Concept: Francophone Manitobans Cohort - Methodology

Introduction

This concept provides information on how Francophone Manitobans were defined and identified for research in the MCHP deliverable Health and Healthcare Utilization of Francophones in Manitoba by Chartier et al. (2012). The concept describes the methods used to identify Francophones, including the data sources and the type of information that is available in the data. The linkable survey and administrative databases available in the MCHP Data Repository provide the ability to develop a Francophone cohort that allows MCHP researchers to investigate, analyze and report on the health and healthcare utilization of Francophone Manitobans compared to other Manitobans.

Information in this concept is taken directly from the Francophone deliverable and provides links to more detailed information available in the report.

Definition of Francophone

Statistics Canada has derived a variable about the first official language spoken, based on a series of survey questions about language. This an important variable because it includes respondents whose first language is not necessarily French, but use French to access services in the Canadian context (Forgues, E. & Landry, R., Defining Francophones in Minority Situations: An Analysis of Various Statistical Definitions and their Implications, 2006) - accessed September 6, 2012).

From Statistics Canada, a commonly used definition for the term Francophone is an individual who reported French as their mother tongue, who reported that French was the first official language spoken, or who reported that French was the language most commonly used in their home. This definition includes people who learn French as children or who utilize it at home. It also includes a growing number of immigrants who may have a language other than either official Canadian language as their mother tongue, but who communicate more effectively in French than in English.

Data Sources Used to Identify Francophone Manitobans

The MCHP Data Repository was investigated and reviewed in order to develop an operational definition for Francophone Manitobans that could be used for this research. In order to do this, MCHP looked at language-related indicators and found two major sources: 1). the 2006 Canadian Census survey and 2). linkable administrative databases in the MCHP Data Repository containing language-related indicators.

1. 2006 Canadian Census Survey

The 2006 Canada Census data provides summary information about the demographic, social and economic characteristics of the Canadian population. Using the Statistics Canada definition above, MCHP was able to identify 50,250 Francophones in Manitoba from the 2006 Census data.

IMPORTANT NOTE: The Canadian Census data does not allow the ability to link Census data with individual level data in the MCHP Repository databases, so we can only use aggregate-level descriptive information from the 2006 Census to describe the Francophone Manitoban population.

The following links to information in the Francophone deliverable describe the profile, distribution and socio-demographic characteristics of the Francophone population in Manitoba compared to the rest of the Manitoba population, according to data from the 2006 Canadian Census:


• Figure 2.1 provides a Map of French Language Services in Manitoba
  
• Figure 2.2 provides a map of the Distribution of Francophone Population by RHA District in Manitoba
  
• Table 2.1 identifies the Distribution of Francophones in Manitoba,
  
• Figure 2.3 compares Francophones with all other Manitobans in an Age and Sex Profile of Manitoba illustrated using a population pyramid
  
• Table 2.2 provides a Description of Francophones in Manitoba compared to all other Manitobans on gender, age categories, employment status, education levels, average income values, and the percent that are immigrants and aboriginal, including Metis.

2. Linkable Administrative Databases in the MCHP Data Repository

To develop a Francophone cohort, all administrative databases in the MCHP Data Repository were reviewed for language-related indicators. This investigation revealed three surveys as well as a number of health and education administrative databases that contain language-related indicators.

The three surveys used in the identification of Francophone Manitobans included:


1. several versions of the Canadian Community Health Survey (CCHS),
   
2. the National Population Health Survey (NPHS), and
   
3. the Manitoba Heart Health Survey (MHHS).

In the surveys, a Francophone was defined as a respondent who reported French as their mother tongue, who reported that French was the language most commonly used in their home, or whose first official language spoken was French. For the Manitoba Heart Health Survey (MHHS), one question was utilized: What language did you first speak in childhood?

In the health and education administrative databases, a Francophone was defined as someone who indicated French as a preferred language for services, whose maternal language was French, or who attended a facility where French is the main language used (i.e., attending a school in the Division scolaire franco-manitobaine (DSFM), certain child care centres, and personal care homes).

Several health and education databases contained language-related indicators. These databases and the type of language related-data used to identify Francophone Manitobans included:


• Education - individuals who were in the français (FL1) program at one point in time in their schooling. This program offers all courses in French and is intended for students who are fluent in French.
  
• Red River College - individuals who previously attended a high school where only the français (FL1) program is offered.
  
• Manitoba Immunization Monitoring System (MIMS) - individuals who indicated that they wanted their correspondence in French.
  
• Child Care - children whose parents indicated that French was their preferred language or who attended a facility that offered francisation or who attended a facility that is part of the francophone school division, the Division scolaire franco-manitobaine.
  
• Early Development Instrument (EDI) - children, who according to the Kindergarten teacher, had French as their primary language. (Note that in schools other than schools from the Division scolaire franco-manitobaine, teachers are sometimes unaware that the child's primary language is French.)
  
• Personal Care Home - individuals who at one point in time were residents in Foyer Valade (St. Vital) or Foyer Youville (Ste. Anne), or their primary language was French.
  
• Home Care - individuals who indicated that French was their primary language.

Method of Developing The Francophone Cohort

The Francophone Cohort was developed from individual-level, de-identified linked data from surveys (CCHS, NPHS and MHHS) and health and education administrative databases in the MCHP Data Repository. These data sources are described above. The administrative databases in the MCHP Data Repository can be linked together using methods that do not identify specific individuals, but allow relevant data to be combined for research purposes. For more information on this linkage process, please read the record linkage / data linkage glossary term.

