Early experience with medical assistance in dying in Ontario, Canada: a cohort study

We recognize the letters from Drs. Violette and Bouchard, Dr. Ferrier and Dr. Coelho all seek further clarification about how palliative care involvement was defined. We direct readers to our appendix 2 for details about what was available in the MAID database, and how this information was obtained.

There were two separate data fields where palliative care was relevant: (1) the specialty/subspecialty of the MAID assessors and providers, and (2) the involvement of a palliative care provider at the time of the MAID request. For both questions, the assessment was made by a nurse investigator working for the Office of the Chief Coroner of Ontario. The specialty/subspecialty of the MAID assessors and providers was determined with reference to the registry of the College of Physicians and Surgeons of Ontario and the College of Nurses of Ontario; however, some assessors and providers were designated as palliative care providers (even in the absence of a professional certification or designation) when it was known they were providing palliative care within a palliative care unit or palliative care team. The decision that a palliative care practitioner was involved was based on a thorough review of the notes and assessments and a telephone conversation with the MAID provider. This assessment was not based on any billing information, and it was not based solely on the word of the MAID provider.

In response to comments about credentials, we do not feel that it would be appropriate to judge the quality of palliative care on the basis of the credentials of the provider, especially in Canada. There was no accredited Palliative Care training in Canada prior to 1999, and the first subspecialty training programs were only accredited 2016. Many of Canada’s foremost palliative care providers have been providing exceptional palliative care and palliative care instruction without the benefit of formal Palliative Care credentials.

We were not able to compare the prevalence of palliative care involvement in the MAID cohort to any similar measure in the Ontario decedent cohort, as this level of detailed review is not available in the latter. However, as explained in the discussion, others have examined the prevalence of palliative care involvement among all Ontarians,[1] and this number appears to be less than 50% even using the most broad definition of palliative care.

We acknowledge the comments of Drs. Graham, Goligher and Gallagher about the possible interpretations of the data we presented. To be clear, we are not denying that there are problems with access to Palliative Care in Canada, nor are we denying the existence of socioeconomic vulnerability. These are clearly issues, but based on our findings, they are very unlikely to be driving the provision of MAID in Ontario. Indeed, access is hard to measure, but given that three quarters of patients were being followed by a palliative care provider at the time of the request, it is hard to argue that poor access was a driving factor in MAID provision. We will not dispute the personal experience of some providers, but the data collected by the coroner’s office suggests that these reports may not reflect the broader experience of a population.

MAID remains a controversial subject in the medical community, and the findings of our study do not resolve this controversy. But for those who are keen to address the underlying drivers of MAID, our study offers some insight into what those might be.

References

1. Brown CR, Hsu AT, Kendall C, et al. How are physicians delivering palliative care? A population-based retrospective cohort study describing the mix of generalist and specialist palliative care models in the last year of life. Palliat Med 2018;32:1334-43.

We commend our colleagues for their research on early experiences of MAiD in Ontario [1]. However, we have to question their interpretation that people are not requesting MAiD due to lack of access to palliative care because of the inadequacies of their data.

As clinicians, we find that palliative care is often consulted at the time of a MAiD request. The study does not reveal how the authors determined that a patient received palliative care nor how long prior to the MAiD request that palliative care was initiated. The authors use an Ontario study[2] to compare data on palliative care access in MAiD recipients and interpreted their data showing access by 74.4% of MAiD recipients to the other study’s 47% as evidence that MAiD requests were not due to lack of access to palliative care. A 2017 Ontario study[3] found that 38.8% had their first palliative care access in the last month prior to death, and 12.0% in the second month prior to death. This study concluded there is a large variation in the intensity and timing of care, with many receiving little care, and a significant proportion of care initiated and delivered close to death. This conclusion is supported by a 2018 Canadian Institute for Health Information study[4] that demonstrated most palliative care services were received only in the last month of life and that over 80% receiving palliative care in hospital were admitted through emergency, demonstrating that palliative care is not accessed early or systematically in Canada. The same study examined 349 records of MAiD in hospital between June 2016 and March 2016 and found 70% were identified as palliative on their last hospital admission where they received MAiD.

Previous research [5,6] has illustrated that individuals with months of high levels of disease burden (physical, emotional and spiritual/existential distress) and the convergence of certain psychosocial factors leads to depression and hopelessness and ultimately to a desire for hastened death. Providing palliative care to those who have already been suffering for months and thus end up distressed and suffering enough to request hastened death is most often providing palliative care too late. This is not even to mention the significant minority of 22.8% of MAiD recipients in the study by Downar et al[1] who apparently had no palliative care involvement whatsoever at any time before medically-assisted death.

We find the suggested conclusion, from less than transparent data, is unsupported and more of a political statement than research.

