Statistics Canada
Symbol of the Government of Canada

Findings

Warning View the most recent version.

Archived Content

Information identified as archived is provided for reference, research or recordkeeping purposes. It is not subject to the Government of Canada Web Standards and has not been altered or updated since it was archived. Please "contact us" to request a format other than those available.

First Nations (North American Indians), Métis and Inuit are the three major Aboriginal groups in Canada.  Research has consistently shown that First Nations have a much shorter life expectancy than that for Canada as a whole, and are at increased risk for causes of death that occur more frequently at younger ages, such as injuries and suicides.1-16 The pattern is similar for Inuit.15,17,18  However, mortality patterns among Métis are largely unknown.19

The Métis people of Canada were originally descended from unions between European men and North American Indian women, primarily in Western Canada.  The Métis have their own distinct culture and traditions.20 While there is no single definition of Métis, the term generally includes any person of mixed North American Indian and European ancestry who self-identifies as Métis.19

On the 2006 census, individuals who self-identified as Métis represented about one-third (34%) of the total Aboriginal population of Canada.21 Compared with their population size, Métis have been under-represented in Aboriginal health research.22,23

Mortality data for Métis are difficult to generate.  Information on the Aboriginal ancestry, identity or status of decedents is not routinely collected on death registrations.  Because, until recently, there was no comprehensive national registry of Métis, their mortality cannot be studied directly—as has been done for Registered Indians.  And because most Métis people do not live in areas where they constitute a high proportion of the total population, their mortality cannot be studied indirectly with an area-based approach24—as has been done for Inuit.17,25

The 1991 to 2001 Canadian census mortality follow-up study provides an unprecedented opportunity to examine mortality among Métis.  The primary objective of this paper is to fill the data gap with regard to mortality among Métis adults; results for Registered Indians are presented to provide context.  The specific objectives are to determine the risk of premature death among Métis and Registered Indians aged 25 or older, to calculate their remaining life expectancy and probability of survival to age 75, and to identify the causes of death most responsible for excess mortality compared with non-Aboriginal adults. 

Methods

Data sources

The Canadian census mortality follow-up study consists of a 15% sample (2,735,152) of the non-institutionalized population aged 25 or older, all of whom were enumerated via the 1991 census long-form questionnaire.  This cohort was tracked for mortality from June 4, 1991 to December 31, 2001.  Details of the construction and contents of the census mortality database have previously been reported.26

Briefly, because the electronic census data files did not contain names, but names were needed to find the corresponding deaths, creation of the census mortality database required two linkages.  First, the census file was linked to a nominal list (name) file (abstracted from tax-filer data and then encrypted).  Then, the census plus encrypted name file was matched to the Canadian Mortality Database using methods of probabilistic record linkage,27 an approach similar to that used for other mortality follow-up studies at Statistics Canada.28 The resulting file contained anonymous information on demographic characteristics, socio-economic status, activity limitations, disability, and for individuals who died during the study period, cause and date of death.

Eligibility

Only people who were enumerated by the 1991 census long-form questionnaire, had attained age 25 by census day (June 4, 1991), and who were usual residents of Canada were eligible to be part of the mortality follow-up study cohort.

Data quality reports found that the 1991 census missed 3.4% of Canadian residents of all ages.  Compared with the enumerated population, the missed individuals were more likely to be young, mobile, low-income, of Aboriginal ancestry,29 or homeless.  A total of 78 Indian reserves—representing about 38,000 people—were either not enumerated or incompletely enumerated and so were excluded from the census database30 and could not be part of the follow-up study cohort.  The long-form questionnaire was generally given to 1 in 5 Canadian households.  However, all residents of the Indian Reserves that were enumerated, of many remote and northern communities, and of non-institutional collective dwellings received a long-form.  Because of the necessity of obtaining encrypted names from tax-filer data, only tax-filers could be followed for mortality.

Analytical techniques

For each cohort member, person-days of follow-up were calculated from the beginning of the study (June 4, 1991) to their date of death or emigration (ascertained from the name file and known for 1991 only), or the end of the study (December 31, 2001).  Person-days of follow-up were divided by 365.25 to get person-years at risk.

