Studying patients during the end of life is important, as it has the potential to lead to improvements in care for the dying. For patients who die after a controlled withdrawal of life-sustaining therapies in the intensive care unit, information about the natural history of death and the process of removing life support has additionally led to advances in practice for deceased organ donation. However, this unique population of severely critically ill and imminently dying patients has been difficult to study, largely due to assumptions made by research teams and ethics boards alike about the logistical difficulties of obtaining consent and completing research procedures before or during the process of withdrawal of life-sustaining therapies. In this paper, we describe the ethics substudy of the first prospective observational research study in Canada to obtain consent and collect clinical data on patients during the process of withdrawal of life-sustaining therapies in the intensive care unit. We describe in detail the process of protocol development, review by five institutional research ethics boards and bedside staff satisfaction with the study. We conclude that prospective research on a critically ill and imminently dying population is feasible and can be conducted in an ethical manner. Further information is needed about the experiences and motivations of families and substitute decision makers who provide consent for research on critically ill intensive care unit patients at the end of life.
- Research Ethics
- Third Party Consent/Incompetents
- Care of the Dying Patient
- Research on Special Populations
Statistics from Altmetric.com
- Research Ethics
- Third Party Consent/Incompetents
- Care of the Dying Patient
- Research on Special Populations
Patients who are imminently dying or have recently passed away in the intensive care unit provide valuable information about the dying process that could be used to help improve our understanding of what happens at the time of death. Information gleaned from this time period has been invaluable in informing the practice of palliative care within intensive care medicine,1 and has also helped to improve the practice of deceased organ donation.2
Critically ill, dying patients in the intensive care unit represent a vulnerable research population; they are often intubated and heavily sedated, and are therefore unable to make treatment decisions or provide informed consent for research participation on their own behalf.3 Involvement of these patients in prospective research therefore requires third-party consent from substitute decision makers, usually family members, who may struggle to balance personal and family needs with the wishes of the patient.4 In addition, the intensive care unit is a complex and stressful environment which creates further challenges for families, healthcare teams and research coordinators when conducting prospective research.5 Combined with a ‘visceral reaction’6 to the idea of conducting research on dying patients, the above-mentioned reasons present a challenging set of obstacles to overcome when proposing prospective research on the imminently dying and recently dead.
Concerns regarding consent and appropriate ethical oversight of prospective research on imminently dying and recently dead patients have been recognised as significant obstacles to rigorous investigation of effective organ donor management and organ donation protocols.3 ,5 The Determination of Death Practices in Intensive Care Units (DDePICt) pilot study was the first prospective observational research study in Canadian intensive care units that enrolled imminently dying and recently dead patients. This paper describes the development of the ethical approach for the study, the process of review by four institutional research ethics boards across Canada and staff satisfaction with the study.
The DDePICt pilot study collected information about the feasibility of conducting a prospective observational study in a population of critically ill and imminently dying patients in the intensive care unit. Feasibility measures included collection of information on the process of research ethics board review and approvals, family consent rates and staff satisfaction with the study procedures. The methods and major results of the DDePICt pilot study have been previously described in detail.7
The ethics substudy component of the DDePICt pilot study was conducted by a core group of four people: the pilot study principal investigator, a paediatric intensive care physician (MD, FRCPC), the project coordinator (MSc), the central clinical coordinator (RN, MSc) and the research administrative assistant (BA). The larger ethics substudy group included a Canada Research Chair in Health Law and Policy (LLM), an assistant professor in nursing bioethics (N MSc), a senior Canadian Institutes of Health Research (CIHR) ethics policy advisor (LLM) and a medical lead for organ donation for Canadian Blood Services (MD, FRCPC). The group met in person twice and by teleconference twice per month throughout the study planning and implementation phases.
This ethics substudy describes the processes implemented and the results of these processes, to facilitate the conduct of research at the end of life in critical care.
The DDePICt pilot study involved approaching the families of imminently dying patients for consent to prospective research. Recognising the potential for ethical and emotional challenges within this scenario, study investigators initiated consultations with experts in medical ethics and health law to identify and address potential ethical issues prior to finalisation of the protocol. All study investigators and collaborators convened at a planning meeting, which was followed up with email correspondence and conference calls. During consultations, the study team identified three specific ethical issues: (1) research in a vulnerable population; (2) consent for research in an emotionally charged context; and (3) the potential for disruption of key relationships between intensive care professionals, patients and families. The final version of the protocol included detailed instructions to study staff regarding study procedures that intended to address all three of these concerns within the study procedures.
The study team also anticipated concern from ethics boards regarding the consent document. The original study consent form was drafted with input from ethics and legal consultants to ensure clear, uncomplicated wording regarding study participation. Particular thought was given to the degree of information presented to families—the study team felt that too much information about the technical and quantitative nature of the research could be offensive to families, while too little information on the study purpose could deprive the patient's family of feeling that they had made a significant contribution towards medical research.8 The consent document was presented to patient advisors on the research team to ensure readability and comprehension before completion of the final draft.
