McMaster statistician digs deep into Code Red data and uncovers some startling relationships
By STEVE BUIST
There’s been something gnawing away at Patrick DeLuca for the past five years, an itch he hadn’t been able to scratch.
DeLuca is a statistical whiz and geographic information specialist at McMaster University and one of two researchers who helped put together The Spectator’s highly acclaimed Code Red projects, first launched in 2010.
The series have shown the strong connections that exist between the health of people and their social and economic circumstances.
“One of the interesting things to me is that when you think about everything that came from Code Red, plenty of good things happened, but it was based entirely on a descriptive study,” said DeLuca. “We had maps, some descriptive statistics, good stories and good photos, all melded together.”
But the data had never been subjected to a series of rigorous, heavy-duty statistical tests that could help quantify — or even confirm — these connections between health and wealth.
“I was interested in seeing what the actual relationship was between some of these variables from a statistical point of view,” said DeLuca.
“Obviously, we know that poverty is related to health,” he added. “But the question is how much?”
Quite a bit, as it turns out, based on an article recently published in the academic journal AIMS Public Health, written by DeLuca and a McMaster colleague, Pavlos Kanaroglou.
DeLuca decided to look at one of the most shocking of Code Red’s key variables — the average age at death.
Code Red showed there was a 21-year difference in life expectancy across Hamilton’s 130 neighbourhoods.
“That was the one that kind of struck a chord with a lot of people,” said DeLuca. “The gap was massive.
“It’s kind of the ultimate measure of health, right? Either you’re living or you’re dying.”
What DeLuca did was take about two dozen of the Code Red health, social and economic variables — things such as percentage of people living below the poverty line, urgent hospital admission rates, percentage of single-mom families — and winnow them down to three categories.
One category was defined as poverty, one was defined as low income and working class, and the third was health and aging.
The two researchers then ran some very sophisticated statistical tests on the groups of variables.
What they found was that the three categories accounted for about 42 per cent of the variation in life expectancy across the city.
That may not sound like much, but it’s actually quite startling from a statistical standpoint.
“I’ve seen studies where you’re explaining 12 per cent, maybe 20 per cent,” said DeLuca. “This is 42 per cent.
“We’re talking about almost half of the variation being explained by these three particular factors.”
Of course, that still leaves 58 per cent not explained by the three categories, which isn’t that surprising.
The three categories can’t take into account things such as deaths from accidents or suicides, genetic factors, such as a family history of heart disease, or cancer.
“There’s still a lot we’re not capturing by these Code Red variables,” DeLuca added.
Most importantly, DeLuca has finally scratched his itch and he can now breathe easier. Heavy-duty statistics have backed up Code Red’s findings.
“When Code Red came out, I remember reading somewhere in the comments or someone’s blog, ‘Well, you know, this is just a bunch of maps and there’s no real analysis done here,'” DeLuca recalled. “Well, here it is now.
“The bottom line is this stuff does matter and the social determinants are important and they do, to some degree, explain health in this city.”