Saving lives with location

By on 5 April, 2018

John Snow’s 1854 map of cholera cases in Soho, London that demonstrated spatial clustering of cases around the ‘Broad street pump’.

The role of GIS in community health will be explored at this month’s Locate ’18 & GeoSmart Asia ’18 conference in Adelaide. Dr Neil Coffee and several other speakers will explore GIS advancements in the Health and Community stream scheduled for Wednesday, 11th April.

In 1854, a physician named John Snow was called upon to investigate a cholera outbreak that was decimating the London district of Soho. Prevailing logic in the medical community attributed the means of transmission to ‘miasma’, a hazily-defined airborne pathogen thought to be released by decomposing organic material. A detailed, elegant and widely-accepted theory, it failed to inform any meaningful intervention in the outbreak, and Snow didn’t accept it as the cause – arguing that oral transmission was most likely.

The actions that followed have led John Snow to be dubbed the first epidemiologist. Struggling to convince the medical establishment of his theory, Snow deepened his investigation and turned to a powerful, disruptive technology to test his hypothesis – he made a map. With infections and deaths plotted, Snow was able to demonstrate that cholera cases were clustered around a water pump, now famously known as the ‘Broad street pump’. While he wasn’t able to detect any organic material in the pump’s well, Snow was able to make a critical intervention from his data – he removed the pump’s handle, and the cascade of cases was halted.

His methods would have profound impacts on the field of epidemiology, and his germ-based model of disease transmission, in part proven by his stunning intervention in Soho, eventually became widely accepted by the 1860s.

In 2018, most countries in the world are lagging behind their targets for health-related Sustainable Development Goals (SDGs), the UN’s development agenda for 2030. The UN’s second baseline assessment on SDG progress released in September 2016 found that less than five percent of countries worldwide were on track to meet targets for key public health issues, including tuberculosis, road deaths, suicide and obesity.

Whilst John Snow’s ground-breaking analysis and intervention informed the course of contemporary epidemiology, the blistering pace of technological development and access to spatial data has opened new vistas for public health research, policy and GIS analysis alike.

Dr Neil Coffee.

Dr. Neil Coffee is one of a small set of GIS innovators in Australia, pushing the envelope of health research and location’s place within it. As associate professor of the Centre for Research and Action in Public Health (CeRAPH) at the University of Canberra, he is establishing Australia’s first tertiary course on geo-health, and is currently overseeing a set of projects that interrogate how location data can inform medical interventions and policy positions that lead to real, measurable positive health outcomes – both within and outside of Australia.

One core research interest for Dr. Coffee is the influence of environmental factors on health outcomes. That is, how factors related the built environment and socio-economic development patterns can combine to create liveability conditions that may have a profound impact on lifestyles of their residents, specifically how these factors can affect lifestyle-driven health burdens such as cardiovascular disease, diabetes and obesity.

Dr. Coffee doesn’t mince words on the importance of location in fully understanding the burden of disease.

“Health is one of the biggest spends now for government and it’s getting out of control, because of the rapid increase in the prevalence of chronic diseases not just in Australia, but across the world,” he said.

“It isn’t enough just to know that you know the particular prevalence of cardiovascular disease in Australia. It doesn’t help us focus that funding, it doesn’t help us focus on a particular location where we find there’s an incredibly high prevalence of cardiovascular disease relative to other places. The only way to understand these things is through the use of spatial. If we know both the extent, the location — and the location with the extent that’s relative to it – we can actually start identifying areas where we get a better outcome from the spend of our dollars,” he said.

A flagship undertaking in this area is a new partnership focused on geo-health between the University of Canberra and the Dasman Diabetes Institute (DDI), tackling a specific, thorny public health concern – the alarming prevalence of type 2 diabetes in Kuwait. Dr. Coffee and his team are building a geo-health data lab in Kuwait, within which they plan to study the variability in the disease’s prevalence in the country, to analyse and isolate location-specific factors that may contribute.

Relative location factor (RLF) is a socio-economic status measure that can provide valuable insight in health-focused spatial analysis.

DDI approached the CeRAPH team, and the two entities now have an Memorandum of Understanding for a four-year, multimillion dollar partnership and project.

“They recognised that the spatial dimension was something they needed to understand as well. They know the extent of the problem in Kuwait, but they know nothing about how that’s actually distributed — is it just the same everywhere or are there major areas? Because interventions at the whole population level only have limited success,” Dr. Coffee said.

Once the geo-health centre is established, a local team will be trained and the datasets will be built up, including location of services and health facilities, roads and infrastructure, demographic data and a population monitor tool that will help the team carry out randomised interviews.

“We’ll ask them questions about their health behaviours, so that we can start to establish norms of behaviours for different areas, that we can then start to associate with the health data, and the health environment,” Dr. Coffee said.

The CeRAPH team has a range of operational Australian projects currently, including an initiative to create a liveability index for remote Indigenous communities. However, in the Australian context, Dr. Coffee finds that a persistent barrier often prevents the team conducting the type of analysis that they will be conducting in Kuwait: access to adequately detailed spatial data.

“The main issue for us is the ability to use inner record data at the address level — the actual address level. To make it really make sense of patterns of health, that’s the level we need to be doing the analysis at,” Dr. Coffee said.

The team frequently finds their research constrained by lack of access to address-level data on privacy grounds – even if they will not be publishing results in that level of detail. The potential for re-identification of individuals is oft-cited as the basis for withholding data sufficiently granular for detailed spatial analysis.

“If we’re talking about having anonymized data — we do, but we also have the street address, the argument is that we’ve now got the capacity for re-identification, and that’s quite valid. Now there are ways of randomly perturbing those data spatially in the same way as we would randomly perturb a table of data like the ABS do. But we haven’t really reached a point with Australian data providers where there’s a level of comfort in the capability that that will in fact protect confidentiality,” he said.

The modifiable areal unit problem can significantly alter outcomes.

“We need somehow to get them to be comfortable with the fact that we can take record data at the street level and randomly perturb it spatially such that it doesn’t enable identification – but it does enable us to identify the real patterns.”

Most health, demographic and socioeconomic data in Australia are aggregated to the statistical areas used in the and local government areas (LGAs), and this level of resolution can cause measured outcomes to be subject to the modifiable areal unit problem, which says that both the strength and direction of a statistical association can be manipulated by the way an area is divided.

“Many of those spatial units are very large so they don’t really give us any true understanding of what’s happening internally, distribution-wise. They ignore the fact that some of the issues that we might identify are actually occurring on the borders of two spatial units, and therefore we’re seeing patterns across both that really are quite isolated to a particular location,” Dr. Coffee said.

“So what we often find in Australia out with our analysis is we’re restricted by these spatial units that we know are impacting on the outcome in the first place.”
This issue is compounded by data is not always collected or managed consistently across jurisdictions, such as state departments. Dr. Coffee cites Scotland, New Zealand and Canada as best practice examples, with all available public data managed by a single national body, and individual-level data available for research purposes.

“Many other countries do quite detailed spatial analysis with data with far more availability than we have here in Australia. So we only need to look to some of the overseas examples the ways that we can improve the way data are provided,” he said.

Hear more from Dr Neil Coffee at the Locate ’18 & GeoSmart Asia ’18 conference being held 8-11 April in Adelaide. He will be the lead speaker at the conference’s Health and Community stream.

You may also like to read:


, , , , , ,


Newsletter

Sign up now to stay up to date about all the news from Spatial Source. You will get a newsletter every week with the latest news.

Interview: Tori Murrant, GIS analyst
Having stumbled across the geospatial sector at university, ...
Testing SouthPAN and commercial GNSS services
UNSW surveying students were challenged to put a range of So...
Here’s what’s in our latest issue!
Learn about the metaverse, mapmaking, 3D scanning, RINEX, hy...
Modern Methods of Construction Roadshow
The events will show how the latest software, tools and tech...
Real-time LiDAR mapping system
The Brumby LiDAR rapidly produces point clouds by removing t...