Age- and sex-specific mortality rates by 5-year age groups (at baseline) were used to calculate age-standardized mortality rates (ASMRs) for subgroups, using the total cohort Aboriginal population structure (person-years at risk), both sexes together, as the standard population.  In this case, the Aboriginal population consisted of anyone who indicated a North American Indian, Métis or Inuit ancestry, Registered Indian status, or membership in a North American Indian Band or First Nation.  Corresponding 95% confidence intervals for the ASMRs were calculated as described by Carrière and Roos.31 A similar method was used to calculate confidence intervals for the ASMR rate ratios (RRs) and ASMR rate differences (RDs).

For age-specific analyses, cohort members were categorized by 10-year age group from 25-to-34 to 65-to-74 and 75 or older.  The mortality rates within each of these age groups were age-standardized using 5-year age groups.  For example, the mortality rate in the 25-to-34 age group was age-standardized using the Aboriginal population (as described in the previous paragraph) age distribution for 25-to-29- and 30-to-34-year-olds.  While most analyses used age at baseline (June 4, 1991), life table analyses used age at the beginning of each year of follow-up.

Period life tables for each sex, plus corresponding standard errors and 95% confidence intervals, were calculated according to the method of Chiang.32 These calculations were made after age was transformed from age at baseline to age at the beginning of each year of follow-up, and deaths and person-years at risk were calculated separately for each year (or partial year) of follow-up.  Deaths and person-years at risk were then pooled by age at the beginning of each year of follow-up before calculation of the life tables.

Cox proportional mortality hazard ratios were calculated by sex, first controlling for age (years), and then further controlling for highest level of education (less than high school graduation, high school graduation, postsecondary diploma, university degree), income adequacy quintile (1 to 5), occupation skill level (professional, managerial, skilled/technical/supervisory, semi-skilled, unskilled, no occupation), community size (1 million or more; 500,000 to 999,999; 100,000 to 499,999; 10,000 to 99,999; less than 10,000), and place of birth (Canada or elsewhere).  Place of birth was included in the models to reduce the healthy immigrant effect among non-Aboriginal cohort members.  Detailed definitions of these variables have been previously provided.26  Differences in excess mortality (1 minus the hazard ratio) comparing the two models were interpreted as estimates of the effect of the above-mentioned socio-economic variables on the extent of the disparities between Métis and other cohort members, and between Registered Indians and other cohort members.

The underlying cause of death of those who died during the study period had been previously coded to the World Health Organization’s International Classification of Diseases, Ninth Revision (ICD-9)33 for deaths occurring in the period 1991 through 1999, and to the Tenth Revision (ICD-10)34 for deaths occurring in 2000 or 2001.  For analyses by cause of death, deaths were categorized according to cause groupings established by the Public Health Agency of Canada35 and by the European Union Working Group on Socioeconomic Inequalities in Health.36  Appendix Table A contains the list of ICD codes.

Table A
ICD-9 and ICD-10 codes for selected causes of death

Defining Métis and Registered Indian populations

The Métis population can be defined in several ways from the Census, either using an identity concept or using an ancestry concept.37 Because the 1991 census did not ask respondents to self-identify as an Aboriginal person (North American Indian, Métis, or Inuit), the ancestry approach was used.  Census respondents were asked to which ethnic or cultural group(s) their ancestors belonged.38 From a list of 15 groups, including Métis, respondents were instructed to check as many as applicable.

Registered Indian status was determined by a direct question:  “Is this person a Registered Indian as defined by the Indian Act of Canada?” (Yes, No).  Registered Indians could also report any ethnic origin, including Métis, and if so, such persons could be classified in both groups for this study.  For this analysis, a total of 2,200 cohort members were classified as both Métis and Registered Indians.

