Can Health Care Costs Be Reduced by Limiting Intensive Care at the End of Life?

JOHN M. LUCE and GORDON D. RUBENFELD

Department of Medicine, University of California, San Francisco, San Francisco, California; and Department of Medicine, University of Washington, Seattle, Washington

	INTRODUCTION

TOP INTRODUCTION THE ICU APPEARS TO... RESTRICTIONS ON LIFE-SUSTAINING... CONCLUSIONS REFERENCES

Total health care costs in the United States (U.S.) reached $989 billion in 1995 and now exceed $1 trillion, 14% of the Gross Domestic Product (GDP) (1). Of this total, a disproportionate share is attributable to the care of elderly patients shortly before their deaths. According to Lubitz and Prihoda (2) and Lubitz and Riley (3), 6% of Medicare recipients 65 yr of age and older who died in 1978 and 1988 accounted for 28% of all costs of the Medicare program. In the same two years, 77% of the Medicare decedents' expenditures occurred in the last year of life, 52% of them in the last 2 mo, and 40% in the last month. Inpatient expenses accounted for over 70% of the decedents' total costs.

As concern has risen over the high cost of health care, especially in the last year of life, Fries and coworkers (4), among others, have suggested that such costs might be reduced by decreasing the need and demand for medical services, particularly among terminally ill patients, elderly and otherwise, whose deaths are seemingly imminent. Furthermore, because the terminally ill often are hospitalized and may be candidates for the intensive care unit (ICU), and because the ICU is particularly resource-intensive, reducing use of the ICU among such patients appears to present unique opportunities for cost reduction (5).

A further rationale for limiting ICU care was provided by the Study to Understand Prognoses and Preferences for Outcomes and Risks of Treatment (SUPPORT) (6). This landmark study followed a large cohort of critically ill patients with a predicted 6-mo survival probability of only 52%. The SUPPORT investigators found that despite an intervention designed to improve end-of-life care, many patients who died did so not only at great expense but also after spending at least 10 d in the ICU comatose, receiving mechanical ventilation, with do-not-resuscitate (DNR) orders written 2 d before death, and in pain.

Seven years have elapsed since SUPPORT was published, but the question remains whether health care costs and patient suffering can be reduced by limiting intensive care at the end of life. In this article, we attempt to answer this question first by exploring why the ICU appears to be an ideal place to reduce costs by restricting life-sustaining treatment. We then argue that, although changes in the use of expensive critical care resources near the end of life and efforts to reduce suffering are desirable, they are unlikely to yield significant cost savings. Our article is focused on the U.S., whose ICU environment and economics we are most familiar with. Nevertheless, we hope that readers from other countries with resources such as ours will find our discussion applicable.

	THE ICU APPEARS TO BE AN IDEAL PLACE TO REDUCE COSTS BY RESTRICTING LIFE-SUSTAINING TREATMENT

TOP INTRODUCTION THE ICU APPEARS TO... RESTRICTIONS ON LIFE-SUSTAINING... CONCLUSIONS REFERENCES

Critical Care Is Very Expensive

The major reason that cost reduction should be possible in the ICU is that critical care is extremely expensive. Of the $989 billion spent on health care in the U.S. in 1995, expenditures for hospital care amounted to approximately $350 billion and constituted the largest portion (1). Assuming that ICU costs were 20% of all hospital costs, which they were estimated to be in 1986 (7), these costs were around $70 billion in 1995 or 1% of the GDP. The costs of ICU care probably are higher today, not only because total hospital costs are higher but also because ICU costs may represent a larger fraction of hospital costs, inasmuch as a greater percentage of hospitalized patients are cared for in the ICU.

The high cost of intensive care is reflected in daily ICU costs, which range from $2,000 to $3,000 in many U.S. hospitals (8). As Chaix and colleagues (9) have demonstrated, the ICU costs of individual patients may be equated with the amount of time the patients are cared for in the ICU, their length of stay (LOS). As a result, clinicians and administrators alike may assume that health care, hospital, and ICU costs can be reduced by thousands of dollars simply by decreasing ICU LOS. Such a decrease might involve not only the few patients with a long LOS, who are most expensive on a per-patient basis, but also the larger number of patients who stay in the ICU only a few days.

