Clinical and Financial Evidence for Improving Quality and Efficiency in Your ICU

Print PDF

REMOTE ICU CARE CORRELATES WITH REDUCED HEALTH SYSTEM MORTALITY AND LENGTH OF STAY OUTCOMES

Gregory H. Howell, MD*, Vincent M. Lem, MD and Jennifer M. Ball, RN, BSN, Saint Luke's
Health System, Kansas City, MO
Chest. 2007;132(4):443b-444b.

PURPOSE: Providing around-the-clock intensivist-led care is considered the "gold standard" for improving ICU outcomes. However, the shortage of intensivists limits the current capability to provide this level of care in individual hospitals, let alone in a multihospital system. Our health system implemented the eICU® tele-intensivist program as a mechanism to leverage our limited intensivists, and standardize clinical practice and processes to our seven hospitals. We then evaluated changes in ICU outcomes over time to asses the impact of these programmatic changes.

METHODS: We compared Apache III severity-adjusted ICU and hospital mortality rates and ICU and hospital length of stay (LOS) for this seven-hospital health system (84 ICU beds) over five quarters (2006-2007). Mortality was examined with logistic regression controlling for predicted mortality and LOS was compared with a K-Wallis and nptrend (non-parametric trend analysis) test to look for changes over time.

RESULTS : 3692 ICU patients were severity-adjusted (Apache III score quarterly range 44.5-51.4) and compared across five quarters (Q1 2006 to Q1 2007). Severity-adjusted ICU mortality went from 1.0 to .68, hospital mortality from .95 to .77, ICU LOS from 1.18 to .96 and hospital LOS from 1.09 to .84. Severity-adjusted ICU and hospital mortality (p=0.02 and p<0.001 respectively) and ICU and hospital LOS data (both=p<0.001) were significantly reduced over time.

CONCLUSION: Implementation of a remote ICU care program enabled provision of around-the-clock intensivist monitoring for all ICU patients in our health system. It also allowed us to centralize best practice oversight, and improve compliance of these best practices. These changes in ICU care correlated with reduced mortality and improved operational performance, as reflected in decreases in both ICU and hospital LOS.

CLINICAL IMPLICATIONS: Centralized remote care can be used to leverage intensivist resources across multiple hospitals and this correlates with improved outcomes. ICU and hospital LOS reductions should be associated with financial benefit.

DISCLOSURE: Gregory Howell, No Financial Disclosure Information; No Product/Research
Disclosure Information

EFFECT OF TELEMEDICINE ON MORTALITY AND LENGTH OF STAY IN A UNIVERSITY ICU

Benjamin A Kohl, Jacob T Gutsche, Patrick Kim, Frank D Sites, Edward A Ochroch,
Anesthesiology and Critical Care, University of Pennsylvania, Philadelphia, PA
Crit Care Med. 2007;35(12):A22.

Introduction: The purpose of this study was to evaluate the efficacy of ICU
telemedicine (eICU) in an academic surgical ICU (SICU).

Hypothesis: The addition of an eICU, staffed by board-certified intensivists, to an
academic ICU will decrease patient mortality and length of stay.

Methods: We retrospectively evaluated data from 2,811 patients over 3 years. APACHE
III scores were calculated for all patients and predicted mortality and length of stay were
obtained accordingly. Statistics were performed using StatsDirect statistical software
(England: StatsDirect Ltd. 2005).

Results: Actual ICU mortality and length of stay, as well as hospital mortality and length
of stay decreased after eICU (VISICU, Baltimore, MD) implementation (Table 1). Additionally,
the observed to APACHE III predicted values for all of these variables decreased (Figure 1).

Conclusions: The implementation of a remote ICU system within an academic SICU is
associated with improved patient outcome.

OUTCOMES OF SICU PATIENTS AFTER IMPLEMENTATION OF AN ELECTRONIC ICU ("eICU ®") SYSTEM AND OFF-SITE INTENSIVIST

Vincente Gracias, et al. University of Pennsylvania Health System
Presented at: IATSIC-AAST Conference; August, 2007; Montréal.

Background: Surgical intensive care units (SICUs) which utilize the intensivist model
(i.e., a team dedicated to SICU patients and led by an attending intensivist) have improved
patient outcomes versus non-intensivist models. Intensivist models vary in intensivist
coverage of nights and weekends. We recently implemented an electronic ICU (eICU)
system, in which an off-site intensivist has real-time electronic access to patient bedside
data and plans of care; has visual access to patient rooms; and provides decision support
to the on-site team and home-call intensivist. We hypothesized that the eICU system would
decrease mortality among SICU patients.

Methods: We retrospectively reviewed admissions to an eICU-equipped SICU at our
university hospital for a 12 month period before and after launch of eICU. During both
periods, a SICU fellow and resident(s) were on-site "24-7," and intensivists were on-site on
weekdays and on pager call at night. The off-site eICU intensivist provided coverage from
19:00 to 07:00 on weeknights and all weekend.

Results: 2643 patients were admitted to the SICU during the two-year study period
(Table). After implementation of eICU, hospital mortality decreased significantly, despite
increases in clinical volume and APACHE II scores in the latter period.

Conclusion: An eICU system with off-site intensivist coverage on nights and weekends
was associated with decreased hospital mortality of SICU patients. As one of many
components in the delivery of critical care, the eICU system may improve outcomes by
allowing a SICU to function "at the attending level" 24 hours a day.

MANAGEMENT AND OUTCOMES IN A TELEMEDICINE INTENSIVIST STAFFING PROGRAM (TISP) IN A RURAL HEALTH SYSTEM

Edward T. Zawada, Jr, MD, FCCP, Michael L. Aaronson, MD, Pat Herr, RN, CCRN, David K.
Erickson, MD Avera ICU Research Group Avera Health System, Sioux Falls, SD
Chest. 2006;130(4):226s.

PURPOSE: A TISP was initiated to improve the quality of care and patient safety in
seriously ill patients hospitalized in a rural health care system of 4 main hospitals in the
upper Midwest. The TISP shared the expertise of an experienced intensivist team including
24-hour vigilance of patients for early diagnosis and intervention to correct adverse clinical
trends. An "open" model was chosen in which the attending physicians could choose the
level of consultative management from three categories.

METHODS: Three levels of consultative management were available. Category I required
the telemedicine intensivist team to intervene only for life-threatening emergencies or
to appraise the primary attending of any adverse clinical trend. Category II allowed the
intensivist team to adjust any existing therapy. Category III empowered complete clinical
decision-making to the TISP.

RESULTS : Mortality was reduced 76.5% from that predicted by Apache III severity scoring
for the hospital with the highest number of attending physicians choosing Category III
management. In the hospital with mostly Category I consultation, the mortality was reduced
16 % from that predicted. Reduction of ICU length of stay was 33% vs. -2% in the two
hospitals respectively. There was a significant difference in ventilator days per ventilated
patients between the two hospitals. Significant differences between the two hospitals was
seen in compliance with several evidence-based ICU therapies including DVT prophylaxis,
stress ulcer prophylaxis, use of low tidal volumes, and beta-blocker use in acute coronary
syndrome.

CONCLUSION: In a rural health care system greater discretion by a TISP to supervise and
intervene in seriously ill patients results in improved outcomes.

CLINICAL IMPLICATIONS : In a rural setting where availability of intensivists and
experienced critical care nursing is scarce, telemedicine intensivist consultation can
improve outcomes.

DISCLOSURE: Edward Zawada Jr, None.

REMOTE ICU MANAGEMENT IMPROVES OUTCOMES IN PATIENTS WITH CARDIO PULMONARY ARREST

James P. Shaffer, Michael J. Breslow, Jon W. Johnson, Frank Kaszuba
Crit Care Med. 2005;33(12):A5.

INTRODUCTION: Remote ICU management is a mechanism to provide intensivist
oversight to ICU patients and has been associated with improvements in both mortality and
length of stay.

HYPOTHESIS: The clinical vigilance associated with this care model should reduce the
number of codes and improve outcomes of patients sustaining cardio-respiratory arrest.

METHODS: Health First is an integrated network on Florida's East Coast which activated
a system-wide remote ICU management program (eICU®) in June 2004. We compared
the following parameters within five ICU's pre and post eICU activation: total code events,
codes per patient, codes per patient day, initial resuscitation success, and hospital
discharge rates. Each of the results was analyzed by calculating the 95% confidence
intervals of the odds ratios.

