Brownell – researching population health  

I’m a senior researcher at the Manitoba Centre for Health Policy and we mostly do research using large secondary data bases, administrative data bases which are data bases collected for purposes other than research. For example, the physician data base is collected to pay physicians basically, but it’s got a lot of rich information in it. It’s got information about not only who the physician is and who the patient is but also what the patient saw the physician for.  So there’s a whole collection of these administrative data bases at our centre, and I do research mostly focused on child health and development.  

So, you know, we can look at physician visit patterns among children. From the hospital records we can look at all sorts of birth information. Very important to child development is how healthy you are when you’re born. So we can look at things like low birth weight, prematurity and then start to connect up these data bases. So my focus is primarily child health and development. 

One of the other tools that people use are surveys. There’s the NLYSC — The National Longitudinal Survey for Children and Youth, which provides really rich, in-depth information. But of course, it is a survey. It’s not given to everyone in the population. And what we find when we start looking at surveys, is very often those families who are missed by the survey–particularly a longitudinal survey where you have to keep re-contacting families– are the highly transient ones, the poor, those they can’t contact by phone just to follow up, so we sense that the very, very poor, the very, very challenged families are underrepresented in these surveys.  

The other thing is, in a survey, you’re asked a question, you know, “How’s your child’s health”, “how’s your health”, it’s basically based on an opinion. And what you feel might be really poor health, really excellent health may be different from the next person. So there’s those sort of strikes against surveys. As I say, NLYSC provides very in-depth information that we could never get from our administrative data but what we, what we do get is broad and we get the entire population. And it’s somewhat objective but I won’t say it’s completely objective because, you know, whether or not you go to a doctor could also be determined, you know, if you have an overanxious parent they may take their kid to the doctor more often than a parent who could care less. 

And so the kinds of records we see, okay this child saw the doctor 10 times in the year does that make him sicker than this child who saw the doctor only once. But when you start looking t things like hospitalizations, prescriptions, you’re looking more at okay a doctor has seen this child and said this child is sick enough to be admitted to hospital, or…and so there’s those advantages and the population base covers the entire population. So we get the poorest of the poor, even those who are highly transient, as long as they’re transient within Manitoba when they go to a doctor or hospital, they still have the same personal health information number. And that’s how we find out about them in the data. We don’t have the actual health number but we have an anonymized version of it. And so no matter where they have their visit in the province we can follow them.