Critical care beds are usually in

Priority setting in a hospital critical care unit: Qualitative case study* Jens Mielke, DA, MHSC, MRCP; Douglas K. Martin, PhD; Peter A. Singer, MD, MPH, FRCPC Objective: To describe priority setting for admissions in a hospital critical care unit and to evaluate it using the ethical framework of accountability for reasonableness. Design: Qualitative case study and evaluation using the ethical framework of accountability for reasonableness. Setting: A medical/surgical intensive care unit in a large urban university-affiliated teaching hospital in Toronto, Canada. Participants: Critical care unit staff including medical directors, nurses, residents, referring physicians, and members of a hospital committee that formulated an admissions policy. Interventions: Modified thematic analysis of documents, interviews with participants, and direct observation of critical care unit rounds. Evaluation using the four conditions of Daniels and Sabin s accountability for reasonableness: relevance, publicity, appeals/revisions, and enforcement. Measurements and Main Results: We examined key features and participants views about the priority setting process. Decisions to admit patients involve a complex cluster of reasons. Both medical and nonmedical reasons are used, although the nonmedical reasons are less well documented and understood. Medical directors, who are the chief decision makers, differ in their reasoning. Admitting decisions and reasons are usually explained to referring staff but seldom to patients and families, and nonmedical reasons are seldom surfaced. A hospital critical care admissions policy exists but is not used and is not known to all stakeholders. A formal appeals/revisions process exists, but appeals usually involve informal negotiations. The existence of priority programs in the hospital (e.g., transplantation) adds complexity and heightens disagreement by stakeholders. Conclusion: We have described and evaluated admissions decision making in a hospital s critical care unit. The key lesson of our study is not only the specific findings obtained here but also how combining a case study approach with the ethical framework of accountability for reasonableness can be used to identify good practices and opportunities for improving the fairness of priority setting in intensive care. (Crit Care Med 2003; 31:2764 2768) KEY WORDS: priority setting; admissions; critical care Critical care beds are usually in greater demand than can be met sometimes leading to tragic consequences (1, 2). Prioritizing which patients should be admitted is a serious and ongoing challenge for every critical care unit. Moreover, priority setting in critical care is important because intensive care is expensive (3). A few studies have quantitatively evaluated issues of critical care access and the morbidity and mortality consequences of intensive care unit (ICU) bed shortages (4 *See also p. 2809. University of Zimbabwe Medical School (JM); Collaborative Program in Bioethics (DKM) and Department of Health Policy, Management and Evaluation and the Joint Centre for Bioethics, University of Toronto; Joint Centre for Bioethics (PAS), University of Toronto. Supported, in part, by grant 6606-06-1999/ 2590074 from the Canadian Institute of Health Research, by an Ontario Ministry of Health and Long- Term Care Career Scientist award (DKM), and by a Canadian Institutes of Health Research Investigator award (PAS). Copyright 2003 by Lippincott Williams & Wilkins DOI: 10.1097/01.CCM.0000098440.74735.DE 6), including effects of discharges at night (7). The American Thoracic Society s Bioethics Task Force developed a statement Fair Allocation of Intensive Care Unit Resources (8), and the Society of Critical Care Medicine published guidelines for ICU admission, discharge, and triage (9). Both guidelines describe substantive criteria (e.g., potential benefit) that might be used in setting admissions priorities. But how are priority setting decisions in an ICU actually made? Zussman (10) conducted an in-depth qualitative study of two critical care units over 5 yrs and concluded that, It is all very well and good to develop [priority setting criteria]. But such criteria matter not at all if they are ignored, for what is left out of the predictive models as well as of the ethical reflections on triage is any sense of the socially structured pressures operating on physicians,...the social structures that generate advocacy and disinterest, that generate indifference to some patients and commitments to others. Strosberg and Teres (11) explored gatekeeping in critical care units using illustrative case studies. They describe most gatekeeping decisions in critical care units as being ad hoc and political and not conforming to clear rules. These studies are helpful but limited because the authors did not use an explicit ethical framework to evaluate actual priority setting in critical care. Daniels and Sabin (12) developed an ethical framework for priority setting called accountability for reasonableness. Because it is very difficult for decision makers to reach agreement on substantive criteria for fair priority setting, accountability for reasonableness is an explicit framework for determining what is a fair priority-setting process. Developed in the context of private healthcare organizations in the United States (13), it is also relevant in public healthcare organizations in Canada and elsewhere (14, 15). Accountability for reasonableness has been used to evaluate priority setting at the level of health systems; for example, Ham (16, 17) examined contested decisions in the UK National Health Service, and Martin and 2764 Crit Care Med 2003 Vol. 31, No. 12