RESULTS : Between October 2002 and May 2004 (pre eICU) there were 186 codes in
6,205 patients (21,308 patient days). From June 2004-July 2005 (post eICU) there were
83 codes in 3,954 patients (15,495 patient days). Both codes/patient and codes/patient
day were lower in the post eICU period. The odds ratio (OR) for a code per patient and
per patient day in the post eICU period compared to the pre-eICU period was 0.70 (95%
confidence interval, [CI], 0.54- 0.91) and 0.61 (95% CI, 0.47-0.79) respectively indicating
statistical significance in both of these parameters. Initial resuscitation was successful
in 51.6% in the pre-eICU period and 65.6% in the post-eICU period. The comparative
pre-post OR for 24 hour survival was 0.72, (95% CI 0.44-1.18) which was suggestive of
improvement but not statistically significant.

CONCLUSIONS: Remote ICU management was associated with a significant decrease
in the number of cardio-respiratory arrests occurring in monitored ICU patients. Our data
suggests that this extra layer of support can detect deleterious changes and allow rapid
intervention to prevent detrimental outcomes. Further study in multiple centers employing
this care model will be necessary to provide a more descriptive understanding of how this
system can optimize critical care services.

ICU PROCESS IMPROVEMENT: USING TELEMEDICINE TO ENHANCE COMPLIANCE AND DOCUMENTATION FOR THE VENTILATOR BUNDLE

Brian A. Youn, MD, FCCP, Parkview Hospital, Fort Wayne, IN
Chest. 2006;130:226s-C.

PURPOSE: Outcomes are improved when ventilated patients receive the best practices
combination known as the Ventilator Bundle. However, many ICU's struggle with
implementation and documentation of this best practice guideline. Compliance often falls
short because of incomplete or missing physician orders, missed doses of medications,
failure to use devices, and lack of accurate documentation for applied treatments and/
or contraindications. Despite staff educational efforts, chart audits for compliance and
documentation rarely reach 100%.

METHODS: We hypothesized that physician order compliance and documentation could
be improved using our ICU telemedicine center to implement three components of the
ventilator bundle- head of bed (HOB) elevation, deep vein thrombosis (DVT) prophylaxis,
and peptic ulcer disease (PUD) prophylaxis. We performed this intervention in three
phases. Phase I introduced remote intensivist led, daily multidisciplinary rounds (MDR's)
on all patients. Phase II empowered the MDR remote intensivist to prescribe vent bundle
orders on all patients. Phase III added twice daily RN remote bedside rounds to assist
with documentation. The virtual RN rounds functioned as a reminder for vent bundle
compliance, to confirm medication administration, and to document HOB elevation and
DVT device application when ordered. A daily vent bundle progress note was placed in
the patient chart. Monthly chart reviews were performed on all ICU mechanically ventilated
patients.

RESULTS: Results were compared monthly as percentage compliance per ventilator day.
Percentage compliance improved from baseline to phase III for HOB, DVT, and PUD from
59%, 76%, and 84% to 100%, 100% and 99% respectively. P < .001 for Chi-square and
nptrend analysis.

CONCLUSION: This telemedicine-based performance improvement program enhanced
compliance and documentation with three vent bundle components. Raising awareness
(MDRs) had the least effect on compliance, while writing orders and then documenting that
these orders were carried out showed the greatest effect.

CLINICAL IMPLICATIONS: A centralized tele-ICU program can be instrumental in
achieving greater compliance with quality indicators in the ICU and should be evaluated for
its effect on other ICU best practices.

DISCLOSURE: Brian Youn, None.

CENTRALIZED, REMOTE ICU INTERVENTION IMPROVES BEST PRACTICES COMPLIANCE

Glenn M. Giessel, MD* and Barbara Leedom, RN, Pulmonary Associates of Richmond,
Richmond, VA
Chest. 2007;132(4):444a.

PURPOSE: Payors and quality organizations are calling for greater compliance with ICU
quality measures (best practices) to improve patient outcomes. In spite of these calls,
implementation of best practices remains difficult, particularly in community hospital ICUs
without dedicated intensivists, daily multi-professional rounds or other process-directed
care modalities. We hypothesized that process-directed care from a remote location (eICU®
facility) would enable better compliance with ICU best practices.

METHODS: We compared DVT compliance in two groups during a one month period.
The first group (control) represented 67 beds in 8 ICUs that were not networked into
our eICU facility and the second group (intervention arm) represented 69 beds in 7 ICUs
that were networked into our remote facility. All ICU patients charts were reviewed for
anticoagulant medications and devices (SCDs) ordered and meds administered; and
patients were excluded if they were actively bleeding or coagulopathic (INR>1.5, and/or
platelet count <50K). The intervention arm included the eICU facility nursing staff reviewing
networked patients for DVT prophylaxis. When ICU patients were identified that did not
have DVT prophylaxis in place (meds administered or visual assessment of SCDs) and
were not coagulopathic or actively bleeding the eICU staff either contacted the attending
physician (daytime hours) or the eICU physician wrote the prophylaxis orders (nighttime).
The proportion of patients receiving appropriate DVT prophylaxis in monitored units was
compared to non-monitored units by constructing a contingency table and tested with
Fisher's exact test.

RESULTS : 220 patients charts were reviewed in the control group and 202 charts were
reviewed in the intervention group. DVT prophylaxis compliance was 75% in the control
and 95% in the intervention group. Relative risk of receiving DVT prophylaxis was 1.26
(p<.0001).

CONCLUSION: Improved compliance with DVT prophylaxis is obtained when these
process activities are coordinated from a central location.

CLINICAL IMPLICATIONS: Centralized remote implementation of best practices may
be beneficial for improving compliance for other ICU best practices and in other staffing
environments.

DISCLOSURE: Glenn Giessel, None.

ROLE OF A TELEMEDICINE INTENSIVE CARE UNIT PROGRAM (TISP) ON GLYCEMIC CONTROL (GC) IN SERIOUSLY ILL PATIENTS IN A RURAL HEALTH SYSTEM

Michael L. Aaronson, MD, Edward T. Zawada, Jr, MD, FCCP*, Pat Herr, RN, CCRN Avera
ICU Research Group Avera Health System, Sioux Falls, SD
Chest. 2006;130:226s-A.

PURPOSE: A TISP was launched in 2004 to improve quality of care in rural hospitals. GC
was a priority because it has been shown to improve morbidity, mortality, and ICU length of
stay. We describe the impact on GC after initiation of TISP.

METHODS: Ten rural hospitals are linked by the TISP. The flagship hospital provides the
staff of ten intensive care physicians and fifteen nurses who provide 24-hour supervision
of 61 beds in hospitals separated by as much as 350 miles with populations to as
low as 1,000. The GC protocol was based on the consensus of the Society of Clinical
Endocrinology. Insulin administration is begun when glucose reaches 120 mg/dL to
achieve a goal of 70-110 mg/dL. Human regular insulin is administered with three levels of
aggressiveness depending upon the intensivist's judgment. Modifications occur daily during
remote "glucose rounds." If glucose levels are not at goal and the patient is on the highest
level of insulin administration, insulin glargine is added to the current sliding scale at an
amount equal to 80% of the daily human regular insulin requirement and the sliding scale
is lowered 1 tier. If glucose levels are greater than 150 mg/dL or there is wide variability, an
insulin drip protocol is begun.

RESULTS : Average daily glucose levels fell from 144 mg/dL in the flagship hospital to 124
mg/dL. In the more remote hospitals, average daily glucose levels were reduced from 161
mg/dL to 139 mg/dL.

CONCLUSION: Improved GC in seriously ill patients was achieved by the introduction of
the TISP.

CLINICAL IMPLICATIONS: Improved best practices outcomes can be seen in seriously ill
patients in rural hospitals by a TISP.

DISCLOSURE: Edward Zawada Jr, None.

UTILIZING ROBOTS AND AN ICU TELEMEDICINE PRO GRAM TO PROVIDE INTENSIVIST SUPPORT FOR RAPID RESPONSE TEAMS

Brian A. Youn, MD, FCCP, Parkview Hospital, Fort Wayne, IN
Chest. 2006;130:102s-A.

PURPOSE: The Institute for Healthcare Improvement has identified Rapid Response Teams
(RRT) as an intervention that improves the care of hospitalized patients. Many RRTs utilize
nurses and respiratory therapists because of the limited availability of physician support. We
proposed using our remote intensivist in conjunction with a mobile telemedicine presence
for real-time support to our RRTs.

