Directions: please follow explicitly *** primarily this assignment is filling in the tables- attached all articles to use **** Use the attached "Literature Evaluation Table to complete this assignme

Design: We conducted a 2-year, prospective, cost-effectiveness study in 12 adult ICUs in six hospitals belonging to a large, inte- grated healthcare delivery system.

Setting: Hospitals included a large, urban tertiary referral center and five community hospitals. ICUs included medical/surgical, trauma, neurologic, and cardiac care units.

Patients: The study included 2,953 patients, 18 years old or older, with an ICU stay greater than 24 hours, who were on a ventilator for more than 24 hours and less than 14 days.

Intervention: ABCDE bundle.

Measurements and Main Results: We used propensity score- adjusted regression models to determine the impact of high bundle adherence on inpatient mortality, discharge status, length of stay, and costs. A Markov model was used to estimate the potential effect of improved bundle adherence on health- care costs and quality-adjusted life-years in the year follow- ing ICU admission. We found that patients with high ABCDE bundle adherence (≥ 60%) had significantly decreased odds of inpatient mortality (odds ratio 0.28) and significantly higher costs ($3,920) of inpatient care. The incremental cost-effec- tiveness ratio of high bundle adherence was $15,077 (95% CI, $13,675–$16,479) per life saved and $1,057 per life-year saved. High bundle adherence was associated with a 0.12 in- crease in quality-adjusted life-years, a $4,949 increase in 1-year care costs, and an incremental cost-effectiveness ratio of $42,120 per quality-adjusted life-year. Conclusions: The ABCDE bundle appears to be a cost-effective means to reduce in-hospital and 1-year mortality for patients with an ICU stay. (Crit Care Med 2020; 48:1752–1759) Key Words: cost-effectiveness; critical care; delirium T he provision of critical care is associated with high rates of morbidity and mortality and is a major source of healthcare expenditures in the United States (1). Stud- ies have shown that 20–80% of patients in the ICU develop delirium as a complication of care (2). ICU-acquired delirium is independently associated with increased cognitive and phys- ical impairment, mortality, hospital length of stay (LOS), and healthcare costs (3–10). In a recent study, Vasilevskis et al (11) found that the additional costs of care attributable to ICU de- lirium ranged from $11,132 to $23,497 per patient, and a pre- vious study reported that delirium costs $152 billion dollars annually in the United States (12). Cost-effective, scalable interventions that ameliorate ICU- acquired delirium and facilitate ventilator liberation are important for improving delivery of care and outcomes in crit- ically ill patients. The Awakening and Breathing Coordination, Delirium monitoring/management, and Early exercise/mo- bility (ABCDE) bundle (Supplemental Table 1, Supplemental Digital Content 1, http://links.lww.com/CCM/F811) is an in- terdisciplinary, multicomponent patient safety intervention designed to reduce prevalence of delirium in ICUs by improv- ing collaboration among clinical team members, standardizing care processes, and breaking the cycle of oversedation and pro- longed ventilation (13–18). Studies examining the effectiveness of the ABCDE bundle have shown significant reductions in de- lirium prevalence, ventilator days, coma days, readmission, and in-hospital mortality, and a significant increase in the number of patients who were mobilized out of bed during their ICU stay and discharged home, but few have examined its cost ef- fectiveness (2, 19–23). The objective of this study was to deter - mine the impact of ABCDE processes on inpatient mortality, LOS, discharge status, and direct costs of care (from a payer perspective), with mortality and cost outcomes serving as a basis to evaluate the cost-effectiveness of bundle adherence. DOI: 10.1097/CCM.0000000000004609 *See also p. 1897.

All authors: Center for Clinical Effectiveness, Baylor Scott & White Health, Dallas, TX.

Copyright © 2020 by the Society of Critical Care Medicine and Wolters Kluwer Health, Inc. All Rights Reserved.

Evaluating the Cost-Effectiveness of the ABCDE Bundle: Impact of Bundle Adherence on Inpatient and 1-Year Mortality and Costs of Care* Ashley W. Collinsworth, ScD, MPH; Elisa L. Priest, DrPH; Andrew L. Masica, MD, MSCI Copyright © 2020 by the Society of Critical Care Medicine and Wolters Kluwer Health, Inc. All Rights Reserved.Clinical Investigations Critical Care Medicine www.ccmjournal.org 1753 MATERIALS AND METHODS Overview This cost-effective analysis was a component of a larger ABCDE bundle implementation study that began in July 2012 in 12 ICUs of six Baylor Scott & White Health (BSWH) hospitals in- cluding a large, urban tertiary referral center and five com- munity hospitals. ICUs included medical/surgical, trauma, neurologic, and cardiac care units. The ABCDE bundle incor - porates several individual evidence-based critical care processes.