Considerable Variation Exists in the Use of Intensive Care at the End of Life

One technique health services researchers use to identify treatments that are overutilized is to examine variations in clinical practice. In this regard, investigators at Dartmouth (10) have demonstrated a threefold regional variation in the likelihood of admission to an ICU by Medicare enrollees during the last 6 mo of life. In 10 hospital referral regions, fewer than 20% of Medicare enrollees receive intensive care before death, whereas in 18 hospital regions more than 40% of all Medicare enrollees are admitted to an ICU before dying. While, on average, 11% of Medicare enrollees spend more than 7 d in an ICU in the 6 mo before death, the percentages vary from 3 to 25% among regions.

On Average, Patients Who Die Are More Expensive Than Those Who Survive

The relationship between cost, LOS, and mortality after critical illness has been studied for many years. For example, Detsky and colleagues (11) determined that average hospital charges were less for medical ICU patients with a good chance of survival on admission compared with patients whose survival was unlikely. The care of nonsurvivors involved a significantly higher average LOS and mean charges than the care of survivors in this study. The relationship between increased resource use and mortality is reliable enough that at least one ICU prognostic tool, the Therapeutic Intervention Severity Score, relies on this relationship to predict outcome (12).

Life-Sustaining Treatment Frequently Is Not Cost-Effective

Cost-effectiveness analysis considers both the effectiveness of a health care interventionits ability to do more good than harm when used in the usual circumstancesand the resources required to deliver the intervention (13). Cost-effectiveness usually is described in terms of the cost of an intervention per unit outcome, such as a year of life gained. A quality-adjusted life year (QALY) is the weighted average of the health-related quality of life during a year of increased survival where optimal health has a value of 1 and death has a value of 0 (14). Measures such as a QALY mean little by themselves and are most useful when the cost-effectiveness of one intervention is compared with that of another. Of course, neither or both interventions may be worth the cost, depending on one's point of view (14, 15).

Of the entire SUPPORT cohort, 2,500 patients had acute respiratory failure. A total of 963 of these patients received mechanical ventilation for severe pneumonia or the acute respiratory distress syndrome (ARDS), and only 42 had the ventilator withheld. Of the ventilated patients, 52% survived for at least 6 mo, and 45% survived at least 1 yr. Low-risk patients with an estimated 2-mo survival of greater than 70% at the time of diagnosis had a 1-yr mortality of 39%. Medium-risk patients, with a 51 to 70% estimated 2-mo survival at the time of diagnosis, had a 1-yr mortality of 61%. High-risk patients, with a 50% or less estimated 2-mo survival, had a 1-yr mortality of 79% (16).

The incremental cost per QALY of providing rather than withholding ventilator support and restorative care increased across the prognostic strata for the SUPPORT patients. The cost was $24,000 per QALY among low-risk patients, $44,000 per QALY among medium-risk patients, and $110,000 per QALY among high-risk patients. Thus, although treatment of low- and medium-risk patients appeared to be cost-effective, the incremental cost per QALY for those patients with a less than 50% chance of surviving at least 2 mo after diagnosis did not compare favorably with other medical interventions, such as therapy for severe hypertension or coronary artery bypass surgery.

Schapira and colleagues (17) found that fewer than one-fourth of cancer patients who were mechanically ventilated for acute respiratory failure survived the ICU or spent more than 3 mo at home before dying. Overall, the cost per year of life gained at home was $95,142 for patients with solid tumors and $449,544 for patients with hematologic malignancies. Wachter and coworkers (18) performed a similar analysis of patients with the acquired immunodeficiency syndrome (AIDS) who received mechanical ventilation for severe Pneumocystis carinii pneumonia (PCP). The cost of ICU admission of subsequent hospitalization for the 25% of patients who left the ICU alive averaged $174,787 per year of life saved over the 10 yr of this study. Changes in ICU LOS and survival rates led to higher and lower costs during different periods, however, illustrating how cost-effectiveness may vary over time. In all periods, the cost of care of survivors was significantly more expensive than for those dying before discharge.