METHODS: Our remote tele-intensivist currently covers 44 ICU beds in 5 hospitals with a
combined hospital bed capacity of 752. Intensivist support for the system-wide RRT was
initiated in October 2005 utilizing robots that are brought to the patient's bedside along with
the RRT team. The tele-intensivist can visualize the patient and any bedside monitors and
can be seen and heard by the patient and care-givers in the hospital. We evaluated the
preliminary outcomes and nursing satisfaction to this new RRT methodology.

RESULTS : There were 64 RRT calls from med/surg floors over the first 16 weeks of the
program. Preliminary results found that the remote intensivist provided immediate care
orders in 70% of the cases, and 55% of the cases required transfer to another hospital unit
(ICU or telemetry). Since initiating this program, out of unit cardiac arrests have declined
from a nine-month prior average of 38% to currently 28%. A nine-month prior average of
codes per 1000 discharges has dropped from 11% to 8.7%. Nursing satisfaction scores
averaged 4.7-5.0 (1-5 scale, with 5 strongly agrees) for improved communication and
collaboration and better patient outcomes.

CONCLUSION: Mobile telemedicine units in conjunction with a remote intensivist can
provide expert support to multiple hospitals RRTs concomitantly. Intensivist assessments
and orders provide more timely urgent care interventions. This program has positively
impacted preliminary data on out of unit cardiac arrests, codes per 1000 discharges and
nursing satisfaction.

CLINICAL IMPLICATIONS: Use of telemedicine technology can provide intensivist
coverage to multiple hospitals from a central location and represents a significant capability
for extending intensivist care out to floor-based patients in need.

DISCLOSURE: Brian Youn, None.

SCREENING FOR SEVERE SEPSIS; AN INCIDENCE ANALYSIS

Teresa Rincon, Grace Bourke, Daniel Ikeda, Adam Seiver, Sutter Health, Sacramento, CA
Crit Care Med. 2007;35(12):A257.
Introduction: The Surviving Sepsis Campaign (SSC) and the Institute for Healthcare
Improvement (IHI) recommend a severe sepsis screening process followed by aggressive
implementation of the bundle for the treatment of this complex disease state. Institutions
struggle with implementation of a sepsis screening process. The incidence of this disease
state is unknown.

Hypothesis: We hypothesize that we can define the true incidence of severe sepsis
using an electronic screening tool looking at 161 ICU beds at 10 hospitals.

Methods: An electronic screening tool based on the IHI screening form was developed
and utilized for this process. The tool was linked to a database for rapid analysis. All ICU
patients were screened for severe sepsis upon admission into one of 12 ICUs located
in 10 hospitals by a nurse located in the Sutter eICU center (VISICU®). Patients with
infectious processes who did not meet severe sepsis screening criteria were screened
every 12 hours. Patients without an infectious process were screened every 3 days. Upon
identification of a patient with criteria for severe sepsis, critical care physicians in the eICU
confirmed the diagnosis.

Results: From 1Q 2006 through 2 Q 2007 the Sac eICU performed 37,362 screens on
15,085 patients. 2560 patients were identified as positive for severe sepsis (17% severe
sepsis rate). Of the 15,085 patients 844 (5.6%) met the criteria at time of triage or during
the emergency room (ER) stay, 1336 (8.9%) met criteria upon ICU admit or during the ICU
stay, and 380 (2.5%) met criteria in an area outside the ICU or ER. This process includes a
filter for false positive screens.

Conclusions: The incidence of severe sepsis in an ICU represents a large component
of an ICU population. Our data suggests that the incidence for severe sepsis is higher
than what has been previously reported. Identifying and targeting this population for timely
intervention will have a significant impact on the survival of at risk patients.

IMPROVING COMPLIANCE WITH SURVIVING SEPSIS CAMPAIGN GUIDELINES VIA REMOTE ELECTRONIC ICU MONITORING

Bela Patel, Internal Medicine, Division of Pulmonary, Critical Care and Sleep Medicine, The
University of Texas-Medical School Houston, Houston, TX; Lillian Kao, Surgery, Critical
Care, The University of Texas Health Science Center Houston, Houston, TX; Eric Thomas,
Internal Medicine, The University of Texas Health Science Center Houston, Houston, TX;
Tammy Campos, Memorial Hermann Texas Medical Center, Houston, TX
Crit Care Med. 2007;35(12):A275.

Introduction: Mortality from severe sepsis remains high. The surviving sepsis
campaign guidelines are evidenced based recommendations to improve outcomes.
However, in many centers compliance with the speed of initiation of the 6 hour sepsis
resuscitation bundle is low.

Hypothesis: Integrating the electronic ICU into the multidisciplinary workflow of sepsis
management in an academic 16 bed MICU will improve compliance with the resuscitation
sepsis bundle.

Methods: All patients admitted to the MICU during May 1 to August 15 2007 were
screened for sepsis by the electronic ICU monitoring staff. Compliance with sepsis bundles
was evaluated. Communication verbally or via fax order transmittal was established to
ensure compliance. Interval monitoring of targets (CVP, ScVO2, MAP) with EGDT pathway
implementation occurred approximately every 15-30 minutes.

Results: The overall mean compliance with the resuscitation 6 hour bundle increased
from 3 to 29% during the 15 week period by integrating remote electronic ICU Monitoring
into the sepsis management workflow. The intervention improved compliance with all of
the bundle elements. Serum lactate within 6 hours improved from mean 67% to 90%.
Blood cultures prior to antibiotics improved from 67% to 75%. Antibiotics within time
guidelines improved from mean 26% to 51%. Treatment of hypertension with fluids and/
or vasopressors per guideline improved from mean 39% to 67%. CVP target within time
guideline improved from mean 13% to 45%. ScV02 target within time guideline improved
from mean 4% to 29%.

Conclusions: Remote Electronic ICU monitoring may improve overall compliance with
the resuscitation sepsis bundle.

CENTRALIZED, REMOTE CARE IMPROVES SEPSIS IDENTIFICATION, BUNDLE COMPLIANCE AND OUTCOMES

Teresa Rincon, BSN, CCRN*, Grace Bourke, MBA and Daniel Ikeda, MD, Sutter Health,
Sacramento, CA
Chest. 2007;132(4):557b-558b.

PURPOSE: Sepsis is responsible for 215,000 deaths per year and the Surviving Sepsis
Campaign was initiated to standardize care and improve outcomes in this patient
population. We previously reported on reduced sepsis mortality (CCM 2006, Vol. 34, A2
& A108) in our ICU patients and we hypothesized that the improvement in outcomes
correlated with the development of a centralized process for identifying sepsis patients and
implementing the sepsis bundle in a more timely fashion during the same time period.

METHODS: We screened high risk patients in eight hospitals (118 ICU beds) from our
eICU® center. When patients were identified who met sepsis criteria they were then tracked
for compliance with the sepsis bundle. The eICU physician would either implement the
bundle (order blood cultures, baseline labs, measure lactate and administer antibiotics
within 2 hours) if given the authority (high category of intervention) or would contact the
attending physician and advise for timely bundle implementation. Data was extracted and
analyzed by nptrend and reports were provided back to the facilities on a monthly basis.

RESULTS : 8116 of 8134 ICU admissions were screened from Jan-Nov 2006 (99.8%),
and of those screened 1120 patients met criteria for sepsis (13.8%). Sepsis bundle
implementation showed the following changes over the 11 months: Antibiotics within 2
hours went from 51% to 79% (p<.001), blood cultures drawn before antibiotics from 63%
to 74% (p<.001), lactate measurement from 49% to 55% (p=.07), and baseline labs from
78% to 84% (p=.003).

CONCLUSION: Accurate sepsis identification can be achieved from a central location and
correlates with both improved sepsis bundle compliance and reduced mortality.

CLINICAL IMPLICATIONS: Centralized remote identification of at-risk patients may be
beneficial for improving adherence to best practices for identification and management of
sepsis as well as other common conditions.

DISCLOSURE: Teresa Rincon, No Product/Research Disclosure Information; Consultant
fee, speaker bureau, advisory committee, etc. Eli Lilly Medical Advisory Board and Speaker
Bureau

IMPLEMENTATION OF A STANDARD PROTOCOL FOR THE SURVIVING SEPSIS 6 AND 24 HR BUNDLES IN PATIENTS WITH AN APACHE III ® ADMISSION DIAGNOSIS OF SEPSIS DECREASESMORTALITY IN AN OPEN ADULT ICU

Daniel Ikeda, Saman Hayatdavoudi, John Winchell, Alexandra Rojas, Teresa Rincon, Alan
Yee, Sutter Health, Sacramento, CA
Crit Care Med. 2006;34(12):A108.