Derivation of Métis within the cohort

Within the entire 2.7 million-person cohort of the mortality follow-up study, 19,100 persons indicated some Métis ancestry (Table 1).  Of those, 11,800 reported only Métis ancestry or Métis plus other Aboriginal (with or without any other non-Aboriginal ancestry).  The remaining 7,300 reported Métis and at least one non-Aboriginal ancestry, but no other Aboriginal ancestry.  To simplify terminology, these two groupings will be referred to as “Métis plus other Aboriginal” and “Métis plus non-Aboriginal,” respectively.  The composition of the two groupings was based on an analysis of 1996 census data that cross-classified ethnic origins by Aboriginal identity (not asked in 1991).37 The categories included in the “Métis plus other Aboriginal” grouping were those with any Métis ancestry for which the highest proportion of the category self-identified as Métis.  Altogether, in the 1996 census data, 91% of the “Métis plus other Aboriginal” category self-identified as Aboriginal (73% Métis), and 50% of the “Métis plus non-Aboriginal” category self-identified as Aboriginal (48% Métis).

Table 1
Derivation of the “Metis” subset of entire cohort, non-institutional population aged 25 or older at baseline, Canada, 1991

The socio-demographic characteristics of the entire mortality follow-up study cohort differed from those of the two Métis groupings (data not shown).  The characteristics of the “Métis plus other Aboriginal” group were most distinctive, while those of the “Métis plus non-Aboriginal” group were intermediate between the entire cohort and the “Métis plus other Aboriginal” group.  Therefore, the only results shown in the remainder of this article are those for the former group.  As well, the “Métis plus other Aboriginal” is referred to simply as “Métis.”

Results

The 1991 to 2001 Canadian census mortality follow-up study followed about 2.7 million persons, including 11,800 Métis and 56,700 Registered Indians, for 11 years.  The latter two groups accounted for 120,920 and 577,380 person-years at risk, and 889 and 4,506 deaths, respectively, during that period (Table 2).  Linkage rates to the name file (comparing cohort members to long-form census respondents) were lower for Métis (62% for men and 68% for women) than for non-Aboriginal cohort members (79% and 75%), but not as low as rates for Registered Indians (47% and 59%).  Based on 1991 deaths that could be identified independently in the Canadian Mortality Database and/or the name file, ascertainment of deaths in the cohort (from 1991 to 2001) was estimated at about 97% overall, and 95% to 96% among persons with any Aboriginal ancestry, Registered Indian status or membership in an Indian Band or First Nation.

Table 2
Long-form census respondents, cohort members, linkage rate, deaths and person-years at risk, by sex and population group, non-institutional population aged 25 or older at baseline, Canada, 1991 to 2001

A comparison of the demographic and socio-economic characteristics of Métis cohort members with those of all eligible Métis adults in the census population (weighted) revealed very few differences between the two (Appendix Table B); therefore, the sample of Métis was not biased with respect to those characteristics.  Results were similar for Registered Indians (Appendix Table C).

Table B
Demographic, socio-economic and geographic characteristics of in-scope (eligible) Métis census respondents, compared with Métis cohort members, by sex, non-institutional population aged 25 or older at baseline, 1991

Table C
Demographic, socio-economic and geographic characteristics of in-scope (eligible) Registered Indian census respondents, compared with Registered Indian cohort members, by sex, non-institutional population aged 25 or older at baseline, 1991

The geographic distribution of Métis and Registered Indian cohort members differed from that of all eligible Métis and Registered Indian adults in the census population (weighted).  The cohort had an over-representation of Métis and Registered Indians living in communities with less than 10,000 people, on reserves, or in the territories.  These differences reflected the fact that all residents of participating Indian reserves and many remote and northern communities were enumerated using a long-form questionnaire (rather than the 1 in 5 ratio for most of the household population), and thus, were eligible for the cohort.

Characteristics of Métis and Registered Indian cohort members

Compared with non-Aboriginal members of the cohort, Métis were younger and less likely to be legally married, and more likely to be in a common-law relationship (Table 3).  They were less likely to have completed high school and to be employed, and more likely to be in the lowest income quintile.  Nearly 70% lived in Manitoba, Saskatchewan or Alberta (Prairies), and almost 60% lived in rural areas or communities with less than 10,000 population.  Fewer than 10% lived on an Indian Reserve.

