Portal de Periódicos da CAPES

Is There a Starling Curve for Intensive Care?

2012; Elsevier BV; Volume: 141; Issue: 6 Linguagem: Inglês

10.1378/chest.11-2819

1931-3543

Hannah Wunsch,

Hemodynamic Monitoring and Therapy

Large differences exist in the provision of ICU beds worldwide, with a complicated mix of risks and benefits to the population of having either too few or too many beds. Having too few beds can result in delayed admission of patients to the ICU or no admission at all, with either scenario potentially increasing mortality. Potential societal benefits of having few beds include lower costs for health care and less futile intensive care at the end of life. With added ICU beds for a population, mortality benefit should accrue, but there is still the question of whether the addition of beds always means that more lives will be saved or whether there is a point at which no additional mortality benefit is gained. With an abundance of ICU beds may come the possibility of increasing harm in the forms of unnecessary costs, poor quality of deaths (ie, excessively intensive), and iatrogenic complications. The possibility of harm may be likened to the concept of falling off a Starling curve, which is traditionally used to describe worsening heart function when overfilling occurs. This commentary examines the possible implications of having too few or too many ICU beds and proposes the concept of a family of Starling curves as a way to conceptualize the balance of societal benefits and harms associated with different availability of ICU beds for a population. Large differences exist in the provision of ICU beds worldwide, with a complicated mix of risks and benefits to the population of having either too few or too many beds. Having too few beds can result in delayed admission of patients to the ICU or no admission at all, with either scenario potentially increasing mortality. Potential societal benefits of having few beds include lower costs for health care and less futile intensive care at the end of life. With added ICU beds for a population, mortality benefit should accrue, but there is still the question of whether the addition of beds always means that more lives will be saved or whether there is a point at which no additional mortality benefit is gained. With an abundance of ICU beds may come the possibility of increasing harm in the forms of unnecessary costs, poor quality of deaths (ie, excessively intensive), and iatrogenic complications. The possibility of harm may be likened to the concept of falling off a Starling curve, which is traditionally used to describe worsening heart function when overfilling occurs. This commentary examines the possible implications of having too few or too many ICU beds and proposes the concept of a family of Starling curves as a way to conceptualize the balance of societal benefits and harms associated with different availability of ICU beds for a population. Large differences exist in the provision of ICU beds worldwide. Even among the developed countries of Western Europe and North America, availability of beds varies dramatically, with up to eightfold differences among countries.1Wunsch H Angus DC Harrison DA et al.Variation in critical care services across North America and Western Europe.Crit Care Med. 2008; 36: 2787-2793Crossref PubMed Scopus (482) Google Scholar This large variation brings up the question of whether there is an optimal provision of ICU beds for a given population. Moreover, what are the consequences to society, in terms of both risks and benefits, of having either very few or many beds? There are at least three potential models of the balance between risks and benefits associated with additional provision of ICU beds for a population. The first option is the more-is-better model, where incremental increases in the provision of beds always lead to additional societal benefits that outweigh risks (Fig 1, curve A). The second is the flat model where there is a level of provision that is needed to minimize harm and the ratio of benefit to harm remains stable with any additional delivery of intensive care (Fig 1, curve B). The third model proposed—the harm model—raises the possibility of increasing harms associated with bed provision over a certain number per capita, with little benefit accrued with the added beds (Fig 1, curve C). This idea is similar to the concept of falling off a Starling curve, which is traditionally used to describe compromised heart function in an overfilled heart.2Starling EH The Linacre Lecture on the Law of the Heart. Given at Cambridge, 1915. Longmans, Green & Co, London, England1918Google Scholar It is important to recognize that this conceptual model is not unidimensional (on the y-axis); instead, it includes a broad mix of potential societal benefits and harms extending beyond mortality to encompass areas such as costs and patient and family experiences (Fig 2). Any delivery of intensive care involves a complex mix