Acute care hospitalization by Aboriginal identity, Canada, 2006 through 2008
by Gisèle Carrière, Evelyne Bougie, Dafna Kohen, Michelle Rotermann and Claudia Sanmartin
Differences in health, health determinants, and use of health care services between Aboriginal and non-Aboriginal peopleNote 1Note 2Note 3Note 4Note 5Note 6Note 7Note 8Note 9Note 10Note 11Note 12Note 13Note 14Note 15
Note 16Note 17Note 18Note 19Note 20 suggest that the frequency and nature of acute care hospitalization may vary. However, national information about hospital admissions of Aboriginal people is scarce. In some provinces—Manitoba, British Columbia, Alberta, and Saskatchewan—hospital records contain First Nations identifiers, Note 1Note 2 or were appended for Métis persons.Note 3 In other jurisdictions, Aboriginal identity is not routinely included on hospital records. As a result, national-level data about the hospitalization of Aboriginal people are not available.
Researchers have attempted to address this data gap by estimating hospital use by people in areas with higher versus lower percentages of Aboriginal identity residents.Note 21Note 22 However, because area-based data are subject to misclassification,Note 23 individual-level information is preferable.
This study is based on 2006 Census (long-form) socio-demographic information (including Aboriginal identity) that was linked to the Discharge Abstract Database to create a sample for analysis from all provinces and territories except Quebec. The primary purpose is to provide national figures (excluding Quebec) on acute care hospitalizations of Aboriginal (First Nations living on and off reserve, Métis, Inuit in Inuit Nunangat) and non-Aboriginal people. A secondary objective is to identify the leading diagnostic categories (chapters) of acute care hospitalizations, based on the “most responsible diagnosis.”
Data from the 2006 CensusNote 24 were linked to the Canadian Institute for Health Information’s Discharge Abstract Database (DAD) from 2006/2007 through 2008/2009 for nine provinces (excluding Quebec) and the three territories. The complete census file (excluding Quebec), which contains approximately 23.4 million records, was used for record linkage to the DAD.Note 25
Each year, the DAD consolidates about 3 million hospital records from all acute care facilities, and some psychiatric, chronic rehabilitation, and day surgery facilities in Canada,Note 26Note 27Note 28 except Quebec. Because of the exclusion of Quebec, residents of that province (including Inuit in Nunavik) are not represented in the linked data, nor are hospitalizations in Quebec of residents of other provinces and territories.
Hierarchical deterministic linkage was conducted, based on common identifiers recorded in both the census and the DAD: date of birth, sex, and residential postal code. A validation study concluded that the linked file is suitable for health-related research and is broadly representative of the population of Canada.Note 25
An important limitation is the low rate of census coverage and eligibility to link among individuals who identified as Aboriginal.Note 25 Lower coverage means that Aboriginal people were more likely to be underrepresented in the linked census. Records with lower eligibility for linkage were those lacking sufficient information for a linkage attempt.Note 25 The likely impact is underestimation of hospitalization rates of Aboriginal people and a possible downward bias compared with estimates for non-Aboriginal people.
Linkage was performed in accordance with the Directive on Record LinkageNote 29 and approved by Statistics Canada’s Executive Management Committee.Note 30 Details about the linkage methodology are available elsewhere.Note 25
Long-form census respondents, who represent about 20% of the non-institutional population, provided socio-demographic data, including Aboriginal identity.Note 24 All households in Nunavut, Northwest Territories (excluding Yellowknife), Yukon (excluding Whitehorse), and all Indian reserves and settlements were asked to complete the long-form questionnaire.
The final census cohort eligible for linkage to the DAD consisted of 4.65 million long-form respondents, to whom 1,028,604 acute care hospitalizations were linked during the 2006/2007-to-2008/2009 period. According to a validation study of the linked file, 7.2% to 7.7% of Aboriginal people linked to at least one hospitalization record during this period. The corresponding figures for First Nations were somewhat higher: 7.6% to 8.1%. From 5.0% to 5.4% of non-Aboriginal people linked to at least one hospitalization.Note 25
Appendix Table A contains unweighted counts for the 2006 Census–DAD linked study cohort.
The 2006 Census question on Aboriginal identity was: “Is this person an Aboriginal person, that is, North American Indian, Métis, or Inuit (Eskimo)?” Respondents marked all that applied. Answers were classified as: North American Indian (only), Métis (only), Inuit (only), other Aboriginal (multiple or indeterminate), or non-Aboriginal. The analysis includes only single-identity Aboriginal respondents; about 3% of census respondents reporting other Aboriginal (multiple or indeterminate) identitiesNote 31 were excluded.
