Comfort in palliative sedation (Compas): a transdisciplinary mixed method study protocol for linking objective assessments to subjective experiences

The gold standard for detecting distress is patient self-reporting and several instruments, such as the Visual Analog Scale (VAS) or the Numeric Rating Scale (NRS) for pain, are based on this [6]. However, in the case of CDS, patients are unable to communicate whether or not they are still in distress or still (partially) aware of what is going on with them. Some scales have been developed for non-communicative patients as well, but several problems with those have been reported. One well-documented problem is that such scales seem unable to detect pain and awareness in all patients, e.g. because they depend on inferences made from patients’ motor responsiveness, which in case of CDS are nonexistent [7]. Another problem is that these scales are only partially, and in most cases not at all, validated for dying patients [8‐10]. In the guidelines for palliative sedation it is acknowledged that the efficacy and safety of palliative sedation are not sufficiently understood and therefore it has been concluded that “there is no scale available to assess the patient’s comfort during sedation”. These findings cause an even bigger concern considering the evidence that family members of patients, compared with caregivers, often have different perceptions of a patient’s comfort and their quality of dying during CDS. While family members tend to overestimate pain, caregivers often seem to rather underestimate it [11]. Furthermore, interrater reliability among nurses and physicians often seems poor [12].

In recent years, doubts have risen on whether patients labeled ‘unconscious’ really are completely insensate and unaware. Studies in different types of patients and settings that critically reviewed awareness, consistently reported that persons were, in contrast to what was assumed by the caregivers, not always (completely) unaware. For example, several studies have shown that patients diagnosed with a vegetative state (currently also called ‘unresponsive wakefulness syndrome’) did show some (minimal) clinical signs of conscious awareness in about 40% of the cases [13]. In some cases, a purportedly unconscious patient could even reliably generate appropriate EEG responses to two distinct commands and occasionally was even able to establish basic communication with ‘yes’ or ‘no’ answers using functional Magnetic Resonance Imaging (fMRI) [14]. These findings demonstrate that at least a minority of clinically diagnosed unresponsive patients shows some residual cognitive function and conscious awareness and that even skilled caregivers were not able to recognize [13]. Also, patients with locked-in syndrome may be mistakenly considered unconscious as may some (rare) patients during general anesthesia [15]. In contrast to the setting where surgical or intensive care patients are managed, advanced monitoring equipment is usually lacking in palliative or home care settings. Terminal or palliative sedated patients ultimately die and therefore patient self-reporting is missing.

The above findings show that the ‘traditional’ clinical tools and procedures to assess comfort and awareness in dying non-communicative patients have important methodological limitations. It should be noted that the problems with assessments are not to be ascribed to a lack of competence in the caregivers, but are of a much more fundamental nature: the absence of reliable tools. The developers of guidelines are aware of these limitations. Although guidelines refer to some scales, the recent Belgian guideline rightly stresses that “their utility in palliative care is not proven” [16]. The current guidelines for CDS are limited to suggest “a daily visit of the physician” and to “continue attention for possible expressions of discomfort (e.g. facial expressions, movements…)” [3, 16].

Since the aim of CDS is to give optimal comfort, but not to hasten death, the principle of ‘proportionality’ is a pivotal aspect of this treatment and hence the guidelines state that sedation should be ‘no deeper than necessary to avoid suffering’ [1, 16]. To meet this principle of proportionality, caregivers administer the doses of the drugs so that they are high enough to provide comfort but should not hasten death. Studies have shown that CDS does not usually affect survival time [17]. However, as palliative sedation is considered ‘slow euthanasia’ by some, physicians may be ‘extra careful’ with the use of high doses of sedative medication. Several studies have reported underuse of medicines due to a lack of knowledge, unwarranted beliefs, to avoid the perception of giving ‘excessive’ doses and even because of fear among caregivers of being accused of ‘killing’ the patient [18].