Table 3
Percentage distribution of demographic, socio-economic and geographic characteristics of non-Aboriginal adults, Métis and Registered Indians, by sex, non-institutional population aged 25 or older at baseline, Canada, 1991

Registered Indian cohort members were similar to Métis with respect to age structure and marital status.  Compared with Métis and non-Aboriginal cohort members, Registered Indians were less likely to be employed, and more likely to be in the lowest income quintile and to have lower educational attainment.  In part because of over-sampling on reserves, about 70% of Registered Indian cohort members were living on a reserve at time of the census.

Life expectancy

For Métis adults of both sexes, remaining life expectancy at age 25 was substantially shorter than that of non-Aboriginal cohort members, but longer than that of Registered Indians.

At age 25, a Métis man could expect to live an additional 49.5 years, compared with 52.8 years for non-Aboriginal men in the cohort, a difference of 3.3 years (Figure 1, Table 4).  While Métis women had  longer remaining life expectancy than Métis men, the gap between Métis women and non-Aboriginal women was wider.  A 25-year old Métis woman could expect to live an additional 53.7 years, compared with 59.2 years for non-Aboriginal women, a difference of 5.5 years.

Figure 1
Life expectancy at age 25 of Registered Indians, Métis and non-Aboriginal cohort members, by sex, non-institutional population aged 25 or older at baseline, Canada, 1991 to 2001

Table 4
Life expectancy at age 25 and probability of survival from age 25 to 75 for Registered Indian, Métis and non-Aboriginal cohort members, non- institutional population aged 25 or older at baseline, Canada, 1991 to 2001

For both sexes, life expectancy at age 25 was estimated to be about one year longer for Métis than for Registered Indians (49.5 versus 48.4 years for men; 53.7 versus 52.9 years for women).

Probability of survival to age 75

About 57% of Métis men were expected to survive to at least age 75 (conditional on survival to age 25), compared with 64% of non-Aboriginal men, a 7-percentage-point difference (Figure 2, Table 4).  For Métis women, the corresponding figures were 63% and 79%, a 16-percentage-point difference.

Figure 2
Probability of survival from age 25 to 75 of Registered Indians, Métis and non-Aboriginal cohort members, by sex, non-institutional population aged 25 or older at baseline, Canada, 1991 to 2001

The point estimates for survival to age 75 were lower for Registered Indians.  About 51% of Registered Indian men were expected to survive to age 75, compared with 64% of non-Aboriginal men, a 13-percentage-point difference.  For Registered Indian women, the corresponding figures were 62% and 79%, an 18-percentage-point difference.

Age-specific and age-standardized mortality rates

Table 5 shows age-standardized mortality rates (ASMR) per 100,000 person-years at risk for Métis, Registered Indians and non-Aboriginal cohort members, as well as ASMR rate ratios (RRs) for Métis and Registered Indians, compared with non-Aboriginal cohort members.  Age-standardized mortality rates were significantly higher for Métis men (RR=1.38) and women (RR=1.72), and higher still for Registered Indian men (RR=1.56) and women (RR=1.96).  For Métis and Registered Indians of both sexes, rate ratios were highest in the younger age groups and diminished with advancing age (Figure 3).

Table 5
Deaths, age-standardized mortality rates (ASMRs) per 100,000 person-years at risk, and rate ratios for Métis and Registered Indians compared with non-Aboriginal cohort members, by sex and age group, non-institutional population aged 25 or older at baseline, Canada, 1991 to 2001

Figure 3
Age-specific mortality rate ratios comparing Métis and Registered Indians with non-Aboriginal cohort members, by sex and age group, non-institutional population aged 25 or older at baseline, Canada, 1991 to 2001

Causes of death

Among Métis men, the most common causes of death were circulatory system diseases (32% of the total ASMR), followed by all cancers (23%) and external causes such as suicides and motor vehicle accidents (18%)—a ranking similar to that of non-Aboriginal men (Appendix Table D).  For Métis women, the most common causes of death were all cancers (33%), circulatory system diseases (29%), respiratory system diseases (7%), external causes (6%) and digestive system diseases (6%)—the same ranking as for non-Aboriginal cohort women (Appendix Table E).