Cost Reduction Strategies Are Possible in the ICU

A final reason that the ICU should be the ideal place to reduce costs is that cost savings should be achieveable in that environment. In support of this idea, Esserman and coworkers (19) retrospectively identified ICU patients who either died during admission or within 3 mo of discharge or were in the upper 25% percentile of all critically ill patients in terms of resource use. Were such recipients of "potentially ineffective care" identified prospectively and their ICU stays restricted to 5 d, the investigators estimated, charges at the study hospital could have been decreased by $1.8 to $5 million a year.

After this study, Cher and Lenert (20), demonstrated that "potentially ineffective care" occurred less among Medicare recipients enrolled in California health maintenance organizations (HMOs) than among patients in traditional fee-for-service plans, presumably because HMOs emphasize cost-effectiveness. Similarly, Rapoport and colleagues (21) found that critically ill patients who were enrolled in managed care plans in Massachusetts had lower ICU and hospital LOS and less use of mechanical ventilation than patients with traditional fee-for-service coverage.

Pronovost and Angus (14) have argued that "managed care may indirectly affect end-of-life care by requiring that ICU patients be cared for by critical care physicians." In keeping with this argument, Kollef and Ward (22) demonstrated that patients cared for by nonintensivists in a medical ICU had a higher hospital mortality, a higher ratio of actual to predicted mortality, longer ICU and hospital lengths of stay, and lower ICU mortality than patients cared for by intensivists. The lower ICU mortality and higher hospital mortality of patients cared for by nonintensivists was explained by the fact that such patients were less likely to undergo timely withholding and withdrawal of life support in the ICU.

This and other studies suggest that physicians working in the ICU might be able to reduce costs by promoting the use of advance directives and influencing patients and their families to accept palliative care when they are unlikely to benefit from restorative treatment. Rapoport and associates (23) documented that the issuing of such orders early in the ICU stay could be associated with significant reductions in resource use. Similar reductions presumably should be realized if terminal care were provided in hospices and other non-ICU settings, according to these and other investigators.

	RESTRICTIONS ON LIFE-SUSTAINING TREATMENT ARE NOT LIKELY TO YIELD SIGNIFICANT COST SAVINGS

TOP INTRODUCTION THE ICU APPEARS TO... RESTRICTIONS ON LIFE-SUSTAINING... CONCLUSIONS REFERENCES

Shortening Stays in the ICU May Not Reduce Costs

The assumption that early identification of terminally ill patients whose care can be transferred to alternative settings is intuitively appealing. Nevertheless, ample evidence exists that the strategy of shifting patients from apparently high cost to lower cost settings does not result in appreciable cost savings. For example, in an attempt to reduce health care costs in the United States, patient care was shifted from acute care hospitals to short stay surgical centers and nursing homes between 1980 and 1995. Inpatient acute care hospital stays dropped by 40% during this period. However, instead of declining, overall hospital costs actually increased, in part because a significant proportion of the anticipated savings were due to fixed costs (24).

A fixed cost is the expense a hospital or other institution incurs regardless of the volume of care it provides, for example, electricity and mortgage payments. Variable costs, such as medications, increase with each additional patient. Some costs, for example staff costs, are variable in theory, but in reality are fixed owing contractual arrangements. Average costs indicate the cost of a treatment divided by the number of patients it is provided to. The marginal or incremental cost is the cost of the next use or the cost compared to an alternative. An opportunity cost is the cost of not being able to use a resource for another purpose because it has been used up (15).

The importance of fixed costs in health care accounting was underscored by a recent study (25) in which all expenditures for Cook County Hospital were categorized as either fixed or variable. Fixed costs included such items as capital expenditures, employee salaries and benefits, building maintenance, and utilities. Variable costs included health care worker supplies, patient care supplies, diagnostic and therapeutic supplies, and medications. The total hospital budget in 1993 was $429 million, of which $360 million (84%) was fixed and $69 million (16%) was variable. Thus, the majority of cost in providing hospital services, including those in the ICU, were fixed and not amenable to cost savings through reducing LOS unless beds were closed and personnel were fired.

