- Email: [email protected]
Criteria for intensive care unit admission and severity of illness
CRITICAL ILLNESS AND INTENSIVE CARE: I
Criteria for intensive care unit admission and severity of illness
patients who require perioperative critical care due to demand for beds. The last year has seen further pressures on critical care resources in the UK during the H1N1 flu pandemic. The admission of a patient to critical care takes place after considering the patient’s acute illness and the degree of reversibility in the context of their background medical problems and physiological reserve.
Miriam Baruch Ben Messer
Admission criteria and levels of care In the document ‘Levels of Critical Care for Adult Patients’ published in 2009 by the Intensive Care Society (ICS), patients in hospital are assigned a level of care from 0 to 3 based on an assessment of their clinical needs. This is shown in Table 1. The level of care is not governed by the location of the patient and equally the level of care assigned does not necessarily determine the nurse-to-patient ratio for that particular patient. The identification of patients whose clinical condition is deteriorating or at risk of deteriorating to the point of requiring critical care team review can be done by ward-based medical teams. If the patient is designated level 2 or 3 they will almost certainly be moved to a critical care area. Patients who are currently level 2 but at significant risk of becoming level 3 will often be admitted directly to an ICU where their needs are most easily met. Some patients who may not previously have been considered for admission to critical care are now being considered in certain circumstances. Examples include patients with multiple comorbidities where specific limits of treatment are set and in other patients after the decision to initiate palliative care to facilitate organ donation after cardiac death.
Abstract This article explores the provision and organisation of critical care services in the UK and examines the issues surrounding admission to, discharge from and the withholding of critical care. Critical care expansion in the UK in recent years has centred on the provision of increased numbers of beds and the development of outreach services. Despite the expansion in critical care there is difficulty in matching supply and demand for beds. There remains controversy regarding the effectiveness of outreach in improving outcomes for patients referred to critical care. The discharge of patients from critical care has also come under scrutiny since mortality rates are higher for patients discharged out-of-hours. Patients’ needs following critical care are carefully planned with the base medical teams because readmission to critical care is associated with increased mortality. Scoring systems are used in critical care to compare outcomes between critical care units and to facilitate research but cannot predict outcome for individual patients. The decision that patients will not benefit from critical care admission can be one of the most difficult. Many patients who would not have been considered for organ support previously are now admitted to critical care with pre-determined limits of treatment. Involvement of the base team is integral to making decisions for these patients.
C
C
Keywords APACHE; critical care ethics; critical care scoring systems; Early Warning Scores; ICNARC; Intensive Care Society; levels of critical care; outreach; standardised mortality ratio
Critical care has been organized into levels of care depending upon the level of organ support and observation that an individual patient requires The levels of care are from 0 to 3 with 0 being standard ward care and observation and level 3 being what is traditionally thought of as intensive care
Early warning scores Track and trigger systems are used to identify patients at risk of deterioration. An integral part of these is the early warning score. Early warning scores (EWS) have been developed to increase awareness and recognition of acute illness by assigning a graded score to increasingly deranged physiological parameters within the following domains: respiratory, cardiovascular, urine output and conscious level. The score is based on simple observations which may be different depending on local case mix but usually include respiratory rate, pulse rate, blood pressure, temperature, conscious level and urine output. The scores across the domains are added to give an EWS which can then be entered into an algorithm to establish the most appropriate course of action which is defined by the frequency of subsequent observation and by the experience of the medical review that the EWS should trigger. We include the EWS used in our hospital in Figure 1. During the 1990s there was increasing recognition that care prior to critical care admission was often suboptimal with regard to simple management of airway, breathing, circulation, oxygen
Introduction The provision of critical care beds in England has increased by over 800 in the last 10 years,1 a rise of almost 30%. Figures from January 2011 reveal there are 3747 adult critical care beds in England, 45% in high-dependency units (HDU) and 55% in intensive care units (ICU). Ideally this provision should be able to accommodate peaks and troughs in demand, but it is not uncommon for hospital trusts to postpone major surgery for
Miriam Baruch FRCA is a Consultant in Anaesthesia & Intensive Care at Royal Victoria Infirmary, Newcastle upon Tyne, UK. Conflicts of interest: none declared. Ben Messer MA DICM is a Consultant in Anaesthesia & Intensive Care at Royal Victoria Infirmary, Newcastle upon Tyne, UK. Conflicts of interest: none declared.
SURGERY 30:5
225
Ó 2012 Elsevier Ltd. All rights reserved.