Since the prospective phase of the study was performed, the bundle has been modified slightly to include additional com- ponents (pain assessment/management and family engage- ment), and the ICU liberation approach endorsed by the Society of Critical Care Medicine is now the “ABCDEF bundle” (24). The number of care processes a patient is eligible for on a given day depends on whether or not the patient is ventilated and passes the appropriate screening criteria. Different strategies were em- ployed to improve bundle adherence in study ICUs during the first year of the study including modification of the electronic health record (EHR) to facilitate uptake and documentation of bundle elements, staff training, clinical champions, and monthly perfor - mance reports (25). Given the EHR modifications went live in July 2013, altered the documentation of bundle adherence, and likely improved the reliability and validity of the adherence data, we lim- ited this cost-effectiveness analysis to the observations obtained in the 2 years following the EHR modifications. This study was reviewed and approved by the BSWH Institutional Review Board.

Patients The 2,953 patients admitted to study ICUs from July 2013 to June 2015 who were greater than or equal to 18 years old, had an ICU admission greater than 24 hours, and were on a ven- tilator for greater than 24 hours and less than 14 days were included. Patients were excluded if they were on comfort care; were awaiting a transfer order to a non-ICU bed; had a primary diagnosis of brain tumor, mental disorder, stroke, intracranial injury, or poisoning; or had a hospital stay greater than 30 days.

Study Design We used a prospective, quasi-experimental design to examine differences in bundle adherence on in-hospital mortality, LOS, and cost outcomes. We then conducted an exploratory cost-ef- fectiveness analysis using a Markov model to estimate differ - ences in 1-year costs and quality-adjusted life-years (QALYs) for patients with low and high levels of bundle adherence from a payer perspective.

Outcome Measures We examined differences in in-hospital mortality, LOS, dis- charge status, and direct costs of hospital care among patients with varying levels of bundle adherence. Bundle adherence was calculated as the total number of care processes a patient re- ceived divided by the total number of care processes the patient was eligible for during the ICU stay. Given that the bundle con- sisted of multiple daily care processes, few patients had 100% adherence. Recognizing the potential of partial adherence to improve outcomes, we examined differences between patients with high and low bundle adherence, with high adherence de- fined as receiving greater than or equal to 60% of bundle ele- ments based on the mean level of adherence obtained in sites following ABCDE bundle implementation efforts, rather than using an all-or-none adherence measure (26). We also esti- mated differences in costs and QALYs for patients with low and high bundle adherence in the year following ICU admission.

Data Sources Process measures, demographics, and outcomes data were col- lected from the EHR and administrative databases. The cost of inpatient care was calculated as the direct care cost for each patient and was obtained from the Trendstar clinical costing system. These costs included the costs of any additional patient services, with the exception of overhead or physician fees, asso- ciated with bundle application. The cost of bundle implemen- tation was approximately $165,000 and included salary support (1.65 full-time equivalents) for the project lead, project man- ager, clinical champions, information technology personnel for EHR modifications, and data analysts plus the cost of trainings and visual aids. These sunk costs were excluded from the cost calculation. Costs were adjusted to 2013 dollars using the med- ical component of the consumer price index (27). Postacute care costs were estimated from 2014 Medicare average payments for patients based on discharge status (28).

One-year mortality rates and QALYs based on discharge status were obtained from a 1-year prospective economic evaluation of patients who received prolonged ventilation in an academic medical center ICU (29, 30).

