AACN Advanced Critical Care Volume 26, Number 1, pp. 13-22 © 2015 AACN
A Case for the Use of Validated Physiological Mortality Metrics to Guide Early Family Intervention in Intensive Care Unit Patients Molly F. Searl, RN, BSN, MSN, CCRN
ABSTRACT In the current health care climate a large portion of health care dollars are spent in the final months of life, so ensuring that care provided is in line with the wishes of patients and their families is more critical than ever. On the one hand, surviving families often report that they wish they had been given prognostic information earlier and that, in retrospect, they would have made different treatment decisions if they had been given prognostic information. On the other hand, providers often are reluctant to discuss prognosis for various reasons, not the least of which is the inherently uncertain nature of prognostication. To address this issue, this article reviews pertinent literature about
provider reticence, family preference, and the move toward palliative care and includes a discussion of the various validated mortality-prediction models available. A case is made to use those validated metrics to guide early discussions of palliative or endof-life care for patients who are critically ill. A suggested checklist to facilitate inclusion of prognosis discussions in family meetings is included as well as a case study to illustrate the problem, current practice, and a model for improvement. Keywords: critical care communication, end of life, mortality prediction scores, palliative care, prognosis
vasopressor support before he could be taken for confirmatory computed tomography scan. A bedside echocardiogram showed profound rightsided heart failure, with a nearly akinetic right ventricle. When his son arrived on postoperative day 1, his first question to the health care team
Mr S is a 96-year-old man with a recent history of recurrent falls and consequent orthopedic fractures. Before his most recent hospital admission, he fell and sustained a left-sided femur fracture. While being evaluated for open fixation of this fracture, he had chest pain and was diagnosed with a non–ST-elevation myocardial infarction for which he underwent a cardiac catheterization on hospital day 2. After being cleared for the orthopedic operation by cardiologists, he went to the operating room on hospital day 5 for a left femur fixation. Initial postoperative clinical findings were concerning for a pulmonary embolism, and on his first night after the operation, his clinical condition deteriorated, requiring intubation and
Molly F. Searl is Critical Care Transport Nurse, The Johns Hopkins Hospital, 1800 Orleans St, Baltimore, MD 21287 ([email protected]). The author has no financial disclosures or conflicts of interest to report. The views expressed herein are those of the author and do not necessarily reflect the views of The Johns Hopkins Hospital. DOI: 10.1097/NCI.0000000000000063
13 Copyright © 2015 American Association of Critical-Care Nurses. Unauthorized reproduction of this article is prohibited.
NCI-D-14-00031.indd 13
09/01/15 8:06 AM
SEARL
W W W.A ACNA DVA NCE D CRIT ICA LCA RE .COM
can facilitate a change from strictly curative to either a blended curative-palliative or a comfort-only model of care, both of which have been shown to possibly extend life and improve quality of life for both patients and caregivers.1,7,13 Unfortunately, fewer options currently are available to reduce the stress providers feel when providing a poor prognosis.6 Education and debriefings have been proposed as a means to reduce these feelings of stress, but few data are available on the effectiveness of these interventions.7 With regard to feelings of failure, they are relatively typical for providers who do not have training in either palliative or end-oflife (EOL) care,8 although more attention has been paid in recent years to the idea that the quality of a patient’s death should be given as much credence as the quality of his or her life.1,14 One solution proposed and widely accepted to manage both stress and feelings of failure is the integration of palliative care, through the inclusion of specially trained providers, into ICU teams.15 Alternatively, critical care providers can be given basic training on how to discuss and integrate palliative care into their daily practice. The feasibility of such interventions depends on the size of the institution, funding availability, and staff buy-in15 and is beyond the scope of the current discussion. Finally, providers cite unreliable predictors of prognosis as a reason for not initiating discussions.7 Although no provider can predict with absolute certainty when a patient will die, metrics are available that can help provide a scientifically based prediction of in-hospital mortality for patients who are critically ill.16 Few data are available on the current accuracy or frequency of prognostic discussions in ICUs. Breslow and Badawi17 found that, despite recommendations, only 10% to 15% of ICUs use mortality-predicting metrics. The recommendations cited suggest the use of prognostic metrics as indicators of quality improvement and care standardization, not as guidelines for the treatment of individual patients.17 Clinically, the general trend tends to be that prognostic information, when it is presented, is based not on metrics but on clinician experience or opinion, which can be quite varied.5,17 Research is not consistent with regard to whether discussion of prognosis decreases ICU resource use18 or length of stay,18 or increases the number of do-not-resuscitate orders,4 but when prognosis is discussed, variation in the information
was, “Can my dad make it through this?” The answer was, “We just don’t know.”
C
ommunication between physicians and patients or their families in the critical care setting has been well documented as an area in need of improvement.1–5 Also well documented is the difficult nature of discussing prognosis, specifically the disconnect between a provider’s willingness to discuss prognosis6–8 and the desire of patients and families to be given accurate information about prognosis.9–12 The purpose of this article is to make a case for the use of validated physiological prognostic metrics to aid in the discussion of prognosis in patients who are critically ill, and thus improve communication between the health care team and the families of patients in the intensive care unit (ICU). By using metrics as a guide, providers could more easily initiate early implementation of palliative or comfort care for patients who are critically ill, many of whom often undergo aggressive treatment until the final days of life.1,13 As a means of illustration, the case study described earlier is both an example of current practice and a model for improved communication. The Difficulty in Discussing Prognosis Relatively few studies have investigated the reasons why critical care providers choose not to provide prognostic information to patients or families. Those that have been conducted tend to report 3 recurring themes: (1) the belief that providing a poor prognosis will cause patients to lose hope and potentially deteriorate more quickly than if the information had been withheld6; (2) the concern that discussions of prognosis, especially when the estimated survival time is short, cause a large amount of stress among providers,6 often making them feel as if they have failed in their duty to care for their patient8; and (3) the idea that prognosis is uncertain and there is no reliable way to predict the course of any individual’s disease with absolute certainty.7,8 Regarding clinical deterioration, studies have reported that providing prognostic information to patients, even if the prognosis is grim, actually fosters hope, because the patients have a realistic notion of their disease trajectory and can trust their providers to be honest with them, regardless of the situation.7 Early and frank discussions of prognosis also 14
Copyright © 2015 American Association of Critical-Care Nurses. Unauthorized reproduction of this article is prohibited.
NCI-D-14-00031.indd 14
09/01/15 8:06 AM
VOL UME 2 6 • N U MBER 1 • JANUARY- M ARCH 2015
VA LIDAT E D P H YSIOLOGICA L M ORTA LIT Y M E T RICS IN T H E ICU
presented to patients and families has been cited as a significant barrier to effective EOL care in the ICU2 and can lead to ineffective use of palliative care,19 increased caregiver stress,9 and poor satisfaction as reported by surviving family members.3 The difficulty in providing prognostic information in the critical care setting is multifaceted but can be generally summarized by recognizing that clinicians, be they doctors, nurses, nurse practitioners, or any other members of the health care team, have difficulty reconciling the seemingly mutually exclusive mandates to save lives and provide effective palliative care.19 Often, the treatments with the greatest likelihood of success are those that cause distressing symptoms in patients,20 and those symptoms often are the aspects of EOL that cause the most dissatisfaction among surviving families.3
pare can be addressed through early initiation of family meetings,13 a key part of which should be—based on family surveys—a discussion of prognosis.9 Qualitative studies have shown that the desires described earlier, specifically the desire to have discussions of prognosis, contribute significantly to improving family satisfaction, which is increasingly recognized as a measure of successful treatment.21 Families have reported that their treatment decisions, made on behalf of their loved ones, may have been different had they been informed of the objective prognosis early in the disease course,10 but data are mixed. Both the seminal Study to Understand Prognoses and Preferences for Outcomes and Risks of Treatments (SUPPORT)4 trial and newer research by Daly and colleagues18 report no change in EOL decision making after the initiation of structured family communication that included discussions of prognosis. No information is available about the source—provider experience or physiological metric—of that prognostic information, although Daly and colleagues18 report that Acute Physiology and Chronic Health Evaluation (APACHE) III data were collected and scores were recorded as part of the patient demographic information. Note that mortality scores report predicted survival, but the perception of what survival actually means can be very different between families and providers.10 Specifically, families often consider survival to mean a return to previous functional status,12 whereas survival to providers often means “alive at discharge,” with less focus on the postillness functionality of the patient.12 In fact, with the advent and proliferation of life-sustaining treatments, patients and families tend not to understand that medical intervention can—on numerous occasions—only serve to prolong the dying process12 and is often the source of many distressing symptoms20 identified by family members as a significant source of regret after the death of their loved one.3 Providers must recognize the need to reduce suffering and not allow the potential adverse effects of treatment of distressing symptoms to prevent such treatment.20 As an example, a provider should not expect a patient to tolerate pain (a distressing symptom) to prevent or manage hypotension (a physiological condition), because treatment of the symptom (pain) may exacerbate the condition (hypotension), without first discussing treatment options and possible consequences
“Mr S cannot possibly survive this” was the general consensus on the part of not only the ICU team caring for him but also the primary surgical team on postoperative day 1. This view was expressed by multiple care providers during evening rounds. Mr S’s son was present at the bedside throughout the day, although he was not present during rounds when this discussion occurred. When providers entered the patient’s room after rounds, Mr S’s son asked, “How’s Dad doing?” to which the response was, “He is very sick. He needs a lot of help maintaining his blood pressure and breathing right now.” No discussion of data-driven prognostic information was offered.
The Need to Discuss Prognosis Family Perspective
When surveyed, families of critically ill patients report that their main concerns—after curing the underlying disease—are to ensure that their loved one does not suffer unduly,6 that distressing symptoms are addressed and treated in a timely fashion,20 and that they be allowed to share in decision making9 in an effort to prepare themselves for the possibility that their family member might not survive.10 Issues of suffering and symptom relief can be addressed with the inclusion of palliative care in all critical care settings, regardless of prognosis,15 but this, again, is beyond the scope of this discussion. Shared decision making and time to pre15
Copyright © 2015 American Association of Critical-Care Nurses. Unauthorized reproduction of this article is prohibited.