CRITICAL ILLNESS AND INTENSIVE CARE: I
Levels of critical care for adult patients as described by the Intensive Care Society Level
Criteria
0
C
1
C C C
Examples
Medical and nursing needs can be met with standard ward based care
C
Patients recently discharged from critical care Patients requiring supplementary clinical monitoring Patients referred for critical care team review or being seen regularly by the outreach service
C
C
C
C C C
2
C
Patients receiving single organ support (basic respiratory and basic cardiovascular support together can be considered as level 2)
C
Intravenous fluid and antibiotic therapy 4 hourly observations Patients at risk of clinical deterioration Epidural analgesia or patient-controlled analgesia in use requiring specific nursing observations Tracheostomy in-situ without respiratory support Patients who have a chest drain in-situ Parenteral nutrition Basic respiratory support
C
Mask/hood/tracheostomy continuous positive airway pressure or non-invasive bi-level ventilation FiO2 0.5 delivered via facemask Basic cardiovascular support
C
Use of arterial line for invasive blood pressure monitoring or arterial blood gas sampling Single intravenous vasoactive drug infusion Advanced cardiovascular support
C
Multiple vasoactive/inotropic drug infusions Renal support
C
Acute renal replacement therapy Neurological support Continuous intravenous medication to control seizures
3
C C
Patients receiving advanced respiratory support Patients receiving a minimum of two organs supported
C
Ventilatory support via an endotracheal tube (including tracheostomy)
Table 1
Outreach
therapy and monitoring requirements. In one study of patients admitted to critical care, care was suboptimal in more than half of admitted patients and could have contributed to mortality in one-third of patients.2 Furthermore, cardiac arrest and critical care admission are often preceded by derangements in acute physiology. This has been documented by the NCEPOD report of 2005 which found that 66% of patients admitted to critical care exhibited physiological instability for more than 12 hours prior to admission. EWS have subsequently been developed and introduced often alongside an outreach service. EWS have been validated in both medical and surgical populations and higher scores correlate with a higher hospital mortality and an increased likelihood of critical care admission. The use of EWS as part of a track and trigger system is now a National Institute for Health and Clinical Excellence guideline.3
C
C
C
The development of critical care outreach was one of the key recommendations in the Audit Commission’s Report ‘Critical to Success’ published in 1999. It was envisaged to be part of a service improvement for critical care allowing units to become more integrated within their trusts. ‘Comprehensive Critical Care’ published by the Department of Health in 2000 considered the provision of critical care services in the UK and how the service should be organized and delivered in the future. Outreach featured prominently in the plan and it was thought that it would influence care in three ways: to identify patients who are deteriorating on the wards and facilitate their timely admission to critical care if required to provide continuity of care for patients discharged from critical care to the wards and follow-up after hospital discharge to have a significant role in education and sharing of skills with ward nurses and doctors, allowing them to identify patients in their care at risk of deterioration and assisting in the management of patients recently discharged from critical care. Outreach teams are generally regarded as multidisciplinary, being led by a senior critical care physician and including critical care nursing and physiotherapy staff. The support of level 1 patients on the wards is central to the role of outreach. Triggers
Early warning scores were developed and introduced after studies found that patients admitted to critical care showed physiological deterioration prior to admission Basic clinical observations are noted and correspond to a score dependent upon the degree of abnormality of the observations An overall score determines the frequency of observation and the seniority of medical review required
SURGERY 30:5
226
Ó 2012 Elsevier Ltd. All rights reserved.
CRITICAL ILLNESS AND INTENSIVE CARE: I
Figure 1
for calling the outreach team will be established locally. Particular problems encountered by the outreach team include reviewing patients who are too sick to benefit from critical care and patients who do not wish to be admitted to critical care.
SURGERY 30:5
Making decisions in these groups of patients requires clinicians to be experienced in the assessment of these patients. Discussions with the parent clinical team, patient and patient’s family are often required in these situations.
227
Ó 2012 Elsevier Ltd. All rights reserved.
CRITICAL ILLNESS AND INTENSIVE CARE: I
base teams, the nursing staff and physiotherapy staff as to the appropriateness of discharge. Ideally, the decision is not made on the basis of the need for beds for new admissions. Crucial to safe discharge is the handover of the patient to medical and nursing staff. On our critical care unit, a discharge cannot take place unless there has been handover to a named member of the base medical or surgical team. The aims of discharge are to discharge a patient to a ward where they can be managed safely and the avoidance of unexpected readmission to critical care. There is a substantial impact on mortality associated with readmission to critical care. In 2008, critical care discharges from a large Scottish unit were examined.7 Patients readmitted to critical care had a hospital mortality of 40.2% compared to 8.4% for patients discharged from and not readmitted to critical care. Important predictors of readmission in the study above were age and APACHE II score. Predictors of death following discharge were age, APACHE II and time spent in hospital prior to initial critical care admission. Currently, however, there is no validated scoring system which predicts safe discharge and has a high negative predictive value for readmission. Often, the decision to discharge will be the reverse of the admission decision in terms of the level of care the patient requires according to the ICS levels of care discussed above. In general a patient with care needs less than level 2 no longer requires critical care. However in practical terms, this is complicated by the following factors: The level of care available on the ward. Some wards are more used to providing level 1 care and observations than others both in terms of nursing staffing levels and the experience of medical cover for that ward. Availability of level 1 care beds is variable and often occurs on an ad hoc basis with a ward agreeing to provide ‘special care’ for a patient which involves increasing staffing levels. In our hospital, there are eight designated level 1 care beds confined to the specialities of upper gastrointestinal surgery and plastic surgery. Familiarity of the ward with certain treatment or monitoring devices common in a critical care setting. Examples of this would be tracheostomies, intercostal drains, and lumbar or extraventricular drains. Further care planning from the base team. If a patient has a plan for further treatment in the very near future and will require critical care admission after this, they are often managed in a critical care setting prior to this rather than extra ward transfers. An example would be a patient waiting for a washout of a wound or removal of abdominal packs 48 hours after initial surgery. Intensity of physiotherapy needs. Although not meeting the criteria for level 2 care, some patients, particularly those with tracheostomies requiring regular suction of the airways are kept in a critical care setting. The timing of discharge. Patients are often not discharged at weekends due to reductions in the number and experience of ward medical staff at these times. Importantly, out of hours discharge is also avoided where possible. A 2006 Canadian study of critical care discharges was carried out which extracted data from 47,062 critical care patient discharge events. Hospital mortality was significantly
The key question is: ‘has the establishment of outreach teams in the last 10 years been shown to improve care for patients?’ There have been two randomized controlled trials (RCTs) of the impact of outreach. The Australian MERIT study published in 2005 compared 12 hospitals with an outreach service (termed ‘Medical Emergency Team’, MET) with 11 hospitals without outreach.4 The study did not demonstrate a significant reduction in cardiac arrests, unplanned ICU admissions, or unexpected deaths. A UK study published in 2004, was carried out in a single 800bed district general hospital and aimed to determine the influence of the introduction of an outreach service on in-hospital mortality and length of stay. Wards started as controls prior to a period of training before becoming intervention wards.5 This study demonstrated that the introduction of an outreach service reduced in-hospital mortality (adjusted OR ¼ 0.52; 95% CI 0.32e0.85) when compared to wards without an outreach service. A Cochrane database systematic review published in 20076 was only able to include these two RCTs in the analysis. It concluded that the level of evidence available is insufficient to make informed decisions regarding the introduction of outreach services. There are limitations to the impact that an outreach service and EWS can have on the identification and treatment of sick patients on the wards. Not all critically ill patients exhibit physiological deterioration which can be tracked over a period of time. Additionally, there is always the possibility of sudden collapse, as in a patient with a pulmonary embolus or a dysrhythmia. The introduction of an outreach team should not result in the de-skilling of ward staff to be able to identify and initially manage sick patients. The educational component of outreach should ensure ward doctors and nurses can accomplish this and communicate effectively with the critical care team. It is possible that the combination of outreach, EWS, training of nursing staff, junior medical staff rotating through critical care and the introduction of training courses for medical staff such as CCrISPÒ (Care of the Critically Ill Surgical Patient) may improve outcome.