Table D
Deaths and age-standardized mortality rates (ASMRs) per 100,000 person-years at risk for Métis and Registered Indians compared with non-Aboriginal cohort members, by cause of death, male non-institutional population aged 25 or older at baseline, Canada, 1991 to 2001

Table E
Deaths and age-standardized mortality rates (ASMRs) per 100,000 person-years at risk, for Métis and Registered Indians compared with non-Aboriginal cohort members, by cause of death, female non-institutional population aged 25 or older at baseline, Canada, 1991 to 2001

For Registered Indians, the ranking of causes of death was somewhat different.  Among Registered Indian men, the most common causes were circulatory system diseases (28%), external causes (21%) and all cancers (18%).  For Registered Indian women, the most common causes were circulatory system diseases (26%), all cancers (25%) and external causes (11%).

Tables 6 and 7 show age-standardized rate ratios and rate differences by major cause of death groupings for men and for women, respectively.  (Appendix Tables D and E show the corresponding numbers of deaths.)  Rate ratios for Métis men were elevated for most causes, particularly external causes (RR=2.65) such as drowning (RR=6.94), homicide (RR=4.76), poisoning (RR=3.52) and motor vehicle accidents (RR=3.22), and deaths from infectious (RR=1.74), endocrine (RR=1.86) and digestive (RR=1.93) system diseases.  Rate ratios for Métis women were also elevated, especially for poisoning (RR=6.71), infectious diseases (RR=2.99), digestive system diseases (RR=3.01), mental disorders (RR=2.90), and genitourinary (RR=2.97) and endocrine (RR=2.66) system diseases.

Table 6
Rate ratios and rate differences per 100,000 person-years at risk for Métis and Registered Indians compared with non- Aboriginal cohort members, by cause of death, male non-institutional population aged 25 or older at baseline, Canada, 1991 to 2001

Table 7
Rate ratios and rate differences per 100,000 person-years at risk for Métis and Registered Indians compared with non- Aboriginal cohort members, by cause of death, female non-institutional population aged 25 or older at baseline, Canada, 1991 to 2001

Rate ratios for Registered Indian men were even higher for most causes of death, notably external causes (RR=3.52), mental disorders (RR=3.30), endocrine (RR=3.18) and digestive system (RR=2.76) diseases.  Rate ratios for Registered Indian women were especially high for infectious (RR=5.34), endocrine (RR=5.00) and digestive (RR=4.44) system diseases and external causes (RR=3.66).

Rate differences (RDs) for Métis men compared with non-Aboriginal men were highest for deaths from external causes (RD=89) and circulatory system diseases (RD=55), which together accounted for 67% of the total RD.  Rate differences for Métis women compared with non-Aboriginal women were highest for deaths from circulatory system diseases (RD=67) and cancers (RD=46), accounting for 49% of the total rate difference.

Rate differences for Registered Indian men compared with non-Aboriginal men were highest for deaths from external causes (RD=136) and circulatory system diseases (RD=53), which together accounted for 59% of the total rate difference.  Rate differences for Registered Indian women were highest for circulatory system diseases (RD=70) and external causes (RD=49), accounting for 39% of the total rate difference.

Deaths were also categorized as smoking-related, alcohol-related or amenable to medical intervention (for example, deaths before age 75 from breast and cervical cancer, infectious diseases, cerebrovascular disease, pneumonia or influenza) as defined by the European Union Working Group on Socioeconomic Inequalities in Health.36 Rate ratios for smoking-related deaths were not significantly elevated for Métis men or for Registered Indian men, compared with non-Aboriginal men (Table 6).  But they were significantly elevated for Métis (RR=1.75) and Registered Indian (RR=1.17) women (Table 7).  Rate ratios for alcohol-related deaths were significantly higher among Métis and Registered Indians of both sexes.  Rates of premature death (before age 75) considered amenable to medical intervention were also significantly higher for Métis and Registered Indians of both sexes. 