NCI-D-14-00031.indd 15
09/01/15 8:06 AM
SEARL
W W W.A ACNA DVA NCE D CRIT ICA LCA RE .COM
with the patients and families.20 Avoidance of suffering, or improved quality of life, has been cited as a primary goal for families and patients during illness,22 and research supports the notion that morbidity rather than mortality is a more accurate term to describe the concerns of patients and families.23 Although many researchers have compiled data on objective mortality-prediction measures, data reflecting morbidity are, by necessity, based on surveys of survivors and their families. These surveys, collectively known as health-related quality of life surveys, collect subjective data on a number of variables, including functional status, anxiety, and depression, as well as objective measures, such as delirium and pulmonary, renal, and cognitive function.23 When combined, these measures give a picture of functional status, which is often decreased, after critical illness.23 Data report that patients who require longer and more aggressive care are likely to experience greater functional dependence after critical illness.24,25 Patients with comorbid conditions on admission and patients who are older or require ventilator support for longer periods of time are at highest risk for postillness sequelae.24 These same physiological variables also are used to calculate mortality scores; thus, patients with higher mortality scores on admission are more likely to require aggressive care26 and, therefore, are more likely to experience long-term functional consequences. Therefore, logically, predictors of mortality—that is, prognostic metrics—can be extrapolated to predict postillness morbidity for those patients who survive their critical illness. This postillness morbidity is, as mentioned earlier, generally the most important factor for patients and their families and often is not discussed during family meetings.27
the provider is not completely sure.9 In fact, the tendency of providers to tell families that it is best to “wait and see” before discussing prognosis,11 or the fact that different providers often present different prognoses,2 is actually detrimental to families and limits their ability to process information and make appropriate treatment decisions.10 Conflicting opinions about prognosis and efficacy of treatment often lead to staff burnout related to caring for patients who are chronically critically ill,2 and the hesitancy on the part of providers to discuss prognosis has been shown, in some cases, to actually contribute to a decrease in frequency of family conferences as length of stay increases, when illnesses become more complicated, and the need for realistic discussions of prognosis becomes more pressing.5 Postoperative day 2: Mr S has not improved. His liver function has decreased, demonstrating catastrophic liver injury; his kidney function has deteriorated; and he still requires maximal vasopressor and ventilator support. His son remains at the bedside, and frequently asks how his father is doing. During a discussion with his father’s physician, he relayed the story of his mother’s death approximately 1 year before. She had suffered a relatively minor injury and had developed sepsis as a result. He stated, in passing, that he did not want his father to suffer like that, relying on machines to keep him alive. Again no discussion of prognosis was offered.
Arguably, a disconnect exists in perception that leads to a demonstrable lack of communication between providers and families.3 That disconnect can be traced, at least partially, to discomfort on the part of providers to discuss prognosis in concrete terms, because they are uncomfortable in providing such information without supporting evidence, because there are conflicting opinions on prognosis, or because they are justifiably uncomfortable with or undertrained in how to provide potentially devastating news to patients and their families. The question then becomes whose opinion on prognosis should be used to guide discussions. Should it be the primary service attending, in the case of a surgical ICU patient, who has seen thousands of postoperative liver cancer patients as an example? Or should it be the intensivist who has cared for thousands of critically ill individuals regardless of the underlying pathology? Or should it be the palliative care team,
Provider Perspective
Perhaps the most substantial argument against discussing prognosis, from a provider’s perspective, is uncertainty; even the most accurate prognostic metrics are variable and designed specifically to aggregate mortality information for groups rather than predict mortality for specific patients.17 Research supports the idea that providers are generally less likely to provide specific information about prognosis than any other aspect of health care,6 even though family members report that they generally would prefer a discussion of prognosis, even if 16
Copyright © 2015 American Association of Critical-Care Nurses. Unauthorized reproduction of this article is prohibited.
NCI-D-14-00031.indd 16
09/01/15 8:06 AM
VOL UME 2 6 • N U MBER 1 • JANUARY- M ARCH 2015
VA LIDAT E D P H YSIOLOGICA L M ORTA LIT Y M E T RICS IN T H E ICU
whose job is to focus care on comfort first, either as an alternative to or in collaboration with curative efforts? In an attempt to remove opinion from the already-complicated discussion, the use of valid and proven prognostic metrics is a reasonable and relatively easy solution.
Table 1: Advantages, Limitations, and General Considerations of 3 Prognostic Metricsa Advantages All have well-established discrimination and calibration in general ICU patients APACHE, SAPS, and MPM have been validated in large numbers of clinical trials spanning all critical care settings (ie, surgical, medical, neurological)
Prognostic Metrics Various prognostic metrics that include both general prognostic scores and organ dysfunction scores are available for providers.28 For the purpose of this discussion, only general prognostic metrics are reviewed; the most commonly used metrics are the APACHE, Simplified Acute Physiology Score (SAPS), and Mortality Prediction Model (MPM).29 These 3 metrics have been found to have good discrimination—the ability to identify ICU patients with the highest risk of death—and good calibration—the extent to which the predicted outcome matches observed outcome.17 In brief, these prognostic metrics were developed in an effort to better categorize disease severity and to standardize research in the critical care setting, and as a basis for comparing care across ICUs.16 See Table 1 for a general appraisal of advantages and limitations common to all 3 metrics. A more detailed discussion of each metric follows.
Based on physiological data that, when combined, provide accurate mortality predictions Limitations Not well studied for specific ICU patient populations (eg, AIDS, postpartum) or in ICUs with specialized treatment modalities (eg, ECMO) Some disease processes have expected trajectories that can change scores based on when they are calculated (eg, expected low GCS with new SAH patients could improve with time) Lead-time bias—the change in score based on treatment before data collection—can alter scores, especially in patients who have been transferred from one facility to another General considerations All scores can be affected by improvements in medical technology, implementation of best practice recommendations (eg, early antibiotics for suspected or confirmed sepsis), or changes in practice (eg, new providers rotating through the ICU)
Acute Physiology and Chronic Health Evaluation
Originally proposed in 1981, the APACHE system was designed to provide an objective classification of risk in critically ill patients.30 The author of this study acknowledged the inherent reluctance on the part of physicians to discuss prognosis for individual patients, so this particular tool was designed to refer to groups of patients with similar physiological variables, not to guide individual treatment decisions.30 As a result of polling experts for the variables deemed “important,” Knaus and colleagues30 derived a weighted classification system that allowed comparison of an individual patient to a group of similar patients to predict mortality. This original scoring system, as with all such prediction metrics, was found to have less discrimination over time, largely due to advances in medical technology. Consequently, the APACHE score has undergone many revisions, the most current being the APACHE IV, which includes measures of length of stay and incorporates logistical regression to provide users with a predicted mortality score as a percentage.31
Abbreviations: AIDS, acquired immunodeficiency syndrome; APACHE, Acute Physiology and Chronic Health Evaluation; ECMO, extracorporeal membrane oxygenation; GCS, Glasgow Coma Scale; ICU, intensive care unit; MPM, Mortality Prediction Model; SAH, subarachnoid hemorrhage; SAPS, Simplified Acute Physiology Score. a Based on data from Kelley.16
The physiological values required to calculate an APACHE IV score for Mr S are listed in Table 2; values included therein reflect the worst values in the first 24 hours as stipulated by the scoring system. A primary diagnosis that led to ICU admission also is required for an APACHE score.17 Previous iterations of the APACHE were limited by their proprietary nature, meaning that hospitals were required to purchase costly software programs to access the information; however, APACHE IV is now available online and free of charge at http:// www.apachefoundations.cernerworks.com.31 Other freely available online severity scoring calculators (http://clincalc.com) report APACHE II, 17
Copyright © 2015 American Association of Critical-Care Nurses. Unauthorized reproduction of this article is prohibited.
NCI-D-14-00031.indd 17
09/01/15 8:06 AM
SEARL
W W W.A ACNA DVA NCE D CRIT ICA LCA RE .COM
Table 2: Raw Data Used to Calculate Mr S’s Mortality-Prediction Scoresa Patient Data Temperature, °C Heart rate, beats/min Respiratory rate, breaths/min
On Admission
24 Hours After Admission
36.3
35
93
107
24
30
72/37 (49)
120/49 (66)b
150
0
White blood cell count, 10 /μL
29.7
31
Hematocrit, %
27.2
28.2
Serum sodium, mEq/L
143
142
Serum potassium, mEq/L
3.8
4
Serum bicarbonate, mEq/L
18
19
BUN, mg/dL
26
34
Serum creatinine, mg/dL
0.91
1.67
Serum albumin, g/dL
2.1
1.6
BP (MAP), mm Hg Urine output, mL/d 3
Serum bilirubin, mg/dL
1.1
2.5
Serum glucose, mg/dL
110
139c
GCS
3T
3T
ABG FIO2, %
100
80
ABG PaO2
166.5
93.7
ABG PaCO2
21.5
29.3
7.2
7.4d
APACHE IV (predicted mortality), %
82.5
87.9
SAPS II (predicted mortality), %
88.9
89.7
MPM II (predicted mortality), %
90
87.3
ABG pH
Sources for predicted mortality data APACHE IV: http://www.apachefoundations.cernerworks.com SAPS II: http://clincalc.com/IcuMortality/SAPSII.aspx MPM admission: http://www.sfar.org/scores2/ mpm2_admission2.php MPM 24 hours: http://www.sfar.org/scores2/ mpm2_24_48_722.php Abbreviations: ABG, arterial blood gas; APACHE, Acute Physiology and Chronic Health Evaluation; BP, blood pressure; BUN, blood urea nitrogen; FIO2, fraction of inspired oxygen; GCS, Glasgow Coma Scale; MAP, mean arterial pressure; MPM, Mortality Prediction Model; PaCO2, partial pressure of carbon dioxide; PaO2, partial pressure of oxygen; SAPS, Simplified Acute Physiology Score. a None of the scores has a caveat for the use of medical support beyond ventilators (accounted for in the ABG values) or sedation medication (accounted for in the GCS). The use of vasopressors to sustain blood pressure or insulin to control blood glucose is not noted and thus does not change the weight of any individual scores. The supportive measures noted for some values in the 24 Hours After Admission column are for the edification of the readers and are meant to clarify that some of the values presented, while they appear improved after 24 hours, are in fact with more support than previously. b Norepinephrine (Levophed) infusion at 0.12 mcg/kg per minute and vasopressin infusion at 0.04 units per hour. c Insulin infusion at 5 units per hour. d Sodium bicarbonate drip at 150 mL/h.