Geographical location of census respondents was used to identify Inuit living in Inuit Nunangat and First Nations living on reserve (Indian reserves or settlements) or off reserve. Inuit estimates are provided only for those in Inuit Nunangat, the four Inuit land claim regions―Nunatsiavut (northern coastal Labrador), Nunavik (northern Quebec), Nunavut, and Inuvialuit―which together represent 78% of the total Inuit population. Inuit counts for this analysis exclude Nunavik because hospital discharges for Quebec were not available. As a result, 9,565 Inuit (19% of the total Inuit population)Note 9 were excluded.
The 2006 Census on-reserve population includes all residents in any of eight census subdivision (CSD) types legally affiliated with First Nations Indian bands, as well as other types of CSDs in northern Saskatchewan, the Northwest Territories, and the Yukon that have large concentrations of First Nations people. “On reserve” comprises legally defined Indian reserves, Indian settlements, other land types created by the ratification of Self-Government Agreements, and other northern communities affiliated with First Nations according to criteria established by Indigenous and Northern Affairs Canada.
This analysis pertains to First Nations living on reserve, First Nations living off reserve, Métis living off reserve, Inuit in Inuit Nunangat (excluding Nunavik in Quebec), and the non-Aboriginal population. Throughout the remainder of this report, “Aboriginal” refers to members of these four Aboriginal groups; “Inuit” refers to Inuit in Inuit Nunangat excluding Nunavik; and “Métis” refers only to Métis persons who were not residing on reserves.
The frequency of hospitalizations based on the most responsible diagnosis was compiled for each Aboriginal identity group and for non-Aboriginal people. Individuals could be represented more than once if they were hospitalized multiple times during the 2006/2007-to-2008/2009 period.
Acute care DAD hospital discharge records linking to eligible long-form census respondents were classified based on the person’s census-based Aboriginal identity and geographic location rather than the province that submitted the hospital record. This enabled reporting of hospitalizations in a province different from the respondent’s province of residence at the time of the 2006 Census.
Most responsible diagnosis
Each hospital discharge record contains up to 25 diagnoses and 20 intervention codes based on the International Statistical Classification of Diseases and Related Health Problems, 10th Revision, Canada (ICD-10-CA).Note 32 The “most responsible diagnosis,” which refers to the most significant diagnosed condition and/or accounts for the greatest length of stay, was used to sort hospital records into chapters pertaining to specific diseases or injuries, etiology of the disease, conditions specific to body systems, or to conditions and situations that are risk factors to health.Note 33 The first three characters of each most responsible diagnosis were used to classify hospitalizations by chapter (Appendix Table B).
A frequency ranking procedure was applied to all in-scope linked census–DAD records to determine the most common diagnoses. The highest-ranking chapter codes, in addition to hospitalizations for all chapters combined, with and without pregnancy and child-birth-related hospitalizations, were selected to calculate hospitalization rates for each Aboriginal identity group and for non-Aboriginal people.
Age-standardized hospitalization rates (ASHRs) per 100,000 population, rate ratios (RRs), and 95% confidence intervals were calculated by Aboriginal identity group. To reduce the variation that can occur with small numbers, hospital discharge records for the three fiscal years linked to 2006 Census long-form respondents were combined to compile acute care hospitalizations.
ASHRs used the sum of linked hospitalizations for a given Aboriginal identity group as numerators, divided by the denominator—unweighted person counts from the Census study cohort for the same identity group, multiplied by three (number of DAD years). Age standardization used the direct method, based on the age structure of the national Aboriginal population from the 2006 Census. The following age groups were used: 0 to 9; 10 to 19; 20 to 29; 30 to 39; 40 to 49; and 50 or older.
Age-standardized 95% confidence intervals for the ASHRs and RRs were derived with the Spiegelman method.Note 34 The non-Aboriginal population is the reference for RRs.
ASHRs for all-cause acute care hospitalizations were consistently higher among Aboriginal people than among non-Aboriginal people (Tables 1 and 2). For First Nations people living on reserve, the ASHR was 2.6 times that of the non-Aboriginal population (17,042 versus 6,459 per 100,000 population). ASHRs were somewhat lower among First Nations living off reserve (11,190) and Métis (9,535), but still well above those of the non-Aboriginal population (RRs = 1.7 and 1.5, respectively). Among Inuit, the ASHR for all-cause hospitalizations was twice (13,227) that of non-Aboriginal people. Even when pregnancy- and childbirth-related hospital discharges were excluded, patterns were similar.
