NCBI Bookshelf. A service of the National Library of Medicine, National Institutes of Health.
National Clinical Guideline Centre (UK). Care of Dying Adults in the Last Days of Life. London: National Institute for Health and Care Excellence (NICE); 2015 Dec 16. (NICE Guideline, No. 31.)
5.1. Introduction
The recognition and weighing up of factors that may indicate that someone is in the last days or hours of life are complex and subtle. This can be a difficult task, even for an experienced palliative care clinician. Prognostic tools have been developed to assist clinicians in making a more accurate prognosis, but they are not used in routine clinical practice so clinicians are not familiar with them.
The current approach to recognising imminent dying utilises a range of signs and symptoms that are best observed over days to weeks, if the dying person's clinical course allows such observations. Over a period of days these include multiple organ failure, progressive weakness, reduced mobility and ability to carry out normal activities of daily living, increased periods of sleep, reduced oral intake and a general reduction in cognitive function, awareness and communication (with family or other important people as well as professionals). Changes that may indicate impending death within hours, that have been prioritised for inclusion in this review, include variations in respiratory cycle, weakening of pulse, and shutting down of skin circulation, and noisy respiratory secretions.
A further challenge arises when a person who was thought to be imminently dying, starts to show signs of recovery such as increased alertness and communication, desire for oral intake and improved mobility. Such reversals may be temporary, or may signify a true recovery from the dying process. Therefore it is important to determine the evidence base in this area to implement any necessary changes in clinical management to assist the person with living for a longer period of time, for example, reinstatement of medications, hydration and nutrition that may have been withdrawn.
The ‘More Care Less Pathways’ review30 recommended that clear guidance be issued on the clinical decision-making process at the end of life and, in particular, managing the uncertainties around diagnosing the dying or recovery phases. The Committee chose to ask the following question.
5.2. Review question: What signs and symptoms indicate that adults are likely to be entering their final days of life; or that they may be recovering? How are uncertainties about either situation dealt with?
For full details see review protocol in Appendix C.
This is an integrative review86 which allows for the inclusion of different study designs (experimental, observational as well as qualitative) in order to fully understand an area of concern. The incorporation of qualitative elements (and information from published Delphi consensus surveys) enabled further exploration of these areas. Mixed methodology is often used to capture a wide range of evidence in systematic review, but further to the synthesis of the relevant studies it includes a thematic analysis to provide a conceptual map of the topic (that is, a theoretical framework). The results are presented as a summary, and narrative synthesis captures results that may not be directly apparent from a quantitative or narrative synthesis alone (such as the uncertainties of recognising the signs in the final stages which will be useful for the other topics in this guideline).
Table 12
PICO characteristics of review question.
5.3. Quantitative review: clinical evidence
Seven studies were included in the review;21,33,47,52,61,62,65 these are summarised in Table 13 below. Evidence from these studies is summarised in the GRADE clinical evidence profile below (Tables 14-15). See also the study selection flow chart in Appendix E, study evidence tables in Appendix H, forest plots in Appendix K, GRADE tables in Appendix J and excluded studies list in Appendix L.
Table 13
Summary of studies included in the review.
Table 14
Clinical evidence profile: Diagnostic performance of predictors of mortality.
Table 15
Clinical evidence profile: Prognostic indicators of mortality.
The Eastern Cooperative Oncology (ECOG) performance status was included in the review, which is a scale ranging from 0 (fully active, able to carry on all pre-disease performance without restriction) to 5 (dead).
5.4. Qualitative review: clinical evidence
Three qualitative studies29,51,99 and 5 surveys2,13,22,32,54 were identified. These papers are summarised in Table 16 below. Key findings from these studies are summarised in the clinical evidence summary Table 17. See also the study selection flow chart in Appendix E, study evidence tables in Appendix H, and excluded studies list in Appendix M.
Table 16
Summary of studies included in the review.
Table 17
Themes and sub-themes derived from the evidence.
Two of the qualitative studies interviewed nursing staff with experience in nursing oncological patients on the signs and symptoms that they believe indicate someone with cancer is in the last days of life.29,99 A further qualitative study interviewed junior doctors about prognosis and approach to care decisions when caring for seriously ill hospitalised people with conditions other than cancer.
