Good morning.
My name is Dr. Cordell Neudorf. I'm the current chair of the Canadian Public Health Association and a local medical health officer in Saskatoon. I'm here with our CEO, Debra Lynkowski.
CPHA represents the interests of public health professionals across the country, many of whom work at the local level in the 115 public health departments in Canada.
In addition to our presentation today, we've submitted a brief to the committee on the impact of cancelling the long-form census on health equities and public health.
The primary factors that shape the health of Canadians are not just medical treatments or lifestyle choices. About 80% of what determines our health are things like the socio-economic, physical, and political environments in which we live, work, and play. Research has also found that the quality of these health-shaping environments and conditions is very strongly determined by decisions governments take on a range of different public policy domains.
The information that's been gathered and made available through the long-form census has been essential to understanding the health of our communities and to designing and targeting programs and policies to improve the health and well-being of Canadians at that small-area level, particularly for those most vulnerable and most at risk.
The shift to a voluntary survey like the NHS is of particular concern to the public health community in Canada for a lot of reasons, but primarily it's because the long-form census is really the only reliable, valid, and historical source of this foundational demographic data, down to areas as small as sub-neighbourhoods for specific cities, or for provinces. Alternate local data sources just don't have that historical aspect and the sample size to get down to that level. They don't have the same kinds of response rates as the census has provided.
There's broad agreement amongst statisticians and social scientists that the voluntary survey won't provide data of the same quality and reliability compared to data that was gathered through the mandatory form over the past years.
A voluntary survey basically means that some people answer and others do not, and more people do not answer when it's voluntary. Those people tend to be poor and from marginalized and immigrant communities because of barriers such as language, literacy, disability, and, quite frankly, just the complexity of their lives. First nations, Métis, and Inuit are already underrepresented in current data, and a voluntary move would exacerbate this problem.
The scale and location of the non-response biases can't be completely assessed ahead of time. The estimates from other attempts to generate data in this way have found that in order to maintain statistically accurate analysis, data would have to be generated at a higher geographic level than the former census model. Basically that means it becomes unusable at the neighbourhood level because we can't drill down to that level with confidence and make the kinds of decisions we want to on targeting programs and policies to sub-populations.
We use this data at a local level in public health to generate things like our annual health status report, which drills down, using census data as a model, on subgroups that have certain demographic profiles to see what kinds of differential health outcomes and health disparities are being seen at that sub-neighbourhood level.
Accurate comparisons to past data are essential to measure whether changes we've made to certain health or social policies are having the intended effect of improving the conditions in which Canadians live and work.
The conventional census model provided a critical foundation for the generation of data from other surveys as well. It's basically used to establish sample frames and ensure accurate representativeness of the population. I would compare it to...the existing high-quality data is building your foundation for other surveys on solid bedrock, compared to voluntary surveys that you are now using as a benchmark; it becomes a less stable foundation, basically, on which to design other surveys.
I have a few examples of how we're using this at a local level.
Ongoing research and monitoring is done in a lot of cities to track the depth and breadth of health disparities between areas in the city with higher levels of deprivation compared to those with lower levels. We use an index to do this, which is comprised of data from the long-form census: income, education, employment, and various social factors. Many years have been spent in trying to set up this index in a way that's valid and reliable across the country.
We can then generate aggregate, small-area-level data that we can put our health data on to track how the differential health outcomes play out at a far less intrusive level than if we drilled down to an individual level.
In using this type of study we found health disparities. For example, infant mortality rates in low-income areas of the city were 448% higher than in high-income areas. We're instituting programs to try to intervene at this level. But if we institute major program or policy changes and don't have an ongoing reliable set of data that's collected in the same way, it will be hard to measure the impact of these programs and services. There tends to be a fairly small impact over the short term.
Our recommendations include three options to maintain the continuity of decades worth of accurate and reliable data.
One is to make the NHS survey mandatory through a legislative mechanism like a private member's bill. Two is to postpone the census for a year to allow time to examine and resolve the issues pertaining to the mandatory long-form census. Three is to reinstate the former census model for the 2016 census, with public consultation about mechanisms to maximize compliance.
A portion of the funds allocated should be used to encourage Canadians to respond to this survey to increase awareness about how the data is used and the measures in place to protect personal information.