The apparent economic benefit of alternative care for dying, critically ill patients represents cost shifting rather than cost saving when patients do not die but instead continue to receive care elsewhere. In calculating the potential cost savings of transferring patients from the ICU to other settings, it is essential that the marginal cost of the ICU day is used and not the average cost. This calculation accounts for the fact that although the first day in the ICU may be extremely expensive, requiring one-to-one nursing care and extensive monitoring, subsequent days are likely to consume fewer resources. Actual cost savings are likely to be significantly less when comparing the marginal cost of a saved ICU day with care in a hospital ward or hospice. Care at home is an option, but in many cases patients would be cared for by female family members at a substantial opportunity cost (26).

The Most Expensive Patients Are Those with Indeterminate Outcomes

As discussed earlier, patients who die in the ICU generally are more expensive than those who survive, and the outcomes of both groups of patients should be predictable. Yet, on an individual basis, the most expensive patients are not only those with the longest ICU LOS but also those with an outcome different than what was predicted. This finding was first demonstrated in the study by Detsky and colleagues (11) cited earlier, in which patients with a long LOS who were predicted to die yet lived and patients who were predicted to live yet died were the costliest. More recently, in the SUPPORT study, 68% of patients who received the most expensive and prolonged courses of intensive care had predicted probabilities of survival between 40% and 80% estimated on the day of ICU admission (27).

Clinical prognostication in large groups of patients has been enhanced in recent years by the development of scoring systems such as SUPPORT and the Acute Physiology and Chronic Health Evaluation (APACHE) model on which it is based. Yet these systems are not accurate enough to be used in making life-or-death decisions at the bedside. Lynn and colleagues (28) observed that the median predicted chance of survival for 2 mo on the day before actual death was 0.17 and was 0.51 1 wk before death for the entire SUPPORT cohort using the scoring system derived for that study. Median prognoses varied substantially among diseases: the median for patients with congestive heart failure was a 0.62 chance of living 2 mo on the day before death, whereas patients with coma had a 0.11 chance. Median prognostic estimates were not much different when given by physicians or by APACHE.

The results of SUPPORT and other studies indicate that the most expensive ICU patients cannot easily be predicted. Furthermore, we cannot readily know in advance who will die. According to Lynn and associates (28), "Using statistical estimates of prognosis to designate a category of `terminally ill' patients for public policy purposes is unavoidably arbitrary, will often be contested, and will have differential effects upon those dying with different diseases." Estimates of cost savings by limiting care to dying patients in the ICU presume a perfect knowledge of outcomes only available retrospectively.

The Most Cost-Ineffective Care Is Provided to a Small Number of Patients

Although ICU care is not cost-effective for some patients, the total number of critically ill patients for whom such care is cost-ineffective is small compared with the entire U.S. patient population. As a result, the costs generated by such patients, although large on a per-patient basis, are small in comparison to the costs generated by patients who are not critically ill. For example, even if life-sustaining treatment had been withdrawn from all patients in the SUPPORT study with a predicted 2-mo survival of less than 1%, it would have resulted in only a 13% reduction in costs for the entire cohort. Nearly all of this savings would have been attributable to care that would have been withdrawn from 12 patients, including younger patients and those with religious convictions not to have life support withdrawn (29). This cost savings would be minimal compared with eliminating other highly cost-ineffective interventions that may be provided more frequently with less grave consequences, for example, routine magnetic resonance imaging studies for patients with equivocal neurologic complaints (30).

