Abstract
Background/Objectives:
Concerns have repeatedly been raised about end-of-life decision-making when a patient with diminished capacity is represented by a professional guardian — a paid official who is appointed by a judge. Such guardians are said to choose high-intensity treatment even when it is unlikely to be beneficial or to leave pivotal decisions to the court. End-of-life decision-making by professional guardians has not been examined systematically, however.
Design:
Retrospective cohort study.
Setting:
Inpatient and outpatient facilities in the VA Connecticut Healthcare System.
Participants:
Decedent patients represented by professional guardians who received care at Connecticut VA facilities from 2003–13 and whose care in the last month of life was documented in the VA record.
Measurements:
Through chart reviews, we collected data about the guardianship appointment, the patient’s preferences, the guardian’s decision-making process, and treatment outcomes.
Results:
There were 33 patients with professional guardians who died and had documentation of their end-of-life care. The guardian sought judicial review for 33% and there were delays in decision-making for 42%. In the last month of life, 29% of patients were admitted to the intensive care unit, intubated, or underwent cardiopulmonary resuscitation, while 45% received hospice care. Judicial review and high-intensity treatment were less common when information about the patient’s preferences was available.
Conclusion:
Rates of high-intensity treatment and hospice care were similar to older adults overall. Since high-intensity treatment was less likely when the guardian had information about a patient’s preferences, future work should focus on advance care planning for individuals without an appropriate surrogate.
Keywords: Surrogate decision-making, end-of-life care, court-appointed guardians, incapacitated patients
Introduction
Health professionals often turn to the court to appoint a professional guardian when a patient has prolonged incapacity and no suitable surrogate.1 This is an individual with no prior relationship to the patient who is asked to make medical decisions. When pivotal choices arise near the end of life, case studies and clinical experience have led experts to suggest that professional guardians respond in two potentially problematic ways. One is by selecting high-intensity treatment regardless of the likely benefit because the guardian perceives this to be the “safer” course.2–4 The other is by avoiding complex decisions and leaving them to the court, whether or not the law requires the guardian to do so.5 The result is that high-intensity treatment is delivered by default while health professionals wait for a judge to make a decision.6, 7
These issues have not been examined systematically, however. The primary reason is that data about guardianship are not collected by courts, by health systems, or in population-based surveys.8 Although several investigators have examined decision-making for hospitalized patients who have no surrogate decision-maker at all,9 or are waiting for a guardian to be appointed,10, 11 there is no empirical evidence about what actually happens in the clinical setting when decisions must be made by a professional guardian for a patient who is near the end of life.12 The need for such evidence has become particularly urgent because the number of individuals requiring professional guardians is expected to grow dramatically over the next 20 years as increasing numbers of older adults develop dementia and lack traditional surrogate decision-makers.13–15
In previous work, our team developed a method to identify a population-based sample of patients with professional guardians using data from the Department of Veterans Affairs (VA).16 Our objective in the present study was to characterize end-of-life decision-making processes and treatment outcomes for patients in this sample who died. We also sought to determine whether there was an association between the availability of information about a patient’s preferences and certain decision-making processes and outcomes, such as judicial review and high-intensity treatment.
Methods
Terminology and Background
Most states use the term “guardian” to refer to an individual appointed by the court to oversee the affairs of another adult. A few employ the term “conservator” instead. For consistency, we use “guardian” throughout. There are two types of guardians. A guardian of person makes medical and custodial decisions, while a guardian of estate manages financial matters. We use “guardian” to mean “guardian of person” and refer to a professional guardian when the court selects a stranger for this role. Because of substantial state-to-state variation in how guardianships function, including differences in the availability and structure of guardianship offices,17 the work background of professional guardians, and laws governing guardianship and end-of-life decisions,5 we limited the study to a single state.
Study Population
Our approach to identifying patients with professional guardians has been described previously.16 In brief, we obtained access to electronic notes and administrative data for patients who received care at Connecticut VA facilities from 1/1/2003–12/31/2013. Using SQL Server 12.0 (Microsoft Corp., Redmond, WA), we searched for forms of “guardian” and “conservator” in standardized note titles and in structured fields containing information about relationships between patients and their next of kin. We reviewed the charts of patients flagged through those searches and determined whether a professional guardian made medical decisions. We identified patients who died using the VA Vital Status File.