Finally, as part of the study start-up, the core research team visited each site that was to enrol patients into the DDePICt study. During the site visits, the principal investigator presented the study design, and the rationale for studying critically ill patients after withdrawal of life-sustaining therapies. Staff at some sites expressed concern regarding study procedures, particularly the requirement to have patients monitored for 30 min following the declaration of death when usual practice would be to remove all physiological monitoring. The site visit gave intensive care unit staff the opportunity to voice their opinions and ask questions. After their concerns had been addressed, staff expressed a more positive attitude and understanding towards the study.
Ethics board approval process
The DDePICt pilot study was submitted sequentially for ethical review at hospital ethics boards at all five study sites, beginning with the central site. All five sites approved the study protocol within 3 months of their original submissions. No site required more than one set of revisions to the original protocol and consent form for ethics approval. A single, common protocol was used across sites, with individual centres creating site-specific study procedures to guide enrolment and data collection as required.
Ethics boards at study sites had an average of 1–2 questions or concerns regarding the pilot study (table 1).
Three out of five site ethics boards requested that the consent form emphasise that standard patient care, including the family's ability to interact with dying patients, would not be influenced by study participation (table 1). Two out of five sites suggested that descriptions of the purpose of the study (to determine the feasibility of collecting information about the declaration of death and the physiology of death during and after the dying process) should be removed to avoid confusion or distress. Four out of five ethics boards felt that the language used in the consent form to describe death and the dying process was inappropriate in some way, with comments suggesting terminology was ‘too harsh’ (study site 5), ‘technical or dismissive’ (study site 2).
Staff satisfaction with study procedures
A 14-question staff satisfaction questionnaire, developed by the study team, was administered to all members of the bedside clinical care team following completion of study procedures. A copy of the questionnaire was published in Appendix I of the pilot study publication.6 The questionnaire intended to capture bedside staff perceptions of how the study affected families and whether the study procedures interfered with quality end-of-life care. The study team developed the questionnaire using input and review from critical care physicians and nurses. The survey went through several drafts before being finalised but was not piloted on healthcare professionals prior to study initiation.
A total of 76 questionnaires were completed for 37 patients, with a range of 0–3 questionnaires completed by bedside staff for each patient, and a range of 7–26 questionnaires completed at each site (table 2). Bedside staff did not return any completed questionnaires for four enrolled patients. Not all questionnaires were fully completed, resulting in a denominator of <76 for some domains (median 74, 97% complete). The majority of questionnaires were completed by nursing staff (52%) followed by physicians (24%).
The majority of staff who completed the questionnaire felt comfortable caring for a patient who was in the study (67/74, 90.5%), and most had no problem leaving the monitors in place for 30 min after the declaration of death (59/74, 79.7%). The vast majority of staff also felt that the study did not have a negative influence on patient care (66/75, 88.0%), and felt that carrying out the study procedures was not time consuming (65/75, 86.7%). Finally, 68.9% (51/74) of staff felt that the study was important to conduct, and 52.9% (38/72) felt that the study was acceptable to families.
A minority of staff indicated some discomfort with discussing study procedures with the patient's family (6/76, 8.2%). Approximately one third of staff members who responded to the survey indicated that discussions with families about study procedures were not applicable to their experience (25/76, 34.2%). Comments were provided on 17 (22.3%) of the questionnaires. Staff comments generally fell into one of three categories: suggestions for improvements of study procedures, concerns with impacts on patient and family care, or support and general interest in the study (table 3).
No family members or substitute decision makers spontaneously approached the research team to express discomfort with the study procedures during or after the study. No substitute decision makers requested the withdrawal of any patient's participation or removal of their data from the study. However, substitute decision makers and families were not approached or followed up to ask about their experiences of consenting to research on behalf of the patient, or their experiences of the study procedures.
The DDePICt pilot study was the first prospective, quantitative observational study conducted in a critically ill, imminently dying and recently dead population in intensive care units in Canada. On the basis of enrolment rate and the consent rate of 87%, we demonstrated that recruitment into a prospective study during the end of life in the intensive care unit is feasible.7 The involvement and ongoing collaboration of an interprofessional team to develop study design, data collection materials and consent forms was an integral component of this study's success. We obtained strong buy-in for the study protocol from site staff, universal approval by hospital ethics boards and had a high rate of intensive care unit staff satisfaction despite the study's focus on a potentially sensitive topic.