Although the life-years lost by forgoing a more expensive procedure such as bypass surgery may be understood only in the abstract, the life lost by transferring a patient from the ICU to a palliative care setting is personally felt by caregivers and family members. When standard cost-effectiveness analysis is applied, as it was by the Oregon Health Services Commission, many minor treatments were funded at the expense of forgoing life-saving treatments. Yet this ranking was criticized as bringing to light "a fundamental and irreconcilable conflict between cost-effectiveness analysis and the powerful human proclivity to rescue endangered life: the `Rule of Rescue' "(31). Saving costs in the ICU by reducing potentially life-saving treatments may actually be more difficult than choosing to eliminate other more effective treatments because society has such difficulty in withholding rescue treatments when available.

Just as intensive care is cost-ineffective for only a minority of patients, so do the majority seem to benefit from the ICU. Most of the SUPPORT patients with acute respiratory failure, for example, survived for at least 6 mo, and 72% of them reported that their quality of life was good, very good, or excellent (16). Similarly, of the 25% of AIDS patients with PCP who survived the ICU in the study by Wachter and coworkers (18), most were alive a year later. Danis and associates (32) have shown that patients and families of patients who have been treated in the ICU often are willing to undergo intensive care again even to achieve one more month of life.

Cost Reduction Strategies May Not Work

The study of patients receiving "potentially ineffective care" conducted by Esserman and coworkers (19) suggested that considerable savings could be realized if such patients were identified and their ICU stays reduced. The potential of this approach appeared to be confirmed when Cher and Lenert (20) demonstrated that care of patients with "potentially ineffective care" was less among patients enrolled in California HMOs. Yet the same study showed that HMO members were more likely to die within 100 d of discharge, suggesting that limiting potentially ineffective care through managed care may limit potentially effective care to critically ill patients (33).

Furthermore, it remains unclear that managed care really produces cost savings by limiting intensive care. Angus and coworkers (34) recently analyzed discharge data on adult hospitalizations in Massachusetts to determine whether resource consumption varied among patients enrolled in commercial fee-for-service, commercial managed care, fee-for-service Medicare, and Medicare-sponsored managed care. Unadjusted ICU LOS was shorter for managed care patients, but payer status had no independent effect on length of stay when age, comorbidities, severity of illness, and other patient-related factors were adjusted for.

Although managed care could reduce ICU costs by favoring the use of intensivists, and although several studies suggest that intensivists deliver more effective and inexpensive care than nonintensivists, it should be noted that these findings are not consistent. SUPPORT patients with chronic obstructive pulmonary disease (35) or congestive heart failure (36) who were cared for by pulmonologists or cardiologists enjoyed no demonstrable improvement in end-of-life care or survival compared with patients cared for by generalists, and care given by cardiologists was associated with greater costs and resource use.

We have noted that intensivists and nonintensivists alike might be able to reduce costs by promoting the use of advance directives and promoting palliative care. This approach certainly is desirable from a clinical standpoint. Furthermore, the idea that the use of advance directives would be associated with more frequent withholding and withdrawal of life support makes empiric sense. Nevertheless, the presence of advance directives has not been shown to influence the use of life-sustaining therapies in several studies. For example, increasing the documentation of preexisting advance directives was not associated with a reduction in hospital resource use in SUPPORT (37).

Advance directives are more common today than they were at the time of SUPPORT (38). Nevertheless, most patients neither have advance directives nor consider them important, especially if they do not anticipate becoming critically ill. Furthermore, intensivists frequently continue life-sustaining therapy despite patient or surrogate wishes that it be discontinued, as demonstrated by Asch and coworkers (39). Given this situation and the uncertainty of prognostication discussed earlier, it seems unlikely that even wider use of advance directives can appreciably reduce the costs of ICU care.

A final strategy, one linked with the increased advance directives, is to provide terminal care in settings other than the ICU. In analyzing this strategy, Emanuel (40) reviewed prior studies relevant to the use by patients of advance directives and hospice and determined that such use could save 25 to 40% of health care costs of the patients during the last month of life compared with the costs incurred through hospital care. Yet these savings would decrease to 10 to 17% of costs over the last 6 mo of life, when fewer of the patients would require hospitalization, and decrease further to 0 to 10% over the last year of life.