Singer (18) examined priority setting for health technologies in Canada. However, to our knowledge, this promising framework has not yet been applied to admission decisions in critical care. The purpose of this study is to describe priority setting for admissions in a hospital critical care unit and evaluate it using accountability for reasonableness. METHODS Design This was a qualitative case study. A case study is an empirical inquiry that investigates a contemporary phenomenon within its reallife context (19). This method is appropriate for this research because priority setting in critical care units is complex and contextdependent and involves social processes. Setting The setting for this study was the medical and surgical ICU (MSICU) of a large urban university-affiliated hospital. It is a 16-bed closed unit, staffed by critical care specialists. Other critical care units in the hospital (cardiovascular and coronary care units) were not examined. In 1998, hospital management recognized that the institution could no longer admit all patients who presented for admission to the MSICU, and so MSICU admission guidelines were developed. These guidelines included the creation of priority levels favoring in-hospital patients and those in priority programs (Table 1). Sampling The study was carried out between November 2001 and March 2002. Theoretical sampling was used. This technique is commonly used in qualitative research and uses prior knowledge of the setting to focus on those documents, individuals, and observational settings that may provide information relevant to the emerging findings. Data Collection Three primary data sources were used. The first source was key documents (e.g., the hospital guidelines for admission to the MSICU, minutes of the guideline committee meetings). The second source was semistructured interviews with 20 key informants, using an interview guide based on previous prioritysetting case studies but modified according to emerging findings (available on request); all interviews were audiotaped and transcribed (Table 2 describes the interview participants). The third source was direct observation of morning rounds and other discussions in the MSICU by one of the investigators (JM) a total of 18 hrs. Data Analysis Although this was not a grounded theory study, we adapted grounded theory analytic techniques because they are specifically designed for analyzing complex social processes (such as priority-setting processes) (20, 21). First, using modified open coding, we examined the data and identified components of the priority-setting process (e.g., medical criteria or negotiation ). Then, using modified axial coding, we organized these components under overarching themes. Because this study was guided by an explicit conceptual framework, the themes were the four conditions of accountability for reasonableness (described subsequently). We addressed the validity of our findings in five ways (22). First, we triangulated data from three different sources (documents, interviews, and observations) to maximize comprehensiveness and diversity (23). Second, two primary researchers coded the raw data and agreed on the coding list. Third, members of an independent interdisciplinary research group, consisting of a philosopher, nurse, hospital administrator, and bioethicist, enhanced the reflexivity in the analysis by participating in the data analysis. Thus, the role of prior assumptions and experience, which can influence any inquiry, were acknowledged and examined. Fourth, all research activities were rigorously documented to permit a critical appraisal of the methods (24). Fifth, a draft of the findings was distributed to a subgroup of six participants, and comments were invited as a member check. The participants verified the accuracy of the report and the reasonableness of the findings. The Conceptual Framework: Accountability for Reasonableness Daniels and Sabin (12, 25) developed an ethical framework for legitimate and fair priority setting called accountability for reasonableness. A goal of priority setting is justice. However, because no societal consensus exists regarding substantive principles of justice, a key goal is procedural justice that is, a legitimate and fair process (26, 27). According to accountability for reasonableness, an institution s priority-setting decisions may be considered legitimate and fair if they satisfy four conditions: relevance, publicity, appeals/ revisions, and enforcement described in Table 3. Research Ethics Approval for this study was obtained from the Committee on