In for the long haul

McGill hosts international conference on longitudinal studies

Hannah Hoag | International experts, government policy advisers and data managers, and McGill researchers filled the Faculty Club Ballroom January 26 and 27 to discuss the future of Canada's long-term social and health surveys.

The conference is part of an ongoing effort to improve the design of, and access to social statistics in Canada. Conference organizers, including McGill's Céline Le Bourdais, Canada Research Chair in Social Statistics and Family Change, asked experts to present their assessments of past Canadian longitudinal studies and provide direction for their future.

"The idea was to bring experts from other countries," said Le Bourdais. "Let's talk about everything: content, how do you plan, finance. How do you link to administrative data? How do you ensure success? What were the failures?"

The conference was funded by Statistics Canada, the Social Sciences and Humanities Research Council and the Canadian Institutes for Health Research.

Longitudinal surveys can tell a lot about a population and how it is changing. A survey that polls individuals only once provides only a snapshot of the population at a particular time. A longitudinal survey is more like a feature-length movie. When researchers conduct a longitudinal survey, it can go on for years, even decades. One British survey has been following a group since 1946.

In Canada, longitudinal surveys conducted by Statistics Canada have been used to gather information about child development, the challenges youth face making the switch from school to work, the adjustments made by new immigrants to Canada, emerging public health issues, and changing income and work conditions.

John Frank, scientific director for the Institute of Population & Public Health, part of the Canadian Institutes of Health Research, framed the importance of the conference in terms of health and economics. "Some diseases are much less likely to be killers than they were 30 years ago, like heart disease. But other things are replacing those causes of death and some of them will have enormous costs of care and may be preventable," he said. "We can use what we know from these studies to generate projections of disease rates and their burden on the system."

Beyond the financial burden of disease, these surveys can indicate the number of professionals, such as radiation oncologists, we may need by 2030, or whether we'll need a more robust homecare program. They help policy makers address problems and craft solutions before they're needed.

The conference also brought to light issues around access to data and the design of surveys so that the data are useful to many researchers. "After a decade of experience, it's time to start taking stock," said Garnett Picot of Statistics Canada.

Survey design is key to producing an informative study. For Le Bourdais (who studies family dynamics), a survey that interviews only one person in a household doesn't provide information about the family. "These days family cannot be put under a single household, with divorce and recomposition, you have children moving between households, yet we don't have any information on what is connecting the household and what resources are moving between the households."

By bringing together all those involved to identify the questions that need to be answered and the best tools for success, "we will be able to better design surveys to answer the questions we are facing now," said Le Bourdais.