Clearly, ICU costs can be saved by closing beds or not opening them in the first place, as is the approach taken in countries with limited health care resources (41). This would lead to implicit or, less commonly, explicit rationing of intensive care. Yet the price of such rationing would be to dash the expectations of patients and sacrifice lives that could be saved or at least prolonged through ICU admission. In fact, this process has proven to be politically untenable in the U.S. and has led to growing dissatisfaction in Canada (31, 42).

	CONCLUSIONS

TOP INTRODUCTION THE ICU APPEARS TO... RESTRICTIONS ON LIFE-SUSTAINING... CONCLUSIONS REFERENCES

High cost and high mortality make the ICU an obvious location to reduce health care costs by identifying terminally ill patients and promptly instituting palliative rather than restorative care. Nevertheless, as we have tried to demonstrate, there is little evidence that terminally ill patients can be readily identified, even a few days before death in some instances. Furthermore, even if the terminally ill could be identified, it is doubtful that significant cost savings could be realized by denying their access to the ICU. Finally, at least one study suggests that indiscriminate efforts to deny access can actually increase overall mortality.

That said, we stress that providing timely and intense palliative care throughout the health care system is desirable in its own right, independent of potential economic implications. The SUPPORT study, among others, has illustrated that end-of-life care should be improved for all ICU patients, not only those who are likely near death. In countries with adequate resources such as the U.S., all critically ill patients should receive a therapeutic trial of intensive care unless they or their surrogates specify otherwise, or in the absence of available information on the patients' wishes, such care can be deemed not to be in their best interests. When restorative efforts fail, they should give way to a palliative approach if patients or surrogates agree, not because this approach will save moneywhich is unlikely to be the case for the reasons we have givenbut because palliation will be of the greatest benefit regardless of its impact on health care costs.

	Footnotes

Correspondence and requests for reprints should be addressed to John M. Luce, M.D., Division of Pulmonary and Critical Care Medicine, San Francisco General Hospital, 1001 Potrero Avenue, Room 5K1, San Francisco, CA 94110. E-mail: john_luce{at}sfgh.org

(Received in original form September 14, 2001 and accepted in revised form November 29, 2001).

	References

TOP INTRODUCTION THE ICU APPEARS TO... RESTRICTIONS ON LIFE-SUSTAINING... CONCLUSIONS REFERENCES

1. Levit KR, Lazenby HC, Braden BR, Cowan CA, McDonnell PA, Sivarajan L, Stiller JM, Won DK, Donham CS, Long AM, Stewart MW. National health expenditures, 1995.  Health Care Financing Rev 1996; 18: 175-214 . [Medline]

2. Lubitz J, Prihoda R. The use and costs of Medicare services in the last 2 years of life. Health Care Financing Rev 1984; 5: 117-131 . [Medline]

3. Lubitz JD, Riley GF. Trends in Medicare payments in the last year of life. N Engl J Med 1993; 328: 1092-1096 [Abstract/Free Full Text].

4. Fries JF, Koop CE, Beadle CE, Cooper PP, England MJ, Greaves RF, Sokolov JJ, Wright D. Reducing health care costs by reducing the need and demand for medical services. The Health Project Consortium. N Engl J Med 1993; 329: 321-325 [Free Full Text].

5. Bayer R, Callahan D, Fletcher J, Hodgson T, Jennings B, Monsees D, Sieverts S, Veatch R. The care of the terminally ill: morality and economics. N Engl J Med 1983; 309: 1490-1494 [Abstract].

6. SUPPORT Principal Investigators. A controlled trial to improve care for seriously ill hospitalized patients: the Study to Understand Prognoses and Preferences for Outcomes and Risks of Treatments (SUPPORT). JAMA 1995;274:1591-1598.

7. Jacobs P, Noseworthy TW. National estimates of intensive care utilization and costs: Canada and the United States. Crit Care Med 1990; 18: 1282-1286 [Medline].

8. Noseworthy TW, Konopad E, Shustack A, Johnston R, Grace M. Cost accounting of adult intensive care: methods and human and capital inputs. Crit Care Med 1996; 24: 1168-1172 [Medline].