Patient Preferences, End-of-Life Decision-Making Processes, and Treatment Outcomes
We conducted chart reviews to collect data about end-of-life decision-making processes and treatment outcomes. The first step was to determine whether the VA record contained information about end-of-life care, since many veterans receive care outside the VA system. We excluded patients whose care in the last month of life was primarily delivered outside the VA or in a VA facility outside Connecticut.
For the remaining patients, we used a structured instrument to gather information. First, we collected data about the patient and guardian, including the primary reason for the patient’s diminished capacity, the professional guardian’s background, and the length of the guardianship. Next, we determined whether information about the patient’s preferences was present. If advance directives were not scanned into the chart, we looked for evidence that the patient had discussed treatment goals with the guardian or with involved family and friends. We collected data concerning the decision-making process at the end of life, including information about involvement of the court and hospital ethics committee and about difficulties and delays resulting from the guardianship. Finally, we ascertained whether high-intensity treatment occurred, including intensive care unit (ICU) transfer, intubation, and cardiopulmonary resuscitation (CPR). We also identified patients who received hospice care.
In addition to abstracting these selected elements, we constructed an overall narrative of the patient’s end-of-life care. We reviewed progress notes to capture, to the extent possible, a more complete story of the medical decisions that arose and how these decisions were made. These narratives included information about the patient’s medical and psychosocial history, interactions between health care providers and the professional guardian, and the events prior to death. We include selected narratives to provide the reader with more detailed examples corresponding to the results of the structured review.
The research plan was approved by the institutional review boards at Yale University and the VA Connecticut Healthcare System.
Statistical Analysis
We used univariate statistics to describe the population and the prevalence of characteristics ascertained in chart reviews. We then used Fisher’s exact test to examine associations between the availability of patient preferences, the decision-making process followed by the guardian, and treatment outcomes. Analyses were performed using SAS, version 9.4 (SAS Institute, Cary, NC).
Results
Among 134,241 patients who received care at Connecticut VA facilities from 2003–2013, 97 (0.1%) had professional guardians and died during the study period. For 61 patients, end-of-life care primarily occurred outside the VA system; 3 patients died outside Connecticut. For the remaining 33 patients, the VA chart documented care in the last month of life in detail.
Characteristics of these 33 patients appear in Table 1. Their median age at death was 71 years (IQR, 60–81) and 58% were 65 years or older. All were men, and 60% were nursing home residents. The most common reasons for incapacity were dementia (37%) and psychotic disorders (31%). The professional guardian was an attorney for 89% of patients. A median time of 2.75 years (IQR 0.5, 5.75) elapsed between the guardian’s appointment and the patient’s death, with 2 patients (9%) who had professional guardians appointed in the last month of life. Some information about preferences was available in 23 of 33 cases (70%). More than a third of patients had completed a living will. Others had spoken to the guardian or to family members about their wishes regarding life-sustaining treatment. In 10 of 33 cases (30%), the professional guardian had no information about the patient’s preferences.
Table 1.
Characteristics of Decedent Patients with Professional Guardians
| Characteristic | Patients with professional guardians (n=33) |
|---|---|
| Age, years, median (IQR) | 71 (60–81) |
| White, n (%) | 28 (84) |
| Male, n (%) | 33 (100) |
| Nursing home resident, n (%) | 20 (60) |
| Other | 5 (15) |
| Social worker | 4 (11) |
| Length of guardianship, years, median (IQR) | 2.75 (0.5, 5.25) |
| From family member involved with care | 4 (12) |
Numbers cannot be added because information was available from multiple sources for some patients.
Characteristics of the end-of-life decision-making process and treatment outcomes for these patients are shown in Table 2. The majority of patients (73%) were hospitalized in the last month of life. The medical team had difficulty communicating with the guardian for 22% of patients. The ethics committee was involved for 6%. The guardian sought judicial review for 33%. There were delays in decision-making for 42%, ranging from 1 day to more than 2 weeks. Of the 33 patients, 29% experienced high-intensity treatment, while 45% received hospice care.
Table 2.