C
C
C
C
Outreach has been introduced to assist ward teams in the management of unwell patients both prior to and subsequent to critical care admission Outreach teams have an important role in the support and education of ward medical and nursing staff in managing critically ill patients There are limitations to outreach services which may in part explain the conflicting evidence regarding its effectiveness Outreach complements early warning scoring systems and, together, they may be an effective intervention to reduce hospital mortality
Discharge from critical care Discharge from critical care is a multidisciplinary decision which, rather like admission decisions, takes into account the patient’s physiology, chronic health status, level of dependency and future treatment planning from the base team. Advice is taken from the
SURGERY 30:5
228
Ó 2012 Elsevier Ltd. All rights reserved.
CRITICAL ILLNESS AND INTENSIVE CARE: I
Commonly used scoring systems Severity score
Year
Patients
Units
Geographical area
Timing of score
APACHE II17 APACHE III18 APACHE IV19 SAPS 220 SAPS 321
1985 1991 2006 1993 2005
5815 17,440 110,558 13,152 16,784
13 40 104 137 303
First 24 hours First 24 hours First 24 hours First 24 hours Admission
MPM II022 ICNARC23
1993 2007
19,124 216,626
137 163
USA USA USA Europe, North America Europe, South America, Central America, Australasia Europe, North America UK
Admission First 24 hours
APACHE, Acute Physiology and Chronic Health Evaluation; SAPS, Simplified Acute Physiology Score; MPM, Mortality Probability Model; ICNARC, Intensive Care National Audit and Research Centre.
Table 2
higher for night-time discharges at 11.8% compared to 9.0% for daytime discharges. The odds ratio for death with night-time discharge was 1.35.8 The availability of outreach support for discharged patients. In the resource limited environment of critical care, patients requiring slightly higher levels of care and more frequent observation can be safely discharged with appropriate outreach support to the wards which would rely on the 24-hours availability of an outreach service.
C
C
C
C
C
There are many scoring systems in operation in the UK and although they have been widely used for purposes of comparative audit and are used to define patients eligible for research projects, they cannot predict outcome of critical illness in individual patients. Scoring systems can predict the probability of a binary event (survival or death) but cannot predict whether an individual patient will survive or not. The commonly used scoring systems are summarized in Table 2. There are several reasons why scoring systems may produce inaccurate, invalid or non-transferable data: Physiological and diagnostic data have to be collected and there are inherent inaccuracies in this process. In some scoring systems, data are collected from the first 24 hours of critical care admission and high scores may reflect very unwell patients or poor initial treatment. Local organization of services is important with some wards able to deliver a higher level of care and treatment prior to critical care admission. All physiological scores measure a response to an insult rather than the size of the insult itself and this response will vary markedly between patients. An immeasurable but important concept is that of biological age which is not considered in scoring systems. Neither are social deprivation nor geographical region. As improvements in care occur and admitting policies change, there is inevitably a change in the mortality of a cohort of patients with similar physiology scores over time. Exclusion criteria from scoring systems can exclude up to 15% of critical care patients. Finally, many systems developed do not account for the diagnosis causing the physiological disturbance. In some situations, an identical degree of physiological disturbance can be associated with a far lower mortality if the underlying condition is more easily treatable. The Intensive Care National Audit and Research Centre (ICNARC) model9 was developed to compare risk-adjusted outcomes between UK intensive care units and on the basis of a large patient cohort, they developed a risk prediction model to be applied to UK critical care patients. Importantly, a diagnostic weighting was included, they made no exclusions of patients
Discharge from critical care is often a complex decision requiring the opinions of several teams involved in a patient’s care There are reasons why patients are kept in a critical care environment which do not always reflect their level of care requirements There is no scoring system able reliably to predict readmission to critical care Patients readmitted to critical care have a significantly increased hospital mortality Effective handover is key to a safe discharge and to ensure continuity of care
Scoring systems Scoring systems are a means of quantifying the severity of illness of critically unwell patients. They take into account the severity of the acute physiological derangement, age, chronic health status and some have a weighting for the diagnosis associated with the critical illness. Scoring systems have potential value in a critical care setting for the following purposes: to evaluate the outcomes of one critical care unit compared to others with similar case mixes to give a standardized mortality ratio (comparative audit) to allow enrolment of patients with similar degrees of physiological derangement into clinical trials to facilitate admission decisions and assist communication about likely prognosis.
SURGERY 30:5
229
Ó 2012 Elsevier Ltd. All rights reserved.
CRITICAL ILLNESS AND INTENSIVE CARE: I
with certain pathologies and the source of admission was considered. Finally, ICNARC is continually updated with data from the participating critical care units enabling an up-to-date assessment of the risk model. The importance of including a diagnosis in the scoring system is highlighted in the following worked example. Consider two patients without significant past medical history, one with a diabetic ketoacidosis (DKA) and one with faecal peritonitis (FP). Their physiology is detailed in Table 3 below and entered into an APACHE II model which is widely used in the UK. An ICNARC-predicted mortality is also included. It can be seen that the predicted mortality varies markedly in the ICNARC model, despite the similarity of the acute physiological derangement and the identical APACHE II scores.
C
C
C C
the care of the dying patient, the General Medical Council quotes the duties of a doctor. These are listed in italic text below with the ethical principles specific to this setting in bold text.
C
C
C
C
C
Scoring systems are useful to compare outcomes between critical care units and allow enrolment in studies Scoring systems do not predict an individual’s risk of mortality from critical illness There are several scoring systems available internationally There are problems inherent in the construction of an ideal scoring system that is valid and applicable to UK populations
Unfortunately, the principles outlined above are necessarily vague and often do not provide a definitive course of action in the real-life setting of a critically unwell patient. The problem in defining a patient too sick to benefit from critical care admission is in the difficulty in defining futility and the best interests of the patient. The definition of futility is difficult because modern critical care is highly successful at normalizing physiology and the advent of extracorporeal life support and transplantation medicine offers potential solutions to previously unsurvivable short or long term physiological situations. Defining futility also requires 100% certainty of an adverse outcome which is impossible in clinical medicine. Defining best interests is also difficult and what is an acceptable outcome to some patients may be completely unacceptable to others. Two high-profile cases of patients in long-term persistent vegetative states have shown two opposite approaches to lifeprolonging treatments from the respective families.10,11 Overall, the decision to withhold admission to critical care has to be taken as part of a multidisciplinary team and broadly speaking must comply with the ethical guidelines that are available to medical practitioners. Factors to consider include the patient’s acute physiological status, degree of reversibility of illness and background physiological reserve. The patient’s views must also be considered when the patient has the capacity to participate in the process and when this is not possible, their beliefs and views based on the testaments of their family and friends are also important. Alleviating suffering is crucial and active and palliative treatments are by no means mutually exclusive and comfort and dignity are critical regardless of the direction of the treatment even when this direction is unclear at critical care referral. It is also important to take into account, the possibility and availability of a definitive management of the underlying condition which is taken with the support of the base medical team. The outcome of critical illness is often unpredictable and decisions are based on the best available evidence and the experiences of the critical care physician. Likely disability and dependence following critical illness is also important to consider in the knowledge of the patient’s wishes under these circumstances. Part of the role of critical care physicians is to be a part of a team which takes these difficult decisions and to support the patient, family and base medical team in the decision-making process.
Patients too sick to benefit from ICU Decisions relating to withholding critical care services from patients are among the most difficult that the intensive care team has to make. Historically, the four ethical principles which should be taken into clinical practice are: beneficence, nonmalevolence, autonomy and justice. In the specific context of withholding treatment and end of life decisions, dignity and honesty are additional important principles. In its guidelines for
Variable
Patient 1 (diabetic ketoacidosis)
Patient 2 (faecal peritonitis)
Age Haematocrit (%) White cell count ( 109 cells/litre) Temperature ( C) Mean arterial pressure (mmHg) Heart rate Respiratory rate Sodium (mmol/litre) Potassium (mmol/litre) pO2 (kPa) pH Creatinine (mm/litre) Glasgow Coma Scale APACHE II score APACHE II predicted mortality (%) ICNARC-predicted mortality (%)
48 30 17 33.8 50 120 34 148 6.1 8.9 7.01 112 14 21 38.9 5.6
60 24 26 35.9 68 124 34 142 3.2 8.9 7.37 216 14 21 38.9 65
Bold values represents the importance of the diagnosis in risk prediction.
Table 3
SURGERY 30:5
Make the care of your patient your first concern e beneficence, non-malevolence Protect and promote the health of patients and the public e beneficence, non-malevolence Provide a good standard of practice and care e beneficence, non-malevolence Treat patients as individuals and respect their dignity e beneficence, autonomy, dignity Work in partnership with patients e autonomy, dignity, honesty
230
Ó 2012 Elsevier Ltd. All rights reserved.
CRITICAL ILLNESS AND INTENSIVE CARE: I
C
C
C
C
2 McQuillan P, Pilkington S, Allan A, et al. Confidential inquiry into quality of care before admission to intensive care. BMJ 1998; 316: 1853e8. 3 Available at: http://www.nice.org.uk/nicemedia/live/11810/35950/ 35950.pdf (accessed on 18 Apr 2011). 4 MERIT study investigators. Introduction of the medical emergency team (MET) system: a cluster randomised controlled trial. Lancet 2005; 365: 2091e7. 5 Priestley G, Watson W, Rashidian A, et al. Introducing Critical Care Outreach: a ward-randomised trial of phased introduction in a general hospital. Intensive Care Med 2004; 30: 1398e404. 6 McGaughey J, Alderdice F, Fowler R, Kapila A, Mayhew A, Moutray M. Outreach and Early Warning Systems (EWS) for the prevention of Intensive Care admission and death of critically ill adult patients on general hospital wards (Review). Cochrane Database Syst Rev 2007. Issue 3. Art. No.:CD005529. doi:10.1002/14651858.CD005529.pub2. 7 Campbell AJ, Cook JA, Adey G, Cuthbertson BH. Predicting death and readmission after intensive care discharge. Br J Anaesth 2008; 100: 656e62. 8 Priestap FA, Martin CM. Impact of intensive care unit discharge time on patient outcome. Crit Care Med 2006; 34: 2946e51. 9 Harrison DA, Parry GJ, Carpenter JR, Short A, Rowan K. A new risk prediction model for critical care: The Intensive Care National Audit & Research Centre (ICNARC) model. Crit Care Med 2007; 35: 1091e8. 10 Miles S. Informed demand for “non-beneficial” medical treatment. N Engl J Med 1991; 325: 512e5. 11 Available at: http://news.bbc.co.uk/onthisday/hi/dates/stories/ november/19/newsid_2520000/2520581.stm (accessed on 18 Apr 2011).
Deciding which patients are too sick to benefit from critical care is one of the most challenging decisions critical care practitioners face There is guidance from the General Medical Council in the form of general ethical principles Patients’ wishes, acute physiological derangement, physiological reserve and reversibility of illness should be taken into account when making these decisions Teamwork is essential in the decision-making process
Conclusion This article has explored some of the areas surrounding evaluation of patients prior to critical care admission and the complexities of critical care discharge planning. The differences in several commonly used critical care scoring systems have been highlighted. Demand for critical care provision continues to rise and ensuring optimum care for all level 2 and 3 patients within a hospital can be challenging. A
REFERENCES 1 Department of Health. Critical care beds. Available at: http://www.dh. gov.uk/en/Publicationsandstatistics/Statistics/Performancedataand statistics/Beds/DH_077451 (accessed 19 Apr 2011).
SURGERY 30:5
231
Ó 2012 Elsevier Ltd. All rights reserved.
Criteria for intensive care unit admission and severity of illness
patients who require perioperative critical care due to demand for beds. The last year has seen further pressures on critical care resources in the UK during the H1N1 flu pandemic. The admission of a patient to critical care takes place after considering the patient’s acute illness and the degree of reversibility in the context of their background medical problems and physiological reserve.
Miriam Baruch Ben Messer
Admission criteria and levels of care In the document ‘Levels of Critical Care for Adult Patients’ published in 2009 by the Intensive Care Society (ICS), patients in hospital are assigned a level of care from 0 to 3 based on an assessment of their clinical needs. This is shown in Table 1. The level of care is not governed by the location of the patient and equally the level of care assigned does not necessarily determine the nurse-to-patient ratio for that particular patient. The identification of patients whose clinical condition is deteriorating or at risk of deteriorating to the point of requiring critical care team review can be done by ward-based medical teams. If the patient is designated level 2 or 3 they will almost certainly be moved to a critical care area. Patients who are currently level 2 but at significant risk of becoming level 3 will often be admitted directly to an ICU where their needs are most easily met. Some patients who may not previously have been considered for admission to critical care are now being considered in certain circumstances. Examples include patients with multiple comorbidities where specific limits of treatment are set and in other patients after the decision to initiate palliative care to facilitate organ donation after cardiac death.
Abstract This article explores the provision and organisation of critical care services in the UK and examines the issues surrounding admission to, discharge from and the withholding of critical care. Critical care expansion in the UK in recent years has centred on the provision of increased numbers of beds and the development of outreach services. Despite the expansion in critical care there is difficulty in matching supply and demand for beds. There remains controversy regarding the effectiveness of outreach in improving outcomes for patients referred to critical care. The discharge of patients from critical care has also come under scrutiny since mortality rates are higher for patients discharged out-of-hours. Patients’ needs following critical care are carefully planned with the base medical teams because readmission to critical care is associated with increased mortality. Scoring systems are used in critical care to compare outcomes between critical care units and to facilitate research but cannot predict outcome for individual patients. The decision that patients will not benefit from critical care admission can be one of the most difficult. Many patients who would not have been considered for organ support previously are now admitted to critical care with pre-determined limits of treatment. Involvement of the base team is integral to making decisions for these patients.
C
C
Keywords APACHE; critical care ethics; critical care scoring systems; Early Warning Scores; ICNARC; Intensive Care Society; levels of critical care; outreach; standardised mortality ratio
Critical care has been organized into levels of care depending upon the level of organ support and observation that an individual patient requires The levels of care are from 0 to 3 with 0 being standard ward care and observation and level 3 being what is traditionally thought of as intensive care
Early warning scores Track and trigger systems are used to identify patients at risk of deterioration. An integral part of these is the early warning score. Early warning scores (EWS) have been developed to increase awareness and recognition of acute illness by assigning a graded score to increasingly deranged physiological parameters within the following domains: respiratory, cardiovascular, urine output and conscious level. The score is based on simple observations which may be different depending on local case mix but usually include respiratory rate, pulse rate, blood pressure, temperature, conscious level and urine output. The scores across the domains are added to give an EWS which can then be entered into an algorithm to establish the most appropriate course of action which is defined by the frequency of subsequent observation and by the experience of the medical review that the EWS should trigger. We include the EWS used in our hospital in Figure 1. During the 1990s there was increasing recognition that care prior to critical care admission was often suboptimal with regard to simple management of airway, breathing, circulation, oxygen
Introduction The provision of critical care beds in England has increased by over 800 in the last 10 years,1 a rise of almost 30%. Figures from January 2011 reveal there are 3747 adult critical care beds in England, 45% in high-dependency units (HDU) and 55% in intensive care units (ICU). Ideally this provision should be able to accommodate peaks and troughs in demand, but it is not uncommon for hospital trusts to postpone major surgery for
Miriam Baruch FRCA is a Consultant in Anaesthesia & Intensive Care at Royal Victoria Infirmary, Newcastle upon Tyne, UK. Conflicts of interest: none declared. Ben Messer MA DICM is a Consultant in Anaesthesia & Intensive Care at Royal Victoria Infirmary, Newcastle upon Tyne, UK. Conflicts of interest: none declared.
SURGERY 30:5
225
Ó 2012 Elsevier Ltd. All rights reserved.