The Francophone cohort development involved a two-step process.

1. Francophones Identified Through Language Indicators from Surveys and Administrative Databases

The first step was a review of the surveys and health and education administrative databases for language indicators (process described above), in order to identify Francophone Manitobans.

Initial results based on the language indicators in the surveys and administrative databases identified 19,396 individuals for the Francophone cohort. Tables identifying the specific numbers found in each of the surveys and MCHP databases are provided in Table 3.1 and Table 3.2 in the Francophone deliverable.

2. Francophones Identified Through Family Linkage

The second step in developing the Francophone cohort involved identifying family members, or first degree relatives of the 19,396 Francophones identified through the language indicators in the first step of the process. First degree relatives included parents, children, and siblings. For Francophones born before 1952, spouses were also added. We decided not to include second degree family members (grandparents, uncles, cousins) because of the greater uncertainty about their language group. See the Cautions/Limitations section below for more details on "levels of certainty".

The first degree relatives were identified at Manitoba Health through the Health Registration Numbers contained in the Registry. The Registry contains longitudinal demographic histories for every individual who has registered for the Manitoba Health Services Insurance Plan (MHSIP) since 1970. From 1970 to 1984, individuals were found using a combination of family registration number (REGNO) / (REGNO_CODE) , date of birth, and sex. From 1984 and onward, an individual Personal Health Identification Number (PHIN) was assigned to each provincial resident.

Typically, registration numbers are assigned to the male head of the family, and the remaining family members (spouse and children) use that same registration number. When an individual reaches eighteen years of age, he or she receives his or her own family registration number. For females who marry and report this event to Manitoba Health, they are given the choice of changing to the registration number of their husband or keeping their own number.

Family linkage is more successful for younger individuals, because of the way REGNO was assigned. For example, it is only possible to identify siblings in the 2008 population file who are 56 years and younger. A person must be under 18 in 1970 or later in order to ever be identified as a dependent, which puts a lower limit on birth year of 1952.

Different methods were used to link family members, as described below:


• The Registry begins with families as of 1970 (and will only include children born in 1952 and later).
  
• Mothers and fathers are identified through different methods. Mothers were found through hospital birth records for Manitoba births after 1970. Fathers were identified through the registry data alone, specifically the male family head of a child's birth REGNO.
  
• Children were identified by finding dependents associated with each male and female family head. Single-parent families have the children assigned to whichever parent can be identified.
  
• Siblings were not identified directly, but rather as being dependents of the same family head at birth. Siblings were defined as those sharing a female family head, for those having one identified, and male family head for those who do not. Female family head were utilized first because 98.8% of children had one, while male family head was only found for 78.3%. Single parents are more often female; and in the event of divorce the mother usually gets custody, either legally or practically.
  
• Spouses were identified through a "spouse pair flag" found the in the Registry. These flags exist for Spouse or Common-Law Spouse who report their marriage to Manitoba Health, so some spouses will not be found.

For more information on MCHP's work related to family linkage, please read the Family Structure History concept.

The Final Francophone Cohort

The family linkage process found an additional 27,558 Francophones that were added to the cohort. The Francophone Cohort contained 46,954 individuals after the family linkage was completed. Of this number, 40,600 remained in the cohort for the research study period 1998-2008.

It is acknowledged that not all of the family members included in the Francophone cohort are French-speaking or would consider themselves Francophones. It is estimated that about two-thirds of the additional members would be French speaking and those who are not French speaking can be considered part of the Francophone community through family ties.

Development of the Francophone cohort allowed MCHP researchers to investigate, analyze and report on the health and healthcare utilization of Francophone Manitobans. Table 3.3 from the deliverable provides the Distribution of the Francophone Cohort by Regional Health Authority (RHA). Table 3.4 from the deliverable provides the Demographics of the Francophone and Matched Cohort Groups used in the comparison of health and health services in this research.

For more information on the methods used to create the Francophone cohort and the types of statistical methods selected for use in this research, please read Chapter 3: Methods of the Health and Healthcare Utilization of Francophones in Manitoba deliverable.

Validation of the Francophone Cohort

To validate the data, we compared the health indicator rates of the Francophone Cohort and the smaller survey sample of Francophones. We found no differences between the rate ratios of these two data sources in 75 of the 76 indicators. This increased our confidence that the results found in this study are valid.

For more information on the validation process, please read the section titled Comparing Results from the Francophone Cohort and Matched Cohort to the Representative Survey Sample in the deliverable. The glossary term Statistical Testing in the deliverable describes the statistical methods employed to compare the health-indicator rates.

Cautions / Limitations

• While the Francophone definition from the administrative data is more uncertain than a definition obtained through direct questions in the survey sample, it is reasonable to assume that individuals who requested services in French or attended a facility where French was the main language spoken were likely Francophones. Many Francophones do not request French services or attend these facilities, so it was not possible to identify all Francophones using the language-related indicators in the administrative databases.
  
  
• Levels of Certainty - while some issues were identified in the development of the cohort, MCHP categorized three levels of certainty in our identification of Francophones.
  
  
  1. The highest level of certainty is when an individual has met the definition of being a Francophone, either through a survey or through one of the administrative databases.
     
  2. The second level would include people who are very likely to be Francophone as they attended a school in the Francophone School Division or lived in a predominantly Francophone personal care home.
     
  3. The third level includes people where there is a reasonably high probability of being a Francophone as a result of being a first degree relative of someone who was identified in the first two categories.