Use of Human Subjects of the University of Toronto and the hospital s research ethics board. Written informed consent was obtained from each individual before being interviewed. All raw data were protected as confidential and were available only to the research team. No individual participants were identified. RESULTS The results have been organized according to the four conditions of accountability for reasonableness (relevance, publicity, revisions, enforcement), and key points are illustrated with verbatim quotes from participants. Relevance Decisions to admit patients to the critical care unit involve a complex cluster of reasons. Multiple reasons, which participants described as medical or nonmedical, combined in particular ways in each particular case. The participants in the decision-making process are the ICU physicians (staff physicians, also called medical directors) and also fellows and residents. A wider range of participants were involved in developing the admissions Table 1. Hospital Guidelines for Admission to the Medical Surgical Intensive Care Unit (Excerpt) Priority 1 (not in order of priority) Priority 2 In-house medical and surgical emergencies Critically ill patients in the emergency departments Patients considered to be in a priority program, specifically neurosciences, transplantation, oncology, cardiovascular diseases, and the patient ambulatory care program Patients referred from another institution who require urgent management only available at, or specifically suited to, the expertise offered at the hospital intensive care units All other patients (the medical director of the unit may decide to upgrade the status of a particular patient) Crit Care Med 2003 Vol. 31, No. 12 2765

policy including three members of the hospital s Community Advisory Committee, physicians, administrators, legal counsel, and bioethicists. Both the ICU physicians and those who refer to them agree that the patients who should be admitted to the critical care unit are those who require the specialized services of the unit and will benefit from them referred to as medical reasons: That s because of respiratory failure and the need for mechanical ventilation. And then I suppose the second most common reason would be, ah, hypotension and the need for intravenous, ah, inotropic support. In addition, nonmedical reasons regarding admissions decisions were cited, including the following: Table 2. Study Participants (n 20) Availability of a bed and nursing staff: If we think that the patient would benefit from coming here the key decision, the key, um, um, factor would be the availability of a bed and a nurse. Predominantly a nurse. Family pressure: I think, she [the patient] came because the seed had been planted with the family. There you have the family, all three of which are crying and saying: Save her! Save her! You know, it would be a lot more difficult to just say: Okay, we ve done all this in the Emergency Room. We re going to just stop now. Hospital policy: The Guidelines for Admission to the MSICU state priorities. However, clinicians understanding and application of these priorities varied, most notably by placing transplant patients above other priority programs contrary to the policy: If there were two patients who needed to come into the ICU and one of them was a transplant patient and the other one was something else, the transplant patient would take priority. However, medical personnel felt some concern about basing priorities on hospital programs: If you have any questions in your mind about that then I think the process is not entirely fair because there is some priority given to the transplant or the neurosurgery case. You know, some of us feel, I think, that there is some unfairness there. The perception held by many non-icu hospital staff is that the ICU physicians use well-defined medical criteria for their decisions: And they have their set of clinical criteria which, you know, is like the law. It s very black and white and it s very easy to deal with. So, you know, if this patient is on a ventilator and, you know, blah, blah, blah...this has got to be a patient who s ready for the ICU and they will admit the patient. And if there s actually something they can do for the patient in the ICU, they will admit the patient. These criteria are hard and Medical Personnel Nursing Personnel Others ICU directors (staff physicians, intensivists), 5 ICU nurse manager, 1 Committee members, 3 ICU fellows, 1 ICU nurse, 2 ICU administrator, 1 ICU residents, 2 Patient relations rep, 1 Referring physicians, 4 Transplant surgeon, 1 Vascular surgeon, 1 Internal medicine, 2 Total, 5 Total, 12 Total, 3 ICU, intensive care unit. they re well-defined and they re based on standard of practice. However, ICU physicians themselves perceive these decisions to be a negotiated process: The admission of patients to the unit is a negotiated process. Most of the time it functions...implicitly. Physicians who make the admission decisions, and others who observe them, agreed that there are differences in the way they are made: I think you could show that case scenario to some intensivists and they