INTRODUCTION: Numerous studies have shown that individual components of the
Surviving Sepsis 6 and 24 hr Bundles decrease mortality.

HYPOTHESIS: We hypothesized that use of a protocol implementing the Surviving Sepsis
6 and 24 hr Bundles in patients with APACHE III® admission diagnosis of sepsis will show a
measurable decrease in mortality in an open adult Intensive Care Unit (ICU).

METHODS: In this prospective study we used a protocol to manage 266 consecutive
patients admitted to a tertiary community hospital 24 bed open adult ICU from 7/1/2004
- 6/30/2006 with an APACHE III admission diagnosis of Sepsis. The historical control
cohort was 48 consecutive ICU patients admitted between 1/1/2004 - 6/30/2004 with an
APACHE III admission diagnosis of Sepsis. The protocol implemented the surviving sepsis 6
and 24 hr Bundle guidelines, using pre-printed order sets and shared patient management
by critical care physicians located in the Sutter eICU (VISICU®), a remote electronic
monitoring unit.

RESULTS: The actual ICU mortality was 40.07% in the control period, compared to
18.86% for the study period (x2 = 28.98, p < 0.001). APACHE III (Cerner®) predicted ICU
mortality was 24.18% for historical control vs. 23.11% for the study group. Divided into
6-month intervals the actual ICU mortality was 22.27%. (7/1/04 - 12/31/04), 16.34%
(1/1/05 - 6/30/05), 17.21% (7/1/05 - 12/31/05) and 16.22% (1/1/06 - 6/30/06). An
estimated 56 lives were saved over this 30-month period

CONCLUSIONS: Utilization of a protocol applying the Surviving Sepsis 6 and 24 hr bundle
guidelines in patients with an APACHE III admission diagnosis of Sepsis was associated
with a significant sustained decrease in mortality compared to a historical control in a
tertiary community hospital open adult ICU.

ECONOMIC IMPACT OF eICU ® IMPLEMENTATION IN AN ACADEMIC SURGICAL - ICU

Benjamin A Kohl, Frank D Sites, Jacob T Gutsche, Patrick Kim, Anesthesiology and Critical
Care, University of Pennsylvania, Philadelphia, PA
Crit Care Med. 2007;35(12):A26.

Introduction: We have recently shown an improvement in mortality and length of stay
after implementing eICU (VISICU, Baltimore, MD) in a large academic surgical ICU. The
purpose of this study is to measure the economic impact of this transition.

Hypothesis: Implementation of eICU in an academic surgical ICU, allowing round-theclock
intensivist oversight, will decrease ICU and hospital costs.

Methods: We retrospectively compared a random sample of 189 patients pre-eICU
to 2,622 patients 3 years post eICU using a multiplier of 13.87 to normalize populations.
Assumptions based upon published literature include an average surgical ICU cost per day
of $1,500-$2,000 and an average daily cost on a general floor of $500-$600. Because
of the disparate sizes in populations a multiplier of 13.87 was used to standardize the
numbers. There was no significant change in practice paradigm during the time period.

APACHE III scores were used to calculate predicted length of stay in ICU and hospital.
Results: An almost 10% reduction in ICU stay and 20% reduction in floor stay occurred
after implementation of eICU. This translated into a savings of $706,272-$941,697 for the
ICU and $2,134,339-$2,842,940 for the floor

Conclusions: Implementation of an eICU in an academic SICU resulted in significantly
reduced costs.

THE ECONOMIC AND CLINICAL VALUE OF A REMOTE INTENSIVE CARE UNIT

John A Hitt, Ed Zawada, Pat Herr, Brian Pederson, Avera McKennan Hospital, Sioux Falls, SD
Crit Care Med. 2007;35(12):A20.

Introduction: Remote monitoring of intensive care patients is increasingly prevalent
and we have previously demonstrated care improvement and reductions in mortality
(Chest130;226S-a,226S-b).

Hypothesis: The cost benefit for remote monitoring needs further study and an
economic and clinical benefit can be shown.

Methods: Avera McKennan Hospital (AMH) is a 490 bed community teaching hospital
in Sioux Falls, SD. AMH is the hub of a network of smaller community hospitals in 4
states. AMH established a remote monitored intensive care unit (RMU) in 2004 using
VISICU technology. The AMH RMU monitors 67 beds (27 at AMH; 40 across network).
We accounted for all the start up and ongoing annual costs allocated to AMH for the RMU
network. We modeled costs savings using an estimated $1600 cost per intensive care unit
(ICU) day for a one year period before and after RMU implementation and compared it to
our hospital accounting system (TSI). We scored charts retrospectively in from the preperiod
and prospectively in the post-period using the APACHE III severity scoring system.

Results: AMH accounting systems were used to calculate initial setup costs for RMU
at AMH: 1) VISICU Software $950,000 2) Network setup $400,000, 3) ICU modifications
$140,000. Annual costs for the RMU program: 1)license .fees $200,000, 2)RMU staffing
$1,250,000(14 FTEs), 3)Miscellaneous $100,000. Clinical benefit Pre: Post length of stay
(LOS) decreased 1.13:0.60(observed vs. expected p<.001). ICU predicted days in post
period were 5487 and actual days were 2826. Annual savings estimate calc 5487-2826
(days avoided) = 2581 x $1600 per ICU day = $4,100,000. The TSI estimated annual cost
savings was $450,000-650,000.

Conclusions: Cost savings with RMU is demonstrable. The magnitude of the savings
and resultant return on investment can vary greatly depending on the assumptions made,
accounting methods used. Decreases in mortality (cost per life saved) must be valued in
addition to financial savings and support the cost effectiveness of RMU. Simultaneous care
improvements contributed to the savings.

FINANCIAL BENEFIT OF A TELE-INTENSI VIST PROGRAM TO A RURAL HEALTH SYSTEM

Edward T. Zawada, MD, FCCP*, Pat Herr, RN, CRRN, David Erickson, MD and John Hitt,
MD, Avera ICU Research Group Avera McKennan Hospital & University Health Center,
Sioux Falls, SD
Chest. 2007;132(4):444.

PURPOSE: Providing around the clock intensivist care to a rural health system represents
a significant staffing challenge. Our health system implemented a tele-intensivist program
in 2004 to leverage our limited intensivist staff and improve clinical outcomes. We have
previously reported on the clinical benefits of our program (Chest Vol. 130:226S). However,
the current healthcare environment requires that new technologies also save money to be
sustainable. Length of stay (LOS) is the single most important determinant of hospital cost,
and we hypothesized that this care delivery model would also reduce length of stay across
our health system.

METHODS: This study compared severity-adjusted LOS (APACHE-III) one year before
and two years following implementation of the tele-intensivist program. For the pre-period,
200 randomly selected ICU patients (50 charts from each quarter for four quarters prior to
program activation) from the tertiary (24 beds) and each of 3 regional hospitals (10 beds,
10 beds, and 6 beds) were compared to continuous APACHE-III scoring in the post period.
Data were analyzed using a rank sum test on the difference of expected and observed
LOS.

RESULTS: ICU LOS ratios (observed/expected) pre and post were 1.13 and 0.60 (-46.8%)
in the tertiary hospital, 1.35 to 0.86 (-36.4%), 1.42 to 0.93 (-34.7%) and 0.96 to 0.89
(-7.6%) in the regional hospitals. Hospital LOS ratios were 0.62 to 0.53 (-21%) in the tertiary
hospital, 0.79 to 0.63 (-20.3%), and 0.67 to 0.62 (-7.4%), and 0.79 to 0.80 (1.4%) in the
regional hospitals. Both ICU and hospital LOS were reduced (p<0.001) and across the
health system were associated with an annual reduction in 4146 ICU days and 572 hospital
days.

CONCLUSION: Remote telemedicine intensivist staffing reduces severity-adjusted ICU
and hospital lengths of stay and is associated with a substantial number of saved days
across the health system.

CLINICAL IMPLICATIONS: Further analyses are required to determine the etiology of
saved days, but based upon LOS reduction our tele-intensivist program demonstrates a
financial benefit.