For First Nations people living on reserve, the highest ASHR was for conditions related to “pregnancy, childbirth, and the puerperium” (Table 1). “Diseases of the digestive system” ranked a distant second, followed by “injuries, poisoning, and other consequences of external causes,” “diseases of the respiratory system,” and “diseases of the circulatory system.” “Mental and behavioural disorders” and “endocrine, nutritional, and metabolic diseases” had the sixth and seventh highest ASHRs among First Nations living on reserve.
Because the rank order of leading causes varied slightly by Aboriginal identity, the order of RRs did as well. The causes with the highest RRs for First Nations living on reserve were “endocrine, nutritional, and metabolic diseases” (RR = 4.9), “mental and behavioural disorders” (RR = 3.6), “diseases of the respiratory system” (RR = 3.3), and “injuries, poisoning and other consequences of external causes” (RR = 3.2) (Table 2, Figure 1). In addition, ASHRs for “pregnancy, childbirth and the puerperium” and for “diseases of the digestive system” were more than twice those of non-Aboriginal people.
For First Nations people living off reserve, ranking of ASHRs yielded the same leading causes of acute care hospitalization as for First Nations living on reserve. The highest RRs among First Nations living off reserve were for “mental and behavioural disorders” (RR = 2.7) and for “endocrine, nutritional, and metabolic disorders” (RR = 2.7). As well, RRs indicating ASHRS at least double those of the non-Aboriginal population were found for “injuries, poisoning and other consequence of external causes” (RR = 2.1) and “diseases of the respiratory system” (RR = 2.1).
Among Métis, the rank order of ASHRS for the leading causes was the same as for First Nations people. And for all leading causes, RRs showed that ASHRs were elevated among Métis relative to non-Aboriginal people. The highest RRs were for “mental and behavioural disorders” (RR = 2.1), “endocrine, nutritional and metabolic disorders” (RR = 2.1), and “diseases of the respiratory system” (RR = 1.9).
Among Inuit, the leading causes of hospitalization were the same as those for other Aboriginal identity groups, but the order of ASHRs (aside from birth-related, which ranked first) was different. Hospitalizations due to “injuries, poisoning and other consequences of external causes” ranked second, followed by “diseases of the digestive system” and “diseases of the respiratory system.” Relative to the non-Aboriginal population, the highest RR among Inuit was for causes relating to “mental and behavioural disorders” (RR = 3.3). RRs for hospitalizations due to “diseases of the respiratory system” and “injuries, poisoning and other consequences of external causes” indicated ASHRs about two and a half times those of non-Aboriginal people (RR = 2.7 and RR = 2.5, respectively).
Results from the 2006 Census–DAD linkage reveal that the leading causes of acute care hospitalization were the same for First Nations, Métis, Inuit, and the non-Aboriginal population. However, ASHRs among Aboriginal people were almost invariably higher than (often double or triple) those of the non-Aboriginal population. After the leading cause of hospitalization (pregnancy-related), the rank order of the most frequent causes of hospitalization varied somewhat across Aboriginal identity groups. This emphasizes the importance of examining hospitalization separately for First Nations people living on and off reserve, Métis, and Inuit.
The results are consistent with provincial patterns of hospitalization rates for First Nations in Western Canada,Note 2 particularly for injuries, diseases of the digestive system, and diseases of the respiratory system. Results are also consistent with rates for Registered First Nations in Manitoba,Note 1 Métis with diabetes in Ontario,Note 35 and Inuit children with lower respiratory tract infections.Note 36Note 37 In addition, the findings are similar to those of area-basedNote 18Note 19 and person-levelNote 13Note 14Note 15Note 16 studies of premature mortality. Elevated ASHRs that resulted among Inuit for “injuries, poisoning and other consequence of external causes” and for “diseases of the respiratory system” were anticipated, based on previous area-based hospitalization and person-level mortality analyses.Note 9Note 18Note 22Note 36Note 37
Because rates were age-standardized, the disparities in ASHRs between Aboriginal and non-Aboriginal people are not due to variations in the age structure of the populations. To a considerable extent, high ASHRs reflect the health status of the two populations, notably, the higher prevalence of poor health,Note 4 chronic conditionsNote 5Note 6Note 8Note 9Note 36Note 37 and unintentional injuries,Note 10Note 12Note 13 and the shorter life expectancy of Aboriginal people.Note 14Note 15Note 16Note 17Note 18 Elevated RRs for some causes of hospitalization could be expected, given the higher prevalence of specific chronic conditions among the Aboriginal population such as diabetes mellitus,Note 4Note 6 asthma,Note 8Note 9, and gallstones.Note 38Note 39
Other factors may be involved in producing higher ASHRs for Aboriginal populations. These might include socioeconomic disadvantage or underlying health determinants that, elsewhere, have included processes of colonization.Note 40Note 41 Additional adjustment or multivariate analysis to account for the role of other factors could clarify suggested associations between higher rates of hospitalization and Aboriginal identity.