Five surveys reporting descriptive data around recognising dying were found which both supported the themes found in the qualitative reviews and provided further information. Two of these were Delphi studies, 1 of which focused on nurses' opinions of gastrointestinal cancer patients54 whilst the other Delphi included both the multiprofessional team and a lay person's opinions on people dying from any cause.32
There were 2 prospective observational studies13,22 and 1 retrospective study2 which investigated the factors that affect prognostic accuracy of doctor's assessments of dying people. These included all causes of deaths.
5.4.1. Summary of included studies
5.4.2. Summary of themes
5.5. Economic evidence
Published literature
No relevant economic evaluations were identified.
See also the economic article selection flow chart in Appendix F.
5.6. Evidence statements
Clinical - quantitative
The quantitative evidence review found that there is moderate quality evidence from observational studies using multivariate analysis of people with terminal cancer admitted to a palliative care unit, reporting Eastern Cooperative Oncology Group (ECOG) score, fatigue and desaturation. One study of 374 people showed a reduced ECOG score as a predictor of mortality within 7 days, OR 3.45 (1.65, 7.20). An associated study of people aged 65 years and over (n=459) supported this finding, OR 2.02 (1.40, 2.92). A low quality study of 93 people in the same setting determined fatigue and desaturation as predictors of mortality within 2 weeks, HR 5.90 (2.04, 17.03) and HR 3.30 (1.42, 7.66), respectively.
An increased triage pulse (greater than or equal to 110 bpm) and increased triage respiration (greater than 28/min) was identified as a predictor of mortality within 2 weeks, RR 4.92 (1.42, 17.09) and RR 12.72 (3.08, 52.49), respectively (low quality evidence). This was from 1 observational study (n=122) using multivariate analysis of people with cancer presenting to an emergency department with acute dyspnoea as a primary or secondary complaint.
Moderate quality evidence from a diagnostic observational study of 357 people with terminal cancer admitted to a palliative care unit, indicated that clinical signs and symptoms (palliative prognostic score, Richmond agitation scale, death rattle, apnoea periods, respiration with mandibular movement, peripheral cyanosis, Cheyne-Stoke breathing and pulselessness of radial artery) have a high specificity (81.3% - 99.2%) and varying sensitivities (11.3% - 64.0%) for diagnosing mortality within 3 days. Two large, but low quality, diagnostic retrospective observational studies of people admitted to hospital for more than 24 hours (n=42701) and presenting at the emergency department (n=71453), showed that laboratory tests can diagnose mortality within 2 days (sensitivity 36.2 - 71.8%, specificity 64 - 84.7%). Area under the curve for these tests ranged from 0.53 - 0.80, indicating very poor to moderate test accuracy.
Clinical - qualitative
Qualitative evidence indicated several themes around healthcare professionals' experiences in recognising adults that are entering their final days of life or who may be recovering. Moderate quality evidence from 5 studies (2 qualitative studies, n=33; 2 Delphi studies, n=324; and 1 observational study, n=474) indicated that physical changes, including cardiovascular changes, deterioration of physical condition, reduced oral intake, worsening pain and skin changes, were observed. Two moderate quality qualitative studies of 33 healthcare professionals identified presentation of spiritual and psychosocial changes, such as social withdrawal, changes in mood and changes in spiritual experience.
The theme of difficulty in recognising dying was found to include the following subthemes; complexity of recognising dying (2 interviews and 1 survey of moderate quality, n=285) and factors that affect prognostic accuracy (2 surveys of moderate quality, n=719). The dying trajectory was recognised as variable in length of time (1 study of moderate quality, n=15).
Little evidence was identified for managing uncertainty for those entering the last days of life or who may be recovering. One low quality qualitative study (n=8) was identified that explored junior doctors' perceptions on how they would manage people differently if they thought they were going to die.
Economic
No relevant economic evaluations were identified.
