Appendix 1 — Population estimates

In summary, the population estimates for the health regions in this publication were prepared as follows. First, the 1996 population estimates were based on the 1996 Census, adjusted for net undercoverage. Secondly, for non-census years, the Census-based population estimates were adjusted for changes in the population, primarily using administrative data. Lastly, subprovincial/subterritorial population estimates were controlled to sum to the population estimates at the provincial/territorial level prepared by Statistics Canada.

Statistics Canada methodology

For health regions in all provinces/territories except for Quebec, Alberta and British Columbia, the method is that of Statistics Canada, described below. These estimates are based on census data and on population estimates by census division (CD) produced by Demography Division.

Population universe

The population included in these estimates is based on the 1991 and 1996 Census of Canada. The universe, therefore, corresponds to the census universe, which includes Canadian citizens and landed immigrants with a usual place of residence in Canada (or abroad on a Canadian Forces base, attached to a diplomatic mission or aboard merchant vessels) and non-permanent residents. For a complete description of the census universe, refer to the 1996 Census Dictionary, Statistics Canada Catalogue no 92-351-XPE.

Census data are adjusted to take into account net census undercoverage and incompletely enumerated Indian reserves. Components of population change occurring between census day and July 1, 1991 or 1996 were also taken into account.

Method for deriving demographic estimates by health regions

Population estimates at the census division (CD) level are produced annually by Demography Division, using the component method, which accounts for changes in the number of births and deaths, as well as intraprovincial, interprovincial and international migration. The CD population estimates are prorated to the provincial population estimates.

Demographic estimates by health region are derived from these estimates using the following steps:

Calculation of conversion factors:

Health Statistics Division has created a file linking enumeration areas (EAs), census subdivisions (CSDs) and census divisions (CDs) for each health region.

The 1996 Census population by CDs (EAs or CSDs in case of split CDs) was adjusted for late changes in geography and for partially enumerated Indian reserves to ensure EAs added up to the CD.

In cases where health regions split CDs, the proportion of population in each CD split is used to derive 'conversion factors' in order to allocate the CD level estimate to the appropriate health region.

Application of conversion factors to CD demographic estimates:

Conversion factors were applied to CD population estimates for 1995 to 1998 allowing conversion of CD population estimates into HR population estimates. The same approach used to obtain the total population estimates was used to establish age and sex estimates by HR.

Evaluation of the method

To assess the quality of demographic estimates by HR, two evaluations were performed. First, an estimate was produced for July 1, 1996 with a postcensal CD demographic estimate based on the 1996 Census which was compared with the 1996 Census counts. This comparison indicated that the conversion factors were providing results comparable with the census counts by HR.

Second, the same method was also used to generate estimates for 1996 by HR from a postcensal CD estimate based on the 1991 Census. The results were compared with the actual 1996 counts as derived from the 1996 Census, thereby allowing an evaluation of the accuracy of the method over a five-year period. The results indicated that the methodology adequately accounted for the demographic changes.

For more information on the method used to produce health region or CD estimates, contact Daniel Larrivée of the Demography Division of Statistics Canada at (613) 951-0694 or e-mail daniel.larrivee@statcan.gc.ca.

Provincial methodologies

For health regions in Quebec, Alberta and British Columbia, the administrative files used for adjusting the 1996-based population estimates for non-census years differ from those used by Statistics Canada in the description above. For these provinces, the health region population estimates were supplied by the Institut de la statistique du Québec, Alberta Health and Wellness, and BC Stats.

Quebec

Estimates by health regions for Quebec, prepared by the Institut de la statistique du Québec, are based on projections derived from different data sources including the registration file for insured people of the "Régie de l'assurance-maladie du Québec", birth registration and Statistics Canada CD estimates. The estimates are initially prepared for the Centres locaux de services communautaires (CLSC) territories and then aggregated into Health Regions.

For information on the methodology used to derive the health region population estimates in Quebec, for 1995 to 1998, contact the Institut de la statistique du Québec, Direction des statistiques sociodémographiques at (418) 691-2406 or demographie@stat.gouv.qc.ca.

Alberta

Statistics Canada used population estimates by health regions prepared by the Alberta Health and Wellness department. These estimates are derived from the Alberta Stakeholder Register File containing information on individuals who are entitled to basic medical services under the Alberta Health Care Insurance Plan (AHCIP). Population growth as derived from this file for each of Alberta's health regions was then applied to Demography Division's estimates for 1996 to account for annual changes in the population. The estimates were further adjusted to correspond to the provincial estimates produced annually by Demography Division.

For more information on Alberta's estimates, refer to Calculating Demographic and Epidemiological Quantities in Alberta by Geo-Political Areas, Health Surveillance, Alberta Health and Wellness, March 2001 or contact Donald Schopflocher of the Health Strategies Division, Health Surveillance in Alberta at (780)422-4630 or e-mail at donald.schopflocher@health.gov.ab.ca.

British Columbia

The methodology used by British Columbia to derive small area populations by gender and age group is divided into two parts.

A Regression Approach, specifically the Difference-Correlation Method (DCM), is the primary method underlying the sub-provincial population estimates. A secondary method, known as Proportional Allocation (PA), is also used to estimate the population for certain classes of areas. Both these methods use information derived from a set of indicators obtained from administrative files that are symptomatic of regional population changes.

In essence, the British Columbia small area population estimation model works as follows. Beginning with the most recent federal census (in this case the 1996 Census of Canada), each region's share of provincial population is adjusted up or down according to the current share of the provincial total of a weighted combination of residential hydro connections and/or Old Age Security recipients. Estimates of the population living in municipalities along with that portion of the population living outside the municipality but within the regional district (i.e., unorganized area), are controlled at the provincial level by a British Columbia population estimate prepared by Statistics Canada. Regional district population estimates are derived by summing the municipal and unorganized area population estimates. Local health areas are also controlled at the provincial level, and in order to ensure consistency, the local health area population estimates within each regional district are then tied to the regional district population estimates.