Others have suggested that higher ASHRs may reflect less access to primary care services.Note 42 On-reserve First Nations people primarily live in rural areas. While patterns of health care use in rural versus urban Canada are due, in part, to differences in health determinants,Note 43 the availability of health services also plays a role.Note 44Note 45Note 46
This study has a number of limitations that should be considered in assessing the findings.
ASHRs should not be interpreted as representing the prevalence of specific health conditions. Rather, results represent health conditions that require acute care hospitalization. Furthermore, these findings reflect the disproportionate health burden borne by Aboriginal people as a result of disparities in accessibility and availability of health services.
Results and conclusions pertain only to the Aboriginal identity groups analyzed in this report. People not enumerated by the 2006 Census were excluded, notably, residents of 22 Indian reserves and settlements.Note 47 Validation of the linked files used in this study showed lower coverage of populations in the territories and of younger age groups, characteristics pertinent to the Aboriginal population.Note 25 Eligibility rates for linkage to the DAD were lower among people who identified as Aboriginal, individuals of lower socioeconomic status, rural/farm residents, and residents of Nunavut and British Columbia. As a result, Aboriginal people are underrepresented, and their ASHRs are likely undercounted. In particular, a downward bias affects eligibility to link to the DAD for First Nations people on reserves in Ontario, Alberta, and British Columbia (data not shown). Comparisons between First Nations living on and off reserve should consider this bias.
Another major shortcoming is that hospitalizations in Quebec were not available.
No adjustment was made for deaths of study cohort members; therefore, populations at greater risk of death within follow-up are underrepresented. Moreover, Aboriginal people have a greater risk of premature mortality than do non-Aboriginal people,Note 15Note 16Note 17Note 18 so it is possible that ASHRs for Aboriginal people are artificially low. This decreased “eligibility” for hospitalization may have created a downward bias in the compilation of hospitalization rates for Aboriginal people.
The analysis pertained only to acute care hospitalization; findings are not generalizable to other types of hospitalization such as day surgery and psychiatric services, or to health service use generally. In addition, after 2005, mental health hospitalizations in Ontario were not comprehensively reported to the DAD, but instead, to the Ontario Mental Health Reporting System. Therefore, acute care mental health hospitalizations are underreported in this study.
A cautionary note is warranted about ASHRs for “diseases of the circulatory system,” which were only slightly elevated for Aboriginal people. Cardiovascular diseases have been identified as important contributors to person-years of life lostNote 14Note 15 and a growing health risk for Aboriginal people.Note 48Note 49 Nonetheless, previously reported hospitalization rates for circulatory system diseases among First NationsNote 2 also did not rank high.
Linkage of the 2006 Census and hospital administrative data makes it possible to report acute care hospitalization for all Canada except Quebec for First Nations living on and off reserve, Métis, and Inuit. For each Aboriginal identity group, the diagnoses that most frequently resulted in hospitalization were birth-related, digestive diseases, injuries, respiratory diseases, and mental and behavioural disorders. ASHRs were almost always higher for Aboriginal people—often double to three times those for non-Aboriginal people. However, the ranking of causes according to the extent to which ASHRs differed from those of the non-Aboriginal population varied by Aboriginal identity group.
This information is relevant to health policy and service delivery planning related to the health conditions that place Aboriginal people at increased risk of hospital admission. Future analyses could use the linked 2006 Census information to adjust for confounders beyond age to model associations between demographic, socioeconomic characteristics, and location of residence in order to explain differential hospitalization.
This study was sponsored by the First Nations and Inuit Health Branch (FNIHB), Health Canada. The authors acknowledge FNIHB for their financial support and for their input and feedback on the conception and design of this study, analysis and interpretation of the data, and review of earlier versions of the manuscript.