There remains a well-known [4,7,8,9] need for earlier, equitable access to palliative care for all Canadians with a serious illness.

References

1. Downar J, Fowler RA, Halko R, Davenport Huyer L, Hill AD, Gibson JL. Early experience with medical assistance in dying in Ontario, Canada: a cohort study. CMAJ. 2020 Feb 11.
2. Brown CR, Hsu AT, Kendall C, et al. How are physicians delivering palliative care? A population-based retrospective cohort study describing the mix of generalist and specialist palliative care models in the last year of life. Palliat Med 2018; 32: 1334-43.
3. Tanuseputro P, Budhwani S, Bai YQ, et al. Palliative care delivery across health sectors: a population-level observational study. Palliat Med 2017; 31(3): 247–257.
4. Canadian Institute for Health Information. Access to Palliative Care in Canada. Ottawa, ON: CIHI; 2018. Accessed February 19, 2020 at https://www.cihi.ca/sites/default/files/document/access-palliative-care-...
5. Rodin G, Lo C, Mikulincer M, Donner A, Gagliese L, Zimmermann C. Pathways to distress: The multiple determinants of depression, hopelessness, and the desire for hastened death in metastatic cancer patients. Social Science & Medicine 2009; 68: 562–569
6. Lo C, Zimmermann C, Rydall A, Walsh A, Jones J, Moore M, Shepherd F, Gagliese L, Rodin G. Longitudinal Study of Depressive Symptoms in Patients with Metastatic Gastrointestinal and Lung Cancer. Journal of Clinical Oncology 2009; 28(18): 3084-3089 DOI: 10.1200/JCO.2009.26.9712
7. Tung, J., Chadder, J., Dudgeon, D., Louzado, C., Niu, J., Rahal, R., & Sinnarajah, A. Palliative care for cancer patients near end of life in acute-care hospitals across Canada: a look at the inpatient palliative care code. Current Oncology 2019; 26(1). doi:http://dx.doi.org/10.3747/co.26.4563
8. Constante A, Lawand C, Cheng C. Access to Palliative Care in Canada. Healthcare Quarterly 2019; 21(4): 10-12. doi:10.12927/hcq.2019.25747
9. Xiong, B., Freeman, S., Banner, D. et al. Hospice use and one-year survivorship of residents in long-term care facilities in Canada: a cohort study. BMC Palliat Care 2019; 18, 100 doi:10.1186/s12904-019-0480-z

Downar et al.1 have provided a descriptive analysis of patient characteristic for patient who have received Medical Assistance in Dying (MAiD) in comparison to patients who have died from any cause.

One of the author’s stated intentions was to address the concern that patients may seek MAID as a consequence of poor access to palliative care. Therefore, how access to palliative care was defined and captured in this study is of paramount importance to any associations implied.

In the absence of a clearly accepted and standardized definition for quality of palliative care delivery, it is critical that the definition used is at minimum systematic, transparent and reproducible. In this regard the authors have not provided their definition or objective criteria by which palliative care delivery was judged. The authors also do not provide an internal control to validate the approach used. They applied their definition for palliative care delivery only to patients who received MAiD and not to a similar control matched group of patients who died without MAiD.

If a very broad definition of access to palliative care were applied then it would appear as though access to palliative care among patients who received MAID were high, namely 75%. However, this may not address the actual concern that patients accessing MAID do not have access to adequate or meaningful palliative care provided by sufficiently qualified practitioners. This concern is substantiated by Dr. Downar’s own words from a recent Tweet, “We used the judgment of the RN investigators based on their review of the chart to arrive at 75%. Indeed, that might refer to any type or quality of PC [palliative care]”.2

Consequently, the observation of high prevalence of access to palliative care among patients may have been a forgone conclusion based on a broad and subjectively applied criteria for palliative care involvement. It is especially critical that investigators in this domain focus on patient important definitions for research, ideally with patient involvement from the outset. When one desires palliative care, one desires quality palliative care not merely a construct called “access to palliative care”. The use of surrogate constructs for key definitions in this way can be misleading to public discourse.