Hazard ratios controlling for socio-economic variables

Age-adjusted all-cause mortality hazard ratios comparing Métis and Registered Indians to non-Aboriginal cohort members are shown in Appendix Tables F and G.  Both Métis men and women had elevated hazard ratios (1.34 and 1.71, respectively).  Controlling for income adequacy, educational attainment, occupation skill level, community size and immigration reduced the hazard ratios substantially (to 1.11 and 1.51, respectively), which suggests that 68% and 28% of the differences in hazards were explained by these socio-economic variables.  Age-adjusted hazard ratios for Registered Indian men and women of 1.50 and 1.93, respectively, were reduced to 1.19 and 1.66 when controlling for the socio-economic variables, which suggests that 62% and 29% of the differences were attributable to those variables.

Table F
Unadjusted and adjusted all-cause mortality hazard ratios for Métis and non-Aboriginal members, by sex, non- institutionalized population aged 25 or older at baseline, Canada 1991 to 2001

Table G
Unadjusted and adjusted all-cause mortality hazard ratios for Registered Indians and non-Aboriginal cohort members, by sex, non-institutionalized population aged 25 or older at baseline, Canada 1991 to 2001

Comparative perspective

In 1996, the mid-point of the follow-up period, remaining life expectancy at age 25 for Métis men was about what it had been for all men in Canada in 1986 (a 10-year lag) (Appendix Table H).  For Métis women, remaining life expectancy at age 25 in 1996 was about what it had been for all women in 1973 (a 23-year lag).

Table H
Life expectancy at age 25 and percentage surviving to age 75 (conditional on surviving to age 25), Canada, various years

For Registered Indian men aged 25 in 1996, remaining life expectancy was about what it had been for all men in Canada in 1979 (a 17-year lag), and for Registered Indian women, about what it had been for all women in Canada in 1968 (a 28-year lag).

Discussion

This is the first in-depth study to examine mortality patterns for a large sample of Métis adults across Canada.  Métis and Registered Indians in this cohort had higher mortality rates, shorter life expectancy, and lower probability of survival to age 75, compared with the non-Aboriginal cohort members.  This pattern of higher mortality is consistent with that previously described for First Nations in Canada.3-5,15  Elevated rates of premature mortality (dying before age 75) among Aboriginal people have also been reported in other studies.5,10

Mortality rate ratios were highest in the younger age groups and diminished with advancing age—also noted in earlier studies.1,3,10  Part of this can be explained by the very high death rates from external causes (such as suicides and motor vehicle accidents) among Aboriginal youth and young adults.3,4,10 

Differentials in mortality varied by cause of death.  For some causes, Métis and Registered Indians had much higher death rates, while for others such as cancer, death rates were either similar to or only slightly elevated compared with the non-Aboriginal members of the cohort. 

Rate ratios for Métis and Registered Indian men were highest for deaths related to endocrine and digestive system diseases and external causes such as motor vehicle accidents, poisoning, drowning and homicide.  Earlier studies have found these types of deaths to be more common among First Nations3,4,7,8,18 and to represent a major component of premature mortality.10 The higher rate ratios for deaths from endocrine system diseases likely reflect the higher prevalence of diabetes in the Aboriginal population.3,19,39-42

Rate ratios for Métis and Registered Indian women were especially elevated for mental disorders and endocrine, digestive and genitourinary system diseases. For Métis women, rate ratios for external causes, although elevated, were less pronounced than for Métis men.  For Registered Indian men and women, rate ratios for external causes were similarly elevated.  Compared with non-Aboriginal cohort women, Métis and Registered Indian women had elevated rate ratios for all cancers combined and for most specific cancer sites (with the notable exception of breast cancer).  These results differ from previous research on cancer mortality that tended to show excess mortality only for cancer of the cervix,13,44 gallbladder6,14 and kidney.45

Métis and Registered Indians, especially women, had particularly elevated rates for alcohol-related deaths.  These results for Registered Indians are consistent with previous research.2,4,46 age-standardized mortality rates for premature deaths (before age 75) that are amenable to medical care were elevated for Métis and Registered Indians, accounting for nearly one in five excess premature deaths for Métis men and for Registered Indians of both sexes, and nearly one in ten excess premature deaths for Métis women.

Rate differences provide a different perspective by considering how rare or common a cause of death is, whereas rate ratios focus on relative risk.  Rate differences were particularly high for deaths from circulatory system diseases and external causes, which together accounted for 67% of all excess mortality (the total rate difference) among Métis men, and about 59% among Registered Indian men.