Initial concerns expressed by researchers regarding the perceived willingness of ethics boards to approve this project were not supported by this study. The responses and review timelines of the ethics boards at the five study sites were consistent with expected timelines for critical care research.9 One ethics board did raise concerns about the person obtaining consent, and recommended a deviation from common ethical guidelines stating that treating physicians should not be involved in the process of consent,10 instead arguing that the physician's close relationship with the family would make the consent process more acceptable in this unique situation. This is in contrast to literature suggesting that families do not feel that physician involvement in the informed consent process is useful11 or that physician approach for informed consent may be perceived as coercion.12
One ethics board expressed concern that patient participation in the study may increase the probability of the physician or the patient's family witnessing an autoresuscitation event (spontaneous return of circulation). The board was concerned that this physiological response might be observed on bedside monitors that would have been turned off and removed if the patient was not participating in the study. However, the pilot study protocol was strictly observational, and was concerned only with observing events that may already be naturally occurring in an intensive care unit environment. The study did not require any changes to how withdrawal of life-sustaining therapies was conducted and allowed monitors to be turned to privacy mode (in which the monitor remains on but no physiological waveforms can be seen) if this was usual practice. The bedside care team and patient's family did not have access to the physiological waveform monitoring unless this was already part of usual practice at the hospital. The probability of an autoresuscitation event occurring or being witnessed by practitioners or patients' families was therefore not impacted by the study.
In general, ethics boards' approval of the study protocol seemed to indicate an overall acceptance of this minimal risk observational study enrolling imminently dying patients. However, the requested deviations from common ethical principles and the focus on risks of harm and use of language already common in standard clinical care appeared to hint at a desire for additional protection for imminently dying patients and their families. Similar accounts of paternalistic ‘gatekeeping’ for protocols and projects involving patients at the end of life have also been described in the literature.13–15 Further investigations into the perspectives and experiences of patients and families with end-of-life research in the intensive care unit are warranted to determine whether the populations in question desire or value these additional ethical safeguards.
The study team's experience with high patient enrolment appeared to indicate that study procedures involving observational data collection on dying and recently dead patients were acceptable to the families of critically ill, dying patients. The consent rate of 87% exceeded that expected for a critically ill population in the intensive care unit.5 One comment from a staff satisfaction questionnaire suggests that the higher consent rate may have been due to a need for families to feel that their loved ones' death had not been in vain. The comment on the questionnaire suggested that one family was grateful for a chance to participate in this study ‘…in the absence of being able to donate organs’ (table 3). Published accounts of research in patients with terminal cancer have likewise documented a positive response from patients and families to research participation at the end of life.15 ,16
Bedside staff appeared to be generally satisfied with the study and appeared to accept and cooperate with the required data collection procedures. Comments from the survey indicated that staff were interested in the study outcome and were pleased to be involved in this novel area of research. Negative feedback was related to the flow of information surrounding the study procedures and engagement with the family, rather than concerns with the study purpose or project design.
Prospective research on a critically ill, dying population is novel, and we have a lack of insight on which risks to patients and families are most important to consider when developing and reviewing these protocols. Guidelines for the enrolment of recently dead patients in research state that although research on dead patients does not warrant the same degree of ethical oversight required for living (and thus potentially seriously harmed) patients, it nonetheless requires continued respect for the families, estates and bodies of the dead.17 ,18 Ethical principles required for research enrolment of patients who are imminently dying, but not yet declared dead, are more unclear.19 Some have suggested that research involving dying patients should be held to the same ethical review standards as those used for living patients.20 Others have stated that dying patients are such a vulnerable and special population that they should only be involved in studies where at least some degree of benefit is expected to the patient8 or not be included in research at all.21
The success of the DDePICt pilot study has demonstrated that prospective observational research in a population of critically ill, imminently dying and recently dead patients can be conducted in an ethical manner with review and oversight from existing research ethics boards. The success of this study indicates that prospective observational research at the end of life in the intensive care unit is feasible and can be conducted with prior consent from patients' substitute decision makers. The inclusion of critically ill, imminently dying and recently dead populations in prospective research is feasible and necessary to improve evidence-based care of dying patients. Collection of prospective data from this population has the potential to yield information that will be invaluable for the evolution of safe and successful organ donation protocols, and that will help to advance the science of evidence-based palliative and end-of-life care in the intensive care unit.
The DDePICt pilot ethics substudy had several limitations. The first was a lack of feedback from substitute decision makers who provided and those that refused consent for participation. However, it was considered unethical by research ethics boards to ask families who did not agree to participate for their rationale. Second, we did not have a total denominator for the staff satisfaction survey, and therefore were unable to comment on whether staff who did not return a questionnaire had a different experience or perspective of the study protocol than those who returned it. Finally, we did not survey research ethics board chairs to determine the reasoning behind their decisions and therefore did not fully appreciate any potential concerns they may have had. This would be an interesting area for future research.
The inclusion of critically ill, imminently dying and recently dead populations in prospective research is feasible and necessary to improve evidence-based care of dying patients, including potential organ donors. Responses from institutional ethics boards, consent rates from patients' families and comments from healthcare providers indicate that this research is acceptable, feasible and has minimal impact to routine end-of-life care in the intensive care unit.
Contributors AvB, LH and SD: participated in the conception, design, analysis, interpretation, drafting and the critical revising for important intellectual content of this manuscript and have seen and approved the final version. RW and JC-E: participated in the conception, design, analysis, interpretation and the critical revising for important intellectual content of this manuscript and have seen and approved the final version. KM: participated in the drafting and the critical revising for important intellectual content of this manuscript and has seen and approved the final version.
Competing interests None declared.
Provenance and peer review Not commissioned; externally peer reviewed.
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.