of trade-offs between the benefits and harms that will differ based on population, culture, and finances. This commentary examines the range of possible benefits and harms associated with differing provision of ICU beds for a population, with particular exploration of whether curve C (Fig 1) is plausible.Figure 2Schematic diagram of the individual potential societal benefits and harms associated with different provision of ICU beds for a population.View Large Image Figure ViewerDownload Hi-res image Download (PPT) There are potentially large societal harms associated with too little delivery of intensive care for a population. The majority of low-income countries have very few ICU beds. But often the overall health-care infrastructure, not just the delivery of intensive care, is unavailable. This fact makes it difficult to examine the true impact of a lack of ICU beds. However, other countries, such as the United Kingdom, provide universal health coverage, and yet have approximately one-sixth the ICU beds (three per 100,000 population) compared with the United States (∼20 per 100,000 population).1Wunsch H Angus DC Harrison DA et al.Variation in critical care services across North America and Western Europe.Crit Care Med. 2008; 36: 2787-2793Crossref PubMed Scopus (482) Google Scholar Studies of intensive care in Britain, described later, allow us to understand at least some of the impact of such low provision of ICU beds per capita.3Wunsch H Linde-Zwirble WT Harrison DA Barnato AE Rowan KM Angus DC Use of intensive care services during terminal hospitalizations in England and the United States.Am J Respir Crit Care Med. 2009; 180: 875-880Crossref PubMed Scopus (131) Google Scholar, 4Hutchings A Durand MA Grieve R et al.Evaluation of modernisation of adult critical care services in England: time series and cost effectiveness analysis.BMJ. 2009; 339: b4353Crossref PubMed Scopus (81) Google Scholar, 5Metcalfe MA Sloggett A McPherson K Mortality among appropriately referred patients refused admission to intensive-care units.Lancet. 1997; 350: 7-11Abstract Full Text Full Text PDF PubMed Scopus (143) Google Scholar, 6Goldfrad C Rowan K Consequences of discharges from intensive care at night.Lancet. 2000; 355: 1138-1142Abstract Full Text Full Text PDF PubMed Scopus (248) Google Scholar The most obvious potential harm results from withholding intensive care from individuals because of a lack of ICU beds. Data used to examine the consequences of denied admission are surprisingly difficult to tease apart because of the complex nature of triage decisions in general,7Azoulay E Pochard F Chevret S PROTOCETIC Group et al.Compliance with triage to intensive care recommendations.Crit Care Med. 2001; 29: 2132-2136Crossref PubMed Scopus (156) Google Scholar variation in the ability to care for patients with higher severity of illness outside an ICU,8Lieberman D Nachshon L Miloslavsky O et al.Elderly patients undergoing mechanical ventilation in and out of intensive care units: a comparative, prospective study of 579 ventilations.Crit Care. 2010; 14: R48Crossref PubMed Scopus (43) Google Scholar and different patient and family preferences regarding care.9Barnato AE Herndon MB Anthony DL et al.Are regional variations in end-of-life care intensity explained by patient preferences? A study of the US Medicare population.Med Care. 2007; 45: 386-393Crossref PubMed Scopus (358) Google Scholar Moreover, the difficulties of capturing the at-risk population, triage decisions, and outcomes for patients who have not received intensive care are logistical challenges, sometimes referred to as the denominator problem.10Wunsch H A triage score for admission: a holy grail of intensive care.Crit Care Med. 2012; 40: 321-323Crossref PubMed Scopus (5) Google Scholar Nevertheless, a number of studies from the United Kingdom demonstrate that some patients are not admitted to the ICU because of bed shortages.11Sinuff T Kahnamoui K Cook DJ Luce JM Levy MM Values Ethics and Rationing in Critical Care Task Force Rationing critical care beds: a systematic review.Crit Care Med. 2004; 32: 1588-1597Crossref PubMed Scopus (226) Google Scholar In a study by Metcalfe et al,5Metcalfe MA Sloggett A McPherson K Mortality among appropriately referred patients refused admission to intensive-care units.Lancet. 