would say: Sure, I d bring that patient. Others would say, I think they ll be fine [i.e., do not need admission]. Amongst our directors...you can see very different people who get admitted to the ICU. The primary decision makers, the medical directors, are aware of differences among themselves regarding admission decisions but accept this as inevitable, given differences in experience, attitudes, and familiarity with the hospital. None would say any of the others was wrong in making a decision one way or the other, but they also would not usually consult a colleague in making even quite difficult decisions. The differences are coloured a little bit by your...your level of comfort, um, and your knowledge of the place where you work. Some referring physicians have great concern with the variability of admission decisions: When I phone up and I say I ve got a call from Hospital X that they have a ruptured aneurysm...and they say Well, we do not have a nurse to look after the patient. But if they get a call a half-hour later that there s a liver, they will...they will obviously go to the end of the earth to do the liver. But the aortic aneurysm has less priority, even though you ll die of your aortic aneurysm...our perception is that, be it true or not, that they will go the extra mile for a lung transplant or a liver or a kidney/pancreas. And I can understand that, but it s sort of basically saying, okay, Table 3. Four Conditions of Accountability for Reasonableness Condition Relevance Publicity Appeals/revisions Enforcement Description Priority-setting decisions must rest on reasons (including evidence and principles) that fair-minded participants can agree are relevant to meeting context-specific goals under resource constraints. Fair-minded participants are stakeholders who are disposed to decision making according to rules of mutual cooperation and can involve managers, clinicians, patients, and consumers in general. Priority-setting decisions and their rationales must be publicly accessible. The priority-setting process must include a mechanism for challenge and dispute resolution regarding priority-setting decisions, including the opportunity for revising decisions in light of further evidence or principles. There must be voluntary or public regulation of the process to ensure that the first three conditions are met. 2766 Crit Care Med 2003 Vol. 31, No. 12

well, you re...you re at the bottom of the barrel, so pick another apple. Publicity Regarding individual patients, ICU physicians communicate their decisions and reasons primarily to the referring physicians. Occasionally, but not always, the family is told. The ICU resident would have come down, done the consult and said to the ward team, No. Or they may have said to the family, en passant, Sorry, no, and then disappeared and then the family would have said, Why? In regard to the publicity of the admissions policy, the minutes of the committee formed to create the guidelines portray an inclusive, transparent process. The minutes state an intention to circulate the guidelines widely, including to all doctors and nurses in the hospital. At the time of our study, there was little awareness of written guidelines in the hospital: I do not know if there is a policy. It seems reasonable that there would be, but I have never come (across) it in my one month of ICU. Healthcare staff were unanimous in suggesting that public education about intensive care would assist the admissions process: The average Ontarian has absolutely no clue how the system works. They think they do but they have absolutely no clue. Intelligent, empowered, well-placed people come into the hospital expecting...what they see on television, perhaps, you know. And they are stunned that it doesn t work that way. Appeals/Revisions The Guidelines for Admission to the MSICU state: Any patient or relative who is denied admission to the Unit, or any physician, may challenge the decision. In this circumstance, the individual is entitled to a full explanation of the guidelines and their implementation by a representative of Patient Relations. However, formal appeals are rare. Differences of opinion between hospital staff are negotiated informally: We could in theory get into some conflict where the referring physician would, you know, really feel strongly. There is a mechanism of appeal available for that. There is a second intensive care person who could be called in and review the case and help make that decision. Differences between hospital staff and patients or families usually are addressed informally. Occasionally, the Patient Relations Department intervenes using techniques of conflict resolution to negotiate outcomes. Communication issues are commonly at the root of the appeal: So it s perceived to be bad care or a bad attitude or lack of communication, but really it s a communication botch-up. Enforcement There was no indication of administration support for enforcing the conditions of accountability for reasonableness or equivalent concepts. In addition, during the period of observation and in the interviews, we detected no formal effort to enforce the admissions