In modal hospitals and modal homes for the elderly, forty patients will be selected. In each setting 20 participants will be included consecutively, which should result in 40 detailed case studies. Patients will be deliberately selected to gain maximum insight and reflect variability regarding medical conditions.

For this prospective study methods from social sciences (qualitative data collection based on interviews, observation), quantitative assessments (scales to assess awareness and discomfort) and neurophysiological data will be combined. Hence, a combination of methods that can be considered fairly objective (e.g. EEG and physiological parameters) and rather subjective methods (e.g. observations by family members) will be used to assess discomfort and awareness. Therefore, the following data will be collected:

3. Assessment by 2 trained investigators using 1 scale that is mentioned in the Flemish palliative sedation guideline and 3 other scales that have been proposed in the literature [16]:

4. Neurophysiological assessments of distress and level of awareness using fMRI has been successfully used to detect awareness and distress that could not be detected by thorough behavioral assessment [10, 13, 14, 24, 25]. Issues of availability and extra stress incurred by patients when transferred are making the use of fMRI for practical and ethical reasons not feasible.

EEG has recently been shown to provide a feasible and methodologically sound alternative for bedside detection of awareness and pain [26‐28]. Although results from these methods need to be interpreted with caution, they can provide important indicators that are especially valuable for assessments in non-communicative patients. Data will be collected with a Neurosense monitor that can easily be used and set up. It displays two frontal EEG signals, and calculates a number of parameters including the bilateral WAVcns index (Wavelet Anesthetic Value for Central Nervous System), ranging from 100 (awake) to 0 (flat EEG). The lower the index, the lower the likelihood of consciousness. The recommended WAVcns for general anesthesia with low probability of consciousness is between 60 and 40. For technical details on the respective measurements of awareness and distress we refer to Schulz E et al. and Cruse D et al. [24, 29].

Additionally, heart rate variability (HRV) will be measured using an ANI monitor. HRV is the variation in time intervals between heartbeats and reflects the effect of the vagus nerve on the heart, which is inhibited during pain. It has already been widely studied in relationship to pain, and seems to be a promising biomarker [30]. However, studies with palliative sedated patients show contradicting results thus far [31, 32]. A promising non-invasive technique is the continuous monitoring of HRV transformed into an analgesia nociception index (ANI, 0–100), which assesses parasympathetic activity as a possible measure of nociception [33]. The ANI provides greater stability than raw indices of HRV. A recent study showed that ANI is effective in detecting pain in deeply sedated critically ill patients [34]. The analgesia nociception index is a non-invasive tool based on the analysis of the respiratory fluctuations of heart rate that mainly reflect the variability in the parasympathetic tone. The NeuroSense and ANI monitor record the EEG and ECG signal continuously, enabling a quantitative assessment of these parameters. The data can be used to assess the level of sedation but also to register negative and positive emotion-related activities such as the voice of loved ones and distress.

6. Background information based on the patient’s medical file, observation and interviews.

Data will be collected every day, starting from the day when palliative sedation is initiated until the patient dies. Neurophysiological data will be collected continuously. Video and audio registrations will preferably take place non-stop (in consultation with the family) but may be interrupted during family visits to respect privacy.

A previous study learns that a correlation of 0.45 was found between Bispectral Index (BIS) monitoring and the Ramsay scale [35]. A similar correlation can be expected between the WAVcns and the VAS in the present study. However, in order to become a successful measure of sedation, we require the correlation between both measures to be significantly higher than 0.30. Although EEG monitoring would allow continuous data collection, the VAS will be completed only once or twice a day by relatives (more often by trained investigators) to keep burden as low as possible. As we expect patients can be monitored during 2–3 days on average, some 5 VAS scores (from relatives, more from staff) will be available per patient. This results in a close to 80% probability for our study to find a correlation higher than 0.30 if the actual correlation is 0.45.