1997; 350: 7-11Abstract Full Text Full Text PDF PubMed Scopus (143) Google Scholar 817 patients referred for admission to intensive care were tracked for 90 days. Of these, 168 were refused admission to an ICU at that time, and of these, 94 (56%) did not receive intensive care specifically because of the lack of available beds. Overall, the patients who were refused admission were noted to have a higher 90-day mortality (46%) than patients who received intensive care (37%). Sinuff et al11Sinuff T Kahnamoui K Cook DJ Luce JM Levy MM Values Ethics and Rationing in Critical Care Task Force Rationing critical care beds: a systematic review.Crit Care Med. 2004; 32: 1588-1597Crossref PubMed Scopus (226) Google Scholar systematically reviewed 10 studies and found an overall odds of death of 3.04 (95% CI, 1.49-6.17) for patients refused admission compared with those admitted to an ICU. However, many of these studies had the recurring problem of mixing together patients refused admission specifically because of a lack of beds (inappropriate refusal) with patients who were refused admission for other reasons (appropriate refusal), such as perceived futility of intensive care. These patients will naturally have high death rates and confound data interpretation. The impact of having too few beds may not always be complete refusal of admission to an ICU but, rather, delayed admission. A study comparing medical admissions to the ICU in the United States and United Kingdom demonstrated that the mean time to arriving in the ICU after admission to the hospital was substantially longer in the United Kingdom.12Wunsch H Angus DC Harrison DA Linde-Zwirble WT Rowan KM Comparison of medical admissions to intensive care units in the United States and United Kingdom.Am J Respir Crit Care Med. 2011; 183: 1666-1673Crossref PubMed Scopus (179) Google Scholar Many more patients were also admitted from the wards rather than directly from an ED. Because data suggest that early intervention may improve outcomes, particularly for patients with severe sepsis and septic shock,13Kumar A Roberts D Wood KE et al.Duration of hypotension before initiation of effective antimicrobial therapy is the critical determinant of survival in human septic shock.Crit Care Med. 2006; 34: 1589-1596Crossref PubMed Scopus (4497) Google Scholar, 14Rivers E Nguyen B Havstad S Early Goal-Directed Therapy Collaborative Group et al.Early goal-directed therapy in the treatment of severe sepsis and septic shock.N Engl J Med. 2001; 345: 1368-1377Crossref PubMed Scopus (8211) Google Scholar these delays may be detrimental for patients. For example, in a study of patients with community-acquired pneumonia in the ED, patients who had delayed transfer to the ICU (defined as transfer to the wards and then to the ICU on day 2 or 3 of the hospital stay vs directly from the ED) had substantially increased hospital mortality (OR, 2.07; 95% CI, 1.12-3.85).15Renaud B Santin A Coma E et al.Association between timing of intensive care unit admission and outcomes for emergency department patients with community-acquired pneumonia.Crit Care Med. 2009; 37: 2867-2874Crossref PubMed Scopus (119) Google Scholar A similar study by Chalfin et al16Chalfin DB Trzeciak S Likourezos A Baumann BM Dellinger RP DELAY-ED study group Impact of delayed transfer of critically ill patients from the emergency department to the intensive care unit.Crit Care Med. 2007; 35: 1477-1483Crossref PubMed Scopus (785) Google Scholar of patients who stayed in the ED for > 6 h vs those transferred to the ICU in < 6 h found that delayed admission increased the risk of hospital death and longer hospital stay. Data from the United Kingdom also suggest that having few ICU beds may create pressure at the discharge end (ie, forcing a premature discharge in order to allow admission of another patient). The frequency of this practice in UK ICUs was quantified in a 1990s study that specifically examined time of and reason for discharges. The study demonstrated that 6% of UK patients in the ICU were being discharged at night, with 43% of these discharges specifically labeled as an early discharge due to shortage of ICU beds.6Goldfrad C Rowan K Consequences of discharges from intensive care at night.Lancet. 2000; 355: 1138-1142Abstract Full Text Full Text PDF PubMed Scopus (248) Google Scholar These patients experiencing premature nighttime discharge had substantially increased hospital mortality rates compared with patients discharged during the day. Minimizing the use of intensive care does have some potential societal benefits, although at very low levels of provision, these benefits are very unlikely to outweigh the substantial concerns regarding increased mortality. One benefit may be that it improves the experience of dying patients by minimizing exposure to the discomforts of intensive care at the end of life. Although one in five Americans who die receive intensive care, only one in 20 experience intensive care in the United Kingdom.3Wunsch H Linde-Zwirble WT Harrison DA Barnato AE Rowan KM Angus DC Use of intensive care services during terminal hospitalizations in England and the United States.Am J Respir Crit Care Med. 2009; 180: 875-880Crossref PubMed Scopus (131) Google Scholar The percentage of deaths that involve intensive care among children and young adults is similar in both countries, but the percentages diverge for elderly patients, with very few age > 85 years receiving intensive care before death in the United Kingdom. It is, of course, impossible to say whether some of these patients in the United Kingdom might have received benefit from intensive care. However, in