guidelines. DISCUSSION This is the first study to describe actual admission decision making in a critical care unit and evaluate it using accountability for reasonableness. Previously, accountability for reasonableness has been used to evaluate priority setting at the level of health systems (15, 16, 28) To date, it has not been used to evaluate priority setting for admissions to critical care units, and so our study fills this gap in the literature. Our study demonstrates that it is feasible to describe and evaluate priority setting in critical care units using the methods we described and by doing so to generate lessons for improving the fairness of the priority setting process. In the absence of agreement regarding the correct set of admission criteria, fairness is a key priority-setting goal in critical care. Other critical care units may find the specific findings of this study helpful; also, it is likely that they would benefit from using these methods to improve priority setting in their own context. Lessons Relevance. Despite perceptions to the contrary on the part of other hospital staff, intensivists consider both medical and nonmedical factors when making admissions decisions. The existing hospital admissions policy is partially understood and used sometimes (e.g., by elevating the status of transplant admissions) but not at other times (e.g., emergency department and oncology patients have the same standing as transplant patients according to the guidelines but are not prioritized in the same way). Many actual admission decisions conflict with the policy either because the policy is not known or because the guidelines do not fit the intensivists understanding of the hospital s priorities. At this hospital, the perceived primacy of the transplant program over other priority programs, for example, would lead to intensivists bending over backward for transplant patients in favor of patients from other equally entitled programs. No guidance is provided by the policy when two or more similarly needy patients from priority programs present simultaneously. Moreover, different intensivists make different independent judgments. Consequently, in many cases, whether a patient is admitted to the ICU seems to depend on who is making the decision that day. These findings correspond to previous studies by Zussman (10) and Strosberg and Teres (11). Intensivists independence probably has its roots in a sense of autonomy that intensivists have about professional decisions, and it is what is expected of them when trained as intensivists. How to reshape their decision making so that it is more participatory and more consistent between intensivists is an important question that is beyond the scope of this article. Publicity. The hospital s admissions policy is not widely known within the hospital and not known at all outside the hospital. Decisions and their reasons are only occasionally discussed with patients or their families. There is clearly room for greater transparency or explanation regarding reasons for admissions, in both a specific (explanations to families) and a general sense (widespread awareness of the guidelines). Appeals/Revisions. A formal avenue of appeal/revision does exist but is rarely used. Informal discussions are most often used. Clinicians make different decisions about who should be admitted to the MSICU but do not have opportunities to debate their reasons. Enforcement. Ideas for enforcement were included with the original guidelines document, including provision for a mandatory review 6 months after adoption of the guidelines. This has not, however, materialized, nor is there monitoring of admissions decisions. Accountability for reasonableness provides an explicit ethical framework that can be used to identify good practices and recommendations for improvement. In this case study, we identified two good practices: a) Priority setting decisions are made according to criteria that many Crit Care Med 2003 Vol. 31, No. 12 2767

The key lesson of our study is not only the specific findings obtained here but also how combining a case study approach with the ethical framework of accountability for reasonableness can be used to identify good practices and opportunities for improving the fairness of priority setting in critical care. agree are relevant to the context of a hospital ICU; and b) there is a formal appeals/revisions mechanism for addressing challenges to priority-setting reasoning. Recommendations for improving priority setting that we identified in this case study include the following: a) There should be regular discussion of admission decisions at meetings of the entire ICU team to increase understanding of relevant rationales and how they should be applied; b) decisions and their reasons should be communicated to the patient and/or family (depending on the patient s status and the family s involvement); c) the hospital should provide a forum for reviewing the ICU admissions guidelines and increase efforts to enhance its staff s understanding of the guidelines (e.g., via grand rounds, unit rounds, and the hospital s intranet) and to increase public understanding