Community Characteristics and MAiD Uptake in Ontario: An Ecological Study
Michaela Kelly, MSc
Ellen Wiebe, MD, University of British Columbia, Department of Family Practice
ABSTRACT
Background: MAiD accounts for 2.0% of deaths in Canada, but uptake varies widely across the country. Little research has been done to identify community characteristics that contribute to the differences in MAiD uptake. Objectives: To investigate the relationship between community characteristics and MAiD uptake in order to provide insight for service planning in jurisdictions with new MAiD laws. Methods: This was an ecological study comparing population-level data from each Local Health Integration Network (LHIN). The number of MAiD deaths as a proportion of total deaths was calculated with data from the Office of the Chief Coroner in Ontario. Proportion over 65-years-old, proportion with a university degree, proportion of immigrants, proportion with English as a mother tongue, population density and median income were calculated using 2016 Census data. Proportion of decedents that accessed palliative care and the ratio of MAiD providers to decedents were calculated with data from published reports and presentations. Data was collected for each LHIN and compared. A correlation coefficient was calculated for each community characteristic to quantify its linear relationship to the uptake of MAiD. Linear regression was used to further analyze these relationships. Results: MAiD uptake ranged from 0.9 to 3.8%. After accounting for potential confounding effects, the proportion over 65 years old, the proportion of mother tongue English-speakers, the proportion of immigrants, and the number of MAiD providers per 1000 decedents were most strongly associated with MAiD uptake. Only having an older population was associated with MAiD uptake at a statistically significant level (p<0.05). On average, for every 1% increase in the proportion of population over 65 years, MAiD deaths accounted for 0.33% more of total deaths (95% CI: 0.07 to 0.58; p=0.02). Conclusions: These findings suggest that in addition to individual-level factors identified in previous literature, some community characteristics may also impact the uptake of MAiD.
Submission: July 18, 2021 | Publication: August 16, 2021
DISCLOSURES: The authors have no potential conflicts of interest to declare.
ATTRIBUTIONS: This project was conceived by MK and EW. MK gathered the data, and both authors analyzed the data and wrote and edited the manuscript.
CITATION: Kelly, M and Wiebe, E (2021). Community Characteristics and MAiD Uptake in Ontario: An Ecological Study. Canadian Health Policy, August 2021. ISSN 2562-9492 www.canadianhealthpolicy.com
[Download PDF to access exhibits.]
BACKGROUND
Since 2016, medical assistance in dying (MAiD) has been legal in Canada (Government of Canada, 2016). In 2019, MAiD accounted for 2.0% of all deaths in Canada, but this proportion varied widely across the country (Health Canada, 2020). Previous research has investigated the individual-level demographic factors associated with having MAiD in Canada (Downar et al, 2020), but little research has been done to identify community characteristics that contribute to the differences in MAiD uptake between regions.
The objective of this study was to investigate the relationship between community characteristics and MAiD uptake. We aimed to identify predictors of high uptake of MAiD at a community level to provide insight for service planning in jurisdictions with new MAiD laws.
METHODS
Design
This was an ecological study comparing population-level data from each Local Health Integration Network (LHIN) in Ontario.
Setting
Ontario has over 14 million residents and makes up 39% of the Canadian population (Statistics Canada, 2021). At the time of data collection and analysis, health care services in Ontario were delivered through 14 LHINs across the province (FIGURE 1). This model has recently undergone changes and has been renamed Home and Community Care Support Services (Ontario Ministry of Long-Term Care, 2021). We continue to use the term “LHIN” in this paper to maintain clarity about the regional areas we compared and the timing of our data collection.
Data Collection
Outcome variable: The outcome of interest, MAiD uptake, was defined as the number of MAiD deaths as a proportion of total deaths during a one-year period from September 2017 to August 2018. We calculated this with summary data reported by the Office of the Chief Coroner for Ontario (Ministry of Health and Long-Term Care et al, 2018).
Predictor variables: We measured eight community-level characteristics in each LHIN: six demographic characteristics and two healthcare resource characteristics. We used data from the 2016 Census (Statistics Canada, 2016) to calculate the following demographic characteristics for each LHIN:
- proportion over 65-years-old
- proportion with a university degree
- proportion of immigrants
- proportion with English as a mother tongue
- population density
- median income
We used summary data from published reports to calculate two healthcare resource characteristics for each LHIN:
- access to palliative care was defined as the proportion of decedents who received palliative care in the last 12 months of life between 2010 and 2012 and was measured using summary data from Health System Performance Research Network (Tanuseputro et al, 2012).
- access to MAiD care was defined as the number of physician MAiD providers per 1000 decedents and was calculated with summary data from the Office of the Chief Coroner’s Office of Ontario of the number of physician MAiD providers in each LHIN as of 2019.