Characteristics of Decision-Making Processes and Treatment Outcomes
| Characteristic | Patients with professional guardians (n=33) |
|---|---|
| Hospitalized in the last month of life, n (%) | 24 (73) |
| Medical team had difficulty communicating with guardian, n (%) | 8 (22) |
| Ethics committee involved in decisions | 2 (6) |
| Judicial review sought by guardian, n (%) | 11 (33) |
| Because the guardian and medical team disagreed | 2 (6) |
| CPR | 6 (17) |
| Hospice care in the last month of life, n (%) | 15 (45) |
Numbers cannot be added because patients could receive multiple treatments.
The Figure shows the relationships between the availability of patients’ preferences, the decision-making process, and treatment outcomes. For patients whose preferences were not known, there was considerable heterogeneity in processes and outcomes. Guardians sought guidance from the court in 6 cases but not in 4 others; 2 of these 4 patients received high-intensity care, while 2 did not. When guardians had information about patients’ preferences, there was less variation. When such information was available, professional guardians were less likely to seek judicial review (22% versus 60% of patients; p=.049) and high-intensity care was less common (26% versus 70%; p=.026). Judicial review was independently associated with high-intensity care (67% versus 33% of patients), but the difference was not statistically significant (p= .065).
Figure.
Patient Preferences, Decision-Making Processes, and Treatment Outcomes
Among the paths in the Figure, the most common (14 of 33 patients) was that the professional guardian had information about the patient’s preferences, made decisions without judicial review, and agreed to low-intensity end-of-life care. An example (Patient A) appears below:
Man in his 90s with Alzheimer’s disease who was represented by a professional guardian because of concerns about financial abuse by a caregiver. The guardian visited him at home every two weeks and eventually located a living will stating that the patient did not want his life artificially prolonged. The guardian authorized a do-not-resuscitate/do-not-intubate (DNR/DNI) order. Two years later, the patient developed difficulty eating and drinking. The guardian agreed with his primary care doctor’s recommendation for hospice services. Hospice was consulted and he died a week later.
The second most common path (5 of 33 patients) occurred when a professional guardian did not have information about the patient’s preferences and sought judicial review, with the patient receiving high-intensity care. One example (Patient B) involves a veteran with multiple chronic medical conditions and significant functional decline who died in the ICU:
Man in his 50s with schizophrenia, HIV-associated dementia, chronic kidney disease, cirrhosis, and multiple amputations that left him wheelchair-bound. He had a brother, a sister, and a child, but was estranged from all of them. He never completed a living will. He was admitted to a nursing home and a lawyer was appointed as his guardian five years before his death. There were attempts to discuss his preferences regarding life-sustaining treatment, but the patient did not appear to understand these questions. He developed worsening renal insufficiency. A dialysis catheter was placed and the patient immediately became bacteremic. This led to a meeting about his code status in which the guardian “admitted to not knowing what was actually involved in coding someone” but stated that he believed the “chance of success in a code was 50–50.” He asked for information indicating otherwise to be faxed to him and ended the meeting. The patient remained full code. He suffered a hemorrhagic stroke and was intubated, admitted to the ICU, and placed on vasopressors. Brainstem reflexes were minimal, then absent. The guardian was “not comfortable making decisions about DNR status” or withdrawal of care. He told the medical team that he would be able “to get a [court] hearing on this in 6–7 days.” An expedited hearing was arranged a day later. The judge ordered care withdrawn and the patient died.
The third most common path (4 of 33 patients) involved a professional guardian who had some information about a patient’s preferences and made decisions without judicial review, but with the patient receiving high-intensity care. A representative narrative (Patient C) involves a patient with a prolonged hospitalization:
Man in his 60s with vascular dementia who had completed a living will directing that life-sustaining treatment be withheld if he developed a terminal illness. His most significant relationship was with a girlfriend whom he had appointed as his health care agent. She relinquished this role after he began to drink heavily and became verbally abusive to her. A lawyer was named his guardian. He was admitted to the hospital with severe C. difficile colitis and his guardian consented to a total colectomy. He was intubated peri-operatively, could not be extubated, and developed renal failure. The nephrology fellow spoke to the guardian, who consented to hemodialysis. Several weeks later, he became bacteremic. At this point, the chart records: “need to discuss … weaning life support.” The guardian was contacted and affirmed that “the vet would not want any heroic measures done for him.” He was eventually extubated and made DNR/DNI, but remained in the ICU with intermittent tachyarrythmias and worsening pulmonary edema. The surgical team met with the guardian “to discuss [the patient’s] ultimate goals.” They agreed to continue current measures, without escalation of care. On post-operative day 42, he became bradycardic and hypotensive during dialysis and died.