5.7. Conceptual framework
The evidence identified from the quantitative and qualitative reviews has been summarised graphically in a conceptual framework shown in Figure 3. This reflects the themes identified from the qualitative review along with the evidence from the quantitative review. The Committee were keen to represent the dying trajectory and the potential for improving within this framework. The Committee considered it an important tool for bringing together the mixed methods review and aided formulation of recommendations.
Figure 3
Conceptual framework for recognising dying.
5.8. Recommendations and link to evidence
| Recommendations |
|
|---|---|
| Relative values of different outcomes | The Committee designed the protocol for this review on the symptom categories as described in Domeisen et al., 2013,32 and felt critical outcomes included:
Prognostic and diagnostic outcomes were prioritised for inclusion in the review with confounding factors such as treatments that may supress levels of consciousness or artificial organ support, such as ventilation. Although biochemical markers were not specifically included in the scope in relation to their role in recognising dying, the Committee recognised that many people, particularly those being cared for in hospital, will be having laboratory tests. They requested that, where available in the evidence reviewed, this information should be captured and presented. The literature search was performed around recognising dying and signs and symptoms, and any laboratory test data were presented to the Committee. The Committee noted that there are tools which can help clinicians to prognosticate if a person has years or months (and possibly just weeks) to live – these tools exist for cancer and Chronic Obstructive Pulmonary Disease (COPD). These tools are not sensitive enough for use in our remit, that is, to recognise when a person is shifting into the last days or hours and therefore prognostic tools were excluded from the evidence review. |
| Trade-off between clinical benefits and harms | The quantitative review identified evidence of a range of clinical signs and symptoms that may indicate imminent mortality, such as the Eastern Cooperative Oncology Group (ECOG) scale, death rattle, apnoea periods, respiration with mandibular movement and peripheral cyanosis, although sensitivity was low. Weak evidence was identified for laboratory tests for diagnosing imminent death. It is noted that the majority of the studies were conducted in specific populations (for example, people with terminal cancer). The Committee discussed the trade-off of having a high sensitivity versus a high specificity of identifying imminent death and considered that both were very important, but that a high specificity is key, so that nearing death is not mistakenly diagnosed. The clinical signs and symptoms identified in the review are non-invasive tests or measures and therefore should not cause any harm to the dying person. Benefits of correctly recognising imminent death may allow opportunity for shared decision making and allow valuable time between the dying person and those important to them. No harms were identified for using signs or symptoms for recognising when a person is entering the last days of life. |
| Trade off between net health benefits and resource use | No economic evaluations were identified for strategies that recognised when the individual was entering the dying phase. Such strategies will have economic consequences as once it is recognised that an individual is entering the dying phase, they will receive different treatment that will impact resource use. Correctly predicting when an individual is in the dying phase is integral to patient outcomes to ensure protocols are in place and unnecessary interventions are not initiated. Most of the symptoms used to predict when an individual is entering the dying phase do not require any equipment or tests for detection and can be gathered from examining the person. These signs and symptoms will likely have been gathered through regular monitoring anyway. The Committee stressed the need for improved communication between healthcare professionals and that a specialist should be consulted when there is great uncertainty. However, in most cases, this assessment should be completed by clinicians, therefore it is unlikely there are increased upfront costs incurred for recognising dying, apart from within community settings where there could be some additional costs if the clinician has to be called to do the assessment. An increase in the number of correctly predicted cases could reduce downstream costs as they prevent unnecessary interventions being initiated. |
| Quality of evidence | Low to moderate quality prognostic and diagnostic outcomes were identified for the quantitative review. The Committee was not surprised that several signs and symptoms were highlighted as predictors of mortality, such as ECOG status, triage pulse and triage respiration, especially given the specific populations of the studies (people with terminal cancer and people with terminal lung cancer with acute dyspnoea, respectively). Other signs and symptoms were of interest to the Committee and gave high specificities, but low sensitivities for diagnosis of mortality within 3 days,47,48 such as noisy respiratory secretions, apnoea periods, and respiration with mandibular movement, peripheral cyanosis, Cheyne-Stoke breathing and pulselessness of radial artery. No evidence could be pooled given the variation in outcomes. Moderate to high quality evidence across themes were identified in the qualitative review. These included the following main themes
