Today, the Canadian Institutes of Health Research (CIHR) announced funding for the establishment of the SPOR Canadian Data Platform. To mark the occasion, we sat down with Dr. Ted McDonald to discuss the unique challenges and opportunities that administrative health data present in the Maritimes, and the exciting new plans for connecting data across the country.
Can you tell the readers a bit about yourself?
I’m a Professor of Economics at the University of New Brunswick—and a Maritimer myself. My Master's and PhD are both from the University of Melbourne in Australia, and my first job was at the University of Tasmania. After many years overseas, this job came up and we moved to Fredericton. That was back in 2001, so we’ve been here quite a while now. Over the last nine or ten years, I’ve been focused on developing the data infrastructure that was needed to support better health research in the province.
What sparked your interest in this area of work?
There’s a great need for data and research infrastructure in the Martimes, particularly in New Brunswick where we have some of the biggest challenges including an aging population and very high rates of chronic disease coupled with fiscal constraints—no province needs evidence-based policy and evaluation more than New Brunswick. And because we are a small province, we recognized early on with the Maritime SPOR SUPPORT Unit that we needed to develop economies of scale across our provinces—collaborating, sharing expertise, pooling resources.
Let’s step back a bit, what kinds of data are you dealing with?
Administrative data are data that are collected by government or public bodies in the course of delivering a government service. The main types of data would be information on health care use, education outcomes, or provincially-funded social services. Every time a person receives a service or a service is provided by government, information is collected on the person who receives the service and what the service is—that’s what we mean by administrative data. Almost every province has a provincial data platform that receives and provides administrative health data for research.
What are some of the challenges in working with administrative health data?
There’s a lot of unrealized potential here. In New Brunswick, there’s a ton of information collected, but it’s often very difficult to access and link various data. This is a problem for provision of health care at the individual level but it also creates challenges every time the government needs to make a policy decision. Consider New Brunswick’s Early Learning Centres, which are a form of subsidized, structured child care. That kind of policy has implications not just for early childhood outcomes but also subsequent education outcomes and then transitions into training and the work force, as well as health effects for children and parents. Such a policy initiative has a lot of different dimensions and potential impacts that can really only be looked at when you are able to combine data from multiple sources. That principle was the genesis for setting up the New Brunswick Institute for Research, Data and Training (NB-IRDT). From conception, NB-IRDT was intended to be a repository for all government data. It’s been a very long, challenging and time-consuming process—negotiating the legal, privacy and ethics requirements, data documentation, and all the practical arrangements—but we’ve made a lot of progress.
Can you give an example of how these data resources are important for health policy?
One recent project that is nearing completion is called CHIP–the COPD Health Information Platform. Chronic Obstructive Pulmonary Disease (COPD) is the number one cause of hospitalization in New Brunswick, after childbirth. It’s reaching a crisis point in terms of costs and consequences for the individuals, their families and the health system. An important dimension for diagnosis is pulmonary function testing, but those data are not actually linked to other health data in any systematic way for research and evaluation—that’s what CHIP did. With resources from a private sector funder, AstraZeneca, and the New Brunswick Health Research Foundation (NBHRF), we went out to each of the nine pulmonary function clinics in New Brunswick to collect data on all pulmonary function testing, which was then linked securely and anonymously with the individuals’ other health records including hospitalizations, pharmaceutical use, and ER visits. Now we’re able to study patterns and outcomes around COPD in multiple dimensions.
This seems like it would be important from a patient perspective too.
Absolutely. We can answer questions that are crucial for patient care, like: Of the people who receive pulmonary function testing and have been diagnosed with COPD, are they getting appropriate treatment for that condition? We can start to better understand the characteristics in terms of patients’ COPD experience, socio-economic determinants, variation in outcomes, as well as having a framework for rigorously evaluating new approaches, treatments and therapies for COPD. This is just one example of the possibilities when you have the infrastructure in place to be able to receive linked data and make them available for research. And, I think that people expect that the data that are being collected on them in the course of getting health care are being used to improve their own healthcare and the healthcare of their fellow citizens, while respecting privacy. They may assume that more can happen then is actually happening because of the obstacles to data sharing—technical, legal, procedural. And, that gets us to the SPOR Canadian Data Platform.
Yes, tell me a bit about the SPOR Canadian Data Platform.
Just as there are challenges within provinces and sharing data, there are even bigger challenges to sharing data across provinces. At the bigger picture level, Canada is unusual in that we have 13 different health systems. Each province or territory has to follow the Canada Health Act but within that they’re free to innovate and adapt. So, from a scientific perspective, you can learn a lot from different provinces by being able to compare approaches and outcomes across jurisdictions. The SPOR Canadian Data Platform aims to give us the ability to conduct this sort of interprovincial research. We are not yet at a point that provinces are sharing person-level data with each other, but what the SPOR Canadian Data Platform will do is to establish what’s called a “distributed analysis model,” where the data don’t actually move but where these interprovincial comparisons can still be made.
So, in the “distributed model” data doesn’t move, but the lessons do.
Yes, the data stays in the provincial data platforms and we build common data standards, common data definitions, and common data algorithms—so we’ve got the same look and feel in terms of clinical and administrative data sets across the different provinces. Then we can do parallel analysis and produce comparable results. It’s a practical alternative to being able to pool data on individuals across provinces. The SPOR Canadian Data Platform will also have a single point of entry which will help get multi-jurisdictional projects up and running much more quickly and efficiently. This platform will be especially useful, for example, if you’re studying rare diseases and you don’t have enough sample size to give you the statistical power to have reliable estimates of outcomes. If you can combine data through distributed analysis to include a few tens of cases each from multiple provinces, then you have something that’s more robust—and that can be facilitated through the SPOR Canadian Data Platform. At its core, it’s about facilitating interprovincial research on health issues. And for smaller provinces, it puts us at the table with everyone else.
What role will the provincial health data platforms in the Maritimes play in this new country-wide data platform?
We’re each a member of the new SPOR Canadian Data Platform, and will have full-time staff dedicated to the effort. But, I think one of our comparative advantages is that the three provinces are already used to working together and collaborating through the MSSU. We can take that experience to the Canadian Data Platform and demonstrate some leadership there. Speaking of the SPOR programs, the SPOR networks and certainly the SPOR SUPPORT Units are bringing important, valuable infrastructure investment into our small provinces. We just don’t have the large scale, well-funded research infrastructure that the large provinces do. Given that we have some of the greatest need, I think these kinds of programs suit us well and really demonstrate our ability to use these limited resources most efficiently for the benefit of our citizens.
Dr Ted McDonald is a Professor of Economics at University of New Brunswick in Fredericton, and also holds several administrative positions related to data infrastructure including Academic Director of the Statistics Canada Research Data Centre, New Brunswick; Chair of the Academic Council for the Canadian Research Data Centre Network; New Brunswick lead for the Maritime SPOR SUPPORT Unit; and Director of the New Brunswick Institute for Research, Data, and Training (NB-IRDT). Dr. McDonald also sits on the Executive of the CIHR-funded SPOR Canadian Data Platform.
Maritime Data Platforms
To learn more about the Maritime Data Platforms, visit:
- Health Data Nova Scotia (HDNS)
- New Brunswick Institute for Research, Data and Training (NB-IRDT)
- Secure Island Data Repository (SIDR)
Big Data in New Brunswick: Exploring the Research, Policy Development, and Community Planning Opportunities
May 1, 2019 @ 3:00 - 4:30 pm
Saint John, NB or Online