CRITICAL ILLNESS AND INTENSIVE CARE: I
Levels of critical care for adult patients as described by the Intensive Care Society Level
Criteria
0
C
1
C C C
Examples
Medical and nursing needs can be met with standard ward based care
C
Patients recently discharged from critical care Patients requiring supplementary clinical monitoring Patients referred for critical care team review or being seen regularly by the outreach service
C
C
C
C C C
2
C
Patients receiving single organ support (basic respiratory and basic cardiovascular support together can be considered as level 2)
C
Intravenous fluid and antibiotic therapy 4 hourly observations Patients at risk of clinical deterioration Epidural analgesia or patient-controlled analgesia in use requiring specific nursing observations Tracheostomy in-situ without respiratory support Patients who have a chest drain in-situ Parenteral nutrition Basic respiratory support
C
Mask/hood/tracheostomy continuous positive airway pressure or non-invasive bi-level ventilation FiO2 0.5 delivered via facemask Basic cardiovascular support
C
Use of arterial line for invasive blood pressure monitoring or arterial blood gas sampling Single intravenous vasoactive drug infusion Advanced cardiovascular support
C
Multiple vasoactive/inotropic drug infusions Renal support
C
Acute renal replacement therapy Neurological support Continuous intravenous medication to control seizures
3
C C
Patients receiving advanced respiratory support Patients receiving a minimum of two organs supported
C
Ventilatory support via an endotracheal tube (including tracheostomy)
Table 1
Outreach
therapy and monitoring requirements. In one study of patients admitted to critical care, care was suboptimal in more than half of admitted patients and could have contributed to mortality in one-third of patients.2 Furthermore, cardiac arrest and critical care admission are often preceded by derangements in acute physiology. This has been documented by the NCEPOD report of 2005 which found that 66% of patients admitted to critical care exhibited physiological instability for more than 12 hours prior to admission. EWS have subsequently been developed and introduced often alongside an outreach service. EWS have been validated in both medical and surgical populations and higher scores correlate with a higher hospital mortality and an increased likelihood of critical care admission. The use of EWS as part of a track and trigger system is now a National Institute for Health and Clinical Excellence guideline.3
C
C
C
The development of critical care outreach was one of the key recommendations in the Audit Commission’s Report ‘Critical to Success’ published in 1999. It was envisaged to be part of a service improvement for critical care allowing units to become more integrated within their trusts. ‘Comprehensive Critical Care’ published by the Department of Health in 2000 considered the provision of critical care services in the UK and how the service should be organized and delivered in the future. Outreach featured prominently in the plan and it was thought that it would influence care in three ways: to identify patients who are deteriorating on the wards and facilitate their timely admission to critical care if required to provide continuity of care for patients discharged from critical care to the wards and follow-up after hospital discharge to have a significant role in education and sharing of skills with ward nurses and doctors, allowing them to identify patients in their care at risk of deterioration and assisting in the management of patients recently discharged from critical care. Outreach teams are generally regarded as multidisciplinary, being led by a senior critical care physician and including critical care nursing and physiotherapy staff. The support of level 1 patients on the wards is central to the role of outreach. Triggers
Early warning scores were developed and introduced after studies found that patients admitted to critical care showed physiological deterioration prior to admission Basic clinical observations are noted and correspond to a score dependent upon the degree of abnormality of the observations An overall score determines the frequency of observation and the seniority of medical review required
SURGERY 30:5
226
Ó 2012 Elsevier Ltd. All rights reserved.
CRITICAL ILLNESS AND INTENSIVE CARE: I
Figure 1
for calling the outreach team will be established locally. Particular problems encountered by the outreach team include reviewing patients who are too sick to benefit from critical care and patients who do not wish to be admitted to critical care.
SURGERY 30:5
Making decisions in these groups of patients requires clinicians to be experienced in the assessment of these patients. Discussions with the parent clinical team, patient and patient’s family are often required in these situations.
227
Ó 2012 Elsevier Ltd. All rights reserved.
CRITICAL ILLNESS AND INTENSIVE CARE: I
base teams, the nursing staff and physiotherapy staff as to the appropriateness of discharge. Ideally, the decision is not made on the basis of the need for beds for new admissions. Crucial to safe discharge is the handover of the patient to medical and nursing staff. On our critical care unit, a discharge cannot take place unless there has been handover to a named member of the base medical or surgical team. The aims of discharge are to discharge a patient to a ward where they can be managed safely and the avoidance of unexpected readmission to critical care. There is a substantial impact on mortality associated with readmission to critical care. In 2008, critical care discharges from a large Scottish unit were examined.7 Patients readmitted to critical care had a hospital mortality of 40.2% compared to 8.4% for patients discharged from and not readmitted to critical care. Important predictors of readmission in the study above were age and APACHE II score. Predictors of death following discharge were age, APACHE II and time spent in hospital prior to initial critical care admission. Currently, however, there is no validated scoring system which predicts safe discharge and has a high negative predictive value for readmission. Often, the decision to discharge will be the reverse of the admission decision in terms of the level of care the patient requires according to the ICS levels of care discussed above. In general a patient with care needs less than level 2 no longer requires critical care. However in practical terms, this is complicated by the following factors: The level of care available on the ward. Some wards are more used to providing level 1 care and observations than others both in terms of nursing staffing levels and the experience of medical cover for that ward. Availability of level 1 care beds is variable and often occurs on an ad hoc basis with a ward agreeing to provide ‘special care’ for a patient which involves increasing staffing levels. In our hospital, there are eight designated level 1 care beds confined to the specialities of upper gastrointestinal surgery and plastic surgery. Familiarity of the ward with certain treatment or monitoring devices common in a critical care setting. Examples of this would be tracheostomies, intercostal drains, and lumbar or extraventricular drains. Further care planning from the base team. If a patient has a plan for further treatment in the very near future and will require critical care admission after this, they are often managed in a critical care setting prior to this rather than extra ward transfers. An example would be a patient waiting for a washout of a wound or removal of abdominal packs 48 hours after initial surgery. Intensity of physiotherapy needs. Although not meeting the criteria for level 2 care, some patients, particularly those with tracheostomies requiring regular suction of the airways are kept in a critical care setting. The timing of discharge. Patients are often not discharged at weekends due to reductions in the number and experience of ward medical staff at these times. Importantly, out of hours discharge is also avoided where possible. A 2006 Canadian study of critical care discharges was carried out which extracted data from 47,062 critical care patient discharge events. Hospital mortality was significantly
The key question is: ‘has the establishment of outreach teams in the last 10 years been shown to improve care for patients?’ There have been two randomized controlled trials (RCTs) of the impact of outreach. The Australian MERIT study published in 2005 compared 12 hospitals with an outreach service (termed ‘Medical Emergency Team’, MET) with 11 hospitals without outreach.4 The study did not demonstrate a significant reduction in cardiac arrests, unplanned ICU admissions, or unexpected deaths. A UK study published in 2004, was carried out in a single 800bed district general hospital and aimed to determine the influence of the introduction of an outreach service on in-hospital mortality and length of stay. Wards started as controls prior to a period of training before becoming intervention wards.5 This study demonstrated that the introduction of an outreach service reduced in-hospital mortality (adjusted OR ¼ 0.52; 95% CI 0.32e0.85) when compared to wards without an outreach service. A Cochrane database systematic review published in 20076 was only able to include these two RCTs in the analysis. It concluded that the level of evidence available is insufficient to make informed decisions regarding the introduction of outreach services. There are limitations to the impact that an outreach service and EWS can have on the identification and treatment of sick patients on the wards. Not all critically ill patients exhibit physiological deterioration which can be tracked over a period of time. Additionally, there is always the possibility of sudden collapse, as in a patient with a pulmonary embolus or a dysrhythmia. The introduction of an outreach team should not result in the de-skilling of ward staff to be able to identify and initially manage sick patients. The educational component of outreach should ensure ward doctors and nurses can accomplish this and communicate effectively with the critical care team. It is possible that the combination of outreach, EWS, training of nursing staff, junior medical staff rotating through critical care and the introduction of training courses for medical staff such as CCrISPÒ (Care of the Critically Ill Surgical Patient) may improve outcome.