Statistical Analysis We conducted a univariate analysis to examine unadjusted differences in patient characteristics and outcomes. We com- pared differences in continuous variables and outcomes that did not violate normality assumptions with independent t tests and differences in categorical variables and outcomes with chi- square and Fisher exact tests. Because patients with greater severity of illness and risk of mortality were more likely to have low levels of bundle ad- herence, propensity score adjustment was used to reduce the impact of selection bias on the association between bundle ad- herence and the outcomes of interest. The propensity score, the conditional probability of a patient having high bundle adher - ence, was determined from a multivariable logistic regression model based on findings from the literature (Supplemental Table 2, Supplemental Digital Content 2, http://links.lww.com/ CCM/F812). Propensity score-adjusted effects of bundle ad- herence on inpatient mortality and discharge status were mod- eled using logistic regression. A generalized linear model with a log link function and a gamma distribution was used to model direct costs due the highly skewed nature of the data (31). All statistical analyses were conducted using SAS 9.4 (SAS Institute, Cary, NC). Statistical significance was indicated at the α less than 0.05 level. Copyright © 2020 by the Society of Critical Care Medicine and Wolters Kluwer Health, Inc. All Rights Reserved. Copyright © 2020 by the Society of Critical Care Medicine and Wolters Kluwer Health, Inc. All Rights Reserved. Collinsworth et al 175 4 www.ccmjournal.org December 2020 • Volume 48 • Number 12 Cost-Effectiveness Analysis Potential patient life-years saved were calculated by estimating the number of life-years lost for each patient who died. Life expectancy was projected based on the age and sex of the pa- tient using the Social Security Administration’s actuarial life tables for 2010 (32), discounted based on the 5-year survival for patients discharged from ICUs compared with the general population (33). We calculated life-years saved as the differ - ence in projected life expectancy and the age of the patient at the time of death. We used recycled predictions to estimate the effect of high versus low adherence on outcomes. Outcomes were predicted from the modeled equations based on every patient having high adherence and every patient having low adherence. The difference between these two predictions constituted the pre- dicted mean differences in outcomes between groups. We gen- erated 1,000 bootstraps of this process to estimate the mean differences in outcomes and ses of these statistics. Bootstrap estimates obtained were used as inputs in the Markov model (Fig. 1) created with TreeAge Pro (TreeAge Software, LLC, Williamstown, MA) along with the 1-year mor - tality risks, QALYs, and costs of care obtained from the litera- ture (Supplemental Table 3, Supplemental Digital Content 3, http://links.lww.com/CCM/F813). Life expectancy estimates for the patients who died in the year following discharge were based on LOS averages for each discharge location. We assumed a life expectancy of 30 days for patients who died after being discharged home or to home health. We calculated the 1-year incremental cost effectiveness of high bundle adherence as the ratio of incremental healthcare costs in the year following ICU admission to the incremental effects (QALYs).

RESULTS A total of 2,953 eligible patients received care in the study ICUs from July 2013 to June 2015. After excluding patients with missing data, we found that 1,710 (57.9%) had high (≥ 60%) bundle adherence. Patients in the low adherence group had significantly higher all patient refined diagnosis re- lated groups Severity, Risk of Mortality, and Acute Physiology and Chronic Health Evaluation II scores indicating greater ill- ness severity (Table 1). Among the 684 patients who died in the low (< 60%) bundle adherence group, 431 (63.0%) were eligible for 10 or fewer bundle elements. Of the 318 patients who died in the high adherence group, only 57 (17.9%) were eligible for 10 or fewer bundle elements. After risk-adjustment using propensity scores, patients with bundle adherence greater than or equal to 60% had decreased odds of mortality (0.28) (Table 2). Patients with higher levels of bundle adherence had significantly increased odds of being discharged home, to home health, inpatient rehabilitation, and to a skilled nursing facility. Hospital LOS and direct costs were significantly higher in patients with bundle adherence greater than or equal to 60%, after risk adjustment. Rates of risk-adjusted compliance varied across the 12 study ICUs, but patients in cardiac ICUs were significantly less likely to have high bundle adherence compared with patients in medical/ surgical ICUs (Supplemental Table 1, Supplemental Digital Content 1, http://links.lww.com/CCM/F811). The mean effect of ABCDE bundle adherence greater than or equal to 60% on inpatient mortality and costs obtained from the bootstrap analysis was a reduction in mortality (48% vs 22%) and a $4,949 increase in direct inpatient costs (Supplemental Table 3, Supplemental Digital Content 3, http:// links.lww.com/CCM/F813). Potential life-years saved were estimated at 14.3 years per patient. Based on the inpatient mortality rates observed in the included ICUs, the incremental cost-effectiveness ratio (ICER) was calculated as $15,077 per life saved and $1,057 per life-year saved (Table 3). The 95% CI per life saved calculated by applying Fieller’s method from the bootstrap estimates was $13,675–$16,479. In the exploratory cost-effectiveness analysis using a Markov model and QALY and cost inputs from the literature to estimate potential differ - ences in outcomes and costs at 1 year for the study population based on discharge status, we found high bundle adherence (≥ 60%) was associated with a 0.12 increase in QALYs and a $4,949 increase in costs (Table 3). Based on these differences, the ICER was calculated as $42,120 per QALY. One-way sensi- tivity analysis indicated that the ICERs were most sensitive to the probability of being discharged home and the cost of hos- pitalization (Fig. 2). DISCUSSION The ABCDE bundle has been identified as a patient safety in- tervention for critically ill patients that mitigates ICU delirium and is associated with reductions in mortality as well as other deleterious outcomes (20, 34). We found that higher levels (> 60%) of bundle adherence were associated with consider - ably lower risk-adjusted odds of mortality (odds ratio = 0.28).

The results of our survival and cost analysis indicate that use of the ABCDE bundle is a cost-effective strategy for reducing mortality in ICU patients. The mortality reduction we observed in patients with higher bundle compliance is similar to findings in studies by Barnes-Daly et al (20) (odds of hospital survival increased by 7% for every 10% increase in total bundle compliance) and Pun et al (34) (hazard ratio, 0.32; 0.17–0.62 for mortality within 7 d with complete bundle compliance). Patients in our study with high bundle adherence had an increased likelihood of being discharged home or to other care facilities. Direct in- patient costs were higher for patients with higher adherence.

We did not directly examine the sources of incremental cost difference according to bundle adherence level, but a portion of that increased cost likely stems from bed charges related to the higher LOS in the greater than 60% adherence group.

Furthermore, patients with higher bundle adherence had sig- nificantly better likelihood of surviving to discharge, and that may have changed their inpatient spend trajectory (additional testing, monitoring, or other therapeutic interventions which would not have been indicated in the low adherence group). Based on the differences in inpatient mortality and costs, the ICER for high adherence to the ABCDE bundle was $15,077 per life saved and $1,057 per life-year saved. The estimated ICER for Copyright © 2020 by the Society of Critical Care Medicine and Wolters Kluwer Health, Inc. All Rights Reserved.Clinical Investigations Critical Care Medicine www.ccmjournal.org 1755 high bundle adherence in the year following hospital admission was $38,687 per QALY. These estimates are below the threshold of $50,000 per life-year or QALY frequently used as to assess the cost-effectiveness of health interventions in the United States (35). Not surprisingly, the ICERs were most sensitive to the probability of being discharged home and the cost of hospital- ization. The costs of being discharged home or to home health as opposed to a nursing facility are much lower from a health insurer’s perspective, and we observed a wide range in the per - centage of patients being discharged home across other ABCDE bundle studies. Although being discharged home may be linked to additional societal costs, we did not have the data needed to examine such costs. Hospital costs in patients with high adher - ence were approximately $4,000 greater than those with low ad- herence and served as the main source of differences in costs for patients discharged home. We used bootstrap estimates from our model to determine the range for hospital costs, as these costs were not available from previous ABCDE bundle studies. Few studies have examined the cost-effectiveness of the ABCDE bundle. Awissi et al (36) examined the cost-effectiveness Figure 1. Markov tree. Copyright © 2020 by the Society of Critical Care Medicine and Wolters Kluwer Health, Inc. All Rights Reserved. Copyright © 2020 by the Society of Critical Care Medicine and Wolters Kluwer Health, Inc. All Rights Reserved. Collinsworth et al 175 6 www.ccmjournal.org December 2020 • Volume 48 • Number 12 TABLE 1. Characteristics of Patients Admitted to Study ICUs Characteristic Bundle Adherence < 60%, n = 1,243 ≥ 60%, n = 1,710 p Age, mean ( sd) 61.1 (15.1)61.7 (15.6)0.3652 Gender (male), n (%) 696 (56.0) 971 (56.8)0.6691 Race, n (%) White 825 (66.4)1,168 (68.3) 0.2415 Black 364 (29.3)469 (27.4) Asian 41 (3.3)45 (2.6) Other 13 (1.1)19 (1.6) Ethnicity, n (%) Hispanic 169 (13.6)194 (11.4)0.0693 Insurance, n (%) Private 163 (13.1)252 (14.7)0.4679 Medicare 666 (53.6)909 (53.2) Medicaid 71 (5.7)73 (4.3) Other federal 114 (9.2)156 (9.1) Self-pay 127 (10.2)182 (10.6) Other 102 (8.2)138 (8.1) Risk factors