DISCLOSURE: Edward Zawada, No Financial Disclosure Information; No Product/
Research Disclosure Information

REMOTE ICU CARE CORRELATES WITH REDUCED HEALTH SYSTEM MORTALITY AND LENGTH OF STAY OUTCOMES

Gregory H. Howell, MD*, Vincent M. Lem, MD and Jennifer M. Ball, RN, BSN, Saint Luke's
Health System, Kansas City, MO
Chest. 2007;132(4):443b-444b.
Please see abstract on page 2

THE EFFECT OF INSTITUTING AN ICU ELECTRONIC MEDICAL RECORD ON PROFESSIONAL FEE BILLING

Joseph J Hine, Pulmonary and Critical Care Medicine, Medical College of Wisconsin,
Milwaukee, WI
Crit Care Med. 2006;34(12):A21.

INTRODUCTION: Electronic medical record (EMR) systems have been hypothesized to
improve the quality of documentation, patient care and patient safety. However, despite
the financial benefit resulting from better documentation, this aspect of EMRs has not been
examined in an intensive care unit (ICU) setting. The purpose of this study was to determine
the effect of instituting an intensive care-specific EMR on capturing professional fee (profee)
billing charges.

HYPOTHESIS: ICU EMR will improve pro-fee billing charges.

METHODS: This was a single center, retrospective study performed in the medical
intensive care unit (ICU) of the Medical College of Wisconsin. A retrospective analysis
of submitted critical care time related pro-fee charges (CPT codes 99291 and 99292),
was performed (Pre: Oct - July 2004/2005 and Post: Oct - July 2005/2006). Prior to the
initiation of the EMR (VISICU, Inc) critical care time related pro-fees were captured via
manual chart abstraction by professional fee abstracters. There was no change in attending
coverage or coding staff during the study period. The results for two ten month periods
were analyzed using nonparametric rank sum test.

RESULTS : Each month was analyzed for total patient admissions and critical care time
related pro-fee billing charges. The EMR was introduced in October of 2005. Prior to the
institution of EMR, the average monthly pro-fee billing charge was $174,000. After initiation
of the EMR, the average pro-fee billing charge rose to $227,000/month (31% increase,
p=.004).

CONCLUSIONS: The addition of an EMR to an academic, medical intensive care unit
service significantly increased the professional fee billings. Institution of an EMR in ICUs
should increase revenue capture for intensivists, pulmonary departments and/or hospitals.

THE EFFECT OF INSTITUTING AN ICU ELECTRONIC MEDICAL RECORD ON BILLING AND COMPLIANCE

Benjamin A. Kohl, MD, University of Pennsylvania, Philadelphia, PA
Chest. 2006;130:112s-A.

PURPOSE: Electronic medical record (EMR) systems have been shown to improve
the quality of patient care and patient safety. Despite compelling evidence of return on
investment, EMRs have not been universally accepted. The purpose of this study was to
determine what effect instituting an intensive care specific EMR in an academic medical
center has on capturing billable encounters (BE).

METHODS: This was a single center, retrospective study occurring in the surgical intensive
care unit (ICU) of the University of Pennsylvania. A retrospective analysis of all BE was
performed through the study period. Prior to the initiation of the EMR, BE were captured
via manual chart abstraction by professional fee abstracters certified by the American
Academy of Professional Coders (AAPC). There was no change in attending coverage or
coding staff during the study period.

RESULTS: Each year was divided into quarters for analysis. The EMR was introduced in
the second quarter of 2005. CPT code 99291 designates critical care services provided
for between 30 and 74 minutes (after which it is billed as 99292). Prior to the institution of
EMR, the average number of CPT 99291 being captured was 935.4 (range 836-1136).
After initiation of EMR, the average number of CPT 99291 being captured rose to 1663.6
(range 1275-2266). The total number of billable events which were captured was 4,382
prior to the EMR and 4,937 after introduction of EMR. The documentation supported
critical care code billing in 55% of the encounters prior to initiation of the ICU EMR and
77% afterwards. When comparing these numbers to the total BE, this change is statistically
significant by Fisher exact test at p < 0.0001 [OR 2.61, CI 2.39-2.85].

CONCLUSION: The addition of EMR to an academic medical center surgical intensive
care unit significantly increased the capture of billable critical care services as measured by
CPT 99291. CLINICAL IMPLICATIONS: Institution of EMR in academic ICUs may increase
hospital revenue by properly capturing billable events.

DISCLOSURE: Benjamin Kohl, None.

CRITICAL -CARE NURSES ' JOB SATIS FACTION AND ITS EFFECT ON
RETENTION
Wanda F Lewis, eICU, Memorial Hermann Healthcare System, Houston, TX
Crit Care Med. 2007;35(12):A22.
Introduction: The purpose of this quantitative correlational study was to examine the
relationship between job satisfaction of critical-care nurses and its effect on retention of
nurses.

Hypothesis: The demographic and employment variables under study were age, years
as a registered nurse, years as a critical-care nurse, years as a registered nurse, years as
a critical-care nurse, years in current unit, gender, ethnicity, salary, and education. The
null and alternative hypotheses were derived for the first and second research questions.

1. What are the relationships between job satisfaction and demographic and employment variable among critical-care registered nurses?
2. To what extent can critical-care registered nurses' retention intention be predicted by job satisfaction?

Methods: A quantitative correlational study, one-way ANOVAs, Tukey's HSD, F static,
and logistic regression were used to derive relationships between job satisfaction and
retention of the critical-care registered nurse. Where there was a statistical significance,
the-one-way ANOVAs were followed-up with Tukey's HSD test. The population consisted
of 200 hundred critical-care nurses who were members of the American Association of
Critical-Care Nurses.

Results: Satisfaction with pay was the most important area of satisfaction, followed
by satisfaction with autonomy. Results indicated that the satisfaction scales were not
predictive of whether or not an individual would stay in critical-care nursing for 5 years. One
satisfaction scale, satisfaction with task requirements, was related to plan on staying-with
the current organization for 2 years, with higher level of satisfaction with task requirements
substantially increased the likelihood that the individual planned to stay.

Conclusions: The current study has shown a significant relationship between pay and
compensation and autonomy as attributes that determine job satisfaction of practicing
critical-care registered nurses, there is a statistical significant relationship exists between
pay and compensation and autonomy as measured by the IWS questionnaires. Retention
was predicted by satisfaction with task requirements.

THE IMPACT OF A TELE-ICU PROVIDER ATTITUDES ABOUT TEAMWORK AND SAFETY CLIMATE

Eric J Thomas, Ming Ying L Chu-Weininger, Joseph Lucke, Laura Wueste, Medicine,
University of Texas Health Science Center at Houston, Houston, TX; Lisa Weavind, Janine
Mazabob, Memorial Hermann Healthcare System, Houston, TX
Crit Care Med. 2007;35(12):A145.

Introduction: Little is known about how a tele-ICU may affect ICU physicians and
nurses in the outlying units. The tele-ICU may impact communication and teamwork for
better, or for worse. In addition, the tele-ICU should result in changes that improve the
quality and safety. Our goal was to measure provider attitudes about teamwork and safety
climate in three intensive care units (ICUs) before and after the implementation of remote
monitoring by intensivists using telemedicine technology (tele-ICU).

Hypothesis:

Methods: The design was a controlled pre tele-ICU and post tele-ICU cross-sectional
survey of physicians and nurses in ICUs in three hospitals. The outcomes were teamwork
and safety climate scores (TWS and SCS) measured by the Safety Attitudes Questionnaire.

Results: The mean (SD) TWS score was 69.7 (25.3) and 78.8 (17.2), pre and post tele-
ICU, respectively (p = 0.009). The mean SCS score was 66.4 (24.6) and 73.4 (18.5), pre
and post tele-ICU, respectively (p = 0.045). While SCS scores within the ICUs improved,
the overall SCS scores for these hospitals decreased from 69.0 to 65.4. The hospitals were
not administering the teamwork portion of the survey prior to tele-ICU. Three of the nonscaled
items were significantly different pre and post tele-ICU at p<.001. The item means
(SD) pre and post tele-ICU were: 'others interrupt my work to tell me something about
my patient that I already know' 2.5 (1.2) and 1.6 (1.3); 'I am confident that my patients
are adequately covered when I am off the unit' 3.2 (1.3) and 4.2 (1.1); and 'I can reach a
physician in an urgent situation in a timely manner' 3.8 (1.2) and 4.6 (0.6).

Conclusions: Implementation of a tele-ICU was associated with improved teamwork
climate and safety climate, especially among nurses. Providers were also more confident
about patient coverage and physician accessibility, and did not report unnecessary
interruptions.

ECONOMIC IMPLICATIONS OF DATA COLLECTION FROM A REMOTE CENTER UTILI ZIN G TECHNOLO GICAL TOOLS

Teresa Rincon, Barbara Welcher, Dana Srikanth, Adam Seiver, Sutter Health, Sacramento,
CA
Crit Care Med. 2007;35(12):A161.

