Introduction

As part of MCHP’s efforts to analyze family structure, we have developed a methodology to identify spouses and children in Immigration, Refugees and Citizenship Canada (IRCC) Registry Permanent Resident Data (IRCC-PR), using techniques similar to the %RegRelation macro which finds these relationships in the Manitoba Health Insurance Registry. The %ircc_relation macro links family members who arrived as landed immigrants in Canada on a shared application. They are linked via a common application number, and the individual relationships within the applicant family are established by the family status variables for each member of the family. Unlike the insurance registry, which can provide a longitudinal view of the dynamics of family structure, tracking not only marriages and children, but also new births and separations over time, the relationships identified by IRCC-PR data are more of a one-time snapshot, taken when the family immigrated to Canada.

Data Sources

The data used for this concept came from the following Repository databases:

• Immigration Refugees and Citizenship Canada (IRCC) – Registry Database Permanent Resident data
• Manitoba Health Insurance Registry data (used for comparison)

Limitations

The following limitations were identified during development of this concept:
• Longitudinal relationship structure dynamics are not captured in IRCC-PR data. This would require combining these data with data from other registries.
• The two family status variables in legacy landing data lack specificity, and do not allow one to make a distinction between different dependent family members. The vast majority of dependents are children, but in some cases a dependent may be some other family member. This can lead to some ambiguity in family structure analysis, when the assumption is that dependents are children of the principal applicant and/or spouse of the applicant. However, we found very high levels of agreement in family relationships identified by individuals in both the insurance registry and the IRCC-PR data.

Methods - Using the SAS® macro

The SAS® macro can be invoked as follows:


%ircc_relation(cohort=, acqdt=, acqset=, debug=, fstlanddt=, lstlanddt=, lib=);

It uses named parameters, so you only need to specify the ones that do not define the value you want to use by default.


cohort specifies dataset with scrphin values for whom to extract data, or you can use cohort=Canada to get all landing records, even if not linked by scrphin. By default, all linkable scrphin values are read.
acqdt specifies the exact acqdt value to use in the IRCC datasets. The default value is latest current date, 25aug2025 as of this writing.
acqset specifies the exact acqset value to use in the IRCC datasets, which overrides the acqdt value if one is given. Only one of acqdt or acqset should be specified.
debug specifies (via 1 or debug ) whether to output debug information. By default, no debugging data is shown. debug=notes will leave SAS options as they are, usually allowing notes in the listing. debug=stats is like debug=notes, but the macro will also output stats as it does in debug mode.
fstlanddt specifies the earliest landing_dt value to be read from the data. There is no lower limit by default, so the earliest dates available in the data will be used.
lstlanddt specifies the most recent landing_dt value to be read. No upper limit by default.
lib specifies an alternate library name from which the IRCC-PR datasets are read, rather than the registry library.

The macro produces a PRRELCA dataset if cohort=Canada was specified, yielding relationship data for all immigrants to Canada, and a PRRELMB dataset in any case, for linked immigrants to Manitoba. These datasets contain variables similar to the ALLREL dataset produced by %RegRelation: RELATION and RELTYPE show relationship type, RELUCID shows the IRCC UCI value for the identified relative, and RELPHIN shows the scrambled PHIN of the identified relative if linked in the Manitoba insurance registry.

Sample SAS code for this analysis will be provided here shortly.

Results

In a comparison between the %ircc_relation and %RegRelation macros, we found that for Manitoba immigrants whose relationships were identified by both macros, there was a 99.7% agreement on relationship type, and a 99.23% agreement on both relationship type and sex of both individuals identified in the data from the two sources. When comparing immigrants identified only in non-legacy landing data, agreement was even better: 99.8% on type, and 99.77% on both type and sex, due to the better specificity of family status in the post-1999 data. However, the high degree of aggreement even in the legacy data suggests that the vast majority of dependents identified by the legacy BASE_FAMILY_STATUS_ENG variable are in fact dependent children of the principal applicant.