C
C
C
C
Outreach has been introduced to assist ward teams in the management of unwell patients both prior to and subsequent to critical care admission Outreach teams have an important role in the support and education of ward medical and nursing staff in managing critically ill patients There are limitations to outreach services which may in part explain the conflicting evidence regarding its effectiveness Outreach complements early warning scoring systems and, together, they may be an effective intervention to reduce hospital mortality
Discharge from critical care Discharge from critical care is a multidisciplinary decision which, rather like admission decisions, takes into account the patient’s physiology, chronic health status, level of dependency and future treatment planning from the base team. Advice is taken from the
SURGERY 30:5
228
Ó 2012 Elsevier Ltd. All rights reserved.
CRITICAL ILLNESS AND INTENSIVE CARE: I
Commonly used scoring systems Severity score
Year
Patients
Units
Geographical area
Timing of score
APACHE II17 APACHE III18 APACHE IV19 SAPS 220 SAPS 321
1985 1991 2006 1993 2005
5815 17,440 110,558 13,152 16,784
13 40 104 137 303
First 24 hours First 24 hours First 24 hours First 24 hours Admission
MPM II022 ICNARC23
1993 2007
19,124 216,626
137 163
USA USA USA Europe, North America Europe, South America, Central America, Australasia Europe, North America UK
Admission First 24 hours
APACHE, Acute Physiology and Chronic Health Evaluation; SAPS, Simplified Acute Physiology Score; MPM, Mortality Probability Model; ICNARC, Intensive Care National Audit and Research Centre.
Table 2
higher for night-time discharges at 11.8% compared to 9.0% for daytime discharges. The odds ratio for death with night-time discharge was 1.35.8 The availability of outreach support for discharged patients. In the resource limited environment of critical care, patients requiring slightly higher levels of care and more frequent observation can be safely discharged with appropriate outreach support to the wards which would rely on the 24-hours availability of an outreach service.
C
C
C
C
C
There are many scoring systems in operation in the UK and although they have been widely used for purposes of comparative audit and are used to define patients eligible for research projects, they cannot predict outcome of critical illness in individual patients. Scoring systems can predict the probability of a binary event (survival or death) but cannot predict whether an individual patient will survive or not. The commonly used scoring systems are summarized in Table 2. There are several reasons why scoring systems may produce inaccurate, invalid or non-transferable data: Physiological and diagnostic data have to be collected and there are inherent inaccuracies in this process. In some scoring systems, data are collected from the first 24 hours of critical care admission and high scores may reflect very unwell patients or poor initial treatment. Local organization of services is important with some wards able to deliver a higher level of care and treatment prior to critical care admission. All physiological scores measure a response to an insult rather than the size of the insult itself and this response will vary markedly between patients. An immeasurable but important concept is that of biological age which is not considered in scoring systems. Neither are social deprivation nor geographical region. As improvements in care occur and admitting policies change, there is inevitably a change in the mortality of a cohort of patients with similar physiology scores over time. Exclusion criteria from scoring systems can exclude up to 15% of critical care patients. Finally, many systems developed do not account for the diagnosis causing the physiological disturbance. In some situations, an identical degree of physiological disturbance can be associated with a far lower mortality if the underlying condition is more easily treatable. The Intensive Care National Audit and Research Centre (ICNARC) model9 was developed to compare risk-adjusted outcomes between UK intensive care units and on the basis of a large patient cohort, they developed a risk prediction model to be applied to UK critical care patients. Importantly, a diagnostic weighting was included, they made no exclusions of patients
Discharge from critical care is often a complex decision requiring the opinions of several teams involved in a patient’s care There are reasons why patients are kept in a critical care environment which do not always reflect their level of care requirements There is no scoring system able reliably to predict readmission to critical care Patients readmitted to critical care have a significantly increased hospital mortality Effective handover is key to a safe discharge and to ensure continuity of care
Scoring systems Scoring systems are a means of quantifying the severity of illness of critically unwell patients. They take into account the severity of the acute physiological derangement, age, chronic health status and some have a weighting for the diagnosis associated with the critical illness. Scoring systems have potential value in a critical care setting for the following purposes: to evaluate the outcomes of one critical care unit compared to others with similar case mixes to give a standardized mortality ratio (comparative audit) to allow enrolment of patients with similar degrees of physiological derangement into clinical trials to facilitate admission decisions and assist communication about likely prognosis.
SURGERY 30:5
229
Ó 2012 Elsevier Ltd. All rights reserved.
CRITICAL ILLNESS AND INTENSIVE CARE: I
with certain pathologies and the source of admission was considered. Finally, ICNARC is continually updated with data from the participating critical care units enabling an up-to-date assessment of the risk model. The importance of including a diagnosis in the scoring system is highlighted in the following worked example. Consider two patients without significant past medical history, one with a diabetic ketoacidosis (DKA) and one with faecal peritonitis (FP). Their physiology is detailed in Table 3 below and entered into an APACHE II model which is widely used in the UK. An ICNARC-predicted mortality is also included. It can be seen that the predicted mortality varies markedly in the ICNARC model, despite the similarity of the acute physiological derangement and the identical APACHE II scores.
C
C
C C
the care of the dying patient, the General Medical Council quotes the duties of a doctor. These are listed in italic text below with the ethical principles specific to this setting in bold text.