Introduction: The use of complex predictive models to predict survival is widely
accepted. Results from these can lack statistical significance due to low patient volume.
Lack of resources and funding, potential for error with manual data collection processes
and educational deficits may also decrease utilization.

Hypothesis: We hypothesize that centralized remote data collection methods utilizing
advanced technology can improve efficiency, accuracy and costs without increasing tasks
and resources at the bedside.

Methods: In 2006, the Bay Area eICU® hub, began a pilot utilizing an APACHE III®
data collection methodology imbedded in a software application tool. This tool is used for
remote Teleintensive care monitoring of adult ICU patients. The Bay Area eICU center was
able to score 60% of the total patient population (containing the likelihood of ascertainment
bias to an acceptable level). 100% of patients at 4 hospitals in 2006 were scored at the
Sacramento eICU hub, but a mixed process of remote data collection and on site chart
analysis was required. Over 1700 charts were reviewed at these sites in 2006. In addition to
regular eICU staff, 2 full-time clerical and 1.5 full-time RN was necessary for data collection.

Results: After implementation of an imbedded APACHE III data collection tool in 2Q 07
at the Sac hub, a decrease in two clerical and 1.5 RN was achieved by 3Q 07. Despite a
decrease in resources the eICU center achieved APACHE I11 scoring for 161 ICU beds at
10 hospital campuses with an average capture rate of 78.59%. Annual per bed licensing
fee was also decreased by 50% and an overall reduction was achieved for a total estimated
savings of $132,859 for 2007. Projected savings for 2008 is; $318,248,

Conclusions: We have found that utilizing a technological tool for APACHE III data
collection has enhanced resource use while decreasing overall costs at a large hospital
system.

RESIDENT PERCEPTION OF THE EDUCATIONAL AND PATIENT CARE VALUE FROM REMOTE TELEMONITORING IN A MEDICAL INTENSIVE CARE UNIT

Adan Mora, MD*, Saadia A. Faiz, MD, Todd Kelly, MD, Richard J. Castriotta, MD, FCCP and Bela Patel, MD, FCCP, The University of Texas Medical School at Houston, Houston, TX
Chest. 2007;132(4):443a.

PURPOSE: To assess residents' perception of remote telemonitoring with regard to the
educational value it may contribute in their residency training and to improved patient care.

METHODS: An anonymous electronic survey was sent to 133 residents who train in
the medical intensive care units (MICU) affiliated with The University of Texas Medical
School at Houston. One MICU has telemonitoring provided by fellows and academic or
private intensivists via VISICU system of eICU©. The other MICU does not have eICU©
involvement but is staffed by the same cohort of residents.

RESULTS : Ninety-six residents (72%) responded to the survey, including internal medicine,
internal medicine/pediatrics, emergency medicine, anesthesia and preliminary residents
responded. Sixty nine (71.9%) had telemonitoring experience. Of those with telemonitoring
experience, a majority of residents perceived telemonitoring improves patient care (82.3%),
and improves the care they deliver to patients while on call (73.8%). The events/interactions
in which at least 60% of the residents believed telemonitoring was helpful or of some
benefit were: ventilator management (70%), initial management of an unstable patient
(64%), code supervision (64%), management of acute respiratory change (62%), blood gas
interpretation/ acid base management (62%), early goal directed therapy and guidance
(61%) and respiratory failure recognition (60%). It was least helpful with end of life issues
(45%) and supervision on line placement (42%). 62% of residents preferred to train in a unit
with remote telemonitoring. Upon completion of residency, 66.7% of residents expressed a
desire to have remote telemonitoring involved in the care of their patients.

CONCLUSION: Remote MICU telemonitoring in a residency training program was
perceived by residents to have a substantial impact in their education and to improve
patient care.

CLINICAL IMPLICATIONS: Remote telemonitoring contributes to bedside residency
education in critical care medicine and is perceived by residents to improve patient care.

DISCLOSURE: Adan Mora, No Financial Disclosure Information; No Product/Research
Disclosure Information

FELLOWSHIP EDUCATION IN REMOTE TELEMONITORING UNITS

Saadia A. Faiz, MD, Anthony Zachria, DO, Liza Weavind, MD and Bela Patel, MD University of Texas at Houston Health Science Center, Houston, TX
Chest. 2006;130:113s-A.

PURPOSE: To address the shortage of intensivists, remote telemonitoring units have
evolved and provide monitoring by intensivists. This survey aims to evaluate the experience
of fellows exposed to this new modality of critical care.

METHODS: An anonymous electronic survey was sent to all VISICU unit medical directors
and four Pulmonary & Critical Care program directors to enlist their fellows.

RESULTS: Sixteen fellows (13 pulmonary & critical care, 1 critical care, 1 trauma, 1 other)
responded. All were part of a university based teaching program using VISICU and from
2 major cities: Houston, Kansas City. Most had experience via their fellowship program
as a one month rotation, while three were moonlighters. Research opportunities were
available to most (14/16). Most worked with both private and academic physicians. Eleven
felt that the rotation was a good educational experience, but only nine felt that it should
be a formal rotation. They felt that it improved their knowledge base (9/16), enhanced
their communication skills (9/16), and reinforced the importance of professionalism (7/16).
Fifteen fellows felt their exposure would be helpful after their training was completed.
During the rotation, the majority worked with an intensivist, although four worked alone. In
comparison to the ICU, some felt more exhausted (6/16), some felt the same (5/16), and
some felt less exhausted (5/16). In the future, most would consider working as a part-time
intensivist (14/16), but few would consider working full time (4/16). Most of the respondents
would want to work in a place with remote telemonitoring units (14/16), and they all felt
it improved patient care. Of note, thirteen fellows felt it served to further protect against
medical liability.

CONCLUSION: Rotations in remote telemonitoring units should be included in training
curriculum. The experience enhances skills, prepares for the future, and ameliorates
communication and professionalism. Fellows feel it improves patient care and will likely be a
part of their post-graduate practice.

CLINICAL IMPLICATIONS: Formal training for critical care fellows in remote
telemonitoring units may bridge the nationwide shortage of accessible intensivists.

DISCLOSURE: Saadia Faiz, None.

DELIVERING INTENSIVIST SERVICES TO PATIENTS IN MULTIPLE STATES USING TELEMEDICINE

Dellice Dickhaus, Advanced ICU Care, St Louis, MO
Crit Care Med. 2006;34(12):A24.

INTRODUCTION: A shortage of intensivists has been identified nationwide. This shortage
can be felt most acutely by community hospitals. However, research has shown that ICU's
where intensivists manage the patient's care, there are improved outcomes.

HYPOTHESIS: The need for critical care expertise in community hospitals can be met by
bringing intensivists to patients utilizing telemedicine technology. This can be done from a
central location which has a rich population of intensivists.

METHODS: A free-standing eICU® operations center was established in St. Louis, Missouri
to connect board-certified intensivists to patients in community hospitals. The hospitals
are located in Jefferson City, MO (167 beds) and Weston, WI (86 beds). Licensure, hospital
privileges, and malpractice coverage were obtained for all the intensivists in both states.
The intensivists in St. Louis were connected electronically utilizing T1 lines. Utilizing the
VISICU software package, trended and current patient data were regularly reviewed in
real time. In addition, the eICU staff monitored and evaluated patients visually using a
high-resolution camera and spoke with bedside clinicians and patient's families by 2-way
speakers in each of the patient's rooms. Physician orders by the intensivist were signed
electronically.

RESULTS : Two community hospitals, in two different states are now being served by an
established group of intensivists. Early results at the Jefferson City hospital indicate a 17%
decrease in LOS as well as a decrease in mortality. The hospital in Weston, WI has an
intensivist on-site. The combination of the on-site intensivist in one state, and an intensivists
in another state, who are connected through telemedicine technology has provided
intensivist staffing recommended by Leapfrog.

CONCLUSIONS: This is a viable solution to the shortage of intensivists, particularly
in community hospitals across the United States. There are significant challenges with
licensing, credentialing, and malpractice coverage across state lines, which may limit the
ability to expand this model on a broader scale.

SEVERE SEPSIS (SS ) IS UNDERREPORTED IN THE ICU

Omar Badawi, Randy Holl, Erkan Hassan, Research & Development, VISICU, Baltimore,
MD; Marc Zubrow, Christiana Care Health System, Wilmington, DE; Adam Bress, University
of Maryland School of Pharmacy, Baltimore, MD
Crit Care Med. 2007;35(12):A256.