Discussion
We identified a small sample of decedent veterans with professional guardians and examined end-of-life decision-making processes and treatment outcomes for these patients. Communication difficulties and delays in care were common, as was judicial review of the guardian’s decisions. While 29% of patients received high-intensity treatment in the last month of life, hospice care was delivered to 45%. Professional guardians who had information about patients’ preferences were less likely to involve the court and these patients were less likely to receive high-intensity treatment.
Concerns have been raised about end-of-life decision-making by professional guardians for more than 30 years.4, 6, 18–20 Reports have alleged that guardians choose high-intensity treatment even when it is unlikely to be beneficial or that they avoid decisions and leave them to the court, with high-intensity treatment occurring by default while medical providers await judicial review. To our knowledge, our results provide the first empirical data on this subject and complicate the prevailing view. There were professional guardians in our sample who did just what has been described, and there were patients who received care that was clearly burdensome. But there were also instances in which high-intensity treatment occurred for reasons that are familiar and universal, and have little to do with guardianship: because making decisions for another person is difficult, and death is unpredictable, and it is hard to know in the moment what aggressive treatment will achieve. Moreover, the most common end-of-life experience for patients in our sample involved a professional guardian who worked with the medical team without involving the court and did not advocate for high-intensity treatment. The rates of both ICU utilization and hospice care in our sample were similar to what has been reported for older adults overall.21 One explanation for the difference between previous reports and our findings is that health care providers’ views of medical decision-making under guardianship are affected by an availability heuristic. This is the concept that the perceived frequency of an event is influenced by the ease with which an example comes to mind.22 Because troubling cases are so memorable, providers may believe them to be the norm even though they are relatively unusual.
Our finding that judicial review and high-intensity care were less likely when professional guardians had information about a patient’s preferences calls attention to the heightened importance of advance care planning in this population. Ideally, advance care planning should begin before capacity begins to diminish. Adults who do not have family members or friends to serve as potential surrogates, but who retain the ability to make medical decisions, should be identified and encouraged to complete advance directives.23 Even after a professional guardian has become necessary, there are opportunities for the guardian to learn about values and goals by discussing them with the patient, learning about his or her life and past practices, and engaging with available family members and friends. In 2018, the Connecticut Probate Assembly endorsed this approach in Standards of Practice24 that follow those of the National Guardianship Association25 in instructing guardians to do everything possible to determine what the patient would have wanted and to follow those wishes. Programs exist to prepare guardians to gather such information and serve as surrogates.26 In our sample, the median time between appointment of a professional guardian and a patient’s death was almost three years. This suggests that there is often an opportunity for the guardian to learn about the patient before pivotal decisions must be made.
Even with intensive advance care planning efforts, however, there will be individuals with impaired capacity whose preferences for medical care are not known and whose surrogate decision-maker, out of necessity, is a stranger. For health professionals, the accepted approach in these cases is to follow a best interests standard, weighing the risks and benefits of a treatment and making decisions from the perspective of a generic, reasonable person.27 There are instances in which a patient’s best interests are not served by high-intensity treatment and clinicians feel strongly that such care should not occur.28, 29 By contrast, the National Guardianship Association and the Connecticut Probate Assembly guidelines for guardians take a different approach. They ask guardians to adhere to a “presumption in favor of the continued treatment of the person”25 unless there is evidence that a patient would wish otherwise. The result is a tension between the perceived obligations of health professionals and guardians that can lead to uncertainty and distress for both parties. Collaborative efforts between health professionals, lawyers, ethicists, and national guardianship organizations are necessary to address this issue. A more consistent process needs to be developed that balances the responsibility to protect the rights of this vulnerable population with the desire not to subject individuals to medical care that has little likelihood of benefit.