Analysis of qualitative data (interviews and video) will be done in accordance with the interpretative paradigm, aiming to understand meanings and actions and how people construct them, and follow the principles of Grounded Theory [36]. Coding of the neurological data will be performed by an experienced neurologist.

Video registrations will be coded in events and analysed using nVivo, a tool for qualitative analysis that enables researchers to code and analyze events and also allows importing and correlating video and neurophysiological data. A coding scheme will be developed to code behavioral signs of the patient, treatments (painful acts, medical acts, etc.) and environmental circumstances and changes (visits, noises…).

The involved researchers will intensively cooperate in the interpretation of the data. All data will be used for an overall assessment (using the transdisciplinary mixed methods case study approach) [37]. The analysis will focus on:

The findings and interpretations will then be discussed in a multidisciplinary team consisting of members of the Mental Health and Wellbeing Research Group, the department of Anesthesiology and Perioperative Medicine, the department of Experimental Psychology (all from the Vrije Universiteit Brussel) and the Coma Science Group (University of Liège).

Video/audio registration will only be made after explicit consent of the patient (if possible) and his next-of-kin.

The study protocol is approved by the biomedical ethics committee of the VUB/UZ Brussel (BUN 14320136504) and additional approval will be asked from participating hospitals. Written informed consent will be asked from the patient or his/her substitute decision maker.

We are aware that the data collection is challenging and delicate, especially because we will deal with dying patients and their family. Therefore, special attention will be given to clear communication about the aims of our study. We will explain participants that the study is not invasive and will not hinder the (sedated) patient and that the aim is to minimize the risk for suffering.

To our knowledge this is one of the very few studies which seek to take an in-depth view on the assessment of pain and discomfort in non-communicative palliative sedated patients. To this effect multiple ways of data collection are used, in particular both objective and subjective assessments, with the intent to develop as complete a picture as possible of the subject at hand. By using a transdisciplinary approach throughout the entirety of the study, a high degree of integration is achieved; researchers from different disciplines were involved in all aspects of the study, as were members of the palliative community, patients and family members to decide what is ethically and practically acceptable.

There are a number of limitations to this study, partly because doing research at the end of life is delicate, and partly because of the methodological challenges in dealing with non-communicative unconscious patients. It is likely that more extensive measures of brain activity such as fMRI would yield more sophisticated data, but, more extensive neurophysiological monitoring would be more invasive as compared to a NeuroSense monitor and this would likely be more disturbing for family members.

From a methodological point of view, the greatest challenge will be how to combine the different types of data (neurophysiological, quantitative and qualitative) to assess a phenomenon which is considered to be best assessed by self-reporting. Since the latter is not an option, the best possible way to approximate objectivity is by triangulation, combining different measures that are collected at the same moment and analyzed transversally. However gathering different types of data (e.g. assessments of family members) may prove difficult in such a sensitive setting.

Completion of this project will result in a deeper understanding of how the assessment of (dis)comfort and pain during palliative sedation may be potentially improved by including more objective neurophysiological parameters. Additionally, our findings will clarify if the present practices associated with assessing patient comfort during palliative sedation (based on observational scales) need to be reconsidered. The transdisciplinary approach of the researchers involved in the COMPAS study further ensures a maximum level of ecological validity when studying such complex phenomenon.

The threats to the successful conduct of this study relate mainly to the reluctance of the patient or his substitute decision maker to be bothered with participating in a study at such a sensitive and difficult moment. After all, one of the ideas behind the palliative care concept is to give comfort to the patient and that includes an environment that feels more natural and home-like, so the concept of using monitoring devices in such a setting seems contra-indicated. This concern has been voiced by palliative caregivers. However, we feel confident that with proper and sensible explanation the importance of this study can be made understood. Furthermore every aspect of the study has been designed to be as minimally burdensome as possible. Already a pilot case has demonstrated the feasibility of this design for all actors involved, the protocol for this observational study has been registered retrospectively at ClinicalTrials.​gov (ID NCT03273244) [38].