Introduction: The incidence of SS in ICUs varies from 2% to 11%. Underreporting
may contribute to this variability.

Hypothesis: A greater number of ICU patients meet physiologic criteria (PC) for SS than
receive the diagnosis. The organ dysfunction (OD) present influences documentation of
SS.

Methods: Retrospective, multi-center study using the eICU® Program Network database
for patients in an ICU using software designed to identify systemic inflammation (SI) from
11/06 to 7/07. SI was defined by an algorithm aggregating the degree of abnormality in:
HR, RR, WBC, temperature, INR, glucose, ileus and altered mental status. OD associated
with SS was defined using accepted clinical criteria. Patients met criteria for SS if they
had a documented diagnosis of SS at admission or during the ICU stay (DS group), or met
PC for SS (PCS group). The PCS group had SI, OD and a concurrent infectious diagnosis,
but no diagnosis of SS (ie. Undocumented SS). McNemar's test was used to assess
concordance.

Results: 25,582 patients were included from 52 hospitals. 1,222 (4.8%) of patients had
SS. 558 (2.2%) of these patients (PCS group) were not diagnosed with SS (p<0.01). CV/
shock was documented in 491 (74%) of DS patients. No other single OD was identified
in >5% of DS patients. In contrast, PCS patients had a more diverse set of ODs present
(Table 1).

Conclusions: Nearly ½ of ICU patients who met PC for SS did not have the diagnosis
documented. Clinicians may be underreporting SS by focusing on CV OD. Associating
other ODs with SS would increase reporting, potentially leading to improved treatment,
reimbursement and severity scoring.

ICU LENGTH OF STAY (LOS ) OUTLIERS: INCIDENCE AND IMPLICATIONS

Michael J. Breslow, MD*, Edward Larsen, Robb Fromm, MD and Brian Rosenfeld, MD,
VISICU, Inc., Baltimore, MD
Chest. 2007;132(4):442.

PURPOSE: Most ICU patients have a stay of 2-4 days, after which they are able to leave
the ICU. Although less prevalent, patients with long stays account for a disproportionate
number of ICU days and costs. While there is wide recognition of the large impact of
outliers, little is known about the makeup of this important sub-group of ICU patients.

METHODS: APACHE® III mortality and LOS data were collected from 20 health systems
in the eICU Program Network (154 ICUs) throughout 2006. LOS outliers were patients
with ICU stay > 6 days. Patients were grouped based on predicted ICU mortality: < 10%
(low risk), 10-50% (medium) and > 50% (high). Outlier data in the three risk groups were
examined in aggregate and at the ICU level: ICUs with < 200 patients were excluded from
the ICU level analysis. ICU outlier incidence data were compared to mortality performance
using least squares regression analysis.

RESULTS : 63,865 ICU admissions were included in the analysis. 8149 patients had an ICU
LOS > 6 days (12.7%) and accounted for 50% of all ICU days. The incidence of outliers
in low, medium and high risk patients was 8.6, 28.1 and 33.1%, respectively. 54% of all
outliers were low risk patients. There was considerable inter-ICU variability in the incidence
of low risk outliers (sd = 5.4%). Deaths in low risk outliers exceeded predicted mortality by
400%. There was a positive correlation between ICU standardized mortality ratio and the
incidence of low risk outliers (R = 0.63).

CONCLUSION: More than half of all outliers had predicted mortality < 10%. These low risk
outliers accounted for 25% of all ICU days. They also had a significantly higher mortality
rate than expected. The incidence of low risk outliers varied considerably among ICUs, and
was associated with worse ICU mortality performance.

CLINICAL IMPLICATIONS: These data suggest that high quality ICU care can reduce the
incidence of low risk outliers, and thus have a beneficial effect on ICU resource utilization.

DISCLOSURE: Michael Breslow, No Product/Research Disclosure Information;
Shareholder I am a shareholder in VISICU, Inc., a company that sells ICU software and
services to hospitals; Employee I am an employee of VISICU, Inc., a company that sells
ICU software and services to hospitals; Fiduciary position (of any organization, association,
society, etc, other than ACCP I am a member of the Board of Directors of VISICU, Inc., a
company that sells ICU software and services to hospitals.

GREATER COLLABORATION BETWEEN REMOTE INTENSI VISTS AND ON-SITE CLINICIANS IMPROVES BEST PRACTICE COMPLIANCE

Omar Badawi, VISICU. Baltimore, MD; University of Maryland, School of Pharmacy:
Baltimore. MD; Ealaf Shemmeri, University of Maryland Medical System, Baltimore, MD
Crit Care Med. 2006;34(12):A20.

INTRODUCTION: A primary goal of intensive care unit (ICU) remote management
systems is to improve compliance with best practices. This study evaluates how the level
of partnership between remote intensivists and ICU clinicians affects glycemic control (GC)
and deep vein thrombosis (DVT) prophylaxis.

HYPOTHESIS: A positive relationship exists between the level of partnership of remote
intensivists with ICU clinicians and achieving best practice treatment goals.
METHODS: A retrospective, multi-center evaluation was conducted using the eICU®
Program Network database. Patients were excluded if their attending physician had not
assigned a level of partnership for remote intensivists. Level of partnerships were: Minimal
(intervene only in emergencies); Moderate (intervene on emergencies and implement
therapies consistent with the attending physician's care plan); Intense (full management
authority in patient care). Primary outcomes were number of days with an average daily
glucose < 110 mg/dL and number of at-risk patients administered DVT prophylaxis within
48 hours. The relationship between level of partnership and outcomes were assessed with
the non-parametric test for trend.

RESULTS : A total of 7,222 patients met inclusion criteria representing 14 hospitals and
26 ICUs. DVT prophylaxis initiated within 48 hours of ICU admission was significantly
more common in the intense level of partnership group (80% of patients compared with
75% in the moderate and 68% in the minimal groups; p < 0.001). Tight GC was achieved
significantly more often in the intense level of partnership group (26% of ICU days
compared to 17% in the moderate and 18% in the minimal groups; p < 0.001).

CONCLUSIONS: Greater partnership between remote intensivists and ICU clinicians
significantly improves rates of DVT prophylaxis and tight GC. These data suggest that
compliance with best practices can be

DOES BETTER CARE MEAN SHORTER LENGTH OF STAY?

Brian Rosenfeld, VISICU, Baltimore, MD
Crit Care Med. 2006;34(12):A140.

INTRODUCTION: Standardized mortality ratios-SMR (APACHE® III) can be used to
measure ICU quality of care. Similar correlations between observed to predicted length
of stay (LOS) provide additional insight into ICU performance. Previous work (Knaus et
al.Annals of Int Med 1993) suggested that there was no correlation between hospital
mortality performance and ICU LOS. We re-examined this relationship in light of ensuing
changes in hospital critical care.

HYPOTHESIS: ICUs that deliver better care will have lower lengths of stay.

METHODS: This was a multi-center, retrospective study utilizing data from the eICU®
Program Network. APACHE III algorithms were used to generate SMRs and LOS ratios
for 19 health systems, comprising 126 ICUs and 25,404 patients during 2006. Data for
each health system were examined in aggregate (actual:predicted mortality and LOS ratios
calculated for all patients) for two successive quarters. Least square analysis was used to
assess the correlation between mortality and LOS performance; and the significance was
determined using the regression F test.

RESULTS: ICU mortality and ICU LOS performance were positively correlated (R=.6l, p<
.0001), as was hospital mortality and ICU LOS (R=.52, p<.0001). The correlation between
hospital mortality and hospital LOS was not significant.

CONCLUSIONS: ICUs and hospitals that have better mortality performance have lower
ICU lengths of stay. This suggests that better care may result in shorter LOS. Changes
in patient safety following the Institute of Medicine report, newer care modalities and/or
evidence-based critical care medicine changing the practice patterns of ICU clinicians may
be responsible.

SICKER PATIENTS - DOES BETTER CARE MEAN LONGER STAYS?

Michael Breslow, Rob Fromm, VISICU, Inc., Lutherville, MD
Crit Care Med. 2006;34(12):A126.

INTRODUCTION: Payors, regulators and quality organizations are calling for widespread
use of ICU quality measures. Mortality and length of stay (LOS) metrics are of particular
interest. The APACHE® III algorithms, which demonstrate high precision and calibration
when applied to large populations of patients, can be used to generate standardized
mortality ratios (SMR - actual deaths:predicted deaths) and actual:predicted [A:P] LOS
performance. While there is widespread agreement that high quality care reduces mortality,
the relationship between quality of care and ICU LOS may be more complex.