either country, few people would choose to die in an ICU if given the choice.9Barnato AE Herndon MB Anthony DL et al.Are regional variations in end-of-life care intensity explained by patient preferences? A study of the US Medicare population.Med Care. 2007; 45: 386-393Crossref PubMed Scopus (358) Google Scholar, 17National Audit Office End of Life Care: Report by the Comptroller and Auditor General. HC 1043 Session. National Audit Office, London, England2008Google Scholar Another societal benefit of low delivery of intensive care may be decreased costs of care for the health system as a whole. Although the costs associated with care should not be equated with quality, the two often must be balanced. There is some correlation between the provision of ICU beds per capita and health-care spending per capita across countries in North America and Western Europe.1Wunsch H Angus DC Harrison DA et al.Variation in critical care services across North America and Western Europe.Crit Care Med. 2008; 36: 2787-2793Crossref PubMed Scopus (482) Google Scholar Many studies of intensive care equate decreasing ICU length of stay with cost savings, but these savings are likely only realized if decreasing ICU length of stay leads to decreases in the number of ICU beds and fewer patients cared for in those beds or, alternatively, by having few beds to begin with.18Luce JM Rubenfeld GD Can health care costs be reduced by limiting intensive care at the end of life?.Am J Respir Crit Care Med. 2002; 165: 750-754Crossref PubMed Scopus (173) Google Scholar This is due to the majority of costs in the ICU being fixed costs of care (salaries, costs of equipment, etc), which often are estimated to account for up to 80% to 85% of the total costs.19Roberts RR Frutos PW Ciavarella GG et al.Distribution of variable vs fixed costs of hospital care.JAMA. 1999; 281: 644-649Crossref PubMed Scopus (236) Google Scholar, 20Kahn JM Understanding economic outcomes in critical care.Curr Opin Crit Care. 2006; 12: 399-404Crossref PubMed Scopus (40) Google Scholar Therefore, minimizing the operating costs by having few beds generally will lower overall costs of care. In medicine, the assumption often is that more is better, leading to aggressive adoption of new technology,21Al-Khatib SM Hellkamp A Curtis J et al.Non-evidence-based ICD implantations in the United States.JAMA. 2011; 305: 43-49Crossref PubMed Scopus (210) Google Scholar interventions,22Carson SS Cox CE Holmes GM Howard A Carey TS The changing epidemiology of mechanical ventilation: a population-based study.J Intensive Care Med. 2006; 21: 173-182Crossref PubMed Scopus (159) Google Scholar and increasing costs of care.23Bodenheimer T High and rising health care costs. Part 1: seeking an explanation.Ann Intern Med. 2005; 142: 847-854Crossref PubMed Scopus (232) Google Scholar, 24Bodenheimer T High and rising health care costs. Part 2: technologic innovation.Ann Intern Med. 2005; 142: 932-937Crossref PubMed Scopus (249) Google Scholar In the United States, the number of ICU beds per capita has steadily increased over the past 15 to 20 years, with a 26.1% increase from 1985 to 2000 and an additional 6.5% increase from 2000 to 2005.25Halpern NA Pastores SM Thaler HT Greenstein RJ Changes in critical care beds and occupancy in the United States 1985-2000: Differences attributable to hospital size.Crit Care Med. 2006; 34: 2105-2112Crossref PubMed Scopus (94) Google Scholar, 26Halpern NA Pastores SM Critical care medicine in the United States 2000-2005: an analysis of bed numbers, occupancy rates, payer mix, and costs.Crit Care Med. 2010; 38: 65-71Crossref PubMed Scopus (668) Google Scholar There has been little examination in parallel of the potential societal benefits and harms of this high (and ever higher) provision of ICU beds for a population. The epidemiologic studies on whether higher-intensity care in the United States is associated with improved mortality are conflicting. Romley et al27Romley JA Jena AB Goldman DP Hospital spending and inpatient mortality: evidence from California: an observational study.Ann Intern Med. 2011; 154: 160-167Crossref PubMed Scopus (74) Google Scholar recently examined data from California and found an association between increased hospital spending and decreased risk-adjusted inpatient mortality across six common medical conditions. Similarly, Barnato et al28Barnato AE Chang CC Farrell MH Lave JR Roberts MS Angus DC Is survival better at hospitals with higher "end-of-life" treatment intensity?.Med Care. 2010; 48: 125-132Crossref PubMed Scopus (95) Google Scholar found an association between higher end-of-life treatment intensity and lower short-term mortality among patients in Pennsylvania, suggesting that aggressive, intense care provides mortality benefit. But the study also concluded that there were decreasing marginal returns above average intensity and that those returns were further attenuated with longer follow-up. Other work failed to demonstrate additional