Statistical Analysis
The data were recorded in excel and exported to Stata version 15 for analysis. Descriptive statistics were performed to describe the 14 LHINs. A Pearson correlation coefficient was calculated for each community characteristic to quantify its linear relationship to the uptake of MAiD. Linear regression was used to measures the average change in MAiD uptake per 1 unit increase in a community characteristic, and multivariable linear regression was used to control for the effects of confounding factors. Community characteristics were investigated for their potential confounding effects on the associations of interest if their Pearson correlation coefficient was less than -0.10 or greater than 0.10.
RESULTS
Descriptive Statistics
MAiD uptake varied by 2.9%, ranging from 0.9% of all deaths in the Central West LHIN to 3.8% of all deaths in the South-East LHIN (mean=2.0%) (Fig. 1). Population density ranged from 0.6 to 6412.6 people per square kilometer (mean = 664.5), and median annual income ranged from $29454 to $39555 (mean=$33786.57).
Crude Analysis
Of the demographic characteristics, having an older population had the highest linear correlation with MAiD uptake (r=0.48). On average, for every 1% increase in the proportion of population over 65 years, MAiD deaths accounted for 0.14% more of total deaths in a LHIN (95% CI: -0.02 to 0.31, p=0.08). MAiD uptake was also correlated with language and immigration. The proportion of people who identified English as a mother tongue was positively correlated with MAiD uptake (r=0.45), and the proportion of people who identified as immigrants to Canada was negatively correlated with MAiD uptake (r=-0.47). On average, for every 1% increase in the proportion of people with English as a mother tongue, MAiD deaths accounted for 0.03% more of total deaths in a LHIN (95% CI: -0.01 to 0.07, p=0.10). For every 1% increase in the proportion of immigrants, MAiD deaths accounted for 0.03% less of the total deaths in a LHIN on average (95% CI: -0.06 to 0.01, p=0.09). No demographic factor had a correlation coefficient greater than 0.5 (TABLE 1). None of the crude linear regression estimates were at a statistically significant level (p<0.05), but for the 3 strongest correlations, p<0.10 (TABLE 2).
Of the healthcare resource characteristics, having a high ratio of MAiD providers was correlated with a higher uptake of MAiD (r=0.64). On average, for every increase of 1 provider per 1000 deaths, MAiD uptake accounted for 0.11% more of total deaths (95% CI: 0.03 to 0.19%, p=0.014). The proportion of decedents who received palliative care in the last 12 months of life did not account for much of the variation in MAiD uptake (r=0.12).
Multivariable Analysis
After accounting for potential confounding effects, the proportion of over 65 years old, the proportion of mother tongue English-speakers, the proportion of immigrants, and the number of MAiD providers per 1000 decedents were most strongly associated with MAiD uptake. Only having an older population was associated with MAiD uptake at a statistically significant level (p<0.05) (TABLE 3). On average, for every 1% increase in the proportion of population over 65 years, MAiD deaths accounted for 0.33% more of total deaths in a LHIN (95% CI: 0.07 to 0.57, p=0.02); for every 1% increase in proportion of individuals with English as a mother tongue, MAiD uptake increased by 0.05% (95% CI: -0.01 to 0.12, p=0.07); for every 1% increase in proportion who immigrated to Canada, MAiD uptake decreased by 0.09% (95% CI: -0.003 to 0.02, p=0.06); and for every increase of 1 provider per 1000 deaths, MAiD deaths accounted for 0.12% more of total deaths (95% CI: -0.05 to 0.28, p=0.13) (TABLE 3).
DISCUSSION
These findings suggest that the proportion of the population over 65 years of age was independently associated with uptake of MAiD. It is important to note that the outcome of interest, MAiD uptake, was defined as the number of MAiD deaths as a proportion of all deaths within a LHIN. As such, this observed association cannot be explained by the fact older individuals are more likely to die of any cause because this has already been accounted for. Instead, this finding likely suggests that in retirement communities with more individuals over 65, there may be more social awareness and acceptance of MAiD within these communities. Also, people who move to retirement communities may have different attitudes than people who remain in the communities in which they worked and raised their families.
Our findings also suggest a trend toward lower uptake of MAiD in areas with a higher proportion of individuals who immigrated to Canada. Although it was not found at a statistically significant level, this may be because there were only 14 LHINs to compare. Previous studies have found that people choosing MAiD are more likely to be non-immigrant (Steck et al, 2018). Individuals immigrating to Canada may be more likely to have values in line with their countries of origin, which may not have legal assisted death. Alternatively, barriers to access MAiD that are unique to individuals who have immigrated to Canada may be influencing the lower uptake in these areas. Future research should investigate this further.