HYPOTHESIS: High quality care may result in longer LOS for severely ill patients.

METHODS: APACHE III algorithms were used to generate SMRs and A:P LOS data for
163 ICUs. ICUs with less than 125 scored patients were excluded from analysis. ICUs
were segregated into 3 equal groups, based upon SMR performance. For each ICU, LOS
performance (A:P) was calculated separately for patients with predicted ICU mortality rates
of <15%, 15-50% and >50%, respectively. LOS performance for each risk group was then
compared in the best and worst performing ICUs (by SMR).

RESULTS : 23,188 patients from 74 ICUs were included in the analysis. The 24 ICUs
with the lowest SMRs had A:P LOSs of 0.90, 0.99 and 1.39 in the low, medium and high
mortality risk groups, respectively. The 24 ICUs with the highest SMRs had A:P LOSs of
1.29, 1.45 and 1.02, respectively.

CONCLUSIONS: High quality ICUs, as assessed by low SMRs have lower than predicted
LOS for low risk patients but longer LOS in high severity patients. Poorer performing ICUs
do not exhibit this behavior. We speculate that high quality care may extend life in patients
at high risk of dying.

1. Kohl, B. Effect of Telemedicine on Mortality and Length of Stay in a University ICU. Crit Care Med. 2007;35(12):A22.
2. Rincon, T, Bourke, G, Ikeda, D, et al Screening for Severe Sepsis: An Incidence Analysis. Crit Care Med. 2007;35(12):A257.
3. Patel, B, Kao, L, Thomas, E, et al. Improving Compliance with Surviving Sepsis Campaign Guidelines Via Remote Electronic ICU Monitoring. Crit Care Med. 2007;35(12):A275.
4. Kohl, B, Sites, F, Gutshe, J, et al. Economic Impact of eICU Implementation in an Academic Surgical ICU. Crit Care Med. 2007;35(12):A26.
5. Hitt, J, Zawada, E, Herr, P, et al. The Economic and Clinical Value of a Remote Intensive Care Unit. Crit Care Med. 2007;35(12):A20.
6. Lewis, W. Critical-Care Nurses' Job Satisfaction and Its Effect on Retention. Crit Care Med. 2007;35(12):A22.
7. Rincon, T, Welcher, B, Srikanth, D, et al. Economic Implications of Data Collection From Remote Center Utilizing Technological Tools. Crit Care Med. 2007;35(12):A161.
8. Thomas, E, Chu-Weininger, M, Lucke, J, et al. Impact of a Tele-ICU Provider Attitudes About Teamwork and Safety Climate. Crit Care Med. 2007;35(12):A145.
9. Badawi, O. Severe Sepsis (SS) is Underreported in the ICU. Crit Care Med. 2007;35(12):A256.
10. Giessel, G, Leedom, B. Centralized, Remote Intervention Improves Best Practices Compliance. Chest. 2007;132(4):444a.
11. Howell, G, Lem, V, Ball, J, Pagels, G, et al. Remote ICU Care Reduces Health System Mortality and Length of Stay Outcomes. Chest. 2007;132(4):443b-444b.
12. Breslow, M, Larsen, E, Fromm, R, et al. ICU Length of Stay (LOS) Outliers - Incidence and Implications. Chest. 2007;132(4):442.
13. Rincon, T, Bourke, G, Ikeda, D. Centralized, Remote Care Improves Sepsis Identification, Bundle Compliance and Outcomes. Chest. 2007;132(4):557b-558b.
14. Zawada, E, Herr, P, Erickson, D, et al. Financial Benefit of a Tele-Intensivist Program to a Rural Health System. Chest. 2007;132(4):444.
15. Mora, A, Faiz, S, Kelly, T, et al. Resident Perception of the Educational and Patient Care Value from Remote Telemonitoring in a Medical Intensive Care Unit. Chest. 2007;132(4):443a.
16. Awad, O, Maselli-Caceres, D, Estrada-Y-Martin, R. Patient's Acceptance for the Advanced Monitoristems in Electronic Intensive Care Units (eICU). Chest. 2007;132(4):444b.
17. Gracias, V. Outcomes of SICU Patients After Implementation of an Electronic ICU ("eICU") System and Off-Site Intensivist. Presented at: IATSIC-AAST Conference; August, 2007; Montréal, QC, Canada.
18. Zawada, E, Aaronson, M, Herr, P, et al. Relationship Between Levels of Consultative Management and Outcomes in a Telemedicine Intensivist Staffing Program (TISP) in a Rural Health System. Chest. 2006;130(4):226s.
19. Breslow, M, Fromm, R. Sicker Patients - Does Better Care Mean Longer Stays? Crit Care Med. 2006;34(12):A126.
20. Rosenfeld, B. Does Better Care Mean Shorter Length of Stay? Crit Care Med. 2006;34(12):A140.
21. Breslow, M, Holl, R, Kuan, D. A Vital Sign Alerting System for Managing Large Numbers of Critical Care Patients Remotely. Crit Care Med. 2006;34(12):A79.
22. Badawi, O. Greater Collaboration Between Remote Intensivists and On-Site Clinicians Improves Best Practice Compliance. Crit Care Med. 2006;34(12):A20.
23. 23. Hine, J. The Effect of Instituting an ICU Electronic Medical Record on Professional Fee Billing. Crit Care Med. 2006;34(12):A21.
24. Ikeda, D, Hayatdavoudi, S, Winchell, J, et al. Implementation of a Standard Protocol
    For The Surviving Sepsis 6 and 24 Hr Bundles in Patients With an Apache III® Admission Diagnosis of Sepsis Decreases Mortality in an Open Adult ICU. Crit Care Med. 2006;34(12):A2.
25. Hayatdavoudi, S, Ikeda, D, Seiver, A, et al. Impact of a Protocol Treating Severe Sepsis
    on Renal Function and Survival of Septic Shock Patients in an Open Adult ICU. Crit
    Care Med. 2006;34(12):A18.
26. Ikeda, D, Hayatdavoudi, S, Winchell, J, et al. The Impact of Using a Standard Protocol
    For the Surviving Sepsis 6 and 24 Hr Bundles in Septic Patients on Total ICU Risk Adjusted Mortality. Crit Care Med. 2006;34(12):A108.
27. Dickhaus, D. Delivering Intensivist Services to Patients in Multiple States Using Telemedicine. Crit Care Med. 2006;34(12):A24.
28. Casal, R, Lodato, R, Weavind, L, et al. Utilization of Low Tidal Volume Strategy in Patients With Various Indications For Mechanical Ventilation and in Different ICU Settings Assessed By Remote Telemonitoring. Chest. 2006;130:134s-A.
29. Youn, B, Roche, J. ICU Process Improvement: Using Telemedicine to Enhance Compliance and Documentation for the Ventilator Bundle. Chest. 2006;130:226s-C.
30. Zawada, E, Aaronson, M, Herr, P, et al. Relationship Between Levels of Consultative Management and Outcomes in a Telemedicine Intensivist Staffing Program (TISP) in a Rural Health System. Chest. 2006;130:226s-B.
31. Aaronson, M, Zawada, E, Herr, P. Role of a Telemedicine Intensive Care Unit Program (TISP) on Glycemic Control (GC) in Seriously Ill Patients in a Rural Health System. Chest. 2006;130:226s-A.
32. Faiz, S, Zachria, D, Weavind, L, et al. Fellowship Education in Remote Telemonitoring Units. Chest. 2006;130:113s-A.
33. Faiz, S, Baig, M, Weavind, L, et al. Temporal Variability of Remote Electronic Telemonitoring Unit Interventions Between Weekdays and Weekends. Chest. 2006;130:112s-C-113s-C.
34. Kohl, B. The Effect of Instituting an ICU Electronic Medical Record on Billing and Compliance. Chest. 2006;130:112s-A.
35. Youn, B, Roche, J. Utilizing Robots and an ICU Telemedicine Program to Provide Intensivist Support For Rapid Response Teams. Chest. 2006;130:102s-A.
36. Shaffer, J, Breslow, M, Johnson, J et al. Remote ICU Management Improves Outcomes in Patients With Cardiopulmonary Arrest. Crit Care Med. 2005;33(12):A5.
37. Cowboy, E, Rajamani, M, Shanmugam, N. Impact of Remote ICU Management on Ventilator Days. Crit Care Med. 2005;33(12):A1.