C
C
C
C
C
Scoring systems are useful to compare outcomes between critical care units and allow enrolment in studies Scoring systems do not predict an individual’s risk of mortality from critical illness There are several scoring systems available internationally There are problems inherent in the construction of an ideal scoring system that is valid and applicable to UK populations
Unfortunately, the principles outlined above are necessarily vague and often do not provide a definitive course of action in the real-life setting of a critically unwell patient. The problem in defining a patient too sick to benefit from critical care admission is in the difficulty in defining futility and the best interests of the patient. The definition of futility is difficult because modern critical care is highly successful at normalizing physiology and the advent of extracorporeal life support and transplantation medicine offers potential solutions to previously unsurvivable short or long term physiological situations. Defining futility also requires 100% certainty of an adverse outcome which is impossible in clinical medicine. Defining best interests is also difficult and what is an acceptable outcome to some patients may be completely unacceptable to others. Two high-profile cases of patients in long-term persistent vegetative states have shown two opposite approaches to lifeprolonging treatments from the respective families.10,11 Overall, the decision to withhold admission to critical care has to be taken as part of a multidisciplinary team and broadly speaking must comply with the ethical guidelines that are available to medical practitioners. Factors to consider include the patient’s acute physiological status, degree of reversibility of illness and background physiological reserve. The patient’s views must also be considered when the patient has the capacity to participate in the process and when this is not possible, their beliefs and views based on the testaments of their family and friends are also important. Alleviating suffering is crucial and active and palliative treatments are by no means mutually exclusive and comfort and dignity are critical regardless of the direction of the treatment even when this direction is unclear at critical care referral. It is also important to take into account, the possibility and availability of a definitive management of the underlying condition which is taken with the support of the base medical team. The outcome of critical illness is often unpredictable and decisions are based on the best available evidence and the experiences of the critical care physician. Likely disability and dependence following critical illness is also important to consider in the knowledge of the patient’s wishes under these circumstances. Part of the role of critical care physicians is to be a part of a team which takes these difficult decisions and to support the patient, family and base medical team in the decision-making process.
Patients too sick to benefit from ICU Decisions relating to withholding critical care services from patients are among the most difficult that the intensive care team has to make. Historically, the four ethical principles which should be taken into clinical practice are: beneficence, nonmalevolence, autonomy and justice. In the specific context of withholding treatment and end of life decisions, dignity and honesty are additional important principles. In its guidelines for
Variable
Patient 1 (diabetic ketoacidosis)
Patient 2 (faecal peritonitis)
Age Haematocrit (%) White cell count ( 109 cells/litre) Temperature ( C) Mean arterial pressure (mmHg) Heart rate Respiratory rate Sodium (mmol/litre) Potassium (mmol/litre) pO2 (kPa) pH Creatinine (mm/litre) Glasgow Coma Scale APACHE II score APACHE II predicted mortality (%) ICNARC-predicted mortality (%)
48 30 17 33.8 50 120 34 148 6.1 8.9 7.01 112 14 21 38.9 5.6
60 24 26 35.9 68 124 34 142 3.2 8.9 7.37 216 14 21 38.9 65
Bold values represents the importance of the diagnosis in risk prediction.
Table 3
SURGERY 30:5
Make the care of your patient your first concern e beneficence, non-malevolence Protect and promote the health of patients and the public e beneficence, non-malevolence Provide a good standard of practice and care e beneficence, non-malevolence Treat patients as individuals and respect their dignity e beneficence, autonomy, dignity Work in partnership with patients e autonomy, dignity, honesty
230
Ó 2012 Elsevier Ltd. All rights reserved.
CRITICAL ILLNESS AND INTENSIVE CARE: I
C
C
C
C
2 McQuillan P, Pilkington S, Allan A, et al. Confidential inquiry into quality of care before admission to intensive care. BMJ 1998; 316: 1853e8. 3 Available at: http://www.nice.org.uk/nicemedia/live/11810/35950/ 35950.pdf (accessed on 18 Apr 2011). 4 MERIT study investigators. Introduction of the medical emergency team (MET) system: a cluster randomised controlled trial. Lancet 2005; 365: 2091e7. 5 Priestley G, Watson W, Rashidian A, et al. Introducing Critical Care Outreach: a ward-randomised trial of phased introduction in a general hospital. Intensive Care Med 2004; 30: 1398e404. 6 McGaughey J, Alderdice F, Fowler R, Kapila A, Mayhew A, Moutray M. Outreach and Early Warning Systems (EWS) for the prevention of Intensive Care admission and death of critically ill adult patients on general hospital wards (Review). Cochrane Database Syst Rev 2007. Issue 3. Art. No.:CD005529. doi:10.1002/14651858.CD005529.pub2. 7 Campbell AJ, Cook JA, Adey G, Cuthbertson BH. Predicting death and readmission after intensive care discharge. Br J Anaesth 2008; 100: 656e62. 8 Priestap FA, Martin CM. Impact of intensive care unit discharge time on patient outcome. Crit Care Med 2006; 34: 2946e51. 9 Harrison DA, Parry GJ, Carpenter JR, Short A, Rowan K. A new risk prediction model for critical care: The Intensive Care National Audit & Research Centre (ICNARC) model. Crit Care Med 2007; 35: 1091e8. 10 Miles S. Informed demand for “non-beneficial” medical treatment. N Engl J Med 1991; 325: 512e5. 11 Available at: http://news.bbc.co.uk/onthisday/hi/dates/stories/ november/19/newsid_2520000/2520581.stm (accessed on 18 Apr 2011).
Deciding which patients are too sick to benefit from critical care is one of the most challenging decisions critical care practitioners face There is guidance from the General Medical Council in the form of general ethical principles Patients’ wishes, acute physiological derangement, physiological reserve and reversibility of illness should be taken into account when making these decisions Teamwork is essential in the decision-making process
Conclusion This article has explored some of the areas surrounding evaluation of patients prior to critical care admission and the complexities of critical care discharge planning. The differences in several commonly used critical care scoring systems have been highlighted. Demand for critical care provision continues to rise and ensuring optimum care for all level 2 and 3 patients within a hospital can be challenging. A
REFERENCES 1 Department of Health. Critical care beds. Available at: http://www.dh. gov.uk/en/Publicationsandstatistics/Statistics/Performancedataand statistics/Beds/DH_077451 (accessed 19 Apr 2011).
SURGERY 30:5
231
Ó 2012 Elsevier Ltd. All rights reserved.