Charlson Comorbidity Index, mean ( sd) 5.04 (2.76)4.87 (2.83)0.1108 APR-dRG severity, n (%) 1 0 (0.0)1 (0.1)< 0.0001 a 2 4 (0.3)20 (1.2) 3 93 (7.5)233 (13.6) 4 1,145 (92.2)1,454 (85.1) APR-DRG mortality risk, n (%) 1 4 (0.3)7 (0.4)< 0.0001 a 2 9 (0.7)47 (3.0) 3 158 (12.7)450 (26.4) 4 1,071 (86.2)1,204 (70.5) Acute Physiology and Chronic Health Evaluation II score, mean ( sd) 20.6 (7.0) 18.4 (6.5)< 0.001 a surgical, n (%) 162 (13.0) 280 (16.4)0.0108 a Dementia, n (%) 76 (6.1) 136 (8.0) 0.0514 Alcohol, n (%) 28 (2.3) 35 (2.1)0.7025 Current smoker, n (%) 244 (19.6) 342 (20.0)0.8035 Inpatient mortality, n (%) 684 (54.7) 318 (18.4)< 0.001 a Discharge status, n (%) Home 206 (16.5)637 (37.3)< 0.001 a Home health 40 (3.2)117 (6.8) Hospice 117 (9.4)133 (7.8) Long-term care facility 54 (4.3)124 (7.2) Inpatient rehabilitation facility 57 (4.6)149 (8.7) Skilled nursing facility 86 (6.9)232 (13.6) Length of stay (d), mean ( sd) 9.9 (7.0)12.3 (6.8)< 0.001 a Cost difference ($), n (%) 25,685 (26,370) 31,170 (33,109)< 0.001 a APR-DRG = all patient refined diagnosis related groups.a p < 0.05. Copyright © 2020 by the Society of Critical Care Medicine and Wolters Kluwer Health, Inc. All Rights Reserved.Clinical Investigations Critical Care Medicine www.ccmjournal.org 1757 of a multifaceted care processes for management of seda- tion, analgesia, and delirium and found that mean hospital costs were $933 less in the group of ICU patients treated with a sedation, analgesia, and delirium management protocol (p = 0.022), primarily due to an average 1-day reduction in LOS. Although we found greater adherence to the ABCDE bundle to be associated with an increase rather than a decrease in LOS and inpatient costs, we have found evidence that use of bundle had a statistically significant impact on decreasing in-hospital mortality. This study has several limitations. Because this was a quasi-experimental study, differences in patient characteris- tics may have influenced bundle adherence rates, potentially overestimating the impact of improved bundle adherence on outcomes. This bias may have been due to improper appli- cation of bundle inclusion criteria, poor documentation, or differences in the provision of care for patients who were se- verely ill and had a high risk of mortality. We attempted to control for this selection bias and reduce potential endoge- neity by using a propensity score risk-adjustment approach to estimate the conditional probability of a patient having high bundle adherence. However, risk-adjustment can only account for observed confounders and does not ensure a bal- anced distribution of covariates between groups. In addition, all patients in this study were critically ill, and it is difficult to differentiate levels of illness severity within this population with existing measures. Overall bundle adherence observed during the study remained relatively low, as only 58% of patients received greater than 60% of bundle elements, and among patients who died, bundle process eligibility differed greatly in the low and high adherence groups (63.0% of patients in the low adher - ence group were eligible for 10 or fewer processes compared with 17.9% in the high adherence group). This may indicate that ABCDE bundle elements were not applied to patients who died early in their hospital stay and would not have accrued benefit from the care processes, as well as those perceived as being too acutely ill for the bundle to modify mortality risk.

Removing the ICUs with the highest and lowest levels of compliance, controlling for cardiac ICUs, and controlling for bundle process eligibility at less than five and less than 10 pro- cesses, did not significantly impact the observed odds of inpa- tient mortality at the greater than 60% adherence threshold.