9. Chaix C, Durand-Zaleski I, Alberti C, Brun-Buisson C. A model to compute the medical cost of patients in intensive care. Pharmacoeconomics 1999; 15: 573-582 . [Medline]

10. The quality of care in the last six months of life. http://www.dartmouthatlas. org/99US/toc6.php

11. Detsky AS, Stricker SC, Mulley AG, Thibault GE. Prognosis, survival, and the expenditure of hospital resources for patients in an intensive-care unit. N Engl J Med 1981; 305: 667-672 [Abstract].

12. Cullen DJ, Civetta JM, Briggs BA, Ferrara LC. Therapeutic intervention scoring system: a method for quantitative comparison of patient care. Crit Care Med 1974; 2: 57-60 [Medline].

13. Detsky AS, Naglie IG. A clinician's guide to cost-effectiveness analysis. Ann Intern Med 1990; 113: 147-154 .

14. Pronovost P, Angus DC. Economics of managing death in the ICU. In: Curtis JR, Rubenfeld GD, editors. Managing death in the ICU: the transition from cure to comfort. New York: Oxford University Press; 2001. p. 245-255.

15. Eisenberg JM. Clinical economics: a guide to the economic analysis of clinical practices. JAMA 1989; 262: 2879-2886 [Abstract/Free Full Text].

16. Hamel MB, Phillips RS, Davis RB, Teno J, Connors AF, Desbiens N, Lynn J, Dawson NV, Fulkerson W, Tsevat J. Outcomes and cost-effectiveness of ventilator support and aggressive care for patient with acute respiratory failure due to pneumonia or acute respiratory distress syndrome. Am J Med 2000; 109: 614-620 [Medline].

17. Schapira DV, Studnicki J, Bradham DD, Wolff P, Jarrett A. Intensive care, survival, and expense of treating critically ill cancer patients. JAMA 1993; 269: 783-786 [Abstract/Free Full Text].

18. Wachter RM, Luce JM, Safrin S, Berrios DC, Charlebois E, Scitovsky AA. Cost and outcome of intensive care for patients with AIDS, Pneumocystis carinii pneumonia, and severe respiratory failure. JAMA 1995; 273: 230-235 [Abstract/Free Full Text].

19. Esserman L, Belkora J, Lenert L. Potentially ineffective care: a new outcome to assess the limits of critical care. JAMA 1995; 274: 1544-1551 [Abstract/Free Full Text].

20. Cher DJ, Lenert LA. Method of Medicare reimbursement and the rate of potentially ineffective care of critically ill patients. JAMA 1997; 278: 1001-1007 [Abstract/Free Full Text].

21. Rapoport J, Gehlbach S, Lemeshow S, Teres D. Resource utilization among intensive care patients: managed care vs. traditional insurance. Arch Intern Med 1992; 152: 2207-2212 [Abstract/Free Full Text].

22. Kollef MH, Ward S. The influence of access to a private attending physician on the withdrawal of life-sustaining therapies in the intensive care unit. Crit Care Med 1999; 27: 2125-2132 [Medline].

23. Rapoport J, Teres D, Lemeshow S. Resource use implications of do not resuscitate orders for intensive care unit patients. Am J Respir Crit Care Med 1996; 153: 185-190 [Abstract].

24. Reinhardt UE. Spending more through `cost control:' our obsessive quest to gut the hospital. Health Aff (Millwood) 1996; 15: 145-154 [Medline].

25. Roberts RR, Frutos PW, Ciavarella GG, Gussow LM, Mensah EK, Kampe LM, Straus HE, Joseph G, Rydman RJ. Distribution of variable vs. fixed costs of hospital care. JAMA 1999; 281: 644-649 [Abstract/Free Full Text].

26. Emanuel EJ, Fairclough DL, Slutsman J, Alpert H, Baldwin D, Emanuel LL. Assistance from family members, friends, paid care givers, and volunteers in the care of terminally ill patients. N Engl J Med 1999; 341: 956-963 [Abstract/Free Full Text].