There was also a trend towards a higher uptake of MAiD in LHINs that had a higher ratio of MAiD providers. Previous studies about providers’ experiences indicate that there are many barriers to providing MAiD and a lack of providers in some areas (Khoshnood et al, 2018). It is unclear whether a higher ratio of providers is due to greater uptake of MAiD or greater uptake of MAiD is due to better access to MAiD providers.
We did not find an association between income level and MAiD uptake, but previous studies using neighborhood-level data have found higher income to be associated with having MAiD (Downar et al, 2020; Steck et al, 2014, 2018). We may have found this if we used a smaller discrete unit than a LIHN.
LIMITATIONS
Ecological designs have a few inherent limitations that should be considered when interpreting the findings of this study. Namely, the findings of ecological studies cannot be interpreted at an individual level; we cannot link the particular demographic characteristics with the likelihood of particular individuals having MAiD. Since the demographic characteristics measured in an ecological study represent the summary measure for a given population rather than individual levels, it is also possible that an observed linear association is masking a more complicated, non-linear relationship that can only be identified with individual level data.
One further limitation of this study is that MAiD uptake and the community characteristics were measured over different time periods. We tried to reduce the impact of this by using the most recent community-level data that was available; still, there are some differences in the time periods of the examined variables, and this may have resulted in residual confounding effects or an over or under estimation of association with MAiD uptake. We believe that any potential over or under estimation of association is minor because the demographics and health care services of Ontario are reasonably consistent over time.
CONCLUSIONS
The findings from this ecological study suggest that in addition to individual-level factors identified in previous literature, some community characteristics may also impact the uptake of MAiD. Future research that investigates both individual-level and community-level data in the analysis would be beneficial to further investigate the role of community characteristics on MAiD uptake and MAiD service planning.
APPENDIX
*adjusted for of all characteristics in table. |
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
REFERENCES
Downar, J., Fowler, R. A., Halko, R., Huyer, L. D., Hill, A. D., & Gibson, J. L. (2020). Early experience with medical assistance in dying in Ontario, Canada: A cohort study. Canadian Medical Association Journal, 192(8), E173. https://doi.org/10.1503/cmaj.200016
Government of Canada. (2016). An Act to amend the Criminal Code and to make related amendments to other Acts (medical assistance in dying).
Health Canada. (2020). First annual report on medical assistance in dying in Canada 2019. https://www.canada.ca/en/health-canada/services/medical-assistance-dying-annual-report-2019.html
Khoshnood, N., Hopwood, M.-C., Lokuge, B., Kurahashi, A., Tobin, A., Isenberg, S., & Husain, A. (2018). Exploring Canadian Physicians’ Experiences Providing Medical Assistance in Dying: A Qualitative Study. Journal of Pain and Symptom Management, 56(2), 222-229.e1. https://doi.org/10.1016/j.jpainsymman.2018.05.006
Ministry of Health and Long-Term Care, Office of the Chief Coroner of Ontario, & Health Canada. (2018, October). MAID: Program update and Ontario’s approach to the federal monitoring regime. https://www.ontariofamilyphysicians.ca/files/maid-stakeholder-webinar_-federal-monitoring-regime-final.pdf
Ontario Ministry of Long-Term Care. (2021). Home and Community Care Support Services. https://www.health.gov.on.ca/en/common/system/services/lhin/facts.aspx
Statistics Canada. (2016). Data Products, 2016 Census. https://www12.statcan.gc.ca/census-recensement/2016/dp-pd/index-eng.cfm
Statistics Canada. (2021). TABLE 17-10-0009-01 Population estimates, quarterly. https://doi.org/10.25318/1710000901-eng
Steck, N., Junker, C., Maessen, M., Reisch, T., Zwahlen, M., & Egger, M. (2014). Suicide assisted by right-to-die associations: A population based cohort study. International Journal of Epidemiology, 614–622. https://doi.org/10.1093/ije/dyu010
Steck, N., Junker, C., & Zwahlen, M. (2018). Increase in assisted suicide in Switzerland: Did the socioeconomic predictors change? Results from the Swiss National Cohort. BMJ Open, 8(4), e020992. https://doi.org/10.1136/bmjopen-2017-020992
Tanuseputro, P., Budhwani, S., Bai, Y. Q., & Wodchis, W. P. (2012). Understanding the Provision of End-of-Life and Palliative Care Services in Ontario. Health System Performance Research Network.
Toronto Community Health Profiles Partnerships. (2013). Local Health Integration Networks in Ontario [Map]. http://www.ontariohealthprofiles.ca