Given that discernment of bundle process eligibility was re- liant on extractable structured documentation in the EHR, it is possible that there are unmeasured confounders pertaining to severity of illness in the low adherence group. Accordingly, the degree of mortality reduction attributable to bundle use in our analysis is likely overestimated. In spite of this limitation, our results directionally align with other recent studies showing a risk-adjusted mortality benefit associated with ABCDE bundle adherence.

TABLE 2. The Unadjusted and Adjusted Effect of Bundle Adherence on Inpatient Outcomes Bundle Adherence Threshold 60% Unadjusted (95% CI) Adjusted (95% CI) Inpatient mortality (OR) 0.19 (0.16–0.22) a 0.28 (0.24–0.34) a Discharge status (OR) Home 2.99 (2.50–3.57) a 2.46 (2.02–2.89) a Home health 2.21 (1.53–3.19) a 1.76 (1.18–2.63) a Hospice 0.81 (0.63–1.05) 0.85 (0.64–1.14) Long-term care facility 1.72 (1.24–2.39) a 1.35 (0.94–1.94) Inpatient rehabilitation facility 1.99 (1.45–2.72) a 1.83 (1.30–2.57) a Skilled nursing facility 2.11 (1.63–2.74) a 1.61 (1.21–2.13) a Length of stay (d) 0.64 (0.51–0.76) a 0.57 (0.45–0.69) a Cost difference ($) 5,485 (2,689–8,283) a 4,067 (989–7,144) OR = odds ratio.a p < 0.05.

TABLE 3. Cost-Effectiveness of High Versus Low Bundle Adherence in Terms of Inpatient Costs and Survival Inpatient Costs and Survival Cost Per Patient Incremental Cost Inpatient Survival Rate Incremental Effectiveness Cost/ Effectiveness Incremental Cost- Effectiveness Ratio Low bundle adherence (< 60%) $28,366 0.52 $54,550 High bundle adherence (≥ 60%) $32,256 $3,9200.78 0.26$41,353 $15,077 1-yr care costs and QALYS QALYs Low bundle adherence (< 60%) $34,181 0.2237 $152,799 High bundle adherence (≥ 60%) $39,130 $4,9490.3412 0.1175$115,088 $42,120 QALYS = quality-adjusted life-years. Copyright © 2020 by the Society of Critical Care Medicine and Wolters Kluwer Health, Inc. All Rights Reserved. Copyright © 2020 by the Society of Critical Care Medicine and Wolters Kluwer Health, Inc. All Rights Reserved. Collinsworth et al 175 8 www.ccmjournal.org December 2020 • Volume 48 • Number 12 As we did not have data for patients beyond their inpatient stay, we chose to model the impact of ABCDE bundle adher - ence on 1-year outcomes based on 1-year mortality and QALY estimates obtained from another study. The patient popula- tion of that study was similar to our patient population, but patients were 5 years older on average. While we excluded patients who were on the ventilator for greater than 14 days, the other study included 114 patients (14%) who were ventilated for greater than or equal to 21 days. Thus, the mortality esti- mates obtained from the study likely overestimated the 1-year mortality risk and underestimated QALYs. In addition, more rigorous research is needed to quantify health-related quality of life among ICU survivors (37). Our basic Markov model did not account for readmissions and transitions other than from hospital to home/discharge facility and from discharge facility to home or death. We recognize that there are inherent limita- tions in the Markov model, but have included it as an explor - atory analysis and as a starting point for future research given the current lack of cost-effectiveness studies pertaining to the ABCDE bundle. CONCLUSIONS Based on findings from our study and exploratory analysis, the ABCDE bundle appears to be a cost-effective means to improve outcomes for patients with ICU stays. There is building evi- dence that consistent use of the ABCDE bundle can favorably impact a range of clinical measures, including a reduction in the risk of mortality. Further research is needed to obtain bet- ter estimates of the effects of the ABCDE bundle on total costs of care over extended time periods, including an assessment of societal costs, for patients with an index admission to the ICU.

Current address for Dr. Collin- sworth: 3M, Value Based Solutions Group, Medical Solutions Division, Dallas, TX; Dr. Priest: Baylor Scott & White Health Research Institute, Dallas, TX; and Dr. Masica: Texas Health Resources, Arlington, TX.

Supplemental digital content is available for this article. Direct URL citations appear in the printed text and are provided in the HTML and PDF versions of this article on the journal’s website (http://journals.

lww.com/ccmjournal).