of decision making in its intensive care units (e.g., via the media, the hospital s Web site, and other public forums); and d) the hospital should formally monitor priority setting in the ICU to ensure that the decisions are made fairly (i.e., according to the conditions of accountability for reasonableness). Limitations The main limitation of our study is that it was conducted in a single intensive care unit and the results may not be generalizable. Despite this limitation, our findings conform to our clinical experience in several units. We anticipate that staff of other units will recognize their own processes in our findings. On the other hand, generalizability is not an objective of qualitative case studies such as this the goal is to describe this particular case. The process we have used to describe using case study methods and evaluate using accountability for reasonableness is generalizable and can be used by others to identify good practices and opportunities for improvement in their own context. CONCLUSIONS We have described and evaluated admissions decision making in a hospital s critical care unit. The key lesson of our study is not only the specific findings obtained here but also how combining a case study approach with the ethical framework of accountability for reasonableness can be used to identify good practices and opportunities for improving the fairness of priority setting in critical care. ACKNOWLEDGMENTS Thank you to Mark Bernstein and Neil Lazar for valuable assistance in this project and to Norman Daniels for ongoing help. REFERENCES 1. Bhinder S: Our reformed health-care system has fatal price tag. Toronto Star 2000; January 28 2. Gollom M: Fatal rerouting inquest begins: Teenager died after hospital refused admission. National Post 2000; September 11, p A20 3. Cook D, Giacomini M: The sound of silence: Rationing resources for critically ill patients. Crit Care 1999; 3:R1 R3 4. Parmanum J, Field D, Rennie J, et al: National census of availability of neonatal intensive care. British Association for Perinatal Medicine. BMJ 2000; 321:727 729 5. Metcalfe MA, Sloggett A, McPherson K: Mortality among appropriately referred patients refused admission to intensive-care units. Lancet 1997; 350:7 11 6. Sprung CL, Geber D, Eidelman LA, et al: Evaluation of triage decisions for intensive care admission. Crit Care Med 1999; 27: 1073 1079 7. Goldfrad C, Rowan K: Consequences of discharges from intensive care at night. Lancet 2000; 355:1138 1142 8. American Thoracic Society Bioethics Task Force. Fair Allocation of Intensive Care Unit Resources. Am J Respir Crit Care Med 1997; 156:1282 1301 9. Society of Critical Care Medicine Ethics Committee. Consensus Statement on the Triage of Critically Ill Patients. JAMA 1994; 271:1200 1203 10. Zussman R: Intensive Care. Chicago, University of Chicago Press, 1992 11. Strosberg MA, Teres D: Gatekeeping in the Intensive Care Unit. Chicago, Health Administration Press, 1997 12. Daniels N, Sabin JE: Setting Limits Fairly: Can We Learn to Share Medical Resources? Oxford, UK, Oxford University Press, 2002 13. Daniels N, Sabin JE: The ethics of accountability in managed care reform. Health Affairs 1998; 17:50 64 14. Martin DK, Giacomini M, Singer PA: Fairness, accountability for reasonableness, and the views of priority setting decision-makers. Health Policy 2002; 61:279 290 15. Ham C, Roberts G (Eds): Reasonable Rationing: International Experience of Priority Setting in Health Care, London, UK, Open University Press, in press 16. Ham C: Tragic choices in health care: Lessons from the Child B case. BMJ 1999; 319: 1258 1261 17. Ham C, McIver S: Contested Decisions: Priority Setting in the NHS. London, UK, King s Fund Publishing, 2000 18. Martin DK, Singer PA: Priority setting for health technologies in Canada. In: Reasonable Rationing: International Experience of Priority Setting in Health Care. Ham C, Roberts G (Eds). London, UK, Open University Press, in press 19. Yin RK: Case Study Research: Design and Methods. Thousand Oaks, CA, Sage, 1994 20. Strauss A, Corbin J: Basics of Qualitative Research: Techniques and Procedures of Developing Grounded Theory. Thousand Oaks, CA, Sage, 1998 21. Strauss A, Corbin J: Grounded theory methodology: An overview. In: Handbook of Qualitative Research. Denzin NK, Lincoln YS (Eds). Thousand Oaks, CA, Sage, 1994 22. Altheide DL, Johnson JM: Criteria for assessing interpretive validity in qualitative research. In: Handbook of Qualitative Research. Denzin NK, Lincoln YS (Eds). Thousand Oaks, CA, Sage, 1994, pp 485 499 23. Mays N, Pope C: Qualitative research in health care: Assessing quality in qualitative research. BMJ 2000; 320:50 52 24. Mays N, Pope C: Rigour and qualitative research. BMJ 1995; 311:109 112 25. Daniels N: Accountability for reasonableness. BMJ 2000; 321:1300 1301 26. Daniels N: Four unsolved rationing problems: A challenge. Hastings Center Report 1994; 24:27 29 27. Daniels N: Justice, health, and health care. Am J Bioethics 2001; 1:2 16 28. Ham C, McIver S: Contested Decisions: Priority Setting in the NHS. London, UK, King s Fund Publishing, 2000 2768 Crit Care Med 2003 Vol. 31, No. 12