theoretically assigning 100% of the IT telepresence budget to eMobile still generated a net cost avoidance of $206,000.

Discussion

Our work to date demonstrates eMobile is a clinically effective means for projecting critical care expertise beyond the ICU setting. The data suggest that deployment of eMobile critical care carts with RRTs can provide safe and clinically appropriate patient care that is accepted and endorsed by the bedside staff. This assessment is supported by successful management of a third of patients without ICU upgrade over the 33-month study period, the absence of adverse events related to eMobile care support, and the positive opinions documented from bedside care providers.

As with previously presented data, technical issues are noted with approximately a quarter of eMobile calls.⁹ These largely involved issues with wireless connectivity, audio, and/or video issues. Nevertheless, less than 1% of eMobile calls were abandoned due to technical issues, which speaks to the overall resiliency of the technology. Better defining the exact nature of technical issues continues to be an area of active investigation.

Another area of interest moving forward is the evaluation of subpopulations within the eMobile patient database. An eMobile call for nursing support and monitoring of patients awaiting ICU transfer from the ward or emergency room is a particular area of interest, where the ability of eMobile to support nursing and physician workflows by augmenting bedside resources could potentially generate significant cost avoidance and efficiencies in bedside care. Additional subpopulations of interest in studying the effects of eMobile on quality of care include patients being evaluated for stroke, in-hospital cardiac ischemia, and sepsis.

We strongly believe that the independent financial viability of any telehealth initiative is critical to its long-term growth and sustainability. For the eMobile program, we saw cost avoidance from prevention of ICU upgrades as a means of quantifying the value of the program, thereby defining the value of eMobile in terms of cost savings on patient care.

The financial analysis suggests a significant economic benefit from eMobile in terms of cost avoidance from reduction of ICU stays. This is evidenced by the increase in patients managed without ICU upgrade in 2012 and 2013 compared with 2009 with a subsequent increase in projected costs savings over the 2009 RRT program that did not include eMobile.

Cost savings is further supported by the analysis for FY 2014. Net cost avoidance on an average ICU LOS was well in excess of funds spent by Health First on maintaining the telepresence network. The analysis for FY 2014 took two important considerations into account. Choosing the time frame of FY instead of calendar year allowed for a better alignment of clinical data with financial data available from Health First. The analysis for FY 2014 also excludes all external sources of funding for the telepresence network, including SPARROWnet grants from the State of Florida. The intention of excluding these funds was to evaluate the eMobile program's financial viability as a stand-alone entity.

Althugh promising, the financial analyses are preliminary. They did not include start-up costs for the telepresence network, which was originally established between 2010 and 2011. As the eICU staff supported eMobile in the process of performing their regular duties for ICU patients, the costs related to tele-ICU operations proper were also excluded.

Given that the Health First telepresence network is used by multiple entities within Health First, ranging from clinical care to education and administration, calculation of eMobile representing 70% of telepresence support costs is, to some extent, an estimation based on available data. Subanalyses did demonstrate that eMobile would need to account for greater than 80% of the telepresence budget before a break-even point is reached for cost avoidance on a single ICU day. Additionally, even with eMobile assigned 100% of the IT telepresence budget for FY 2014, a large savings of above the cost of telepresence was still realized.

With their concentrations of both telehealth technology and human expertise, tele-ICU centers offer an ideal foundation for the development and support of mobile care platforms. Synergy among different service lines within a healthcare entity has also been critical to eMobile's evolution and development. Evaluation of virtual triage capabilities using telepresence is a key goal of SPARROWnet's mission to create a telepresence support capability for mass casualty management. Aligning the resources of eICU with the trauma center's SPARROWnet program supported creation of the eMobile for RRT and provided the SPARROWnet initiative with the largest database of virtual triage for critically ill patients within the FETTN. With the continued growth of telehealth technology in healthcare, similar synergies should be cultivated and encouraged across the spectrum of healthcare providers.

The eMobile program represents a significant evolution of the tele-ICU model of telehealth. Linking mobile care with tele-ICU provides new opportunities to optimize delivery of care and to better deploy human and material resources for patient care. Further research will focus on optimizing technical aspects of the program, evaluating distinct patient populations, and refining the cost analysis models.

Footnotes

Disclosure Statement

No competing financial interests exist.

References

Institute for Healthcare Improvement. Getting Started Kit: Rapid Response Teams a How-to Guide. May 2008. Available at http://www.ihi.org/topics/rapidresponseteams/pages/default.aspx (last accessed December 15, 2015 ).

Santamaria

, Tobin

, Holmes

. Changing cardiac arrest and hospital mortality rates through a medical emergency team takes time and constant review. Crit Care Med, 2010; 38:445–450.

Mitchell

, Schatz

, Francis

. Designing a critical care nurse-led rapid response team using only available resources: 6 years later. Crit Care Nurse, 2014; 34:41–55.

Rosenfeld

, Dorman

, Breslow

, Pronovost

, Jenckes

, Zhang

, Anderson

, Rubin

. Intensive care unit telemedicine: Alternate paradigm for providing continuous intensivist care. Crit Care Med, 2000; 28:3925–3931.

Lilly

, Zubrow

, Kempner

, Reynolds

, Subramanian

, Eriksson

, et al. Critical care telemedicine: Evolution and state of the art. Crit Care Med, 2014; 42:2429–2436.

Deslich

, Coustasse

. Expanding technology in the ICU: The case for the utilization of telemedicine. Telemed J E Health, 2014; 20:485–492.

Sadaka

, Palagiri

, Trottier

, Deibert

, Gudmestad

, Sommer

, Veremakis

. Telemedicine intervention improves ICU outcomes. Crit Care Res Pract, 2013; 2013:456389.

Pappas

, Marttos

, Stadler

, Armstrong

. Sparrownet: Florida's emergency tele-network. E-poster presentation at the American Telemedicine Association's 19th Annual Meeting and Exposition, Baltimore, May 2014.

Pappas

, Tirelli

, Shaffer

. Health First mobile cart program: Projecting the tele-ICU to every bedside. Concurrent Session presentation at the American Telemedicine Association's 19th Annual Meeting and Exposition, Baltimore, May 2014.