Supported, in part, by the Agency for Healthcare Research and Quality (R18HS021459) and operational funds from the Baylor Scott & White Health Center for Clinical Effective- ness.

Drs. Collinsworth’s, Masica’s, and Priest’s institution received fund- ing from the Agency for Healthcare Research and Quality for article re- search. Their institution received funding from Boehringer Ingelheim, Mallinckrodt Pharmaceuticals, and the Patient Centered Outcomes Re- search Institute/People Centered Research Foundation for work unrelated to this study. Dr. Priest’s institu- tion received funding from GlaxoSmithKline.

For information regarding this article, E-mail: [email protected] REFERENCES 1. Chang B, Lorenzo J, Macario A: Examining health care costs: Oppor- tunities to provide value in the intensive care unit. Anesthesiol Clin 2015; 33:753–770 2. Kram SL, DiBartolo MC, Hinderer K, et al: Implementation of the ABCDE bundle to improve patient outcomes in the intensive care unit in a rural community hospital. Dimens Crit Care Nurs 2015; 34:250– 258 3. Inouye SK: Delirium in older persons. N Engl J Med 2006; 354:1157– 1165 4. Lundström M, Edlund A, Karlsson S, et al: A multifactorial intervention program reduces the duration of delirium, length of hospitalization, and mortality in delirious patients. J Am Geriatr Soc 2005; 53:622– 628 5. Milbrandt EB, Deppen S, Harrison PL, et al: Costs associated with delirium in mechanically ventilated patients. Crit Care Med 2004; 32:955–962 6. Pandharipande PP, Girard TD, Jackson JC, et al; BRAIN-ICU Study Investigators: Long-term cognitive impairment after critical illness. N Engl J Med 2013; 369:1306–1316 7. Pisani MA, Murphy TE, Van Ness PH, et al: Characteristics associ- ated with delirium in older patients in a medical intensive care unit.

Arch Intern Med 2007; 167:1629–1634 8. Reston JT, Schoelles KM: In-facility delirium prevention programs as a patient safety strategy: A systematic review. Ann Intern Med 2013; 158:375–380 9. Vidán MT, Sánchez E, Alonso M, et al: An intervention integrated into daily clinical practice reduces the incidence of delirium during hospi- talization in elderly patients. J Am Geriatr Soc 2009; 57:2029–2036 10. Vasunilashorn SM, Fong TG, Albuquerque A, et al: Delirium severity post-surgery and its relationship with long-term cognitive decline in a cohort of patients without dementia. J Alzheimers Dis 2018; 61:347–358 Figure 2. Tornado diagram. EV = expected value, ICER = incremental cost-effectiveness ratio. Copyright © 2020 by the Society of Critical Care Medicine and Wolters Kluwer Health, Inc. All Rights Reserved.Clinical Investigations Critical Care Medicine www.ccmjournal.org 175 9 11. Vasilevskis EE, Chandrasekhar R, Holtze CH, et al: The cost of ICU delirium and coma in the intensive care unit patient. Med Care 2018; 56:890–897 12. Leslie DL, Inouye SK: The importance of delirium: Economic and so- cietal costs. J Am Geriatr Soc 2011; 59(Suppl 2):S241–S243 13. Balas MC, Burke WJ, Gannon D, et al: Implementing the awakening and breathing coordination, delirium monitoring/management, and early exercise/mobility bundle into everyday care: Opportunities, chal- lenges, and lessons learned for implementing the ICU Pain, Agitation, and Delirium Guidelines. Crit Care Med 2013; 41:S116–S127 14. Balas MC, Vasilevskis EE, Burke WJ, et al: Critical care nurses’ role in implementing the “ABCDE bundle” into practice. Crit Care Nurse 2012; 32:35–38, 40–47; quiz 48 15. Vasilevskis EE, Ely EW, Speroff T, et al: Reducing iatrogenic risks:

ICU-acquired delirium and weakness–crossing the quality chasm.