mortality benefit or patient satisfaction associated with care in higher-spending regions in the United States.29Fisher ES Wennberg DE Stukel TA Gottlieb DJ Lucas FL Pinder EL The implications of regional variations in Medicare spending. Part 2: health outcomes and satisfaction with care.Ann Intern Med. 2003; 138: 288-298Crossref PubMed Scopus (1120) Google Scholar With regard to mortality, therefore, it remains unclear whether there is a continued increasing benefit or a flattening of the curve with no mortality benefit beyond a certain point. Higher availability of ICU beds for specific patients, such as those undergoing high-risk surgical procedures, may provide substantial mortality benefit. In the United Kingdom, Pearse et al30Pearse RM Harrison DA James P et al.Identification and characterisation of the high-risk surgical population in the United Kingdom.Crit Care. 2006; 10: R81Crossref PubMed Scopus (515) Google Scholar found that high-risk patients representing 12.5% of total surgical procedures accounted for 80% of surgical deaths and noted that only 15% of these patients were admitted to an ICU. A direct comparison of outcomes among high-risk surgical patients in a single hospital in the United States vs a single hospital in the United Kingdom found a fourfold increased risk of death in the UK cohort, even after risk adjustment.31Bennett-Guerrero E Hyam JA Shaefi S et al.Comparison of P-POSSUM risk-adjusted mortality rates after surgery between patients in the USA and the UK.Br J Surg. 2003; 90: 1593-1598Crossref PubMed Scopus (118) Google Scholar The hospitals compared in the two countries served very different patient populations, raising the possibility of unknown confounders. However, the higher use of intensive care in the United States may at least partially explain the differences in mortality. But there may also be a point after which potential harm begins to mount. Like a house officer enthusiastically ordering fluid for an underfilled heart, the benefits for the patient may be large at first, but a point of overload may occur beyond which additional fluid can cause harm. However, it is important to recognize that the harms of aggressive use of intensive care fall across many domains and currently can only be considered in the abstract. There is certainly the possibility of unnecessary use of intensive care. Data from US studies of intensive care demonstrate low overall severity of illness and report mechanical ventilation rates of only 10% to 30%, with many patients admitted to US ICUs purely for monitoring.12Wunsch H Angus DC Harrison DA Linde-Zwirble WT Rowan KM Comparison of medical admissions to intensive care units in the United States and United Kingdom.Am J Respir Crit Care Med. 2011; 183: 1666-1673Crossref PubMed Scopus (179) Google Scholar, 32Zimmerman JE Kramer AA A model for identifying patients who may not need intensive care unit admission.J Crit Care. 2010; 25: 205-213Abstract Full Text Full Text PDF PubMed Scopus (68) Google Scholar In a model developed to identify patients in the United States who may not need ICU admission, 38.5% of all ICU admissions were for monitoring purposes, and only 11.5% of these patients went on to require any form of active treatment (defined using the Therapeutic Intervention Scoring System),32Zimmerman JE Kramer AA A model for identifying patients who may not need intensive care unit admission.J Crit Care. 2010; 25: 205-213Abstract Full Text Full Text PDF PubMed Scopus (68) Google Scholar suggesting that > 30% of ICU patients in the United States never require any ICU-level interventions. Another study of > 240,000 US patients in the ICU found that < 30% of the patients were mechanically ventilated and that < 25% received vasopressors.33Lilly CM Zuckerman IH Badawi O Riker RR Benchmark data from more than 240,000 adults that reflect the current practice of critical care in the United States.Chest. 2011; 140: 1232-1242Abstract Full Text Full Text PDF PubMed Scopus (108) Google Scholar Data are lacking on whether admission of patients to the ICU for monitoring improves outcomes in any way. A number of observational studies of patients undergoing carotid endarterectomy concluded that default admission of patients to the ICU for monitoring purposes did not improve outcomes but increased costs of care and length of hospital stay.34Kraiss LW Kilberg L Critch S Johansen KJ Short-stay carotid endarterectomy is safe and cost-effective.Am J Surg. 1995; 169: 512-515Abstract Full Text PDF PubMed Scopus (84) Google Scholar, 35Back MR Harward TR Huber TS Carlton LM Flynn TC Seeger JM Improving the cost-effectiveness of carotid endarterectomy.J Vasc Surg. 1997; 26: 456-462Abstract Full Text Full Text PDF PubMed Scopus (58) Google Scholar The concerns of iatrogenic complications,36Garrouste-Orgeas M Timsit JF Vesin A OUTCOMEREA Study Group et al.Selected medical errors in the intensive care unit: results of the IATROREF study: parts I and II.Am J Respir Crit Care Med. 2010; 181: 134-142Crossref PubMed Scopus (122) Google Scholar risk of ICU-acquired infections,37Grundmann H Bärwolff S Tami A et al.How many infections are caused by patient-to-patient transmission in intensive care units?.Crit Care Med. 2005; 33: 946-951Crossref PubMed Scopus (117) Google Scholar immobilization,38Schweickert WD Pohlman MC Pohlman AS et al.Early physical and occupational therapy in mechanically ventilated, critically ill patients: a randomised controlled trial.Lancet. 