27. Teno JM, Fisher E, Hamel MB, Wu AW, Murphy DJ, Wenger NS, Lynn J, Harrell FE Jr.. Decision-making and outcomes of prolonged ICU stays in seriously ill patients. J Am Geriatr Soc 2000; 48: S70-S74 [Medline].

28. Lynn J, Harrell F, Cohn F, Wagner D, Connors AF. Prognoses of seriously ill hospitalized patients on the days before death: implications for patient care and public policy. New Horiz 1997; 5: 56-61 [Medline].

29. Teno JM, Murphy D, Lynn J, Tosteson A, Desbiens N, Connors AF Jr,, Hamel MB, Wu A, Phillips R, Wenger N, et al . 1994. Prognosis-based futility guidelines: does anyone win? J Am Geriatr Soc 1994; 42: 1202-1207 [Medline].

30. Mushlin AI, Mooney C, Holloway RG, Detsky AS, Mattson DH, Phelps CE. The cost-effectiveness of magnetic resonance imaging for patients with equivocal neurological symptoms. Int J Technol Assess Health Care 1997; 13: 21-34 [Medline].

31. Hadorn DC. Setting health care priorities in Oregon: cost effectiveness meets the rule of rescue. JAMA 1991; 265: 2218-2225 [Abstract/Free Full Text].

32. Danis M, Patrick DL, Southerland LI, Green ML. Patients' and families' preferences for medical intensive care. JAMA 1988; 260: 797-802 [Abstract/Free Full Text].

33. Curtis JR, Rubenfeld GD. Aggressive medical care at the end of life: does capitated reimbursement encourage the right care for the wrong reason? JAMA 1997; 278: 1025-1026 [Abstract/Free Full Text].

34. Angus DC, Linde-Zwirble WT, Sirio CA, Rotondi AJ, Chelluri L, Newbold RC 3rd,, Lave JR, Pinsky MR. The effect of managed care on ICU length of stay: implications for Medicare. JAMA 1996; 276: 1075-1082 [Abstract/Free Full Text].

35. Regueiro CR, Hamel MB, Davis RB, Desbiens N, Connors AF Jr,, Phillips RS. A comparison of generalist and pulmonologist care for patients hospitalized with severe chronic obstructive pulmonary disease: resource intensity, hospital costs, and survival. Am J Med 1998; 105: 366-372 [Medline].

36. Auerbach AD, Hamel MB, Davis RB, Connors AF Jr,, Regueiro C, Desbiens N, Goldman L, Califf RM, Dawson NV, Wenger N, Vidaillet H, Phillips RS. Resource use and survival of patients hospitalized with congestive heart failure: differences in care by specialty of the attending physician. Ann Intern Med 2000; 132: 191-200 [Abstract/Free Full Text].

37. Teno J, Lynn J, Connors AF Jr,, Wenger N, Phillips RS, Alzola C, Murphy DP, Desbiens N, Knaus WA. The illusion of end-of-life resource savings with advance directives. J Am Geriatr Soc 1997; 45: 513-518 [Medline].

38. Prendergast TJ, Luce JM. Increasing incidence of withholding and withdrawal of life support from the critically ill. Am J Respir Crit Care Med 1997; 155: 15-20 [Abstract].

39. Asch DA, Hansen-Flaschen J, Lanken PN. Decisions to limit or continue life-sustaining treatment by critical care physicians in the United States: conflicts between physicians' practices and patients' wishes. Am J Respir Crit Care Med 1995; 151: 288-292 [Abstract].

40. Emanuel EJ. Cost savings at the end of life: what do the data show? JAMA 1996; 275: 1907-1914 [Abstract/Free Full Text].

41. Zimmerman JE, Knaus WA, Judson JA, Havill JH, Trubuhovich RV, Draper EZ, Wagner DP. Patient selection for intensive care: a comparison of New Zealand and United States hospitals. Crit Care Med 1988; 16: 318-326 [Medline].

42. Donelan K, Blendon RJ, Schoen C, Davis K, Binns K. The cost of health system change: public discontent in five nations. Health Aff (Millwood) 1999; 18: 206-216 [Abstract].