18 Copyright © 2015 American Association of Critical-Care Nurses. Unauthorized reproduction of this article is prohibited.
NCI-D-14-00031.indd 18
09/01/15 8:06 AM
VOL UME 2 6 • N U MBER 1 • JANUARY- M ARCH 2015
VA LIDAT E D P H YSIOLOGICA L M ORTA LIT Y M E T RICS IN T H E ICU
SAPS, and other organ-specific mortality scores. APACHE scores also require a large amount of clinical data, and without automatic chart review processes, data entry can be labor intensive. In addition, when data are not available, the software assumes a normal value, which can potentially lower the calculated score, which was a conscious choice made by the original APACHE authors based on the assumption that if a value is not measured the providers at the time did not suspect it would be abnormal.30 Being the most recent iteration of the APACHE scoring system, the APACHE IV score calculated via the above website is included for Mr S.
and at least one study found it valid in a rural community hospital.35 A major difference between MPM and SAPS/APACHE is that the variables in MPM are binary, that is, marked as present or not present, thus significantly reducing the amount of time and specificity of data to be collected.16 Another difference is that the MPM produces a predicted mortality percentage, rather than a score that is then translated into a percentage.
Based on clinical data for Mr S, his initial APACHE IV score was 146, with a predicted mortality of 82.5% for this hospitalization. Data are included in Table 2.
These metrics have limitations on their effectiveness in certain patient populations or under certain conditions,16 so clinical judgment should guide the timing of score-driven family meetings in these circumstances. See Table 1 for a summary of these issues. Regardless of the metric chosen, the basic underlying supposition is that a set of physiological variables, when combined, can point toward a potential for mortality. Although none of the authors who designed or validated these scoring metrics advocates their use in guiding treatment for specific patients, their use in reducing interprovider disagreement about prognosis cannot be denied.
Based on clinical data for Mr S, his initial MPM predicted mortality rate was 90% for this hospitalization. Data are included in Table 2.
Simplified Acute Physiology Score
From the original APACHE data, Le Gall and colleagues32 created the SAPS, which was based on the same principles as APACHE but sought to streamline and simplify the process by decreasing the number of variables required. As with APACHE scores, SAPS has been validated throughout critical care populations and has undergone 3 revisions, the most recent of which, SAPS 3, includes customized variables for geographic regions and has helped increase discrimination.28 Perhaps the most important difference between SAPS and APACHE is the recognition on the part of Le Gall and colleagues33 that selecting a single primary diagnosis for patients in the ICU with complicated health issues and often with comorbidities is difficult, requiring provider judgment and opinion to determine the primary diagnosis.
By postoperative day 7, Mr S continued to require full ventilator support and multiple vasopressors, and his kidney function had deteriorated to the point where he needed dialysis—a treatment deemed futile by the Renal Service. While his cardiac function and laboratory tests showed some improvement, his respiratory status was complicated by a multidrug-resistant Proteus pneumonia, and he never regained consciousness despite a lack of sedation medications. Attempts were made by the ICU team to procure a do-not-resuscitate order during the day on postoperative day 6, but again with no concrete discussions of prognosis, Mr S’s son continued to cling to the hope that his father might recover. Frustrated with the discussions he had with the ICU team and his perception that the team had “given up without reason,” Mr S’s son requested that the patient be transferred to another facility, which occurred on postoperative day 8.
Based on clinical data for Mr S, his initial SAPS score was 75, with a predicted mortality of 88.9% for this hospitalization. Data are included in Table 2.
Mortality Prediction Model
Finally, in an attempt to remove expert opinion and therefore clinician bias, Lemeshow et al34 created the MPM, which derived variables and weights through statistical regression. The MPM is designed to have data entered at 2 predetermined times, on admission and at 24 hours after admission. As with the other 2 scores, MPM has been validated and found to be accurate across a large number of ICU populations,
Recommendations Intensive care practitioners have a duty to communicate honestly and openly with patients and 19
Copyright © 2015 American Association of Critical-Care Nurses. Unauthorized reproduction of this article is prohibited.
NCI-D-14-00031.indd 19
09/01/15 8:06 AM
SEARL
W W W.A ACNA DVA NCE D CRIT ICA LCA RE .COM
their families about disease prognosis. Rather than continue with the current practice of basing such discussions on individual provider preferences or experiences, validated physiologically based prognostic metrics should be considered. Best practice encompasses accurate diagnosis, effective and timely use of available treatments,26 and discussions of palliative care; although palliative care measures will not affect mortality scores, they will affect satisfaction scores from survivors, which are given nearly equal weight as benchmarks for successful care.21 Although it is ideal, frequent contact between patients and their families and providers is often cost and/or time prohibitive.7 For providers to effect the most positive change, all patients should receive a mortality prediction score on admission to the ICU. The scores presented here have been shown to provide similarly accurate information, and the use of one score over another should be left to the discretion of the individual institution on the basis of staff availability for data entry and common clinical practice. Given the research on effective implementation of family meetings and palliative care, the best practice option would be to integrate a multidisciplinary palliative care team in any such meeting.15 As a secondary, and potentially more feasible, option, specific providers could be trained to have discussions about palliative care and prognosis. If neither of these options is available, any provider can access any of the aforementioned prognostic metrics and open a dialogue with families based on the mortality score. As for which patients need family conferences, those patients with higher severity scores (60% or greater), and therefore a higher probability of mortality, should have a family conference scheduled within the first 48 hours if possible, although timing could change for individual patients with certain illnesses or injuries; see Table 1. For patients with scores in the midrange (30%-60%), family conferences should be initiated within 72 hours of admission. For patients with low mortality scores (< 30%), who have the highest likelihood of returning to function, family conferences could be delayed or even forgone completely. Lead-time bias, or the effect of treatments initiated before data collection, can affect mortality-prediction36 scores, but this should not—as a rule—prevent score calculation or early family meetings. In an effort to address concerns raised by bedside nurses,
specifically a lack of protocols for discussions of prognosis, and the view that families are often given conflicting information about prognosis,2 a checklist incorporating the suggested prognostic levels above is included in Table 3. As with most physiological variables, trends are more significant than individual numbers, so periodic recalculation of prognostic scores is suggested. If the MPM model is used, an automatic recalculation is available, in that online calculators have options for admission data and data at 24, 48, and 72 hours, respectively. Although APACHE and SAPS use the worst values collected in the last 24 hours, no provision is made for recalculation. The checklist in Table 3 includes suggested decision points that prompt
Table 3: Suggested Checklist for Implementing Family Meetings Based on Mortality Predictions Patient Name:
Date:
Time:
Admission APACHE II Score/ Predicted Mortality: Family meeting indicated for > 30%: Within 48 hours (predicted mortality > 60%)
□
Within 72 hours (predicted mortality 30%-60%)
□
Meeting scheduled for/with: Advance directive/living will? Reviewed: □
□ YES
□ NO
□ YES
□ NO
Reevaluate in 72 hours if patient is still in the intensive care unit or with worsening clinical conditiona Repeat APACHE II Score/ Predicted Mortality: Date/Time: Family Meeting Indicated? (predicted mortality > 30% or change in score ≥ 15%) Meeting Scheduled for/with: Abbreviation: APACHE, Acute Physiology and Chronic Health Evaluation. a Changes in condition include but are not limited to: addition of vasopressors, emergent surgery, increase in oxygen requirements, unplanned intubation, decrease in mental status, and cardiac/ respiratory arrest.
20 Copyright © 2015 American Association of Critical-Care Nurses. Unauthorized reproduction of this article is prohibited.
NCI-D-14-00031.indd 20
09/01/15 8:06 AM
VOL UME 2 6 • N U MBER 1 • JANUARY- M ARCH 2015
VA LIDAT E D P H YSIOLOGICA L M ORTA LIT Y M E T RICS IN T H E ICU
recalculation of scores at a specified time or with major changes in patient status, as defined on the checklist. Table 4 includes a suggested script for providers to discuss prognosis based on metrics, using Mr S as an example.
Summary In 1998, the Institute of Medicine defined a good death as “one that is free from avoidable distress and suffering for patients, families, and caregivers; in general accord with patients’ and families’ wishes; and reasonably consistent with clinical, cultural, and ethical standards.”14(p1) Based on that recommendation, and the recognition that EOL in hospitals, especially in critical care, can be unduly and unnecessarily distressing to families,10 a current trend in health care is to provide quality and timely palliative care.1 Considering these factors, practitioners must initiate discussions of palliative care early in a patient’s critical illness, regardless of prognosis,20 but especially when the probability of death is high.10 Arguments can be made that prognostic metrics are not fail-safe or meant to guide treatment decisions for individual patients, but in light of the fact that families often desire prognosis information,2 even if that information is uncertain,9 and given the frequent disagreements between providers on prognosis,2,19,22 a case can be made that using these metrics to guide early discussions of palliative or EOL care could (a) improve family satisfaction with the dying process in hospitals,20 (b) give families more time to comprehend and process the nature of their loved one’s prognosis,10 (c) decrease staff stress when caring for a patient that “everyone knows is going to die,”2 and most importantly, (d) improve the quality of health care at EOL. The suggestions presented here should be subjected to academic scrutiny, but the purpose of this article is to raise awareness of a practice deficit and offer a solution. Likewise, the purpose is not to advocate for the withdrawal or restriction of life-sustaining therapies for patients. Nor is it to advocate the use of prognostic metrics to arbitrarily guide care for individuals. Instead, given the difficult nature of both prognostication and discussions of EOL, this article simply suggests the use of validated physiological metrics to aid in discussions with families and patients in the critical care setting about goals of care. By using established objective prognostic tools, clinicians can ensure that families are given the same information about prognosis, and by initiating discussions about the family’s or the patient’s wishes early, providers can help families prepare for poor outcomes. Most important, with improved communication comes improved quality of care and, for some patients, improved quality of their death.
Table 4: Alternate Script Illustrating Use of Mortality-Prediction Scores in Family Meeting Provider:
Mr S, your father’s condition is very serious. He couldn’t breathe on his own so we had to put in a breathing tube. His blood pressure is very low, so we started him on medications to keep it up. And this morning, we did an echocardiogram, a test that shows us how well his heart is beating, and it showed that half of his heart isn’t working.
Family:
Can my dad make it through this?
Provider:
In a man of his age, with all of the problems he has now, there is an 83% chance of death. We have tools that help us predict if a person will survive a hospital stay. Now, these tools are not 100% accurate. But they do tell us that 83% of people with your father’s issues will die.
Family:
So it’s not for sure he’ll die? He was healthy and independent until this happened.
Provider:
No, we can’t say 100% that he will die, but it is very likely that he will not survive this, and that if he survives, he likely won’t be able to get back to his independent life. Did you father have an advance directive or a living will?
Family:
No, nothing like that.
Provider:
Do you know if he’d want to be kept alive on machines?
Family:
No, he saw how that was for my mom a few years back and he said he didn’t want to be kept alive if there wasn’t a good chance he’d survive.
Provider:
I’m going to give you some time to be with your dad and process what’s happening, I know it’s a lot to take in. But, when I come back, I’d like to talk to you about our goals for your dad and how we can make sure his wishes are honored. That way, we’ll all be on the same page going forward, OK? 21
Copyright © 2015 American Association of Critical-Care Nurses. Unauthorized reproduction of this article is prohibited.
NCI-D-14-00031.indd 21
09/01/15 8:06 AM
SEARL
W W W.A ACNA DVA NCE D CRIT ICA LCA RE .COM
Acknowledgments I thank Marian Grant, DNP, for support and guidance in preparing this article.
18. 19.
REFERENCES 1. National Consensus Project for Quality Palliative Care. Clinical Practice Guidelines for Quality Palliative Care. 2nd ed. Pittsburgh, PA: National Consensus Project for Quality Palliative Care; 2009. 2. Aslakson RA, Wyskiel R, Thornton I, et al. Nurse-perceived barriers to communication regarding prognosis and optimal end-of-life care for surgical ICU patients: a qualitative exploration. J Palliat Med. 2012;15(8):910– 915. doi: 10.1089/jpm.2011.0481. 3. Higgins PC, Prigerson HG. Caregiver evaluation of the quality of end-of-life care (CEQUEL) scale: the caregiver’s perception of patient care near death. PLoS ONE. 2013;8(6). doi:10.1371/journal.pone.0066066. 4. 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(20):1591–1598. 5. DeCato TW, Engelberg RA, Downey L, et al. Hospital variation and temporal trends in palliative and end-of-life care in the ICU. Crit Care Med. 2013;41(6):1405–1411. doi:10.1097/CCM.0b013e318287f289. 6. Hancock K, Clayton JM, Parker SM. Truth-telling in discussing prognosis in advanced life-limiting illnesses: a systematic review. Palliat Med. 2007;21:507–517. doi:10.1177/0269216307080823. 7. Mack JW, Smith TJ. Reasons why physicians do not have discussions about poor prognosis, why it matters, and what can be improved. J Clin Oncol. 2012;30:2715– 2717. doi:10.1200/JCO.2012.42.4564. 8. Borowske D. Straddling the fence: ICU nurses advocating for hospice care. Crit Care Nurs Clin N Am. 2010;24:105–116. doi:10.1016/j.ccell.2012.01.006. 9. Evans LR, Boyd EA, Malvar G, et al. Surrogate decisionmakers’ perspectives on discussing prognosis in the face of uncertainty. Am J Respir Crit Care Med. 2009;179:48–53. doi:10.1164/rccm.200806-969OC. 10. Gillick MR. Decision making near life’s end: a prescription for change. J Palliat Med. 2009;12(2):121–125. doi:10.1089/jpm.2008.0240. 11. Lind R, Lorem GF, Nortvedt P, Hevrøy O. Family members’ experiences of “wait and see” as a communication strategy in end-of-life decisions. Intensive Care Med. 2011;37:1143–1150. doi:10.1007/s00134-001-2253-x. 12. Nelson JE, Angus DC, Weissfeld LA, et al. End-of-life care for the critically ill: a national intensive care unit survey. Crit Care Med. 2006;34(10):2547–2553. doi:10.1097/01. CCM.0000239233.63425.1D. 13. Aslakson RA, Pronovost PJ. Health care quality in endof-life care: promoting palliative care in the intensive care unit. Anesthesiol Clin. 2011;29:111–122. doi:10.1016/ j.anclin.2010.11.001. 14. Field MJ. Approaching death: improving care at the end of life—a report from the Institute of Medicine. Health Serv Res. 1998;33(1):1–3. 15. O’Mahony S, McHenry J, Blank AE, et al. Preliminary report of the integration of a palliative care team into an intensive care unit. BMC Palliat Care. 2010;24(2):154– 165. doi:10.117/0269216309346540. 16. Kelley MA. Predictive scoring systems in the intensive care unit. UpToDate. Waltham, MA: UpToDate; 2014. 17. Breslow MJ, Badawi O. Severity scoring in the critically ill Part 1—Interpretation and accuracy of outcome prediction
20.
21.
22.
23. 24.
25. 26.
27.
28. 29. 30.
31.
32. 33. 34.
35.
36.
scoring systems. Chest. 2012;141(1):245–252. doi:10.1378/ chest.11-0330. Daly BJ, Souglas SL, O’Toole E, et al. Effectiveness trial of an intensive communication structure for families of long-stay ICU patients. Chest. 2010;138(6):1340–1348. Curtis JR, Vincent JL. Ethics and end-of-life care for adults in the intensive care unit. Lancet. 2010;375:1347– 1353. doi:10.1016/S01406736(10)60143-2. Puntillo K, Nelson JE, Weissman D, et al. Palliative care in the ICU: relief of pain, dyspnea, and thirst—A report from the IPAL-ICU Advisory Board. Intensive Care Med. 2014;40:235–248. doi:10.1007/s00134-013-3153-z. Vainiola T, Pettilä V, Rione RP, Räsänen P, Rissanen AM, Sintonen H. Comparison of two utility instruments, the EQ-5D and the 15D, in the critical care setting. Intensive Care Med. 2010;36:2090–2093. doi:10.1007/s00134-0101979-1. Ho LA, Engelberg RA, Curtis JR, et al. Comparing clinician ratings of the quality of palliative care in the intensive care unit. Crit Care Med. 2011;39(5):975–983. doi:10.1097/CCM.0b13e31820a91db. Flaatten H. Mental and physical disorders after ICU discharge. Curr Opin Crit Care. 2010;16:510–515. doi:10.1097/ MCC.0b013e32833cc90b. Unroe M, Kahn JM, Carson SS, et al. One-year trajectories of care and resource utilization for recipients of prolonged mechanical ventilation. Ann Intern Med. 2010;153:167– 175. doi:10.7326/0003-4819-153-3-201008030-00007. Pandharipande PP, Girard TD, Jackson JC. Long-term cognitive impairment after critical illness. N Engl J Med. 2013;369(14):1306–1306 doi:10.1056/NEJMoa1301372. Breslow MJ, Badawi O. Severity scoring in the critically ill. Part 2—maximizing value from outcome prediction scoring systems. Chest. 2012;141(2):518–527. doi:10.1378/ chest.11-0331. Douglas SL, Daly BJ, Lipson AR. Neglect of quality-of-life considerations in intensive care unit family meetings for long-stay intensive care unit patients. Crit Care Med. 2012;40(2):461–467. doi:10.1097/CCM.0b013e318232d8c4. Vincent JL, Moreno R. Clinical review: scoring systems in the critically ill. Crit Care. 2010;14(207). doi:10.1186/cc8204. Keegan MT, Gijic O, Afessa B. Severity of illness scoring systems in the intensive care unit. Crit Care Med. 2011;39(1):163–169. doi:10.1097/CCM.)b)13e3181f96f81. Knaus WA, Zimmerman JE, Wagner DP, Draper EA, Lawrence DE. APACHE—Acute Physiology and Chronic Health Evaluation: a physiologically based classification system. Crit Care Med. 1981;9(8):591–597. Zimmerman JE, Kramer AA, McNair DS, Malila FM. Acute Physiology and Chronic Health Evaluation (APACHE) IV: hospital mortality assessment for today’s critically ill patients. Crit Care Med. 2006;34(5):1297–1310. Le Gall JR, Loirat P, Alperovitch A, et al. A simplified acute physiology score for ICU patients. Crit Care Med. 1984;12(11):975–977. Le Gall JR, Lemeshow S, Saulnier F. A new Simplified Acute Physiology Score (SAPSII) based on a European/North American multicenter study. JAMA. 1993;270(24):2957–2963. Lemeshow S, Teres D, Pastides H, Avrunin JS, Steingrub JS. A method for predicting survival and mortality of ICU patients using objectively derived weights. Crit Care Med. 1985;13(7):519–525. Castella X, Gilabert J, Torner F, Torres C. Mortality prediction models in intensive care: Acute Physiology and Chronic Health Evaluation II and Mortality Prediction Model compared. Crit Care Med. 1991;19(2):191–197. Tunnell RD, Millar BW, Smith GB. The effect of lead time bias on severity of illness scoring, mortality prediction and standardized mortality ratio in intensive care—a pilot study. Anaesthesia. 1998;53(11):1045–1053.
22 Copyright © 2015 American Association of Critical-Care Nurses. Unauthorized reproduction of this article is prohibited.
NCI-D-14-00031.indd 22
09/01/15 8:06 AM
A Case for the Use of Validated Physiological Mortality Metrics to Guide Early Family Intervention in Intensive Care Unit Patients Molly F. Searl, RN, BSN, MSN, CCRN
ABSTRACT In the current health care climate a large portion of health care dollars are spent in the final months of life, so ensuring that care provided is in line with the wishes of patients and their families is more critical than ever. On the one hand, surviving families often report that they wish they had been given prognostic information earlier and that, in retrospect, they would have made different treatment decisions if they had been given prognostic information. On the other hand, providers often are reluctant to discuss prognosis for various reasons, not the least of which is the inherently uncertain nature of prognostication. To address this issue, this article reviews pertinent literature about
provider reticence, family preference, and the move toward palliative care and includes a discussion of the various validated mortality-prediction models available. A case is made to use those validated metrics to guide early discussions of palliative or endof-life care for patients who are critically ill. A suggested checklist to facilitate inclusion of prognosis discussions in family meetings is included as well as a case study to illustrate the problem, current practice, and a model for improvement. Keywords: critical care communication, end of life, mortality prediction scores, palliative care, prognosis
vasopressor support before he could be taken for confirmatory computed tomography scan. A bedside echocardiogram showed profound rightsided heart failure, with a nearly akinetic right ventricle. When his son arrived on postoperative day 1, his first question to the health care team
Mr S is a 96-year-old man with a recent history of recurrent falls and consequent orthopedic fractures. Before his most recent hospital admission, he fell and sustained a left-sided femur fracture. While being evaluated for open fixation of this fracture, he had chest pain and was diagnosed with a non–ST-elevation myocardial infarction for which he underwent a cardiac catheterization on hospital day 2. After being cleared for the orthopedic operation by cardiologists, he went to the operating room on hospital day 5 for a left femur fixation. Initial postoperative clinical findings were concerning for a pulmonary embolism, and on his first night after the operation, his clinical condition deteriorated, requiring intubation and
Molly F. Searl is Critical Care Transport Nurse, The Johns Hopkins Hospital, 1800 Orleans St, Baltimore, MD 21287 ([email protected]). The author has no financial disclosures or conflicts of interest to report. The views expressed herein are those of the author and do not necessarily reflect the views of The Johns Hopkins Hospital. DOI: 10.1097/NCI.0000000000000063
13 Copyright © 2015 American Association of Critical-Care Nurses. Unauthorized reproduction of this article is prohibited.
NCI-D-14-00031.indd 13
09/01/15 8:06 AM
SEARL
W W W.A ACNA DVA NCE D CRIT ICA LCA RE .COM
can facilitate a change from strictly curative to either a blended curative-palliative or a comfort-only model of care, both of which have been shown to possibly extend life and improve quality of life for both patients and caregivers.1,7,13 Unfortunately, fewer options currently are available to reduce the stress providers feel when providing a poor prognosis.6 Education and debriefings have been proposed as a means to reduce these feelings of stress, but few data are available on the effectiveness of these interventions.7 With regard to feelings of failure, they are relatively typical for providers who do not have training in either palliative or end-oflife (EOL) care,8 although more attention has been paid in recent years to the idea that the quality of a patient’s death should be given as much credence as the quality of his or her life.1,14 One solution proposed and widely accepted to manage both stress and feelings of failure is the integration of palliative care, through the inclusion of specially trained providers, into ICU teams.15 Alternatively, critical care providers can be given basic training on how to discuss and integrate palliative care into their daily practice. The feasibility of such interventions depends on the size of the institution, funding availability, and staff buy-in15 and is beyond the scope of the current discussion. Finally, providers cite unreliable predictors of prognosis as a reason for not initiating discussions.7 Although no provider can predict with absolute certainty when a patient will die, metrics are available that can help provide a scientifically based prediction of in-hospital mortality for patients who are critically ill.16 Few data are available on the current accuracy or frequency of prognostic discussions in ICUs. Breslow and Badawi17 found that, despite recommendations, only 10% to 15% of ICUs use mortality-predicting metrics. The recommendations cited suggest the use of prognostic metrics as indicators of quality improvement and care standardization, not as guidelines for the treatment of individual patients.17 Clinically, the general trend tends to be that prognostic information, when it is presented, is based not on metrics but on clinician experience or opinion, which can be quite varied.5,17 Research is not consistent with regard to whether discussion of prognosis decreases ICU resource use18 or length of stay,18 or increases the number of do-not-resuscitate orders,4 but when prognosis is discussed, variation in the information
was, “Can my dad make it through this?” The answer was, “We just don’t know.”
C
ommunication between physicians and patients or their families in the critical care setting has been well documented as an area in need of improvement.1–5 Also well documented is the difficult nature of discussing prognosis, specifically the disconnect between a provider’s willingness to discuss prognosis6–8 and the desire of patients and families to be given accurate information about prognosis.9–12 The purpose of this article is to make a case for the use of validated physiological prognostic metrics to aid in the discussion of prognosis in patients who are critically ill, and thus improve communication between the health care team and the families of patients in the intensive care unit (ICU). By using metrics as a guide, providers could more easily initiate early implementation of palliative or comfort care for patients who are critically ill, many of whom often undergo aggressive treatment until the final days of life.1,13 As a means of illustration, the case study described earlier is both an example of current practice and a model for improved communication. The Difficulty in Discussing Prognosis Relatively few studies have investigated the reasons why critical care providers choose not to provide prognostic information to patients or families. Those that have been conducted tend to report 3 recurring themes: (1) the belief that providing a poor prognosis will cause patients to lose hope and potentially deteriorate more quickly than if the information had been withheld6; (2) the concern that discussions of prognosis, especially when the estimated survival time is short, cause a large amount of stress among providers,6 often making them feel as if they have failed in their duty to care for their patient8; and (3) the idea that prognosis is uncertain and there is no reliable way to predict the course of any individual’s disease with absolute certainty.7,8 Regarding clinical deterioration, studies have reported that providing prognostic information to patients, even if the prognosis is grim, actually fosters hope, because the patients have a realistic notion of their disease trajectory and can trust their providers to be honest with them, regardless of the situation.7 Early and frank discussions of prognosis also 14
Copyright © 2015 American Association of Critical-Care Nurses. Unauthorized reproduction of this article is prohibited.
NCI-D-14-00031.indd 14
09/01/15 8:06 AM
VOL UME 2 6 • N U MBER 1 • JANUARY- M ARCH 2015
VA LIDAT E D P H YSIOLOGICA L M ORTA LIT Y M E T RICS IN T H E ICU
presented to patients and families has been cited as a significant barrier to effective EOL care in the ICU2 and can lead to ineffective use of palliative care,19 increased caregiver stress,9 and poor satisfaction as reported by surviving family members.3 The difficulty in providing prognostic information in the critical care setting is multifaceted but can be generally summarized by recognizing that clinicians, be they doctors, nurses, nurse practitioners, or any other members of the health care team, have difficulty reconciling the seemingly mutually exclusive mandates to save lives and provide effective palliative care.19 Often, the treatments with the greatest likelihood of success are those that cause distressing symptoms in patients,20 and those symptoms often are the aspects of EOL that cause the most dissatisfaction among surviving families.3
pare can be addressed through early initiation of family meetings,13 a key part of which should be—based on family surveys—a discussion of prognosis.9 Qualitative studies have shown that the desires described earlier, specifically the desire to have discussions of prognosis, contribute significantly to improving family satisfaction, which is increasingly recognized as a measure of successful treatment.21 Families have reported that their treatment decisions, made on behalf of their loved ones, may have been different had they been informed of the objective prognosis early in the disease course,10 but data are mixed. Both the seminal Study to Understand Prognoses and Preferences for Outcomes and Risks of Treatments (SUPPORT)4 trial and newer research by Daly and colleagues18 report no change in EOL decision making after the initiation of structured family communication that included discussions of prognosis. No information is available about the source—provider experience or physiological metric—of that prognostic information, although Daly and colleagues18 report that Acute Physiology and Chronic Health Evaluation (APACHE) III data were collected and scores were recorded as part of the patient demographic information. Note that mortality scores report predicted survival, but the perception of what survival actually means can be very different between families and providers.10 Specifically, families often consider survival to mean a return to previous functional status,12 whereas survival to providers often means “alive at discharge,” with less focus on the postillness functionality of the patient.12 In fact, with the advent and proliferation of life-sustaining treatments, patients and families tend not to understand that medical intervention can—on numerous occasions—only serve to prolong the dying process12 and is often the source of many distressing symptoms20 identified by family members as a significant source of regret after the death of their loved one.3 Providers must recognize the need to reduce suffering and not allow the potential adverse effects of treatment of distressing symptoms to prevent such treatment.20 As an example, a provider should not expect a patient to tolerate pain (a distressing symptom) to prevent or manage hypotension (a physiological condition), because treatment of the symptom (pain) may exacerbate the condition (hypotension), without first discussing treatment options and possible consequences
“Mr S cannot possibly survive this” was the general consensus on the part of not only the ICU team caring for him but also the primary surgical team on postoperative day 1. This view was expressed by multiple care providers during evening rounds. Mr S’s son was present at the bedside throughout the day, although he was not present during rounds when this discussion occurred. When providers entered the patient’s room after rounds, Mr S’s son asked, “How’s Dad doing?” to which the response was, “He is very sick. He needs a lot of help maintaining his blood pressure and breathing right now.” No discussion of data-driven prognostic information was offered.
The Need to Discuss Prognosis Family Perspective
When surveyed, families of critically ill patients report that their main concerns—after curing the underlying disease—are to ensure that their loved one does not suffer unduly,6 that distressing symptoms are addressed and treated in a timely fashion,20 and that they be allowed to share in decision making9 in an effort to prepare themselves for the possibility that their family member might not survive.10 Issues of suffering and symptom relief can be addressed with the inclusion of palliative care in all critical care settings, regardless of prognosis,15 but this, again, is beyond the scope of this discussion. Shared decision making and time to pre15
Copyright © 2015 American Association of Critical-Care Nurses. Unauthorized reproduction of this article is prohibited.
NCI-D-14-00031.indd 15
09/01/15 8:06 AM
SEARL
W W W.A ACNA DVA NCE D CRIT ICA LCA RE .COM
with the patients and families.20 Avoidance of suffering, or improved quality of life, has been cited as a primary goal for families and patients during illness,22 and research supports the notion that morbidity rather than mortality is a more accurate term to describe the concerns of patients and families.23 Although many researchers have compiled data on objective mortality-prediction measures, data reflecting morbidity are, by necessity, based on surveys of survivors and their families. These surveys, collectively known as health-related quality of life surveys, collect subjective data on a number of variables, including functional status, anxiety, and depression, as well as objective measures, such as delirium and pulmonary, renal, and cognitive function.23 When combined, these measures give a picture of functional status, which is often decreased, after critical illness.23 Data report that patients who require longer and more aggressive care are likely to experience greater functional dependence after critical illness.24,25 Patients with comorbid conditions on admission and patients who are older or require ventilator support for longer periods of time are at highest risk for postillness sequelae.24 These same physiological variables also are used to calculate mortality scores; thus, patients with higher mortality scores on admission are more likely to require aggressive care26 and, therefore, are more likely to experience long-term functional consequences. Therefore, logically, predictors of mortality—that is, prognostic metrics—can be extrapolated to predict postillness morbidity for those patients who survive their critical illness. This postillness morbidity is, as mentioned earlier, generally the most important factor for patients and their families and often is not discussed during family meetings.27
the provider is not completely sure.9 In fact, the tendency of providers to tell families that it is best to “wait and see” before discussing prognosis,11 or the fact that different providers often present different prognoses,2 is actually detrimental to families and limits their ability to process information and make appropriate treatment decisions.10 Conflicting opinions about prognosis and efficacy of treatment often lead to staff burnout related to caring for patients who are chronically critically ill,2 and the hesitancy on the part of providers to discuss prognosis has been shown, in some cases, to actually contribute to a decrease in frequency of family conferences as length of stay increases, when illnesses become more complicated, and the need for realistic discussions of prognosis becomes more pressing.5 Postoperative day 2: Mr S has not improved. His liver function has decreased, demonstrating catastrophic liver injury; his kidney function has deteriorated; and he still requires maximal vasopressor and ventilator support. His son remains at the bedside, and frequently asks how his father is doing. During a discussion with his father’s physician, he relayed the story of his mother’s death approximately 1 year before. She had suffered a relatively minor injury and had developed sepsis as a result. He stated, in passing, that he did not want his father to suffer like that, relying on machines to keep him alive. Again no discussion of prognosis was offered.
Arguably, a disconnect exists in perception that leads to a demonstrable lack of communication between providers and families.3 That disconnect can be traced, at least partially, to discomfort on the part of providers to discuss prognosis in concrete terms, because they are uncomfortable in providing such information without supporting evidence, because there are conflicting opinions on prognosis, or because they are justifiably uncomfortable with or undertrained in how to provide potentially devastating news to patients and their families. The question then becomes whose opinion on prognosis should be used to guide discussions. Should it be the primary service attending, in the case of a surgical ICU patient, who has seen thousands of postoperative liver cancer patients as an example? Or should it be the intensivist who has cared for thousands of critically ill individuals regardless of the underlying pathology? Or should it be the palliative care team,
Provider Perspective
Perhaps the most substantial argument against discussing prognosis, from a provider’s perspective, is uncertainty; even the most accurate prognostic metrics are variable and designed specifically to aggregate mortality information for groups rather than predict mortality for specific patients.17 Research supports the idea that providers are generally less likely to provide specific information about prognosis than any other aspect of health care,6 even though family members report that they generally would prefer a discussion of prognosis, even if 16
Copyright © 2015 American Association of Critical-Care Nurses. Unauthorized reproduction of this article is prohibited.
NCI-D-14-00031.indd 16
09/01/15 8:06 AM
VOL UME 2 6 • N U MBER 1 • JANUARY- M ARCH 2015
VA LIDAT E D P H YSIOLOGICA L M ORTA LIT Y M E T RICS IN T H E ICU
whose job is to focus care on comfort first, either as an alternative to or in collaboration with curative efforts? In an attempt to remove opinion from the already-complicated discussion, the use of valid and proven prognostic metrics is a reasonable and relatively easy solution.
Table 1: Advantages, Limitations, and General Considerations of 3 Prognostic Metricsa Advantages All have well-established discrimination and calibration in general ICU patients APACHE, SAPS, and MPM have been validated in large numbers of clinical trials spanning all critical care settings (ie, surgical, medical, neurological)
Prognostic Metrics Various prognostic metrics that include both general prognostic scores and organ dysfunction scores are available for providers.28 For the purpose of this discussion, only general prognostic metrics are reviewed; the most commonly used metrics are the APACHE, Simplified Acute Physiology Score (SAPS), and Mortality Prediction Model (MPM).29 These 3 metrics have been found to have good discrimination—the ability to identify ICU patients with the highest risk of death—and good calibration—the extent to which the predicted outcome matches observed outcome.17 In brief, these prognostic metrics were developed in an effort to better categorize disease severity and to standardize research in the critical care setting, and as a basis for comparing care across ICUs.16 See Table 1 for a general appraisal of advantages and limitations common to all 3 metrics. A more detailed discussion of each metric follows.
Based on physiological data that, when combined, provide accurate mortality predictions Limitations Not well studied for specific ICU patient populations (eg, AIDS, postpartum) or in ICUs with specialized treatment modalities (eg, ECMO) Some disease processes have expected trajectories that can change scores based on when they are calculated (eg, expected low GCS with new SAH patients could improve with time) Lead-time bias—the change in score based on treatment before data collection—can alter scores, especially in patients who have been transferred from one facility to another General considerations All scores can be affected by improvements in medical technology, implementation of best practice recommendations (eg, early antibiotics for suspected or confirmed sepsis), or changes in practice (eg, new providers rotating through the ICU)
Acute Physiology and Chronic Health Evaluation
Originally proposed in 1981, the APACHE system was designed to provide an objective classification of risk in critically ill patients.30 The author of this study acknowledged the inherent reluctance on the part of physicians to discuss prognosis for individual patients, so this particular tool was designed to refer to groups of patients with similar physiological variables, not to guide individual treatment decisions.30 As a result of polling experts for the variables deemed “important,” Knaus and colleagues30 derived a weighted classification system that allowed comparison of an individual patient to a group of similar patients to predict mortality. This original scoring system, as with all such prediction metrics, was found to have less discrimination over time, largely due to advances in medical technology. Consequently, the APACHE score has undergone many revisions, the most current being the APACHE IV, which includes measures of length of stay and incorporates logistical regression to provide users with a predicted mortality score as a percentage.31
Abbreviations: AIDS, acquired immunodeficiency syndrome; APACHE, Acute Physiology and Chronic Health Evaluation; ECMO, extracorporeal membrane oxygenation; GCS, Glasgow Coma Scale; ICU, intensive care unit; MPM, Mortality Prediction Model; SAH, subarachnoid hemorrhage; SAPS, Simplified Acute Physiology Score. a Based on data from Kelley.16
The physiological values required to calculate an APACHE IV score for Mr S are listed in Table 2; values included therein reflect the worst values in the first 24 hours as stipulated by the scoring system. A primary diagnosis that led to ICU admission also is required for an APACHE score.17 Previous iterations of the APACHE were limited by their proprietary nature, meaning that hospitals were required to purchase costly software programs to access the information; however, APACHE IV is now available online and free of charge at http:// www.apachefoundations.cernerworks.com.31 Other freely available online severity scoring calculators (http://clincalc.com) report APACHE II, 17
Copyright © 2015 American Association of Critical-Care Nurses. Unauthorized reproduction of this article is prohibited.
NCI-D-14-00031.indd 17
09/01/15 8:06 AM
SEARL
W W W.A ACNA DVA NCE D CRIT ICA LCA RE .COM
Table 2: Raw Data Used to Calculate Mr S’s Mortality-Prediction Scoresa Patient Data Temperature, °C Heart rate, beats/min Respiratory rate, breaths/min
On Admission
24 Hours After Admission
36.3
35
93
107
24
30
72/37 (49)
120/49 (66)b
150
0
White blood cell count, 10 /μL
29.7
31
Hematocrit, %
27.2
28.2
Serum sodium, mEq/L
143
142
Serum potassium, mEq/L
3.8
4
Serum bicarbonate, mEq/L
18
19
BUN, mg/dL
26
34
Serum creatinine, mg/dL
0.91
1.67
Serum albumin, g/dL
2.1
1.6
BP (MAP), mm Hg Urine output, mL/d 3
Serum bilirubin, mg/dL
1.1
2.5
Serum glucose, mg/dL
110
139c
GCS
3T
3T
ABG FIO2, %
100
80
ABG PaO2
166.5
93.7
ABG PaCO2
21.5
29.3
7.2
7.4d
APACHE IV (predicted mortality), %
82.5
87.9
SAPS II (predicted mortality), %
88.9
89.7
MPM II (predicted mortality), %
90
87.3
ABG pH
Sources for predicted mortality data APACHE IV: http://www.apachefoundations.cernerworks.com SAPS II: http://clincalc.com/IcuMortality/SAPSII.aspx MPM admission: http://www.sfar.org/scores2/ mpm2_admission2.php MPM 24 hours: http://www.sfar.org/scores2/ mpm2_24_48_722.php Abbreviations: ABG, arterial blood gas; APACHE, Acute Physiology and Chronic Health Evaluation; BP, blood pressure; BUN, blood urea nitrogen; FIO2, fraction of inspired oxygen; GCS, Glasgow Coma Scale; MAP, mean arterial pressure; MPM, Mortality Prediction Model; PaCO2, partial pressure of carbon dioxide; PaO2, partial pressure of oxygen; SAPS, Simplified Acute Physiology Score. a None of the scores has a caveat for the use of medical support beyond ventilators (accounted for in the ABG values) or sedation medication (accounted for in the GCS). The use of vasopressors to sustain blood pressure or insulin to control blood glucose is not noted and thus does not change the weight of any individual scores. The supportive measures noted for some values in the 24 Hours After Admission column are for the edification of the readers and are meant to clarify that some of the values presented, while they appear improved after 24 hours, are in fact with more support than previously. b Norepinephrine (Levophed) infusion at 0.12 mcg/kg per minute and vasopressin infusion at 0.04 units per hour. c Insulin infusion at 5 units per hour. d Sodium bicarbonate drip at 150 mL/h.
18 Copyright © 2015 American Association of Critical-Care Nurses. Unauthorized reproduction of this article is prohibited.
NCI-D-14-00031.indd 18
09/01/15 8:06 AM
VOL UME 2 6 • N U MBER 1 • JANUARY- M ARCH 2015
VA LIDAT E D P H YSIOLOGICA L M ORTA LIT Y M E T RICS IN T H E ICU
SAPS, and other organ-specific mortality scores. APACHE scores also require a large amount of clinical data, and without automatic chart review processes, data entry can be labor intensive. In addition, when data are not available, the software assumes a normal value, which can potentially lower the calculated score, which was a conscious choice made by the original APACHE authors based on the assumption that if a value is not measured the providers at the time did not suspect it would be abnormal.30 Being the most recent iteration of the APACHE scoring system, the APACHE IV score calculated via the above website is included for Mr S.
and at least one study found it valid in a rural community hospital.35 A major difference between MPM and SAPS/APACHE is that the variables in MPM are binary, that is, marked as present or not present, thus significantly reducing the amount of time and specificity of data to be collected.16 Another difference is that the MPM produces a predicted mortality percentage, rather than a score that is then translated into a percentage.
Based on clinical data for Mr S, his initial APACHE IV score was 146, with a predicted mortality of 82.5% for this hospitalization. Data are included in Table 2.
These metrics have limitations on their effectiveness in certain patient populations or under certain conditions,16 so clinical judgment should guide the timing of score-driven family meetings in these circumstances. See Table 1 for a summary of these issues. Regardless of the metric chosen, the basic underlying supposition is that a set of physiological variables, when combined, can point toward a potential for mortality. Although none of the authors who designed or validated these scoring metrics advocates their use in guiding treatment for specific patients, their use in reducing interprovider disagreement about prognosis cannot be denied.
Based on clinical data for Mr S, his initial MPM predicted mortality rate was 90% for this hospitalization. Data are included in Table 2.
Simplified Acute Physiology Score
From the original APACHE data, Le Gall and colleagues32 created the SAPS, which was based on the same principles as APACHE but sought to streamline and simplify the process by decreasing the number of variables required. As with APACHE scores, SAPS has been validated throughout critical care populations and has undergone 3 revisions, the most recent of which, SAPS 3, includes customized variables for geographic regions and has helped increase discrimination.28 Perhaps the most important difference between SAPS and APACHE is the recognition on the part of Le Gall and colleagues33 that selecting a single primary diagnosis for patients in the ICU with complicated health issues and often with comorbidities is difficult, requiring provider judgment and opinion to determine the primary diagnosis.
By postoperative day 7, Mr S continued to require full ventilator support and multiple vasopressors, and his kidney function had deteriorated to the point where he needed dialysis—a treatment deemed futile by the Renal Service. While his cardiac function and laboratory tests showed some improvement, his respiratory status was complicated by a multidrug-resistant Proteus pneumonia, and he never regained consciousness despite a lack of sedation medications. Attempts were made by the ICU team to procure a do-not-resuscitate order during the day on postoperative day 6, but again with no concrete discussions of prognosis, Mr S’s son continued to cling to the hope that his father might recover. Frustrated with the discussions he had with the ICU team and his perception that the team had “given up without reason,” Mr S’s son requested that the patient be transferred to another facility, which occurred on postoperative day 8.
Based on clinical data for Mr S, his initial SAPS score was 75, with a predicted mortality of 88.9% for this hospitalization. Data are included in Table 2.
Mortality Prediction Model
Finally, in an attempt to remove expert opinion and therefore clinician bias, Lemeshow et al34 created the MPM, which derived variables and weights through statistical regression. The MPM is designed to have data entered at 2 predetermined times, on admission and at 24 hours after admission. As with the other 2 scores, MPM has been validated and found to be accurate across a large number of ICU populations,
Recommendations Intensive care practitioners have a duty to communicate honestly and openly with patients and 19
Copyright © 2015 American Association of Critical-Care Nurses. Unauthorized reproduction of this article is prohibited.
NCI-D-14-00031.indd 19
09/01/15 8:06 AM
SEARL
W W W.A ACNA DVA NCE D CRIT ICA LCA RE .COM
their families about disease prognosis. Rather than continue with the current practice of basing such discussions on individual provider preferences or experiences, validated physiologically based prognostic metrics should be considered. Best practice encompasses accurate diagnosis, effective and timely use of available treatments,26 and discussions of palliative care; although palliative care measures will not affect mortality scores, they will affect satisfaction scores from survivors, which are given nearly equal weight as benchmarks for successful care.21 Although it is ideal, frequent contact between patients and their families and providers is often cost and/or time prohibitive.7 For providers to effect the most positive change, all patients should receive a mortality prediction score on admission to the ICU. The scores presented here have been shown to provide similarly accurate information, and the use of one score over another should be left to the discretion of the individual institution on the basis of staff availability for data entry and common clinical practice. Given the research on effective implementation of family meetings and palliative care, the best practice option would be to integrate a multidisciplinary palliative care team in any such meeting.15 As a secondary, and potentially more feasible, option, specific providers could be trained to have discussions about palliative care and prognosis. If neither of these options is available, any provider can access any of the aforementioned prognostic metrics and open a dialogue with families based on the mortality score. As for which patients need family conferences, those patients with higher severity scores (60% or greater), and therefore a higher probability of mortality, should have a family conference scheduled within the first 48 hours if possible, although timing could change for individual patients with certain illnesses or injuries; see Table 1. For patients with scores in the midrange (30%-60%), family conferences should be initiated within 72 hours of admission. For patients with low mortality scores (< 30%), who have the highest likelihood of returning to function, family conferences could be delayed or even forgone completely. Lead-time bias, or the effect of treatments initiated before data collection, can affect mortality-prediction36 scores, but this should not—as a rule—prevent score calculation or early family meetings. In an effort to address concerns raised by bedside nurses,
specifically a lack of protocols for discussions of prognosis, and the view that families are often given conflicting information about prognosis,2 a checklist incorporating the suggested prognostic levels above is included in Table 3. As with most physiological variables, trends are more significant than individual numbers, so periodic recalculation of prognostic scores is suggested. If the MPM model is used, an automatic recalculation is available, in that online calculators have options for admission data and data at 24, 48, and 72 hours, respectively. Although APACHE and SAPS use the worst values collected in the last 24 hours, no provision is made for recalculation. The checklist in Table 3 includes suggested decision points that prompt
Table 3: Suggested Checklist for Implementing Family Meetings Based on Mortality Predictions Patient Name:
Date:
Time:
Admission APACHE II Score/ Predicted Mortality: Family meeting indicated for > 30%: Within 48 hours (predicted mortality > 60%)
□
Within 72 hours (predicted mortality 30%-60%)
□
Meeting scheduled for/with: Advance directive/living will? Reviewed: □
□ YES
□ NO
□ YES
□ NO
Reevaluate in 72 hours if patient is still in the intensive care unit or with worsening clinical conditiona Repeat APACHE II Score/ Predicted Mortality: Date/Time: Family Meeting Indicated? (predicted mortality > 30% or change in score ≥ 15%) Meeting Scheduled for/with: Abbreviation: APACHE, Acute Physiology and Chronic Health Evaluation. a Changes in condition include but are not limited to: addition of vasopressors, emergent surgery, increase in oxygen requirements, unplanned intubation, decrease in mental status, and cardiac/ respiratory arrest.
20 Copyright © 2015 American Association of Critical-Care Nurses. Unauthorized reproduction of this article is prohibited.
NCI-D-14-00031.indd 20
09/01/15 8:06 AM
VOL UME 2 6 • N U MBER 1 • JANUARY- M ARCH 2015
VA LIDAT E D P H YSIOLOGICA L M ORTA LIT Y M E T RICS IN T H E ICU
recalculation of scores at a specified time or with major changes in patient status, as defined on the checklist. Table 4 includes a suggested script for providers to discuss prognosis based on metrics, using Mr S as an example.
Summary In 1998, the Institute of Medicine defined a good death as “one that is free from avoidable distress and suffering for patients, families, and caregivers; in general accord with patients’ and families’ wishes; and reasonably consistent with clinical, cultural, and ethical standards.”14(p1) Based on that recommendation, and the recognition that EOL in hospitals, especially in critical care, can be unduly and unnecessarily distressing to families,10 a current trend in health care is to provide quality and timely palliative care.1 Considering these factors, practitioners must initiate discussions of palliative care early in a patient’s critical illness, regardless of prognosis,20 but especially when the probability of death is high.10 Arguments can be made that prognostic metrics are not fail-safe or meant to guide treatment decisions for individual patients, but in light of the fact that families often desire prognosis information,2 even if that information is uncertain,9 and given the frequent disagreements between providers on prognosis,2,19,22 a case can be made that using these metrics to guide early discussions of palliative or EOL care could (a) improve family satisfaction with the dying process in hospitals,20 (b) give families more time to comprehend and process the nature of their loved one’s prognosis,10 (c) decrease staff stress when caring for a patient that “everyone knows is going to die,”2 and most importantly, (d) improve the quality of health care at EOL. The suggestions presented here should be subjected to academic scrutiny, but the purpose of this article is to raise awareness of a practice deficit and offer a solution. Likewise, the purpose is not to advocate for the withdrawal or restriction of life-sustaining therapies for patients. Nor is it to advocate the use of prognostic metrics to arbitrarily guide care for individuals. Instead, given the difficult nature of both prognostication and discussions of EOL, this article simply suggests the use of validated physiological metrics to aid in discussions with families and patients in the critical care setting about goals of care. By using established objective prognostic tools, clinicians can ensure that families are given the same information about prognosis, and by initiating discussions about the family’s or the patient’s wishes early, providers can help families prepare for poor outcomes. Most important, with improved communication comes improved quality of care and, for some patients, improved quality of their death.
Table 4: Alternate Script Illustrating Use of Mortality-Prediction Scores in Family Meeting Provider:
Mr S, your father’s condition is very serious. He couldn’t breathe on his own so we had to put in a breathing tube. His blood pressure is very low, so we started him on medications to keep it up. And this morning, we did an echocardiogram, a test that shows us how well his heart is beating, and it showed that half of his heart isn’t working.
Family:
Can my dad make it through this?
Provider:
In a man of his age, with all of the problems he has now, there is an 83% chance of death. We have tools that help us predict if a person will survive a hospital stay. Now, these tools are not 100% accurate. But they do tell us that 83% of people with your father’s issues will die.
Family:
So it’s not for sure he’ll die? He was healthy and independent until this happened.
Provider:
No, we can’t say 100% that he will die, but it is very likely that he will not survive this, and that if he survives, he likely won’t be able to get back to his independent life. Did you father have an advance directive or a living will?
Family:
No, nothing like that.
Provider:
Do you know if he’d want to be kept alive on machines?
Family:
No, he saw how that was for my mom a few years back and he said he didn’t want to be kept alive if there wasn’t a good chance he’d survive.
Provider:
I’m going to give you some time to be with your dad and process what’s happening, I know it’s a lot to take in. But, when I come back, I’d like to talk to you about our goals for your dad and how we can make sure his wishes are honored. That way, we’ll all be on the same page going forward, OK? 21
Copyright © 2015 American Association of Critical-Care Nurses. Unauthorized reproduction of this article is prohibited.
NCI-D-14-00031.indd 21
09/01/15 8:06 AM
SEARL
W W W.A ACNA DVA NCE D CRIT ICA LCA RE .COM
Acknowledgments I thank Marian Grant, DNP, for support and guidance in preparing this article.
18. 19.
REFERENCES 1. National Consensus Project for Quality Palliative Care. Clinical Practice Guidelines for Quality Palliative Care. 2nd ed. Pittsburgh, PA: National Consensus Project for Quality Palliative Care; 2009. 2. Aslakson RA, Wyskiel R, Thornton I, et al. Nurse-perceived barriers to communication regarding prognosis and optimal end-of-life care for surgical ICU patients: a qualitative exploration. J Palliat Med. 2012;15(8):910– 915. doi: 10.1089/jpm.2011.0481. 3. Higgins PC, Prigerson HG. Caregiver evaluation of the quality of end-of-life care (CEQUEL) scale: the caregiver’s perception of patient care near death. PLoS ONE. 2013;8(6). doi:10.1371/journal.pone.0066066. 4. 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(20):1591–1598. 5. DeCato TW, Engelberg RA, Downey L, et al. Hospital variation and temporal trends in palliative and end-of-life care in the ICU. Crit Care Med. 2013;41(6):1405–1411. doi:10.1097/CCM.0b013e318287f289. 6. Hancock K, Clayton JM, Parker SM. Truth-telling in discussing prognosis in advanced life-limiting illnesses: a systematic review. Palliat Med. 2007;21:507–517. doi:10.1177/0269216307080823. 7. Mack JW, Smith TJ. Reasons why physicians do not have discussions about poor prognosis, why it matters, and what can be improved. J Clin Oncol. 2012;30:2715– 2717. doi:10.1200/JCO.2012.42.4564. 8. Borowske D. Straddling the fence: ICU nurses advocating for hospice care. Crit Care Nurs Clin N Am. 2010;24:105–116. doi:10.1016/j.ccell.2012.01.006. 9. Evans LR, Boyd EA, Malvar G, et al. Surrogate decisionmakers’ perspectives on discussing prognosis in the face of uncertainty. Am J Respir Crit Care Med. 2009;179:48–53. doi:10.1164/rccm.200806-969OC. 10. Gillick MR. Decision making near life’s end: a prescription for change. J Palliat Med. 2009;12(2):121–125. doi:10.1089/jpm.2008.0240. 11. Lind R, Lorem GF, Nortvedt P, Hevrøy O. Family members’ experiences of “wait and see” as a communication strategy in end-of-life decisions. Intensive Care Med. 2011;37:1143–1150. doi:10.1007/s00134-001-2253-x. 12. Nelson JE, Angus DC, Weissfeld LA, et al. End-of-life care for the critically ill: a national intensive care unit survey. Crit Care Med. 2006;34(10):2547–2553. doi:10.1097/01. CCM.0000239233.63425.1D. 13. Aslakson RA, Pronovost PJ. Health care quality in endof-life care: promoting palliative care in the intensive care unit. Anesthesiol Clin. 2011;29:111–122. doi:10.1016/ j.anclin.2010.11.001. 14. Field MJ. Approaching death: improving care at the end of life—a report from the Institute of Medicine. Health Serv Res. 1998;33(1):1–3. 15. O’Mahony S, McHenry J, Blank AE, et al. Preliminary report of the integration of a palliative care team into an intensive care unit. BMC Palliat Care. 2010;24(2):154– 165. doi:10.117/0269216309346540. 16. Kelley MA. Predictive scoring systems in the intensive care unit. UpToDate. Waltham, MA: UpToDate; 2014. 17. Breslow MJ, Badawi O. Severity scoring in the critically ill Part 1—Interpretation and accuracy of outcome prediction
20.
21.
22.
23. 24.
25. 26.
27.
28. 29. 30.
31.
32. 33. 34.
35.
36.
scoring systems. Chest. 2012;141(1):245–252. doi:10.1378/ chest.11-0330. Daly BJ, Souglas SL, O’Toole E, et al. Effectiveness trial of an intensive communication structure for families of long-stay ICU patients. Chest. 2010;138(6):1340–1348. Curtis JR, Vincent JL. Ethics and end-of-life care for adults in the intensive care unit. Lancet. 2010;375:1347– 1353. doi:10.1016/S01406736(10)60143-2. Puntillo K, Nelson JE, Weissman D, et al. Palliative care in the ICU: relief of pain, dyspnea, and thirst—A report from the IPAL-ICU Advisory Board. Intensive Care Med. 2014;40:235–248. doi:10.1007/s00134-013-3153-z. Vainiola T, Pettilä V, Rione RP, Räsänen P, Rissanen AM, Sintonen H. Comparison of two utility instruments, the EQ-5D and the 15D, in the critical care setting. Intensive Care Med. 2010;36:2090–2093. doi:10.1007/s00134-0101979-1. Ho LA, Engelberg RA, Curtis JR, et al. Comparing clinician ratings of the quality of palliative care in the intensive care unit. Crit Care Med. 2011;39(5):975–983. doi:10.1097/CCM.0b13e31820a91db. Flaatten H. Mental and physical disorders after ICU discharge. Curr Opin Crit Care. 2010;16:510–515. doi:10.1097/ MCC.0b013e32833cc90b. Unroe M, Kahn JM, Carson SS, et al. One-year trajectories of care and resource utilization for recipients of prolonged mechanical ventilation. Ann Intern Med. 2010;153:167– 175. doi:10.7326/0003-4819-153-3-201008030-00007. Pandharipande PP, Girard TD, Jackson JC. Long-term cognitive impairment after critical illness. N Engl J Med. 2013;369(14):1306–1306 doi:10.1056/NEJMoa1301372. Breslow MJ, Badawi O. Severity scoring in the critically ill. Part 2—maximizing value from outcome prediction scoring systems. Chest. 2012;141(2):518–527. doi:10.1378/ chest.11-0331. Douglas SL, Daly BJ, Lipson AR. Neglect of quality-of-life considerations in intensive care unit family meetings for long-stay intensive care unit patients. Crit Care Med. 2012;40(2):461–467. doi:10.1097/CCM.0b013e318232d8c4. Vincent JL, Moreno R. Clinical review: scoring systems in the critically ill. Crit Care. 2010;14(207). doi:10.1186/cc8204. Keegan MT, Gijic O, Afessa B. Severity of illness scoring systems in the intensive care unit. Crit Care Med. 2011;39(1):163–169. doi:10.1097/CCM.)b)13e3181f96f81. Knaus WA, Zimmerman JE, Wagner DP, Draper EA, Lawrence DE. APACHE—Acute Physiology and Chronic Health Evaluation: a physiologically based classification system. Crit Care Med. 1981;9(8):591–597. Zimmerman JE, Kramer AA, McNair DS, Malila FM. Acute Physiology and Chronic Health Evaluation (APACHE) IV: hospital mortality assessment for today’s critically ill patients. Crit Care Med. 2006;34(5):1297–1310. Le Gall JR, Loirat P, Alperovitch A, et al. A simplified acute physiology score for ICU patients. Crit Care Med. 1984;12(11):975–977. Le Gall JR, Lemeshow S, Saulnier F. A new Simplified Acute Physiology Score (SAPSII) based on a European/North American multicenter study. JAMA. 1993;270(24):2957–2963. Lemeshow S, Teres D, Pastides H, Avrunin JS, Steingrub JS. A method for predicting survival and mortality of ICU patients using objectively derived weights. Crit Care Med. 1985;13(7):519–525. Castella X, Gilabert J, Torner F, Torres C. Mortality prediction models in intensive care: Acute Physiology and Chronic Health Evaluation II and Mortality Prediction Model compared. Crit Care Med. 1991;19(2):191–197. Tunnell RD, Millar BW, Smith GB. The effect of lead time bias on severity of illness scoring, mortality prediction and standardized mortality ratio in intensive care—a pilot study. Anaesthesia. 1998;53(11):1045–1053.
22 Copyright © 2015 American Association of Critical-Care Nurses. Unauthorized reproduction of this article is prohibited.
NCI-D-14-00031.indd 22
09/01/15 8:06 AM