Our study has several limitations. We identified our sample using data from a single state. These data are limited to veterans, who are overwhelmingly male, older than the general population, and more likely to be white.30 Because our findings rely on what practitioners documented in the electronic record, we only included patients whose end-of-life care was delivered primarily through the Connecticut VA. As a result, our sample is small and does not contain patients who spent their last month of life in community settings, without care from VA providers. Despite these limitations, we are unaware of another data source that can be used to study decedent persons with professional guardians. Ours is the first work, to our knowledge, to examine their end-of-life care in a systematic manner.
In conclusion, we found that only a minority of decedent patients with professional guardians received high-intensity treatment and that judicial review and high-intensity treatment were less likely when the professional guardian had some knowledge of a patient’s preferences. These findings suggest a need for robust advance care planning efforts to ensure that the best possible decisions are made when a patient must be represented by a stranger.
Impact statement: We certify that this work is novel. To our knowledge, it is the first systematic examination of end-of-life decision-making processes and treatment outcomes for patients represented by professional guardians.
Acknowledgements
Funding sources: Dr. Cohen was supported by a training grant (T32AG1934) and GEMSSTAR award (R03AG053278) from the National Institute on Aging. All authors were supported by the Claude D. Pepper Older Americans Independence Center at Yale University (P30AG21342).
Sponsor’s role: Dr. Cohen was supported by a training grant (T32AG1934) and GEMSSTAR award (R03AG053278) from the National Institute on Aging. All authors were supported by the Claude D. Pepper Older Americans Independence Center at Yale University (P30AG21342). The funding sources were not involved in the design and conduct of the study; the collection, management, analysis, and interpretation of the data; or the preparation, review, and approval of the manuscript.
Footnotes
Conflicts of interest: No conflicts of interest to disclose.
References
- [1].Karp N, Wood E. Incapacitated and Alone: Health Care Decision-Making for the Unbefriended Elderly. Washington, DC: American Bar Association Commission on Law and Aging, 2003. [Google Scholar]
- [2].Office of Maryland Attorney General. Policy study on Alzheimer’s disease care. 2004. (online). Available at: http://www.marylandattorneygeneral.gov/Pages/HealthPolicy/Alzheimers.aspx. Accessed May 28, 2019 [Google Scholar]
- [3].Hastings KB. Hardships of end-of-life care with court-appointed guardians. Am J Hosp Palliat Care. 2014;31: 57–60. [DOI] [PubMed] [Google Scholar]
- [4].Dying in America: Improving Quality and Honoring Individual Preferences Near the End of Life. Washington, DC: Institute of Medicine, National Academy of Sciences, 2014. [PubMed] [Google Scholar]
- [5].Cohen AB, Wright MS, Cooney L Jr., T. Fried Guardianship and end-of-life decision making. JAMA Intern Med. 2015;175: 1687–1691. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [6].Pope TM. Making medical decisions for patients without surrogates. N Engl J Med. 2013;369: 1976–1978. [DOI] [PubMed] [Google Scholar]
- [7].Pope TM. Legal Briefing: Adult orphans and the unbefriended: making medical decisions for unrepresented patients without surrogates. J Clin Ethics. 2015;26: 180–188. [PubMed] [Google Scholar]
- [8].United States Senate Special Committee on Aging — 108th Congress. Guardianship over the elderly: security provided or freedoms denied? February 11, 2003 (online). Available at: https://aging.senate.gov/imo/media/doc/2112003.pdf. Accessed May 28, 2019. [Google Scholar]
- [9].White DB, Curtis JR, Lo B, Luce JM. Decisions to limit life-sustaining treatment for critically ill patients who lack both decision-making capacity and surrogate decision-makers. Crit Care Med. 2006;34: 2053–2059. [DOI] [PubMed] [Google Scholar]