|
| Other considerations | From the evidence review, the Committee recognised similar factors that they use in their clinical practice to recognise entering the dying phase. They drew on the importance of gathering information from multiple sources in order to do this, including different members of the multiprofessional team. These included a review of the person's medical history and trajectory of symptom deterioration. The Committee recognised that in some people this can be a reflection of a growing need for physiological support, particularly in the intensive care setting. The Committee also discussed the importance of clarifying any change in the dying person's social, spiritual and psychological needs, and also eliciting any goals and wishes they may have, which may be listed in the dying person's advance care plan. The Committee wanted to highlight the importance of basic principles of care when interacting with the dying person in the last days of life, considering the views of the person and those important to them. The Committee wanted to emphasise to those recognising dying that the trajectory also includes potential recovery and improvement and that uncertainty in diagnosing the individual should be taken into account when assessing for potential recovery. The Committee also discussed the reversibility of each individual symptom, for example for a person presenting with progressive weight loss there may be treatable causes that are inhibiting someone from eating.37 The Committee therefore made a consensus recommendation that noted that changes in signs and symptoms could also represent stabilizing of the person's condition , even if temporarily, or that recovery was possible. The evidence review highlighted numerous signs and symptoms that could be used in recognising dying, including fatigue or progressive weight loss. The Committee highlighted that some signs and symptoms may be specific to the last days of life including Cheyne Stokes breathing and noisy respiratory secretions but, whilst specific, they are not universal symptoms. The evidence review suggested functional observations were predictors of mortality; in particular the Eastern Cooperative Oncology Group (ECOG) score. The Committee noted that this was not widely used in the UK, but is similar to the WHO performance scale (also called the Zubrod score). The Committee noted that it is specifically deterioration in the ECOG score that would indicate a likelihood of entering the last days of life, recognising that some disabled people may be at a score of 4 outside of illness. Although not identified in the evidence review, the Committee discussed other scores that may be useful, such as the Barthel Activities of Daily Living Index, Karnofsky Performance Status Scale and the Australia-modified Karnofsky Performance Scale. The value of laboratory tests, such as renal function tests or radiological imaging, in recognising dying was discussed. The Committee noted that, whilst these can be useful tests in practice in an acute setting, these tests may not be appropriate to support recognising dying when people are dying in community settings, as they are invasive and may be considered inappropriate to measure. They chose therefore only to include them in their recommendation if they were available and noted that any data should be used in conjunction with other information of signs and symptoms as discussed above. The Committee made a further consensus recommendation that acknowledged that there may be some circumstances where undertaking clinical tests, even in the last 2-3 days of life, should be undertaken if there was a clinical imperative to do so. That is, the results would directly impact on the care of that person. The Committee felt that such examples would include situations where a full blood count could guide the use of platelet transfusion to avoid catastrophic bleeding. Additionally, measurement of serum electrolytes may helpfully indicate a cause for persistent agitation and seizures. The Committee discussed the evidence base and noted that it was in small and specific populations, such as people with lung cancer, whereas this guidance is looking at a broader population. The Committee recognised that the likely time of death is particularly difficult to anticipate in some chronic conditions, for example dementia, when the disease trajectory is typically variable and there may be a long-standing reduced level of functioning. The Committee also discussed that specialist advice should be sought when there is continued uncertainty or for specialist conditions, for example, in circumstances when an individualised assessment is required for multimorbidity. Colleagues with more experience may include specialist palliative care teams, but these may also include other specialties such as geriatrics, cardiologists or renal physicians. The Committee also felt strongly that reversible conditions should be assessed and noted that some signs and symptoms of improvement may be temporary. This links in to considering the whole disease trajectory and ensuring that there is recognition of recovery as well as when the person may die. From the qualitative review the Committee noted the theme of overestimation of a prognosis by consultants with long-term relationships with people. This is due to consultants not wanting to disrupt their relationship with the person, which may happen as a result of the bad news. They also noted the other extreme, where doctors who have never seen the person before are less concerned about informing the