Rate differences among Métis women were particularly high for circulatory system diseases and cancers, which together accounted for 49% of the excess mortality.  Among Registered Indian women, 52% of excess mortality was attributable to circulatory system diseases, endocrine system diseases, and external causes of death.

Hazard regression analysis showed that substantial shares of these disparities in mortality (roughly two-thirds for Métis and Registered Indian men, and nearly 30% for Métis and Registered Indian women) were associated with their lower levels of income, education, occupation skill, and urban residence, compared with non-Aboriginal cohort members.

Limitations

To be eligible for inclusion in the Canadian census mortality follow-up study, a person had to have been enumerated by the 1991 census long form and been a tax-filer for the year 1990 or 1991.  Under section 87 of the Indian Act, Registered Indians are entitled to a tax exemption for income earned or considered to be earned on a reserve.47

As well as those who did not file a tax return, persons in long-term care facilities, senior’s residences or prisons could not be included in the mortality follow-up study.

Compared with life tables for all Canada (1995 to 1997), the entire cohort had 1 year longer life expectancy for men and 2 years longer life expectancy at age 25 for women.

Linkage rates were lower among Métis (62% for Métis men and 68% for Métis women), compared with non-Aboriginal adults (79% and 75%, respectively).  However, the socio-economic profile of Métis cohort members was similar to that of all Métis long-form census respondents, suggesting that there was likely little or no linkage bias.  Linkage rates for Registered Indians were even lower (47% for men and 59% for women), but again, analysis revealed that the demographic and socio-economic characteristics of those who were and were not linked were similar.

Ascertainment of deaths was estimated to be slightly lower (95% to 96%) among Aboriginal persons, compared with the cohort as a whole (97%).   This would result in a slight downward bias in mortality rates calculated for Métis and Registered Indians, so the true extent of the disparities compared with non-Aboriginal cohort members could be slightly larger than indicated in this study.

Because questions about Aboriginal identity were not asked on the 1991 census, Métis were defined by ethnic origin (ancestry).  However, based on 1996 census data cross-classified by Aboriginal identity, an estimated 73% of persons classified as Métis in this study would likely have identified as Métis. Nevertheless, the classification undoubtedly excluded many persons who might have identified as Métis, without indicating Métis ancestry (impossible to ascertain in 1991), or who reported categories of mixed Aboriginal and non-Aboriginal ancestry where half or less the population were expected to identify as Métis (and whose inclusion in the cohort would have biased the results). From 1996 to 2006, the self-identified Métis population increased by 91%.21,37 However, the Métis in this study were identified based on Métis ancestry reported in 1991; they were not necessarily the same as persons who self-identified as Métis in 2006.

Conclusion

The 1991 to 2001 census mortality follow-up study has made it possible to fill a data gap by examining mortality among the Métis people of Canada.   Métis and Registered Indian adults had higher mortality rates, compared with non-Aboriginal adults.  At age 25, the life expectancy of Métis men was about three years shorter than that of non-Aboriginal men in the study cohort, and that of Métis women, about five years shorter.  Life expectancy for Registered Indian men and women was about four and six years shorter, respectively.  Mortality disparities between Métis and Registered Indians and non-Aboriginal cohort members varied by cause of death.  Many of these deaths were potentially preventable (for example, intentional and unintentional injuries) or amenable to medical care or modification of behavioural risk factors such as smoking or alcohol abuse.  Furthermore, a substantial proportion of these disparities (roughly two-thirds for Métis and Registered Indian men, and nearly 30% for Métis and Registered Indian women) were explained by socio-economic differences.

These results provide baseline data about mortality patterns among Métis and Registered Indian adults that can be used to begin (for Métis) or continue (for Registered Indians) tracking changes over time.

Acknowledgements

Major funding for this study was provided by the Canadian Population Health Initiative, part of the Canadian Institute for Health Information. Additional funding was provided by the Strategic Research and Analysis Directorate of Indian and Northern Affairs Canada. We also acknowledge the key role of Canada’s provincial and territorial registrars of vital statistics who provide the death data for the Canadian Mortality Data Base.