- [10].Bandy RJ, Helft PR, Bandy RW, Torke AM. Medical decision-making during the guardianship process for incapacitated, hospitalized adults: a descriptive cohort study. J Gen Intern Med. 2010;25: 1003–1008 [DOI] [PMC free article] [PubMed] [Google Scholar]
- [11].Ricotta DN, Parris JJ, Parris RS, Sontag DN, Mukamal KJ. The burden of guardianship: a matched cohort study. J Hosp Med. 2018;13: 595–601. [DOI] [PubMed] [Google Scholar]
- [12].Kim H, Song MK. Medical decision-making for adults who lack decision-making capacity and a surrogate: state of the science. Am J Hosp Palliat Care. 2018;35: 1227–1234. [DOI] [PubMed] [Google Scholar]
- [13].Hebert LE, Weuve J, Scherr PA, Evans DA. Alzheimer disease in the United States (2010–2050) estimated using the 2010 census. Neurology. 2013;80: 1778–1783. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [14].Lin IF, Brown SL. Unmarried Boomers confront old age: a national portrait. Gerontologist. 2012;52: 153–165. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [15].Redfoot D, Feinberg L, Houser A. The aging of the Baby Boom and the growing care gap: a look at future declines in the availability of family caregivers. AARP Public Policy Institute, 2013. (online). Available at: http://www.aarp.org/content/dam/aarp/research/public_policy_institute/ltc/2013/baby-boom-and-the-growing-care-gap-insight-AARP-ppi-ltc.pdf. Accessed May 28, 2019. [Google Scholar]
- [16].Cohen AB, Trentalange M, Benjamin AZ, Fried T. Characteristics of patients with professional guardians. JAMA Intern Med. 2019;179: 107–108. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [17].Teaster PB, Wood EF, Schmidt WC, Lawrence SA. Public Guardianship after 25 Years: In the Best Interest of Incapacitated People? Santa Barbara, CA: Praeger, 2010 [Google Scholar]
- [18].President’s Commission for the Study of Ethical Problems in Medicine and Biomedical and Behavioral Research Deciding to Forego Life-Sustaining Treatment: A Report on the Ethical, Medical, and Legal Issues in Treatment Decisions. Washington, DC: U.S. Government Printing Office, 1983. [Google Scholar]
- [19].Gillick MR. Medical decision-making for the unbefriended nursing home resident. J Ethics Law Aging. 1995;1: 87–92. [PubMed] [Google Scholar]
- [20].Breznay J, Paris B. Challenges of guardianship: examining four cases from New York State. Annals of Long-Term Care. 2014;22: 34–38. [Google Scholar]
- [21].Teno JM, Gozalo P, Trivedi AN, et al. Site of death, place of care, and health care transitions among U.S. Medicare beneficiaries, 2000–2015. JAMA. 2018;320: 264–271. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [22].Tversky A, Kahneman D. Availability: a heuristic for judging frequency and probability. Cognitive Psychology. 1973;5: 207–232. [Google Scholar]
- [23].Farrell TW, Widera E, Rosenberg L, et al. AGS Position Statement: Making medical treatment decisions for unbefriended older adults. J Am Geriatr Soc. 2017;65: 14–15. [DOI] [PubMed] [Google Scholar]
- [24].Connecticut Standards of Practice for Conservators. Office of the Probate Court Administrator, State of Connecticut, 2018. (online). Available at: http://www.ctprobate.gov/Documents/Connecticut%20Standards%20of%20Practice%20for%20Conservators.pdf. Accessed May 28, 2019. [Google Scholar]
- [25].National Guardianship Association Standards of Practice. 4th edition, 2013. (online). Available at: https://www.guardianship.org/wp-content/uploads/2017/07/NGA-Standards-with-Summit-Revisions-2017.pdf. Accessed May 28, 2019. [Google Scholar]
- [26].Massachusetts Guardianship Academy. Guardianship Training (online). Available at: https://guardianship.academy/guardianship-training/. Accessed May 28, 2019.
- [27].Buchanan AE, Brock DW. Deciding for Others: The Ethics of Surrogate Decision Making. Cambridge: Cambridge University Press, 1989. [Google Scholar]
- [28].Braun UK, McCullough LB. Preventing life-sustaining treatment by default. Ann Fam Med. 2011;9: 250–256. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [29].Blinderman CD, Krakauer EL, Solomon MZ. Time to revise the approach to determining cardiopulmonary resuscitation status. JAMA. 2012;307: 917–918. [DOI] [PubMed] [Google Scholar]
- [30].US Department of Veterans Affairs. Profile of Veterans: 2016 (online). Available at: https://www.va.gov/vetdata/docs/SpecialReports/Profile_of_Veterans_2016.pdf. Accessed May 28, 2019.