References

1. Downar, J., Fowler, R. A., Halko, R., et al. (2020). Early experience with medical assistance in dying in Ontario, Canada: a cohort study. CMAJ. 2020 Feb 12. [E-pub ahead of print]. http://doi.org/10.1503/cmaj.200016
2. Downar, J. “We used the judgment of the RN investigators based on their review of the chart to arrive at 75%. Indeed, that might refer to any type or quality of PC [palliative care]”. [Twitter]. 2020 Feb 12. Available from: https://twitter.com/jamesdownar/status/1227777599348277248

I find it astonishing that respected researchers could think large data number-crunching can lead to valid conclusions on patient safety in the age of MAiD.
Downar et al, on the basis only of the number of palliative care contacts among patients who die from MAiD, conclude both that palliative care was unable to relieve patients’ suffering, and that a lack of palliative care access was not driving requests. The variability in quality and training among services that call themselves “palliative care” is well known. Nor does the study distinguish between a single visit and full-service palliative care.
The authors also conclude that, as “traditionally vulnerable demographic groups” were underrepresented in the group dying from MAiD, it is “unlikely to be driven by social or economic vulnerability.”
Dr. Balfour Mount tells the story of a patient who was a prominent businessman, diagnosed with advanced cancer. At the admission the patient was silent while his wife explained that he had had enough and pleaded for things to be over quickly. When Dr. Mount later returned to speak with the patient alone, he broke down in tears and confided, “I’m so afraid of being a burden”.
In my work in geriatrics I have seen vulnerability in run-down, bug-infested apartments with no food and no clean sheets. But I have also seen it in palatial houses, where virtually the only visits are from people paid to provide care – while others hover on the periphery waiting for the inheritance. Riches are far from being a protection against coercion to die.
The connection between the data and the conclusions in this study are tenuous at best. To what end? What urgency is there to label the MAiD experiment as safe, prematurely and without evidence?

The authors draw conclusions which could have far reaching impact in the public's perception of MAiD, based on potentially misleading data. Published in the Globe and Mail, “The authors of the new Ontario research say its findings – which are based on reviews of every assisted-dying case in the province over two years – counter fears that the procedure would become a final refuge for patients too poor and vulnerable to access high-quality health services, including palliative care. Instead, the opposite has proved true: It is the affluent, not the marginalized, who most often avail themselves of the assisted-dying law that Canada enacted nearly four years ago.”(1)

The data for this study was based on self reporting of the MAiD provider, and not by the patient. Evidence by self reporting is often biased towards compliance with regulations. Case in point, looking at anonymous questionnaires from Belgium, self reporting led to only 1 in 2 cases being actually reported to authorities as euthanasia. Those not formally reported were correlated with decreased adherence to guidelines and lacking in palliative care involvement.(2)

Given we are on the brink of MAiD expansion without a foreseeable death clause in Canada, a study prematurely declaring no concerns for the vulnerable is alarming and reckless.

Also important to consider, does being affluent preclude vulnerability? Is financial abuse of the elderly a non-issue in Canada?(3) Could poor clinical acumen in helping patients adjust to loss of power or function and high rates of depression still be hidden in these figures?

References:
(1)Grant,K. Wealthier patients more likely to use medical assistance in dying, data show. Globe & Mail. 2020 Feb 11 (newspaper on the internet) Available from: https://www.theglobeandmail.com/canada/article-wealthier-patients-more-l...
(2)Smets T1, Bilsen J, Cohen J, Rurup ML, Mortier F, Deliens L. Reporting of euthanasia in medical practice in Flanders, Belgium: cross sectional analysis of reported and unreported cases. BMJ. 2010 Oct 5;341:c5174. doi: 10.1136/bmj.c5174.
(3)Government of Canada. Elder abuse: It's Time to Face the Reality.(Internet) last accessed on 2020 February 14.
Available from: https://www.canada.ca/en/employment-social-development/campaigns/elder-a...

The excellent research by Downar (1) and colleagues provides vital Canadian data to inform future practice and policy. These findings fit well with my experience in assessing approximately 150 Toronto MAID applicants (2); with one notable exception. Downar et al. report that only 6.6% of families of MAID patients or MAID providers raised concerns post-mortem about difficulties assessing MAID, such as delays in referrals to a willing MAID assessor or provider or lack of clarity on how to make a request. When I routinely asked all MAID patients that I assessed if they had experienced difficulties with their MAID referrals, approximately one-third reported concerns, with some reporting delays of several months. While some of these related to lack of knowledge or information, others were caused by family, partners, friends, spiritual advisors or clinicians actively attempting to delay or prevent their MAID requests. As the authors mention in their study limitations, the third party reports they could analyze post-mortem “may not have represented patients’ experiences accurately”. Provincial and local MAID referral sources need to be better advertised and accessible to reduce delays and suffering.

References:

1. Downar J, Fowler RA, Halko R, Davenport Huyer L, Hill AD, Gibson JL. Early experience with medical assistance in dying in Ontario, Canada: a cohort study. CMAJ. 2020 Feb 11.
2. Stewart DE, Rodin G, Li M. Consultation-liaison psychiatry and physician-assisted death. Gen Hosp Psychiatry. 2018;55:15-19.

Donna E. Stewart CM, MD, FRCPC
University Professor, University of Toronto
Centre for Mental Health, University Health Network