Chest 2010; 138:1224–1233 16. Vasilevskis EE, Pandharipande PP, Girard TD, et al: A screening, pre- vention, and restoration model for saving the injured brain in intensive care unit survivors. Crit Care Med 2010; 38:S683–S691 17. Girard TD, Pandharipande PP, Ely EW: Delirium in the intensive care unit. Crit Care 2008; 12(Suppl 3):S3 18. Hipp DM, Ely EW, Dittus RS: Delirium and long-term cognitive im- pairment in intensive care unit survivors—has your team practiced its\� ABCDEs today? Federal Practitioner 2013; 30:14–18 19. Balas MC, Vasilevskis EE, Olsen KM, et al: Effectiveness and safety of the awakening and breathing coordination, delirium monitoring/ management, and early exercise/mobility bundle. Crit Care Med 2014; 42:1024–1036 20. Barnes-Daly MA, Phillips G, Ely EW: Improving hospital survival and reducing brain dysfunction at seven california community hospitals:

Implementing PAD guidelines via the ABCDEF bundle in 6,064 patients. Crit Care Med 2017; 45:171–178 21. Bounds M, Kram S, Speroni KG, et al: Effect of ABCDE bundle im- plementation on prevalence of delirium in intensive care unit patients.

Am J Crit Care 2016; 25:535–544 22. Pun BT, Balas MC, Barnes-Daly MA, et al: Caring for critically ill patients with the ABCDEF bundle: Results of the ICU liberation col- laborative in over 15,000 adults. Crit Care Med 2018; 47:3–14 23. Trogrlić Z, van der Jagt M, Bakker J, et al: A systematic review of imple- mentation strategies for assessment, prevention, and management of ICU delirium and their effect on clinical outcomes. Crit Care 2015;19:157 24. Society of Critical Care Medicine: ICU Liberation. Available at:

https://www.sccm.org/ICULiberation/ABCDEF-Bundles. Accessed September 23, 2020 25. Collinsworth AW, Masica AL, Priest EL, et al: Modifying the electronic health record to facilitate the implementation and evaluation of a bun- dled care program for intensive care unit delirium. EGEMS (Wash DC) 2014; 2:1121 26. Levy MM, Rhodes A, Phillips GS, et al: Surviving Sepsis Campaign:

Association between performance metrics and outcomes in a 7.5-year study. Intensive Care Med 2014; 40:1623–1633 27. U.S. Bureau of Labor Statistics: Economic News Release: Consumer Price Index Summary. 2014. Available at: http://www.bls.gov/news.

release/cpi.nr0.htm. Accessed September 23, 2020 28. Medicare Payment Advisory Commission: A Data Book: Health Care Spending and the Medicare Program. 2016. Available at: http://www.

medpac.gov/docs/default-source/databook/june-2016-data-book- health-care-spending-and-the-medicareprogram.pdf. Accessed Sep - tember 23, 2020 29. Cox CE, Carson SS, Govert JA, et al: An economic evaluation of prolonged mechanical ventilation. Crit Care Med 2007; 35:

1918–1927 30. Cox CE, Carson SS, Lindquist JH, et al; Quality of Life After Mechan- ical Ventilation in the Aged (QOL-MV) Investigators: Differences in one-year health outcomes and resource utilization by definition of pro- longed mechanical ventilation: A prospective cohort study. Crit Care 2007; 11:R9 31. Glick HA, Doshi JA, Sonnad SS, Polsky D: Economic Evaluation in Clinical Trials. Oxford, United Kingdom, Oxford University Press, 2007 32. Social Security Administration: Period Life Table, 2010. 2014. Avail- able at: https://www.ssa.gov/oact/HistEst/PerLifeTables/2014/Per- LifeTables2014.html. Accessed September 23, 2020 33. Wright JC, Plenderleith L, Ridley SA: Long-term survival following intensive care: Subgroup analysis and comparison with the general population. Anaesthesia 2003; 58:637–642 34. Pun BT, Balas MC, Barnes-Daly MA, et al: Caring for critically ill patients with the ABCDEF bundle: Results of the ICU liberation col- laborative in over 15,000 adults. Crit Care Med 2019; 47:3–14 35. Grosse SD: Assessing cost-effectiveness in healthcare: History of the $50,000 per QALY threshold. Expert Rev Pharmacoecon Out- comes Res 2008; 8:165–178 36. Awissi DK, Bégin C, Moisan J, et al: I-SAVE study: Impact of sedation, analgesia, and delirium protocols evaluated in the intensive care unit: \� An economic evaluation. Ann Pharmacother 2012; 46:21–28 37. Fontela PC, Abdala FANB, Forgiarini SGI, et al: Quality of life in sur- vivors after a period of hospitalization in the intensive care unit: A systematic review. Rev Bras Ter Intensiva 2018; 30:496–507