Feasibility Study on the Use of Provincial/Territorial Medicare Records for Measuring the Level of Inter-provincial and Inter-territorial Migration
4. Conclusion and recommendations
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.
At the beginning of this report, we listed what criteria were needed for a data source to be feasible or viable as a proxy for estimating inter-provincial net migration (Table 1).
An obvious and serious problem in the use of provincial health records has to do with the lag effect – designation of which month for the provinces and territories to put data into is not necessarily the same time as when new residents had arrived in the province or territory. The month assigned to the new resident may relate to the date that they were eligible (normally 3 months after arrival), or based on processing periods (e.g. British Columbia in October and November 2005). To re-iterate, when determining time frames for when people move into another province or territory, the utilization of provincial/territorial administrative data makes it quite difficult to determine that time line since the dates on files normally correspond to the date data have been processed and not a reflection of when people have arrived.
Annex F provides a concrete example of the inconsistent, variable, erratic and unpredictable lag effect between processing, enrolment, eligibility and arrival dates using the Prince Edward Island micro-data for new residents. The enrolment processing periods used are April 2006 and March and April 2007.
This lag effect phenomenon is also evident in the greater-than-one ratios found with the quarterly numbers – a large influx of new residents in provinces in the summer months mean that more data have to be processed which may cause a backlog and not be processed at the same time as people applying for health cards who came from another province or territory. Processing maybe done at a later date especially with data entry staff going on summer vacation; hence, the ratio greater-than-one for the October to December period when new registrants are less and there is time to enter in the backlog of records (as what had happened in British Columbia in 2005). This pattern is, on the whole, consistent if one were to examine those ratios with a value over 1.0 in Table 5. As well, there are those who are new to a province/territory but do not bother getting a health card so we would have no record that they are there unless they have a reason to seek medical attention and are discovered to not have a valid card.
The critical detail of out-migration is not available. Manitoba does give us cancellations (equivalent to out-migrants) but they are not complete and would not correspond exactly with the timing for the in-migrants (in order to calculate net migration).
Most provinces (except for Quebec) and territories provide reports to each other as to the number of new residents who had arrived into their province/territory from other provinces/territories. It may prove useful to tap into this source to get a measure of out migration. However, the timing there is once more an issue when looking at sub-annual data and trying to calculate net migration figures.
Based on this present analysis, it would not be feasible or viable to substitute the provincial and territorial health registration records for our present data sources at this time. As was stated by Rémillard,
« it should be kept in mind that the health insurance file is basically an administrative file and not a population register. It is designed to meet certain needs which are not primarily demographic in nature. 13 …
For the health insurance files to be fully usable, it will be necessary for the players involved to agree on definitions, the population covered, standards, the format of the files, and procedures. Such agreements are not easy to conclude, especially in fields that are under provincial jurisdiction 14 »
Progress has been made (e.g. more provinces and territories are now providing us with data) but there are still gaps in coverage, completeness, consistency, timeliness, reliability, details, uniformity and accuracy.
As stated earlier, it would also be useful to check the quality of our population estimates through having access or a link to all of the provincial/territorial health care registrations on a monthly basis.
Another useful piece of information on the health care registration file would be a variable indicating when a resident first arrived in the province or territory.
These Medicare data though can be used for data quality evaluation but not as a replacement for the CCTB.
To summarize the criteria and the issues:
- Eligibility: Many provinces exclude specific populations and the exceptions vary from one province to the next.
- Completeness: The level of completeness varies from under to over estimation from one province to the next. We do not yet receive data from all jurisdictions.
- Difficult to determine the lag factor (e.g. Prince Edward Island [Annex F] as a concrete example of month of enrolment versus month of arrival versus month of processing).
- Some data very late in arriving (e.g. Yukon).
- Some provinces report 0 migrations on a monthly basis, rendering its quality dubious.
- Some administrative processes lead to "clumping" as registrants.
- The lag effect on the measurement of mobility and Health Care file registrations varies and although it appears to be 3 months it is not accurately measurable (see Annex F).
- Some provinces, by their eligibility rules, delay the actual registration of migrants way beyond the 3 month lag.
Level of detail:
- The serious lack of out-migration data make it not a suitable replacement for the CCTB.
- When we have out-migration figures as in the case of Manitoba the results are not of good quality.
- Unknown origin of new residents for Saskatchewan results in undercoverage of inter-provincial in-migrants.
Lack of individual ages makes it difficult to compare with the CCTB to get specialized cohort groups (e.g. 0 to 17).
13 . Rémillard, p. 31.
14 . Rémillard, pp. 31-32.
- Date modified: