First, let’s discuss the methodology I used here. I simply visited the NCBI’s PubMed and ran a search for the keyword ‘Census’. My hope was that I’d find research articles published using ‘Census’ as a keyword, hopefully finding work that used Census data to zero in on public health issues. I was met with over 20,520 hits. Of a sampling of the first 100, 18 articles were using the word ‘Census’ in its colloquial meaning, with others using it to describe the gathering of genomic information or of studies performed on multiple tissue samplings in a single experiment. Of the 18 papers I decided I’d use papers that were exclusively published within the past 30 days to show how information can be immediately useful when combined with the Ottawa Census and the Ottawa National Household Survey collected in 2011. I used Microsoft Excel to organize the data and run my calculations, Google’s Fusion Tables to prepare the heat-maps and Adobe PhotoShop to make the heat-maps more presentable. In Google’s Fusion Tables, automatic heat-map interval determination was used, as opposed to manual, with a consistent 4 bucket limit for the entire city to make the data easily understandable at a glance. One limit of this approach is that it doesn’t inform us exactly how at risk anyone is from any of the conclusions, but only how at risk anyone is relative to anyone else in Ottawa. Thankfully, that’s all we need to make public health decisions for the City of Ottawa! Here’s what I found:
Orleans and Innes have a high risk for incidence of Heart Disease
Research published last month found a link between the availability of convenience stores and poorer dietary habits among the financially disadavantaged members of the population, that could likely give rise to heart disease. Given this, I personally counted the number of convenience stores in each ward, divided that number by the area of the ward to get a more accurate representation of how concentrated the convenience stores are – a ward with 700 kilometers squared of area and 5 convenience stores won’t encourage the population to eat unhealthy quite as much as a smaller ward with a larger number of stores. I could have correlated this with the self-reported income in the National Household Survey but I chose to not do this for a few reasons: the NHS is voluntary, and also because individual income does not accurately indicate whether someone is living in poverty because they could be living with someone who significantly earns more or may have savings they’re relying on such as in the case of retirement. The city of Ottawa itself prefers to define poverty as primarily a factor of living conditions. As such I considered the sum of all households with 4 or more residents (subtracting families from the equation), and houses that have a reported more than one family co-habitating to be living in poorer conditions. The result is the heat-map above: Orleans and Innes came out on top (even when accounting for the NHS income, Orleans and Innes came in the top group with Somerset, but that data is not shown here), and is caused by both of these Wards both holding many residents in struggling or non-independent conditions and having a high number of convenience stores (Innes had the second highest Convenience Stores per unit Area, sitting at 0.58 stores/kmsq behind Somerset’s 0.62 stores/kmsq). Correcting for age, patient records and historical activity, I’d assume these wards are a good target for ambulance stand-by locations.
Knoxdale-Merivale, Rideau-Goulbourn and Rideau-Rockcliffe are at a higher risk of Very Low Birth Weight infants
Research, that was also conveniently published last Month because my search gave me the results in chronological order with the most recent first, discovered that the concatenation of living in a neighborhood that is financially disadvantaged and being a mother from an ethnic minority increased the risk of having an infant with a Very Low Birth Weight. In order to see where this had a risk of happening in Ottawa, I used the National Household’s Survey (NHS) households with a self-reported income levels below $15,000 (the reason I am using this instead of the poverty definition I came up with in prior is because I figured that if the minority families did not complete the NHS satisfactorily or accurately, then their neighbors will have given us a better indicator of neighborhood quality which is the parameter in question here – additionally the poverty statistic above can be skewed by students which are not the group under study here) normalized to the size of the population in any given ward, and multiplied it by the number of people who reported speaking a non-English, non-French mother tongue, with the logic being that the higher the product, the greater the risk of the results of the research manifesting in the neighborhoods. The three wards named above accounted for three of the wards with the top five self-reported low income in proportion to the size of the ward. They were also three of the top seven wards with the highest number of non-English, non-French mother tongue speakers. In this case, it was the overlap that put these three wards at the top of the risk list. Interestingly, the lowest risk occurs in the wards surrounding the downtown core. This result is significant because it points to these three wards being ideal locations for prenatal care clinics.
Rideau-Vanier and Somerset are likely to have teenage drinking problems
Given the heat-chart above, I thought I’d add a health chart related to more affluent neighborhoods. Well, according to research published 4 days ago, that would be teenage drinking. Where the crime-rate goes up, with data from the City of Ottawa, and immigrant or foreign families are less common, teenage drinking is likely to occur. The research postulates that this is due to proactive parenting, and my charts here do not account for transportation quality (which is supposed to exacerbate the problem), but the results are still fairly conclusive. Even our mid-range risk wards were low risk in the previous chart. The data here was not normalized to account for the percentage of the population that teens make up because of the underlying assumption that travel to these areas is easy, and also because access to alcohol is easier to achieve in these areas where bars are numerous. This is where police patrols need to be concentrated to avoid alcohol poisoning nightmares on particularly problematic nights.
Overall, there are still a few more things I’d like to look at when it comes to wading through the Census data for public health solutions. For example, I need to determine where cobblers and Orthopedists should setup shop given the incidence of foot-related injuries, where ovarian cancer survival rates might demand a center or clinic and where asthma treatment possibilities with difficulty of ambulance access might be, but for the intents and purposes of writing this article for tonight’s Open Data Ottawa meeting, and given the time restraints I’m under, I think the above is sufficient and will make for a good lighthearted presentation.
Conclusion: Census data is cool and there’s so much information to be discovered in it given everything else we know!