person of a poor prognosis or diagnosis. The Committee discussed the importance of monitoring for further changes in the person at least every 24 hours, but that more frequent monitoring may be required as symptoms can change quickly. The attitude of the person was recognised as a very important determinant; especially if they have decided themselves that the time is right for them to die. For example, reversible factors may have been identified, but the person may not want interventions to treat them. An important part of decision making was identified to ensure that the person is asked what they wish and how long they may wish to continue treatment for. The Committee discussed the importance of good communication and shared decision making as being critical components of care (see Chapters 1 and 7). The Committee agreed that it is important that the likelihood that a person is entering the last few days of life is clearly communicated to all concerned including the person (if appropriate), the family and others important to them, as well as to other professionals involved in delivering care. They noted that not all people in the dying phase wished to be informed of their prognosis and, as such, chose to make this point specifically in their recommendations. The uncertainty around recognising the dying phase often lies uncomfortably with many healthcare professionals and the Committee noted that this may lead to poor communication and avoidance of frank discussions with the dying person and others. This approach in turn may give rise to delayed or inappropriate clinical decision-making and cause unnecessary distress. The Committee noted the importance of updating the care plan with any decisions regarding recognising dying. This is of paramount importance to alerting colleagues to the person's deteriorating condition, or possible recovery, so consistent care is given from all involved, preventing unnecessary distress to the dying person in their last days of life. The Committee agreed that managing uncertainty around recognising dying remained a challenge in practice beyond the use of any clinical judgement. The review of the evidence identified potential predictive signs and symptoms for recognising death, but uncertainty still remains. The Committee were interested in the role of the multiprofessional team and how they may be able to manage this uncertainty to reduce its impact on clinical care, shared decision making and communication, and therefore chose to make a research recommendation. The Committee made a separate recommendation around seeking advice from colleagues with more experience of providing end of life care and agreed this may include specialist palliative care teams or other relevant specialties whose input would reduce the uncertainty in recognising dying. |
5.9. Research recommendation
- Question: What can multiprofessional teams do to reduce the impact of uncertainty of recognising when a person is entering the last days of life on clinical care, shared decision-making and communication with the dying person and those important to them?
- Why this is importantIt may be difficult to determine when the dying person is entering the last few days or weeks of life. Predicting the end of life is often inaccurate, and current prognostic tools and models are limited. Some level of uncertainty in recognising when a person is entering the last days of life is likely and is often a challenge to planning care. However, it is crucial to minimise this uncertainty to ensure that it does not prevent key discussions between the healthcare professional and the dying person and those important to them.It is therefore important to identify how the uncertainty of recognising when a person is entering the last days of life influences information sharing, advanced care planning and the behaviour of healthcare professionals. A mixed-methods approach (quantitative and qualitative evidence) is proposed that aims to explore how different multidisciplinary team interventions can reduce the impact of uncertainty on clinical care, shared decision-making and communication, specifically on engaging the dying person and those important to them in end of life care discussions. Multidisciplinary team interventions include any different methods of giving feedback, initiating end of life discussions, record keeping or updating care plans, compared with usual care. Outcomes of interest include quality of life, patient or carer satisfaction, changes to clinical care and identification and/or achievement of patient wishes such as preferred place of death. In addition the barriers and facilitators for the healthcare professionals to manage this uncertainty to best support the dying person and those important to them should be explored.
- Introduction
- Review question: What signs and symptoms indicate that adults are likely to be entering their final days of life; or that they may be recovering? How are uncertainties about either situation dealt with?
- Quantitative review: clinical evidence
- Qualitative review: clinical evidence
- Economic evidence
- Evidence statements
- Conceptual framework
- Recommendations and link to evidence
- Research recommendation
- Recognising when a person may be in the last days of life - Care of Dying Adults...Recognising when a person may be in the last days of life - Care of Dying Adults in the Last Days of Life
- Bacterium K-4b6 16S ribosomal RNA gene, partial sequenceBacterium K-4b6 16S ribosomal RNA gene, partial sequencegi|27464308|gb|AF524858.1|Nucleotide
Your browsing activity is empty.
Activity recording is turned off.
See more...