2009; 373: 1874-1882Abstract Full Text Full Text PDF PubMed Scopus (2332) Google Scholar and communication gaps associated with additional ICU admission and discharge39Horwitz LI Moin T Krumholz HM Wang L Bradley EH Consequences of inadequate sign-out for patient care.Arch Intern Med. 2008; 168: 1755-1760Crossref PubMed Scopus (277) Google Scholar, 40Bell CM Brener SS Gunraj N et al.Association of ICU or hospital admission with unintentional discontinuation of medications for chronic diseases.JAMA. 2011; 306: 840-847Crossref PubMed Scopus (219) Google Scholar may all be underappreciated risks of intensive care, particularly for patients at low risk of death (Fig 2). Many patients and families express a wish to die at home or to die with comfort measures.9Barnato AE Herndon MB Anthony DL et al.Are regional variations in end-of-life care intensity explained by patient preferences? A study of the US Medicare population.Med Care. 2007; 45: 386-393Crossref PubMed Scopus (358) Google Scholar Yet, many die in ICUs in ways that do not match their stated preferences, with ∼20% of Americans receiving intensive care before death.41Angus DC Barnato AE Linde-Zwirble WT Robert Wood Johnson Foundation ICU End-Of-Life Peer Group et al.Use of intensive care at the end of life in the United States: an epidemiologic study.Crit Care Med. 2004; 32: 638-643Crossref PubMed Scopus (902) Google Scholar Although some of this aggressive care is justifiable or may represent a change in preferences, many families of patients who die in the ICU would not choose this end-of-life experience.9Barnato AE Herndon MB Anthony DL et al.Are regional variations in end-of-life care intensity explained by patient preferences? A study of the US Medicare population.Med Care. 2007; 45: 386-393Crossref PubMed Scopus (358) Google Scholar These data suggest that there may be room to improve the delivery of end-of-life care by matching preferences with care. Such alignment does not require any rationing of care but does require more-aggressive measures by health-care workers to address end-of-life preferences early to ensure the most appropriate care.42Wright AA Zhang B Ray A et al.Associations between end-of-life discussions, patient mental health, medical care near death, and caregiver bereavement adjustment.JAMA. 2008; 300: 1665-1673Crossref PubMed Scopus (2174) Google Scholar, 43Zhang B Wright AA Huskamp HA et al.Health care costs in the last week of life: associations with end-of-life conversations.Arch Intern Med. 2009; 169: 480-488Crossref PubMed Scopus (771) Google Scholar With high availability of ICU beds, it often may be easier for clinicians to avoid the question and continue to treat. We know, for example, that it is not just variation in patient preferences that explain large regional variations in end-of-life spending across the United States.9Barnato AE Herndon MB Anthony DL et al.Are regional variations in end-of-life care intensity explained by patient preferences? A study of the US Medicare population.Med Care. 2007; 45: 386-393Crossref PubMed Scopus (358) Google Scholar However, recent data suggest that use of advance directives specifying limitations in end-of-life care do reduce the likelihood of in-hospital death and increase use of hospice care in regions with high overall end-of-life spending.44Nicholas LH Langa KM Iwashyna TJ Weir DR Regional variation in the association between advance directives and end-of-life Medicare expenditures.JAMA. 2011; 306: 1447-1453Crossref PubMed Scopus (257) Google Scholar Finally, interwoven with all of these other aspects of the delivery of critical care is the question of costs. In an idealized system with no cost constraints, spending on intensive care would be a separate concern. However, every health-care system faces budgetary restrictions that are associated with the delivery of care.45Emanuel EJ Where are the health care cost savings?.JAMA. 2012; 307: 39-40Crossref PubMed Scopus (64) Google Scholar In our current system with escalating health-care costs, large numbers of ICU beds represent an expensive fixed cost for hospitals and a large part of societal costs of health care, particularly in the United States.20Kahn JM Understanding economic outcomes in critical care.Curr Opin Crit Care. 2006; 12: 399-404Crossref PubMed Scopus (40) Google Scholar Recent efforts to quantify the costs of intensive care in the United States as a percentage of hospital costs estimated that intensive care represented between 1